Report No. 7 - 2001
Eating Patterns
A Day in the Lives of Nordic Peoples
by
Unni Kjærnes (ed.)
2001
NATIONAL INSTITUTE FOR CONSUMER RESEARCH
PO.BOX 173, N-1325 Lysaker, Norway
2
Eating Patterns
Foreword
This report has been written on the basis of a comparative survey of Nordic
eating patterns: En matdag I Norden, En komparativ undersøkelse av det moderne hverdagslivets matvaner. It has been carried out through a collaboration
between Ass. Professor Jukka Gronow and Research Fellow Johanna Mäkelä
at the Department of Sociology, University of Helsinki (the latter now the
Head of Research at the National Consumer Research Centre, Helsinki), Ass.
Professor Marianne P. Ekström at the Department for Home Economics at the
Göteborg University, Ass. Professor Lotte Holm at the Research Institute for
Human Nutrition, Agricultural University, Copenhagen, and Senior Researcher Unni Kjærnes at the National Institute for Consumer Research
(SIFO) in Oslo. These researchers are together responsible for the overall study design and the plans for the analyses.
The work has been coordinated by Unni Kjærnes at the National Institute for
Consumer Research.
The study has been financially supported by the Joint Committee of the Nordic Social Science Research Council (1996-2000). Furthermore, the economic
contribution from each of the collaborating institutions has been considerable.
In addition to the mentioned researchers, other persons have given important
contributions. In particular, we want to thank research assistants Ellen
Bjørkum and Anders Nyberg at SIFO, whose extensive contribution in data
preparation and analyses are invaluable. This goes also for Anni Jääskeläinen
at the University of Helsinki, who appears even as a co-author in this report.
Professor Elisabeth L. Fürst at the University of Oslo contributed in an early
phase of the project and as a co-author on cooking and gender. Senior researcher Lisbet Berg has contributed with data analyses and co-authorship on
attitudes towards food safety.
4
Eating Patterns
This report has been edited by Unni Kjærnes. However, the authors of each
chapter are responsible for their respective chapters and the underlying analyses.
Lysaker, January 2002
THE NATIONAL INSTITUTE FOR CONSUMER RESEARCH
Table of Contents
Foreword ...................................................................................................................... 3
Table of Contents ......................................................................................................... 5
Summary ...................................................................................................................... 7
1
Introduction ........................................................................................................ 25
Unni Kjærnes, Marianne P. Ekström, Jukka Gronow,
Lotte Holm and Johanna Mäkelä
2
What Did They Eat? ........................................................................................... 65
Johanna Mäkelä, Unni Kjærnes and Marianne Pipping Ekström
3
The Daily Rhythm of Eating .............................................................................. 91
Jukka Gronow and Anni Jääskeläinen
4
The Meal Format .............................................................................................. 125
Johanna Mäkelä
5
The Social Context of Eating............................................................................ 159
Lotte Holm
6
Family Meals.................................................................................................... 199
Lotte Holm
7
The Gendered Division of Cooking.................................................................. 213
Marianne Pipping Ekström and Elisabeth L’orange Fürst
8
Consumers and Food Safety Policies in the Nordic Countries ......................... 235
Lisbet Berg and Unni Kjærnes
9
Discussion and Conclusions ............................................................................. 255
Unni Kjærnes
Literature .................................................................................................................. 285
List of figures and tables .......................................................................................... 301
6
Eating Patterns
Summary
The research questions: grazing or structured meals?
Food habits are eagerly debated these days. A recurring question addresses the
fate of meals, querying whether traditional meal patterns and meal formats are
being disrupted. Meals are seen as the cornerstones of family life, and the underlying discussion is often whether this alleged disruption also indicates an
upheaval in family life and social life in general. Others suggest that even
though alterations in food preferences and cooking may take place, neither
normative regulations nor the role of commensality will necessarily disappear.
Moreover, many studies have presumed that eating habits which include a
family meal taking place at a regular hour every day have been more binding
or compulsory in earlier times in Europe. But there is very little systematic
historical evidence of the real occurrence of such normative or empirical patterns of eating.
This study is about how eating as an everyday activity is structured in modern
societies. We describe the present situation with regard to eating patterns and
meals in the Nordic countries. Are structured meals the dominant feature or
are we seeing tendencies of disruption? What are the impacts of contemporary
family structures and gender relationships? A comparative design is applied in
order to make cultural features more clear, identifying those that are shared
and those that are distinct from one country to the next.
The questions of eating patterns and meals are analysed at two levels. The
first level deals with the existence of structures for the individual, i.e., the existence of ‘meals’ (as opposed to ‘grazing’) – leniency vs. strictness – on the
one hand, and on culturally and socially shared patterns and formats of eating
– variety vs. uniformity – on the other. Second, there are the contradictory
assumptions of eating patterns as a matter of individual choice of lifestyle, or
as strongly influenced by social structure and societal institutions. Our main
concern here has been the influence of conditions with direct relevance for
8
Eating Patterns
everyday routines, such as the organisation of the working day, the type of
family and household, and the place of residence.
Representative surveys in four countries
Our questions called for quantitative empirical testing in representative samples of the population. Data were collected in April 1997 using computer assisted telephone interviews (CATI). One thousand two hundred individuals
above the age of 15 were interviewed in each of the four countries (Denmark,
Finland, Sweden and Norway), yielding a total sample of 4 800. The design of
the questionnaire and the data collection procedures were closely coordinated.
We were interested not in the quantities of food consumed but in the combinations of food and the social context of eating.
The challenge was to formulate questions that could reveal the construction of
meals within this quantitative and comparative setting. Our solution was first
to deconstruct eating into a number of questions and then to reconstruct the
answers into snacks or various types of meals. We needed a concept that
would capture the different aspects of the eating system. Eating events include
every occurrence of eating something (just a beverage or only eating chewing
gum or sweets, etc., were excluded). This is the basic unit of the eating system, which distinguishes between three dimensions: the eating pattern, the
meal format, and the social organisation of eating. The questionnaire consisted of loops where questions on the time, the structure, the content, and the
context of eating events were presented. For each eating event, there was a
basic distinction between hot and cold food. The subsequent questions were
more elaborate for hot than for cold eating events. For practical reasons, the
first eating event had no distinction between hot and cold food.
In the analyses, we have distinguished between five types of eating events,
according to the food components and cooking. These types form a hierarchy
in terms of complexity:
- a snack event (an ice-cream, an apple or a chocolate bar with/without a
beverage),
- a cake event (a cake, pastry, etc.; may also include a snack),
- cold food eating events – a cold meal (sandwich, cold cuts, a salad; may also
include a snack and/or a cake),
- breakfast, i.e., the first eating event of the day (often cold foods, like sandwiches or breakfast cereals, but may also involve cooked dishes like porridge),
- a hot meal (hot indicates that cooking has been involved and that the food is
eaten hot).
Summary
9
‘A proper meal’ is a construction that is used as a tool in our analyses of the
structure and complexity of hot eating events. ‘Proper meals’ were defined as
containing a centre, a staple or/and bread, and vegetables, with trimmings as
an additional option. The idea of ‘proper meals’ is important in Nordic qualitative studies and we wanted to find out how often they occurred in practice.
Flexible eating routines
‘Grazing’ implies that food is eaten more or less at random with regard to
time, place and contents, according to immediate preferences and concerns. In
this study, we found that for the majority in all countries eating was concentrated in relatively few events. Between eighty and ninety per cent had had 35 eating events on the day before (somewhat less on Sundays). The extremes,
with very few or more than 5 eating events per day, represented only the small
percentages that might be expected from the variation and flexibility of ordinary everyday life (illness, extra treats, etc.). The total number of eating
events was somewhat lower in Norway than in the other countries. Presuming
that older age groups or, say, people with lower educational or social status
would be more traditional in their behaviour, one would have expected them
to eat more often and at more regular hours than younger and better educated
urban people. This was, however, generally not the case in our study. In all
countries, living together with others seemed to be the one factor that more
than anything else increased the probability of following socially shared eating patterns.
A majority in all countries also had one or more cold food eating events later
in the day. Ninety per cent or more had eaten at least one hot meal. More than
80 per cent had had one hot meal on the day before in Denmark and Norway,
while ten per cent or less had had two. In Finland and Sweden, on the other
hand, about 40 per cent had had two or even three hot meals on the (week)day
before. This difference was paralleled by the Danes and Norwegians having
had more cold eating events, first of all sandwiches. These three types of
events constitute the core of the Nordic everyday menu, eaten by almost all
respondents on the day before.
In all four countries, there were clearly some peak hours during the day when
a large proportion of the population was eating something. Eating was concentrated at conventional hours: breakfast, lunch and dinner. The most distinct
peaks occurred at breakfast and lunch time on weekdays. Eating hours were
most uniform in Denmark, least so in Finland and Sweden. Additional, but
smaller peaks occurred during mid-morning, mid-afternoon and/or in the evening, but these varied considerably between the countries. Dinner was eaten a
couple of hours later in Denmark and Sweden, compared to Finland and Nor-
10
Eating Patterns
way. Eating hours were quite different and showed much more variation on
Sundays, but again the most uniform pattern was found in Denmark. Still, a
considerable proportion of the informants had had something to eat at any
hour during the day, from 4 in the morning till 12 at night.
As far as the daily rhythm of eating and the average number of different eating events were concerned, Norway and Finland differed quite noticeably
from each other. In some other respects, however, they resembled each other.
In both countries 1) women ate more frequently than men, 2) those who lived
together with others ate more often than those living alone and 3) couples
with children ate more often than couples without children, who ate more
times than single parents with children, who again ate more times than single
persons without children. In Finland, the inhabitants of bigger cities (over
100,000 inhabitants) ate more often than others. However, this was not the
case in Norway.
Denmark was the most homogeneous country. The only significant differences were found between age groups (those who were under 35 ate less frequently than older generations) and between municipalities of different sizes:
unlike in Finland, people living in big cities (over 100,000 inhabitants) ate
less often than others.
In Sweden, as in Norway and Finland, women ate more often (4.0 eating
events) than men (3.8). (This was the only difference that was found in as
many as three countries.) The differences between age groups in Sweden were
similar to those found in Denmark (the middle-aged ate more frequently than
others) with the exception of the oldest age group. Only in Sweden were there
systematic differences between groups of different income levels: the higher
the income, the more often did the respondent eat on average. Only in Sweden
did the occupationally active population eat more often than those at home
(4.0 vs. 3.8).
In Norway, people ate less frequently than in the other countries. Older age
groups ate more frequently than younger ones, the two oldest ones about 40%
more often than the youngest (under 25 years). Women ate more often than
men and the occupationally active ate more often than those who were not
occupationally active. Those living with someone ate more often than those
living alone, probably reflecting more 'order' in the lives of couples and people living together.
Gronow and Jääskeläinen suggest in Chapter 3 that the rhythms of eating represent socially shared routines that have emerged from the need for social co-
Summary
11
ordination, particularly in relation to work and family life. Still, this rhythm is
by no means compulsory (except for those living in some kind of institution),
but represents very flexible conventions with a lot of freedom for the individual. People may establish their own routines and they may make exceptions.
There are also strong cultural elements in these patterns, which are clearly
demonstrated with regard to the cold vs. hot lunch traditions.
Hot or cold food at lunch?
If the cold and hot eating events - one of the basic distinctions in the questionnaire - are considered separately, the Nordic countries fall into two distinct groups. In Denmark and Norway, cold food is eaten relatively more frequently than in Finland and Sweden, which are 'hot food' cultures. The average number of cold eating events (with the exclusion of the day's first eating
event, breakfast) was highest in Denmark (2.0), and second highest in Norway
(1.7), and lowest in Finland and Sweden (1.6). The average number of hot
eating events in both Sweden and Finland was 1.4 and in both Denmark and
Norway 1.0. About four-fifths of all Norwegians and Danes ate hot food once.
One-tenth had not eaten hot food at all during the previous day. In these two
countries, rather few people had eaten hot food more than once (6% in Norway and 12% in Denmark), whereas the respective percentages in Finland
were as high as 43% and in Sweden 40%. In Denmark and Norway there is a
general custom of eating cold food at around noon (what could be called the
lunch hour), whereas Finns and Swedes are more likely to eat hot food at
noon.
Despite some similarities, the differences between population groups were
generally not similar, but varied from country to country. A logistic regression
analysis revealed that in Norway, eating cold food at lunch hour was more
common with people of higher educational status. It is also noticeably more
frequent among those who are occupationally active and among those who
live together with others. The oldest age group (over 65 years) resembles the
other groups more once the differences in their social and occupational status
are taken into account. Denmark was more homogeneous as far as eating cold
food at noon was concerned. Eating a cold food is more common in the middle-aged groups, and lower- and middle-class white-collar workers ate cold
“lunch” more often than both manual workers and higher-class white-collar
workers.
In Sweden and Finland, eating hot food at lunch-time was more common.
There were no statistically significant differences among the social groups in
Finland with regard to eating hot food at noon. In Sweden, both those with
white-collar status and those who were occupationally active (as opposed to
12
Eating Patterns
those who were not employed) reported a higher frequency of eating hot food
at noon in the same model.
Living with someone else was the only factor that increased the probability of
eating hot food at dinner-time in all four countries. Age had some influence in
all countries as well, but not to the same degree. Education and social status
had the opposite effect in Norway compared to Finland. Sweden and Denmark
were more homogeneous in this respect. These results are rather difficult to
interpret except for the fact that those who live together with other adults and
children are more apt to eat a hot food meal in the late afternoon or early evening. It could be presumed that more effort is directed towards more social
than individual eating events.
Simple everyday menus
What have we found out about food and drink patterns in the Nordic countries? The potential range of foods and beverages is extremely wide, and we
must expect all kinds of combinations in a representative study of eating.
However, according to our findings, the range of foods and beverages taken at
breakfast and at cold eating events later in the day was not very large on an
everyday basis. Variety may be significant at the level of labels and flavours.
But this is not reflected when looking at types of foods and dishes as in this
study. The questionnaire included long lists of foods and beverages, of which
several options were hardly selected at all. Even though the number of different dishes served at hot meals is considerable, it has been possible to classify
most of the meals and meal components according to a shared classification
scheme. This picture of limited variation is evident also for the beverages accompanying the food at these events.
Certain patterns could be identified, first of all at the national level. In the
Nordic countries, almost all respondents had had something to eat in the
morning on the day before. A majority in all countries also had one or more
cold food events later in the day. Ninety per cent or more had had at least one
hot meal. More than 80 per cent had had one hot meal on the day before in
Denmark and Norway, while ten per cent or less had had two. In Finland and
Sweden, on the other hand, about 40 per cent had had two or even three hot
meals on the (week)day before. This difference was paralleled by the Danes
and Norwegians having had more cold eating events, first of all sandwiches.
These three types of events – breakfast, sandwiches later in the day, and one
or two hot meals – constitute the core of the Nordic everyday menu, eaten by
almost all respondents on the day before. Their contents and order varied,
however, between the countries. On the other hand, eating events that included only a snack or a sweet pastry were reported to be much more rare:
Summary
13
only 20-30 per cent of the respondents had eaten these types of foods outside
the major eating events identified above.
Quite distinct national meal systems could be identified:
Denmark: breakfast with sandwiches, lunch with sandwiches, quite a few had
a mid-afternoon meal/snack with sandwiches or a cake, a hot dinner
around 6-7pm, and in some cases a snack, a cake or a sandwich later in the
evening. These events represent five clear peaks in the Danish time schedule.
Finland: breakfast with sandwiches or porridge, a hot lunch accompanied by a
salad, a mid-afternoon snack with a cake or a sandwich, hot dinner at
around 5-6 pm, and evening meal with sandwiches. These events also represent five peaks, but the peaks are much less clear compared to the Danish ones.
Norway: breakfast with sandwiches, lunch with sandwiches, a hot meal in the
afternoon (4-5pm), and sandwiches later in the evening. There are four
(not very clear) peaks, with no mid-afternoon peak.
Sweden: breakfast with sandwiches and/or breakfast cereals/porridge, a midmorning snack with sandwiches, a hot lunch, a mid-afternoon cake or
sweet pastry, a hot dinner, and perhaps also a sandwich, cake or snack later
in the evening. All in all, there are 4-5 not very clear peaks (apart from
lunch).
Sandwiches are important components of the eating system
The first eating event on weekdays took place before 10 o’clock in the morning. Very few had skipped breakfast, and the content of that meal was quite
uniform. The main component was open-faced sandwiches. In Norway, very
few had other foods or dishes in addition to or instead of sandwiches. There
was more variation in the other countries, the main alternative or additional
food items being porridge (in Finland and Sweden), breakfast cereals (in
Denmark and Sweden) and yoghurt in all three countries. These foods were
most often accompanied by a hot beverage, first of all coffee. A common replacement or addition was milk (most often in Norway), juice (most in Sweden) or tap water.
Cold meals later in the day were typically open-faced sandwiches. As already
mentioned, the majority had sandwiches for lunch in Denmark and Norway.
This was also the most common alternative to a hot lunch in Finland and
Sweden. Cold eating events in the form of sandwiches were registered in
Norway and Finland also as an evening meal. The Danish pattern included a
(small) peak in mid-afternoon and the Swedish one in mid-morning. Apart
from that, sandwiches had been eaten at most times of the day.
14
Eating Patterns
Cakes and pastries had also been eaten throughout the day. Small peaks could
be identified in Denmark (mid-afternoon, late evening), Finland and Sweden
(both in mid-afternoon). This type of event is well known in Norway as well,
but according to our data it does not seem to be socially coordinated at certain
hours. Minor snacks seem to be even less organised in time, but with a tendency towards more frequent consumption in the evening. This category was
very heterogeneous. Fruits and salty snacks/sweets appeared with about the
same frequency. Minor eating events, consisting only of a cake, some fruit or
sweets, occurred on the average less than once a day, least often in Norway.
Both on weekdays and Sundays most people had not eaten 'sweet pastry' (7080%) or 'snacks' (about 80%) at all in all the countries. As merely a beverage,
a piece of chocolate or chewing-gum alone was not recorded, no conclusion
can be drawn about the frequency and general occurrence of very small 'in
between' snacks on the basis of this study alone. What can be concluded, on
the other hand, is that no respondent relied heavily on such minor snacks; they
were additions to substantial eating events.
Hot meal menus
About two thirds had a hot meal from 3 pm onwards (dinner-time) in all four
countries, although the time schedule varied. There seem to be no fundamental differences between hot meals eaten at different times of the day. Most hot
meals were relatively simple, containing two or three elements, that is a main
ingredient (the most common being meat), a staple (most often potatoes) and,
in about half of the recorded meals, a vegetable side dish. About 25-30 per
cent of the hot meals did not include a separate staple, while quite a few Finnish and Swedish hot meals included bread in addition to potatoes or another
staple. Meat appeared more often on the Danish menu, while a higher frequency of fish was a distinguishing feature of the Norwegian hot meals. Very
few mentioned vegetables as the main ingredient, the lowest frequency appearing in Norway. Few hot meals included more than one course. Starters of
any kind appeared very rarely, while one in four weekday hot meals included
a dessert, mainly fruit or ice-cream.
There was frequent use of dishes with minced meat (and minced fish in Norway) or dishes where different ingredients were combined, that is vegetables
and/or a staple mixed with the meat. Such dishes were least frequent in Denmark, where a higher proportion of the meals included “pure” meat (or fish).
About one in ten hot meals included only one element (pizza, soup, a casserole including several ingredients etc.). A minor proportion of the hot meals
included more than three elements, the highest frequency appearing in Sweden. The variety of vegetables in each country was quite limited. In Norway,
Summary
15
at the extreme, about one third of all hot meals included carrots. The range of
preparation methods was also limited, with pan-frying being the most popular
choice. Relatively “new” preparation methods, like wok, deep-frying or the
use of a microwave oven appeared quite infrequently, the primary alternative
to pan-frying being the traditional methods of oven baking (mainly in Eastern
Finland) and boiling (mainly in Norway).
A significant proportion of the hot meals had a main ingredient classified as
“other”, including a variety of traditional and modern, simple and elaborate
dishes. They ranged from a hot cheese sandwich or porridge, through pizza, to
dishes that may be quite elaborate, like pasta dishes, couscous or a fish soup.
Here the responses included many categories, and it has been difficult to identify clear patterns. This should be interpreted as a “catch-all category” of
dishes that are not easily categorised, rather than as an indication of the deterioration (or modernisation) of hot meals in the Nordic countries.
Some, but not all hot meals were ‘proper meals’
The number of components provides a simple measure of the complexity of
hot meals. In all four countries, people reported eating events that show both
complexity and variation. People still have meals with a structure, but it
should be noted that they also eat simple meals with only a few components,
and some informants did not eat any hot food during the day. In every country, the most typical hot meal had two or three components. Meals with one
component were more common in Norway, whereas in Sweden there were
more meals with four or five components. In Norway and Denmark, no meals
with five components were reported.
A further analysis of the format of hot meals in terms of complexity was made
through the term ‘proper meals’. We found that eating ‘proper meals’ was
quite common in all countries. The ‘proper meal’ combination of centre, staple and vegetables (CSV) was the most common combination on weekdays in
all countries except Finland, where it placed second. More respondents in
Sweden had had at least one ‘proper meal’ during the day before compared to
the other countries; 6 per cent even had two. The Norwegians also had a relatively high proportion of ‘proper meals’, but very few had more than one. The
proportions were lower in Denmark and Finland, but there were some Finns
who had two such meals. These results show that even if people had two hot
meals during the day, both of them were rarely proper ones. More Danish hot
meals consisted of a centre with a staple or a centre with staple and trimmings. Meat was even more popular as a centre when the hot meal in question
was a ‘proper meal’. This was true for all countries except Norway. In Den-
16
Eating Patterns
mark, 85 per cent of ‘proper meals’ had a meat centre. In Norway, fish seemed to be almost as popular as part of ‘proper meals’.
Logistic regression analyses show that only age factors seem to affect the likelihood of ‘proper meals’. In all countries taken together, respondents in the
youngest age group were less likely to eat ‘proper meals’ than the other age
groups. The social groups did not differ from one another. However, when
looking at dinner-time meals alone, we found that higher education and female gender increased the probability of eating ‘proper meals’. The effect was
the opposite for social status, with the middle-class white-collar group having
had ‘proper meals’ in the afternoon/evening less often. The youngest age
group was less likely to eat ‘proper meals’ even in this analysis. People living
alone were less likely to eat ‘proper meals’ than people living in larger households. Furthermore, ‘proper meals’ were more common during weekends than
on weekdays.
When analysed separately, it was revealed that while women in general ate
‘proper meals’ relatively more often, men were more heterogeneous in this
respect. The probability of eating ‘proper meals’ increased among men who
were older, better educated and/or economically active and who lived in families. The effect of social status in Finland was further limited to men, while
there was no significant effect for women. However, men and women living
with children were both more likely to have a proper meal in all four countries.
Most eating took place at home, in the kitchen or at a dining table
It has been suggested that the social context of eating is radically changing.
Modern eating is thought to take place increasingly in isolated circumstances
and less often within the social context of the family. The ways in which
meals are located, timed and socially organised can be understood as reflecting the demands of societal production as well as those of the private sphere.
At almost all times of the day, more people had eaten at home than in any
other place. In each country, large peaks of eating at home occurred in the
morning. There were also peaks indicating eating at home in the middle of the
day, but they were typically smaller. In the evening there were large and
broader peaks again. In Denmark, more than half of the respondents had eaten
only at home (52%), while in Norway and Sweden somewhat less than half
(41% and 43% respectively) had done so.
Older people (aged 45 or more) and those who were not occupationally active
were more likely to be included among those who had eaten exclusively at
Summary
17
home. On average, older age groups had one or more extra eating events in
the home as compared to the youngest age group. The frequency of eating
events in the home depended more on the presence of adults (and children)
than on that of children only. In each country, women tended to eat at home
slightly more frequently than men. In Denmark, people living in large cities
ate at home somewhat less frequently than those living in smaller urban areas.
This trend did not emerge in Norway or Sweden.
In most cases eating at home implied eating in the kitchen or at a dining table.
There was, however, considerable variation, as this included nearly all home
eating in Finland, but only two thirds in Norway. The most frequent alternative was a coffee table/sofa. Very little eating took place at other sites (at a
work desk, in bed etc.) or without sitting down at all. The duration of the eating events varied considerably, probably indicating that events have very variable status. Many eating events were of brief duration, 10-20 minutes, but
the typical duration of eating events also varied greatly between the four
countries, with a tendency towards very brief events (less than 10 minutes) in
Finland and longer events in Denmark.
It was common to engage in other activities while eating, such as watching
television, listening to the radio or reading. About one third of the populations
had done so in relation to all eating events. Watching television while eating
was relatively evenly distributed among different population groups, and this
practice as well as listening to the radio had occurred independently of
whether the eating event was an individual or a social one. Relatively more
people among the younger age groups had not engaged in any such activities
in conjunction with eating events.
Eating at work most often took place in a canteen or separate eating room,
about one fourth to one third at the work desk. Most of those who had eaten at
their place of work on the previous day had done so on at least one occasion
in a canteen or other room that was devoted to this purpose. Eating at a café or
restaurant during work hours is not included here (a practice that was slightly
more frequent in Sweden).
Eating at other places than at work or at home did occur, but not as a dominant feature. Twenty-five per cent had done so on the day before in Denmark
and Norway and 32% in Sweden, including eating in other people’s homes or
at cafés/restaurants. Eating in places like at fast-food outlets or while walking
along the pavement was negligible. A frequency question about eating out (in
restaurants, cafés etc.) confirmed this picture, as only about ten per cent ate
out on a weekly basis, the highest proportion being in Sweden, the lowest in
18
Eating Patterns
Denmark. No significant differences between men and women emerged in
any of the countries. In Norway and Sweden, eating at ‘other places’ occurred
to a greater extent among younger age groups. There were no significant differences between individuals living in different types of households. Contrary
to expectations, our data indicate that individuals who live alone do not tend
to eat at ‘other places’ to a greater extent than those whose households comprise two or more persons.
Eating together with the family or alone?
The social context of eating includes not only the location, but also the company. We found that the eating system included a mixture of events taking
place alone, with other household members, with colleagues and with friends.
A considerable proportion of breakfast eating took place alone, even in nonsingle households. Lunch, however, often took place with family or colleagues. For those who lived with a family, dinner was usually shared with
them. This was the case also for the cold eating event later in the evening.
Eating with friends and others constituted a minor proportion at all hours.
Peaks indicating social eating events were generally higher than those indicating individual events. This suggests that the more ‘important’ eating events,
i.e., events which took place at the most popular/conventional eating hours in
each country, were more often social events. Individual eating events tended
to be somewhat more evenly distributed throughout the day, but it is also clear
that many individual eating events also followed the conventional eating
rhythm.
According to qualitative studies, family meals have great significance and
meaning as symbols of family cohesiveness. Our analysis demonstrated that
more than half (between 54 and 64%) of those living in multiple person
households had at least one hot meal together with the assembled household.
However, although family cohesion may be confirmed through shared meals,
the practice occurred on an every-other-day basis rather than a daily basis.
This was the case both in households with children and in households without
children. Family meals were more frequent among people in the older age
groups and least frequent among the youngest. Aside from that, our data provide no firm basis for drawing conclusions about the future prospects of family meals.
On Sundays, when meals are usually not structured by work hours, family
eating dominated, even for a considerable proportion of those who lived
alone. The Danes had eaten with other family members more often, compared
to the other countries, the Swedes slightly more often alone. Seventy-one per
cent of the Danish sample had had such eating events on the day preceding the
Summary
19
interview, while 67% in Norway and 63% in Sweden had done so. About one
third had had no eating events with family members. This proportion was reduced to about ten per cent in households including at least two adults. Thus
in most cases the reason for always eating alone was that the respondent lived
alone. In Norway, relatively more women than men had eaten in the company
of family members.
Eating in the company of colleagues occurred on weekdays during the day,
but rarely outside regular working hours. Differences between population
groups were large, and varied primarily according to occupational status. Between 58% and 68% of the occupationally active had eaten with colleagues.
Eating in the company of friends or others appeared to play a very small part
in the general picture, indicating that such events were not a frequent everyday habit in the general population and first and foremost a leisure-time activity. The tendency to do so was somewhat more common during the weekend
than on weekdays. The youngest age group (15-24 years) had eaten with
friends to a significantly greater extent than any other age group. Approximately half of the people belonging to this age group had done so. Single
adults had eaten with friends to a greater extent than had other adults whether
or not they lived alone or in households with children.
Few individual eating events outside the home
The two main dimensions of the social context, location and company, have
been combined, distinguishing between events at home and outside the home,
and between individual and social events, respectively. While social events at
home dominated in Denmark, there were slightly more individual eating
events at home in Norway and Sweden. Third in all countries were social
events outside the home, while individual events outside the home constituted
3% to 5%(eating alone at work, a café or a fast-food outlet). Social eating
events lasted considerably longer than the individual ones, particularly compared to individual events outside the home.
In Norway and Sweden, when eating took place at home, slightly more people
had eaten alone than in the company of others. The opposite pattern was obtained in Denmark where eating events at home more often tended to be social
events. Eating outside the home was much more often a social event than an
individual one, and this was the case in each of the four countries in the study.
The location and company in which eating events took place varied between
population groups. Some of these variations follow patterns that might be expected, given the varying living conditions among population groups. Thus,
20
Eating Patterns
those who were living alone tended to have eaten alone to a greater extent
than others, while those who were employed or following a course of education had to a greater extent than others eaten at these locations and in the
company of colleagues or business contacts, etc. These variations appear to be
straightforward reflections of variations in material living conditions. However, those who were living alone did not always eat alone, those living with
family members did not always eat with family members, and not all of those
who were employed had eaten in the company of colleagues or professional
contacts.
There were other variations that were less straightforward but nevertheless
constituted systematic differences: women had more eating events at home
and with family members than had men; men had more eating events outside
the home and with colleagues. Younger age groups had eaten at other places
and with friends to a greater extent; older age groups had eaten at home to a
greater extent. The occupationally active had not only eaten at a place of work
to a greater extent than had others, they had also eaten at other places to a
greater extent as well as in the company of family members, colleagues and
friends. People living in households with one or more other adults had eaten
at home, with family members and with colleagues to a greater extent; single
persons had eaten at the workplace, with colleagues and with friends to a
greater extent. Those with less education had eaten at home to a greater extent, whereas those with further or academic education had more often eaten
at other places and with friends.
Towards eating in social isolation?
Many of the general features of eating patterns were found to be similar in the
countries included in our analysis. This is the case regarding the significance
of home-based eating patterns and the tendency to eat either in the company
of family members or alone, but some significantly different national variations have also been identified. In Denmark, the home and family tend to dominate the picture even more than in the other Nordic countries, and eating
events are typically of longer duration than in the other countries. In Finland,
eating events tend to be of considerably shorter duration. In Norway, eating
while seated in a sofa and while watching television was more widespread
than in other Nordic countries. In Sweden, it is more common to eat out in a
restaurant or café and to have more eating events in the company of other
people during the course of the working day.
These results do not confirm the views put forward in the scientific literature
which suggest that the sociality of eating practices has disappeared and that
eating events are no longer significant events in their own right. On the other
Summary
21
hand, neither do they indicate that all members of the national populations at
issue have an eating pattern that is social in character, in the sense that eating
events include the participation of other people.
Nevertheless, these data also document that when eating takes place in social
isolation, it seems to lose some of the structuring features: such eating events
are of shorter duration than eating events that are shared in the company of
others, and are more frequently undertaken while other activities are occurring
at the same time. Some people reported solely individual eating events, but
the majority of the national populations reported a pattern that included both
individual and social eating events. As long as modern individuals live in
households with other people, and as long as they leave their homes to go to
work or to their place of education, it would seem most likely that they will
continue to engage in eating together with the people around them – be it at
the workplace or in the home. The meal as a social event therefore constitutes
and in all likelihood will continue to constitute a significant element in modern living. The shared patterns are to a large degree retained even among
those who live alone and who are not occupationally active, emphasising the
large impact of these institutions not only on the organisation of everyday life
but also on conventions and cultural norms.
Cooking and gender: a contested issue
The question: "Who did the cooking?" was posed to the respondents of the
survey after they had answered questions about what hot food they had eaten
at home the day before. What seemed to be simple, however, turned out to be
quite complicated.
The results of this study show that cooking is still women's work to a great
degree. Certain changes, however, may be observed which indicate a more
equal sharing of the work involved, most visibly among the youngest and
among those with middle and higher occupational status. However, in all four
countries there was a major inconsistency between men’s and women’s reports of who did the cooking. This may be presented as a gross exaggeration
by men and women of their own contribution, a well-known finding for questions of this type. Yet, the size of the inconsistency, as well as the systematic
variation between social groups, indicates that more substantial explanations
may be required. One topic addressed here is differences according to gender
in the understanding and evaluation of cooking as part of housework.
Among men, Norwegians reported the largest contribution in cooking. According to this criterion, they appeared to be the most modern of the men,
followed by Swedes. However, the self-reporting by men and women differed
22
Eating Patterns
most in Norway. Most traditional, according to the same measures, were the
Finns, both men and women. When it comes to the most modern women in
terms of equal contribution in cooking between the genders, they were to be
found among the youngest in Denmark, Finland and Sweden and among those
with high occupational status in Denmark and Finland.
Opinions about food safety: consumers want protective measures and individual freedom
Current debates about food and eating deal with more than just eating patterns. Much attention is directed towards food safety as a question of consumer protection. Consumer trust has emerged as a key issue that links food
policy and everyday life. As a supplement to the description of Nordic eating
patterns, this study included questions about attitudes towards food safety policies.
There are two classical political solutions that are expected to strengthen consumer safety and security. One is consumer protection, which implies public
regulation of the market. This solution is meant to prevent inferior and harmful products from being marketed. The other strategy has been operationalised
as the commonly expressed opposite to this solution, namely consumer information. By this means, the consumer becomes more capable of actively
choosing between different products and properties, in other words increasing
personal control and individual responsibility.
Our analysis indicates that many people in the Nordic countries think that information and public regulation should be supplementary rather than alternative food policy strategies. One interpretation is that consumers, on the one
hand, want public authorities to ensure a basic level of safety. But, on the
other hand, for their own feeling of security they still want information, education and openness in order to make possible some sort of individual control.
They want a considerable degree of freedom to choose – when this does not
threaten their basic safety. This appears to be a typical attitude among citizens
in and proponents of the Scandinavian, social democratic, welfare state.
Trust in public authorities seems to be a significant condition for confidence
in public information and the regulatory measures of the food system. While
information strategies leave more freedom to choose for the consumer, the use
of information has to be founded on trust in the sender (which in this case is
public authorities). On the other hand, distrust of market institutions, which is
quite common according to our data, is associated with a wish for public interference in the food system. National public authorities are important as trust-
Summary
23
building institutions in the Nordic countries, while scepticism towards both
the market and international institutions seems to be more widespread.
Norwegian consumers seemed somewhat more in favour of public regulations
and information than consumers in the other Nordic countries. More distrust
of market mechanisms and greater trust in public interference in the food
market partly explains why Norwegian consumers more often than Swedish,
Danish and Finnish consumers are in favour of public protective measures to
ensure food safety. However, we cannot claim that the level of trust explains
all differences between the Nordic countries.
Explanations for these national differences are discussed at three levels. Although good, comparative studies are lacking, it may be anticipated that the
experiences differ in terms of concrete food scandals. The second level is related to diverging food policies. Food safety policies are not that divergent in
the Nordic countries, at least in principle. However, again, while Norwegians
have generally been most eager to regulate the food market for various purposes, Danes, with their export-oriented agriculture, have been least willing.
Scepticism towards market solutions seems to be more widespread in Norway, expressing a general anti-liberal attitude towards the food market, while
this topic is more controversial in Denmark. We have seen in our study that
distrust in market actors is associated with a higher wish for information and
regulations. The third level of explanation is related to political trust in general. In studies that are not related to particular issues, Norwegians usually
show a high level of trust, while Finns and (perhaps also Swedes) represent
the other extreme, Denmark taking an intermediate position.
24
Eating Patterns
1
Introduction
Unni Kjærnes, Marianne P. Ekström, Jukka Gronow,
Lotte Holm and Johanna Mäkelä
This study is about how eating as an everyday activity is structured in modern
societies. We want to describe the present situation with regard to eating patterns and meals in the Nordic countries. Is it a state of status quo, where traditional meals dominate? Are we seeing tendencies of disruption? Or do current
eating patterns rather indicate new ways of structuring everyday life? What
are the impacts of contemporary family structures and gender relationships?
Our study seeks to identify general features of modern eating as well as national characteristics within a Nordic setting.
Food habits are eagerly debated these days, one question being the fate of
meals. Market researchers, academics within sociology and anthropology,
nutritionists, and the general public all contribute to the debate. The question
is whether traditional meal patterns and meal formats are being disrupted (cf.
(Murcott 1995). Meals are seen as the cornerstones of family life, and the underlying discussion is often whether this alleged disruption also indicates an
upheaval in family life and social life in general.
Contemporary food habits, therefore, are currently of great interest, and participants in the debate hold very diverging positions. However, food habits are
interesting not only because they receive public as well as individual attention
and concern. They can illustrate or throw light on a number of important sociological questions. Some of the classics within this field – Claude LéviStrauss and Roland Barthes – have called food “a key to the soul of the culture” and “the golden way to the unconscious of society" (Levi-Strauss 1964,
Barthes 1979, see also Fürst 1995). Food and meals are significant as expres-
26
Eating Patterns
sions and affirmations of identity and belonging. This is the case for national
and cultural identity, as well as for personal well-being and self esteem
(Fischler 1988, Murcott 1995). We are, literally, what we eat.
Contemporary eating habits as social practices are influenced by the wider
societal transformation processes that are the core of a long-standing discussion within sociology. The focus is on themes like reflexivity, individualisation, de-traditionalisation, globalisation and regionalisation as well as on gender, identity, class, and family life (e.g., (Ziehe 1989, Beck 1994). Sulkunen
states that “[E]ven in contemporary affluent societies ways of life and associated communities are not completely arbitrary and disorganised, but the
splintered structures of time (accumulation and progress) and space (nation
state and local community), as well as the inner-directedness of contemporary
culture, raise new theoretical and empirical problems in distinguishing historically relevant groupings.” (Sulkunen 1997, 10).
These ideas are also evident in studies of food. Fischler has suggested that we
are in an era of gastro-anomie where regular meals have become increasingly
rare and replaced by irregular eating patterns (Fischler 1988). By the concept
gastro-anomie Fischler refers to a tendency whereby cultural norms for what
should be eaten when and together with whom disappear. Following a collapse of traditional and authoritative external rules about what should be
eaten, the individual faces a splintered, uncertain and confused situation,
where, in the midst of conflicting advice, the individual is left alone, illprepared to make decisions about food consumption.
Fischler is not alone in proclaiming these processes of destructuration of eating, first of all as a consequence of individualisation. Mintz argues that meal
patterns are dissolving (Mintz 1996), while Falk describes a decline of ritual
meals and a growth of food industry (Falk 1994, 23). Though incompletely
investigated, says Burnett, it is highly likely that the meals that are held to be
the very stuff of sociality are in danger of disappearing (Burnett 1989).
Public debates in our countries have also addressed this question, described as
‘beiting’ in Norway1 and, in Sweden, ‘frukostisering’2 or ‘breakfastisation’,
indicating meals that are not cooked and that are selected according to the
fancy of the moment and eaten individually, often while continuing with other
1
‘Beiting’ is a direct translation of the American concept ‘grazing’, the use of which in connection with modern eating seems to have a rather diffuse origin (see chapter 5).
2
‘Frukostiserad’ was first mentioned by the Swedish polling bureau SIFO and the newspaper
Expressen in 1984 (Nyordsboken 2000). It was used in an article about children’s food in 1986
(Nyberg 1986).
Kjærnes et. al: Introduction
27
activities. While the statements put forward in public debates deal with what
people do, moral issues are clearly an underlying issue, but it is often quite
difficult to discern moral issues from practices in these debates.
Statements about a disruption of meals as a significant phenomenon presuppose a certain magnitude. There is a lack of empirical investigations of the
actual practices in modern societies, particularly with regard to the extent of
various eating patterns. Thus we have approached the question of meals by
focusing on eating practices as concretely as possible, rather than on attitudes
or conceptions of ‘normalcy’. Moreover, we have sought to identify general
features which are representative for the populations in each country.
1.1
Studies of meals
1.1.1 Conventions and norms for meals
A number of studies which have focused on meals and the sociality of eating
have given considerable attention to norms and conventions. . Such norms and
conventions are not only related to cuisines in a narrow sense. They include,
as indicated by Murcott in the British research program “The Nation’s diet”,
an array of “cultural conventions governing the micro-politics of social relationships in the food-sharing group” (Murcott 1998, 150). Food, and what we
do to and with it, is proclaimed to lie at the very core of sociality: it signifies
“togetherness” (Murcott et al 1992, 115). It is not only special meals for particular groups that can be approached in this fashion, but also the everyday,
routine meals of home, school or workplace, revealing the symbolic significance of eating in everyday life.
The question “What is a meal?” can obviously be answered from different
perspectives (cf. Mäkelä 1991, see also chapter 4 in this report). A classical
and influential approach to defining meals was worked out by Douglas. Applying a structuralist method, she analysed the structure of daily meals by using classifications and categories of binary oppositions (Douglas 1975). She
identified the same three-partite structure everywhere, in minor eating events
as well as in the format of main meals (starters, main dish, dessert). Moreover,
Douglas and Nicod drew attention to the structural and sensory qualities embedded in the oppositions savoury/sweet, hot/cold, liquid/dry (Douglas and
Nicod 1974). These qualities were essential in defining different types of
meals. Other criteria were complexity, copiousness and ceremoniousness. A
28
Eating Patterns
meal is not, however, the only type of eating. According to Douglas and
Nicod, four terms describe the different forms of eating. A food event is an
occasion when food is eaten, a structured event is a social occasion organised
by rules concerning time, place and sequence of action. Food eaten as part of
a structured event is a meal. A meal observes the rules of combination and
sequence. A snack is an unstructured food event without any rules of combination and sequence. The hot, savoury course of a major meal has the basic
structure of a staple (potato), a centre or main dish (meat, fish or egg), trimmings (vegetables), and a dressing (gravy).
Quite a few studies have been added over the past decades that extend and
complement this model of eating (see also chapter 4 of this report). Many regard food choice as a matter of compromise between a number of individually
and socially determined concerns (Holm and Kildevang 1996a, Caplan et al
1998, 182, Warde and Martens 1998, 144). The ways in which we eat are influenced by a complex range of factors, including individual preferences and
experiences, and group and class membership, as well as how our everyday
life is structured (Ekström 1990).
Eating is not only a matter of food choice, but also a question of when and
where – of time and place. The commensality of eating means that we try to
coordinate our actions. As indicated by Simmel already in 1910, the importance of meals taken together will inevitably lead to temporal regularity
(Simmel 1957). To Simmel, the sociability of eating is related to the refinement of social forms of interaction (Gronow 1997). Rotenberg argues that a
meal is essentially a social affair, "a planned social interaction centered on
food" (Rotenberg 1981, 26). He shows how industrialisation changed time
allocation and the organisation of work and housing. This influenced eating
patterns and transformed the traditional five-meal pattern into a modern threemeal pattern where meals are shorter and less elaborate and fewer meals are
eaten in the family and among friends. Empirically, this is confirmed also
within a Nordic setting. Mäkelä (1995, 1996) found that for working mothers
in Southern Finland, the idea of sharing food with others proved to be essential for the concept of a meal. 3
In addition to the ‘what’, ‘where’ and ‘when’ of meals, people have food and
drink in between or as less conventionally framed alternatives. As Wood has
pointed out, a focus on meals may overlook the fact that, notwithstanding
3
In this group a ‘proper meal’ consisted of three factors: a hot dish, a salad and social company.
Kjærnes et. al: Introduction
29
their importance, meals do constitute only one of several parts of the domestic
food system (Wood 1995).
In spite of social change and conflicts at different levels, important elements
of meal conventions seem to be altered only very slowly (Caplan et al 1998).
One example is the division of labour between the genders, a division that
within a (North) European setting has proved to be deeply embedded in the
conception of eating as such, and – therefore – seems to be very persistent
despite rather comprehensive changes in gender roles in general. The concept
‘a proper meal’ encompasses the ideas of what a meal should contain in order
to be tasteful and healthy (Murcott 1982). The British proper meal consists of
roasted meat, potatoes, (boiled) vegetables, and gravy. A proper meal is to be
served by the wife, for her husband, symbolising the woman as a caregiver
and the husband as a breadwinner. Nordic studies confirm that such ideas are
recognised even within these countries (Jansson 1988, Holm 1990, Ekström
1991, Mäkelä 1995, Bugge and Døving 2000). Wandel et al. suggest that the
ideas of what a proper meal should be represent a kind of ‘ideological lag’ in
our food conventions (Wandel et al 1995).
The existence, and persistence, of such widely shared conventions may explain and shed light on important aspects of our eating practices. The familial
framing of proper eating probably shows both the strengths and the weaknesses of a research perspective concentrating on the existence of structuring
norms and regulations. On the one hand, it underlines the importance of such
structures of eating for commensality and people’s everyday lives. And it
helps us to understand how the absence of a female contribution and a cooked
dinner may “disrupt the social fabric” of domestic life (Wood 1995). On the
other hand, because of the concentration on norms and normality, actual everyday practices may be overseen or misinterpreted.
A recent Norwegian study of young single adults reveals that they all had retained clear norms (or ideals) of proper meals, fitting very well with Murcott’s
(and others’) description. However, their actual experiences when growing up
as well as their present situation often deviated considerably, both with regard
to family life and the incidence of these ideal meals (Guzman et al 2000).
What do these studies of meals tell us about current eating practices? Several
suggest that even though alterations in food preferences and cooking may take
place, neither normative regulations nor the role of commensality will necessarily disappear (Murcott 1998). Bugge and Døving posit that our meal patterns are continuously being transformed, but that these changes are not anywhere near as dramatic and unambiguous as common notions suggest – if
30
Eating Patterns
anything, the opposite is true. Assessed by the situation in Norway, the meal
as a social, moral and family institution is very strong, perhaps even stronger
than before (Bugge and Døving 2000).
However, Aymard et al. (1996) propose that a picture of stability can, at least
partly, be explained by the disciplinary approach adopted in the studies of
everyday eating: “The meal pattern used in European-type societies has been
examined up to this point from an anthropological rather than a sociological
position, which has led us to underline its coherence and possibly to overestimate its stability” (p. 168). Yet, from their French position, Aymard et al.
oppose the idea of a ‘crisis’ of household foodways and the traditional middle-class meal. Statistics from France indicate that while 87.2 per cent of the
meals were taken at home in 1987, the figures were 87.6 per cent in 1980 and
92.2 per cent in 1970. They state that a process of “destructuration”, where
regular meals are replaced by informal and irregular snacking, must be based
on hearsay and rumour (possibly promoted by the food industry), rather than
on actual data4.
While providing important knowledge about the social aspects of eating, studies like those of Aymard et al., Fischler and others, seem insufficient when
confronted with recent statements about an ongoing deterioration of modern
eating patterns. One important problem is that there are very few studies from
which it is possible to make generalisations with regard to the total population. Most of the empirical material about meals has been collected through
qualitative in-depth studies. Moreover, these studies usually focus on certain
meals, in most cases the hot evening meal, rather than studying the eating regime throughout the day.
There are some quantitative survey studies that have included questions about
meal frequency. These studies indicate a certain decline in how often people
eat ‘breakfast’ and ‘dinner’ (Roos and Prättälä 1997, Becker et al 1994,
Wandel et al 1995). First, the “decline” is small and very recent (we don’t
know what people did some decades ago). Second, it is rather problematic to
interpret these studies, because we know little about what people mean by a
‘breakfast’ or a ‘dinner’. Is a ‘dinner’ any food eaten at ‘dinner time’? Or is
food eaten at dinner time not perceived as ‘a dinner’ if the necessary elements
of a ‘proper dinner’ are not included? In other words, these studies do indicate
that eating habits and meals are dynamic. But it is difficult to distinguish be4
They do, however, point out the conflicting and even precarious mix of different social uses
of time that produced today’s standard. This compromise may eventually collapse under the
weight of increased economic pressure and transformations in the labour market, working conditions and work schedules. (Aymard et al 1996)
Kjærnes et. al: Introduction
31
tween the possibilities that routines and conventions are being displaced by
more individualised and arbitrary eating practices (gastro-anomie), on the one
hand, and the emergence of restructured patterns and/or changing conceptualisations, on the other.
1.1.2 Everyday eating
Food represents an intersection between society and the private sphere, the
collective and the individual. Eating is linked to the organisation of society
and social life in a number of ways. It is a question of when to eat, where and
with whom. And eating is not only a social event in itself. The food to be
eaten has to be acquired and prepared in one way or the other, and the eating
is influenced by how this is solved, i.e., the system of food production, distribution and preparation. This study will focus mainly on the conditions for eating and meals, less on the processes and relations associated with food preparation.
Eating takes place as an integrated part of everyday life and thereby contributes to ordering our days into segments: morning, midday, afternoon, evening. By studying this order and rhythm of meals, that is meal patterns, we
study the intersections between the public sphere of production and the private sphere of reproduction, of family and recreation. Aymard, Grignon and
Sabban have stated that “food practices make the greatest contribution to the
structuring of social time and . . . these practices are in turn strongly influenced by the place reserved for them in daily routines as well as by the role
they play in the organization of the latter” (Aymard et al 1996). In this way,
they direct our attention towards the organisation of schedules, the particular
modes in which food preparation and meals interchange with work and other
activities, as part of cyclical calendars as well as throughout the day. These
patterns change over time, but they have, say Aymard et al., a tendency towards regularity in European culture. While we have mentioned above how
these authors have criticised an overestimation of stability and coherence, the
statements about daily schedules may suffer from a similar pitfall. The question remains open whether contemporary eating events do divide the day (or
the year) as clearly as these authors presume (at least outside the French setting).
32
Eating Patterns
Warde takes a more open approach.5 He contends that socio-demographic influence must be considered, but that it is delimited by the situational nature of
food choice. Eating is influenced by our social background and experiences as
well as practical considerations of everyday life. With examples from a Finnish study, the social determinants have different implications in different settings (from (Toivonen 1994). The particular social contexts of eating out can
illustrate this. While a purchased lunch is common in many social groups, fast
food is typical for the young, and those eating in restaurants on other occasions are predominantly from the higher social classes. Different contexts,
therefore, are dominated by different sorts of people. Warde continues: “If
eating is the sum effect of many situated events, the sociologically appropriate
question is whether there is a social logic to the situations in which people
find themselves.” (Warde 1997). As a consequence, however, neither items
nor situations should be observed separately. Rather, the focus should be on
the sequence of situations and bundles of items. By doing so, Warde points to
the role of the organisation and institutionalisation of eating as part of everyday life. The organisation of eating depends on factors like the job and family
situation, as well as by social forces associated with events, like the frequency
of attendance at celebrations or the habit of taking Sunday lunch. We can add
the role of how the provision of food and meals is organised (through shops
for individual purchase, market-based restaurants and fast-food outlets or public catering).
A classical study by Rotenberg shows how eating patterns in Vienna shifted
from the early 1900s, throughout the interwar years, and up till the 1980s
(Rotenberg 1981). Around 1900, the first meal of the day was eaten at home,
the second at work or in a cafe, the third, main meal at home, the fourth again
in a cafe, and the fifth in the evening at home. Meals at home with the family
alternated with meals with colleagues or friends away from home throughout
the day. The condition for this pattern was that the workplace was located
quite close to home. The men did not have to travel far, and the women, who
often contributed by working in at their husband’s workshop, could also tend
to cooking during morning hours. By the 1930s, however, industrialisation
had produced a three-meal pattern, where the men did not go home for the
midday meal, and where socialising with friends took place mainly during
weekends. The family meal was relocated to the end of the working day.
5
Different theoretical positions may be influenced by a number of things, including also (in
this case) differences in conventions and practices related to food and meals in France and Britain, respectively.
Kjærnes et. al: Introduction
33
It should be noted that this is a story about male meals. The process of industrialisation, with its increased division between the public and the private
spheres, had gender implications. Public meals, at work or in cafes, were
mainly for men. Women and children stayed within the private domain, but
Rotenberg’s study implicitly illustrates how domestic life also adapted to the
overall changes.
Similar general shifts in meal patterns were identified in a Finnish study
(Prättälä and Helminen 1990). Urbanisation and industrialisation were found
to be accompanied by a move towards a three-meal pattern and fewer hot
meals.
Everyday life has changed considerably during the past two or three decades.
In particular, the family system and family life have been transformed (BeckGernsheim 1998). In the Nordic countries, most married women are now
working out of home part-time or full-time. The number of divorces has multiplied, and more people live alone. Only a small percentage of the population
have their main source of income from primary (or even manufacturing) industries, and the level of education has increased considerably (see also the
description of the Nordic countries later on in this chapter).
Much of the domestic production of food has been displaced by market and
public provision. Elements of food preparation have been transferred from the
kitchen to the food-processing industry. The number of catered meals is growing in many countries (served at work places, kindergartens, hospitals etc.,
and as ‘meals on wheels’).6 Fast-food outlets and restaurants are increasingly
seen as alternatives to home-cooked food, at least among advertisers and journalists.
Yet, the effect of this displacement should not be overestimated (Fine et al
1996). Food and eating are embedded in social relations, and these inner relations serve to modify and obstruct the process. As the studies mentioned
above show, eating and meals are important for family cohesion and the upbringing of children. This context of close relations does not only comprise
the nuclear family. Serving food at home is also important when relatives and
friends get together.
6
This is not the case in all countries. In Eastern Europe, for example, the number of catered
meals has been drastically reduced following the collapse of socialist governments and its institutions.
34
Eating Patterns
Furthermore, counter-tendencies can be identified, including a technological
development with considerable improvement of (and attention paid to) household appliances. Pointing in a similar direction, the increasing problem of
trust in the food system has led to a stronger emphasis, at least in certain
groups, on fresh, “natural” foods, foods that have not been tampered with,
neither genetically nor industrially (Holm and Kildevang 1996b, Kjærnes
1999). These problems are related not only to food and the complexity of industrial food production as such, but also to political decisions and regulations, so clearly illustrated by the BSE (’mad cow disease’) crisis. The political aspects are reflected in an increasing focus on the domestic origin of the
food in many countries (including the Nordic ones). To many people in the
Nordic countries, domestic food is more ‘safe’, more ‘trustworthy’, etc. (Berg
2000).
In the discussion about meal patterns, we find the distinction between private
and public meals particularly interesting, as it is associated with changes in
the family, but also with the organisation of the working day, the extent of
eating at restaurants and cafés, etc. The extent of private vs. public meals can
indicate the intersection between the public sphere of production and marketing, and the private sphere of family life. This distinction is far from clear-cut
(Fine 2000). In domestic cooking, a considerable part of food preparation and
cooking takes place outside the home, first of all by the food manufacturing
industry, but also extended to take-away outlets. Moreover, the ‘public’ solution may imply a service provided by an employer or public authorities. Alternatively, meals may be bought at market price at a hot-dog stand or in a
fine restaurant.
These various public-private solutions have very diverse implications with
regard to the structure of eating, both in terms of time schedules, commensality, the menu and the organisation and progress of the event as such.
A distinction can also be drawn between individual and social eating. As mentioned above, meals are defined by many authors as a social event, as eating
taking place together with other people. This collective aspect is strongly emphasised in the debate about the dissolution of meal patterns, where the assertion is that processes of individualisation and looser social ties are reflected in
(or promoted by) fewer meals eaten together, with a particular concern for the
family context.
By combining these two dimensions, the private-public and individual-social,
a four-field table can be drawn (Table 1-1). Some examples are used to illustrate the four types of eating. Together, the dimensions are characterised in
Kjærnes et. al: Introduction
35
this study as the ‘social context of eating’. This is further developed in chapters 5 and 6.
Table 1-1 Types of eating events
Private
Individual
Eating taking place alone in
the home
Social
Family meal at home, or a
meal with friends in a private
home
Public
Individual snack on the street,
or lunch at the work-bench
Individual meal at a restaurant
A meal at a restaurant with
company, or lunch with colleagues
1.1.3 Eating embedded in flexible conventions
The section on the idea of a ‘proper meal’ (1.2.1) focused on the cultural conventions of eating, suggesting that eating is strongly socially influenced in
terms of meal grammars and conventions of social interaction. They refer
mainly to relations (and power) within the family, to some degree also class
distinctions. Next, we pointed to the need to fit cooking and eating into our
daily timetables. This not only an individual challenge. The question of where
we are, when and with whom, is affected by the social organisation of society
as well as private (individual and family) priorities. Together, the organisation
of production, working hours and their flexibility, as well as business opening
hours and the timetables in schools and kindergartens create a framework for
the options that each and every one of us have for our daily eating practices.
Both these perspectives seem to predict the persistence of social regularities in
eating, thus contrasting and contradicting the thesis that traditions are increasingly being replaced by reflexive – or anomic - individual actions.
Individualisation means the disruption of strong social regulation. It is often
presented as implying more flexibility and freedom for the individual to
choose according to his or her tastes and preferences. However, many contributions – on both sides – have emphasised negative aspects of complete individualisation, in which each act of choice is reflexively considered (for example, Giddens 1994, Sulkunen 1997). A high degree of unpredictability and
dissolution is extremely impractical within an everyday context, and for the
individual it may create a certain basic uncertainty, even anxiety.
However, say Gronow and Warde, increasing conventionality and individuality are not necessarily social opposites (Gronow and Warde 2001). With the
increasing complexity of modern societies we need both flexibility and daily
Eating Patterns
36
routines. This is supported by Campbell (1996, 149), who contends that life in
modern societies can at the same time become de-traditionalized and more
habitual.
Continuing this line of argument, Gronow and Warde suggest that ‘conventions’ can capture the theoretically and empirically founded persistent social
regularities, while at the same time not ruling out flexibility and individual
freedom. Building on Max Weber, they distinguish conventions from repetitive customs, on the one hand, and externally determined laws on the other.
Conventions are both socially shared and open for modifications by the individual and smaller social groups. Conventions are often taken for granted, but
they can also be subject to reflexive considerations. Conventions are therefore
not necessarily a continuation of tradition; conventions can emerge or change
according to the conditions of contemporary society. Still, this individual
flexibility and general openness for change does not mean that conventions
are devoid of normative considerations. While not being controlling in any
harsh way, sanctions can appear in the form of social approval and disapproval.
From this, we may expect the existence of routines at the individual level and
certain social regularities. Two questions may therefore be distinguished
(Gronow and Warde 2001: 230). First, there is the degree of formalization of
individual lives, a leniency-strictness dimension. Second, there is the degree
of co-ordination within social groups or societies, a variety-uniformity dimension. Both dimensions are relevant in the discussion of disrupted meals.
1.2
The aim of the study
In order contribute to the debate about meals and contemporary eating patterns, this project will analyse differences and similarities not only in food
habits but also their role in modern everyday life. To be able to do this, our
study must be open to both differing cultural classifications and changing
practices. The very first questions in attempting to operationalise and construct a survey are 'What is a meal?' and 'Are there any meals?'
Ethnological and sociological, qualitative studies carried out in the Nordic
countries indicate that in spite of considerable structural and cultural change,
meal patterns are surprisingly stable (see above and later chapters). When
asked in survey studies how often they have breakfast, dinner, etc., most people seem to adhere to the general norms on most days. However, we do not
Kjærnes et. al: Introduction
37
know what this means in terms of contents, commensality, preparation, or the
role of the home and the family for the general population or various population groups.
With this study, we want to improve the understanding of food from a sociocultural perspective. Most of our eating is part of the trivial, non-verbal routines of everyday life, routines that can be hard to identify within a relatively
stable and homogenous culture. A comparative design is therefore applied in
order to make cultural features more clear, identifying those that are shared
and those that are distinct from one country to the next.
Our main aim is to analyse the structure of eating as an integrated and dynamic part of modern everyday life. We ask about contemporary practices.
This question may be analysed at two levels. First, there are the contradictory
assumptions about the structuration of modern eating, as characterised either
by continuity and clear norms and patterns, or by tendencies of dissolution
and anomie. Our main focus will be on the existence of structures for the individual, i.e., the existence of ‘meals’ (as opposed to ‘grazing’) – leniency vs
strictness – on the one hand, and on culturally and socially shared patterns and
formats of eating – variety vs uniformity – on the other.
Second, there are the contradictory assumptions of eating patterns as a matter
of individual choice of lifestyle, or as strongly influenced by social structure
and societal institutions. Our main concern here will be the influence of social
institutions and organisation with direct relevance for everyday routines (organisational background), such as the organisation of the working day, the
type of family and household, and the place of residence. Moreover, we want
to investigate social background factors that may have more indirect influence, through cultural norms and classifications, such as education, generation
and gender. Social stratification variables (class, income) may have an impact
in both these respects.
The country of residence is of relevance both for what we may call the ‘institutionalisation of eating’, i.e.,, various forms of food provision and food policy, and through more general features of social structure and cultural norms.
Our understanding of the institutionalisation of eating as well as cultural
norms can only be conclusive by including even more types of data. The impact of various types of social background data can, however, provide indications of such explanations.
The first level leads us to a set of four questions about the structuration of eating per se. Theoretically, this level relates to the assertion made by Fischler
38
Eating Patterns
and by others, that contemporary eating is dominated by individualisation and
anomie. The comparative design first of all helps us to identify distinct patterns vs. degrees of variation vs. dissolution at the individual level. The question of culturally determined conventions and structure vs. individualisation
and destructuration can be analysed from several perspectives, including the
composition of meals and selection of foods, the daily patterns and the sociality of eating. The following questions will be discussed:
1. To what degree is eating ‘structured’ in the Nordic countries in terms of
commonly shared formats and patterns, and, vice versa, to what degree
does eating appear as anomic or “random” in these respects?
2. To what degree is eating ‘structured’ in the sense that eating is the main
purpose of the event (at a dinner table, no other activities going on simultaneously) in the various Nordic countries?
3. To what degree are eating practices individualised - in general and within
the family context, with regard to the actual situation of eating (alone or
together with others)?
The second level directs attention towards possible associations between eating patterns and social structure. Theoretically, these questions build on contributions which emphasise the impact of social structure in various ways,
including the organisation of work, production and everyday life (Rotenberg
1981). Next, there are the cultural norms and conventions with regard to gendered division of labour, as well as eating and meals as such. The following
questions arise:
4. To what extent are eating routines influenced by general structural features, such as work and family types?
5. Are there nation-specific characteristics which may be related to particular organisational solutions and/or cultural norms? Are there social characteristics that cross national borders?
6. How is the social division of labour with regard to serving food distributed between family/home, work and market institutions (restaurants, cafes, take-away etc.) in the Nordic countries? What is the impact of gender
roles for the division of labour within the family?
Kjærnes et. al: Introduction
1.3
39
The conceptual framework
This introduction has pointed out that most studies of meals are qualitative,
in-depth studies of a small group of people (in most cases based on female
informants). Our research questions require an empirical data base that can be
generalised, i.e., a quantitative design. A challenge in this study has therefore
been to formulate valid and reliable questions that could reveal the structure
of meals within a quantitative and comparative setting. Even though Denmark, Finland, Sweden and Norway in many ways share a common Nordic
cultural background, there are significant differences in food habits as well as
in the vocabulary for eating. The very idea of a comparative perspective contributes to the recognition of differences and similarities in food habits and
stages of the modernisation process in each country.
All this indicated a need for a conceptual model that on the one hand could
catch the degree of structure of various aspects of eating – without presupposing particular structures and conventions. On the other hand, we also needed a
model that could be operationalised in a telephone questionnaire to be used in
different countries.
1.3.1 The eating system
Although familiar with various conceptualisations of eating practices (partly
dealing with the British working class in the 1970s), our preparatory work on
the research questions and their suitable operationalisations involved intriguing discussions about the definition and essence of a meal within the Nordic
setting. We soon realised that 'the Norwegian and Danish dinner meal' does
not necessarily exist in Sweden and Finland, where a hot lunch is far more
common than in Norway or Denmark. Dinner as 'a hot meal one eats after
working hours' could not be taken for granted. Neither would the main meal
of the day imply the same thing in each country. This awareness made us realise that we could not simply use concepts like 'dinner', 'lunch' and 'main meal'
implicitly meaning hot meals at certain hours in our questionnaire. Rather, we
had to strip such culturally-laden concepts of their nation- or culture-specific
garments. In our deconstruction process we decided to ask for ‘eating events’,
i.e., every occasion of eating something (for practical reasons, just a beverage
or only eating chewing gum or sweets, etc, were excluded). Thus, we have
broadened the above-mentioned concept of a ‘food event’ (Douglas and Nicod
1974) to use as our starting point a very inclusive concept, taking account of
various aspects of the eating system. The analysis of the empirical data would
Eating Patterns
40
then involve a reconstruction of what could be called meals, dinners, lunches
or snacks. This way we hope to avoid making false generalisations on the basis of cultural, national or ideological prejudices or “taken-for-grantedness” –
i.e., ortodoxography (Bourdieu 1986).
The reconstruction is based on a model of what we call ‘the eating system’.
The model distinguishes between three dimensions of the eating system: eating pattern, meal format, and the social organisation of eating (figure 1-1).
Thus, the eating system includes eating events, with their particular composition of foods and dishes and their social context, and the way in which various
types of eating events are patterned with regard to chronology, order, etc.
Eating pattern
When?
Social context of
eating events
Where? With whom?
Meal format
What?
Figure 1-1 The eating system
First, we are interested in the schedule of eating events during 24 hours. In our
study, a distinction is made between cold and hot eating events. Anticipating
that hot eating events, in many cases, are more elaborate and socially structured, cold events may be recorded in a more simple manner, even though, at
least sometimes, cold eating events can be elaborately structured, too. The
eating pattern is defined by three different elements: time (the rhythm of eating events), the number of eating events, and the alternations of hot and cold
eating events.
Secondly, the (hot) meal format takes into account both the composition of
the main course (i.e., the centre and trimmings) and the sequence of the whole
meal (i.e., a starter, the main course, dessert). The types of meals are also of
Kjærnes et. al: Introduction
41
interest, particularly variations with regard to the degree of elaboration (from
simple cold snacks/meals to elaborate meals with many courses).
The third dimension of analysis includes aspects of where and with whom
people are eating, and who did the cooking, to define the social organisation
of eating, or the social context. At this level, we also include the duration of
the eating event, and whether other activities were going on simultaneously
(watching TV, reading a newspaper, working at a desk etc.).
Together these three dimensions in our model describe the content or format
and the schedule of eating, the place and the social context and organisation of
eating. These three dimensions do not form a hierarchy. In principle each of
them could be explored separately. Yet, as part of the model, each dimension
has a different perspective to offer.
1.3.2 The analytical model
The way eating is patterned in a society is of course influenced by a complex
interplay between numerous social and cultural forces. In our study we have
focused our interest on two main levels. One we call the level of societal production, a structural level dealing with the overall organisation in society, the
geographical allocation of housing and workplace, the distances between
them, the overall time rhythms of society, including the general organisation
and structure of the working day. Households are also included at this level,
being important for an understanding of the immediate social context for private reproduction (eating). Another level is the cultural level, dealing with
social norms and conventions. These are of course heavily influenced by the
factors included in structural level, but what we think of here are the social
conventions and norms about proper eating rhythms, proper meal formats,
how meals should be socially organized, where eating should take place, with
whom, and what should go on while eating.
Table 1-2 Factors influencing eating patterns
Societal organisation
(‘structure’)
Allocation of housing and work places
Time rhythms
Structure of labour force
Organization and structure of working day
Household types
Social norms and
conventions,
('culture’)
Eating rhythms
Meal format classification
Social organisation of eating
Eating Patterns
42
We will address these two levels of analysis in our study, where various
socio-demographic variables, together with the country of residence, are used
as background factors for differences in societal organisation and social norms
(Table 1-2).
Organisational background
Family/household
Work (type, hours)
Place of residence
ALLOC. OF HOUSING AND WORK
TIME RHYTHMS
STRUCTURE OF LABOUR FORCE
ORG. AND STRUCTURE OF THE
WORKING DAY
Yesterday's eating
Eating patterns
Meal formats
Social setting
Division of labour
Country of
residence
Cultural background
Social status
Education
Gender
Generation
EATING RHYTHMS
MEAL FORMAT NORMS
SOCIAL ORGANISATION OF
EATING
HOUSEHOLD DIVISION OF
LABOUR
Figure 1-2 The analytical model
The links between background factors and eating practices are illustrated in
Figure 1-2. Within this model, the country of residence will involve structural
as well as cultural factors, some of which can only be identified through other
sources of information than a pre-coded quantitative survey (studies of institutional solutions, statistics, discourse analyses, etc.). Such structural factors
include the institutionalisation of lunch-serving outside the home, the provision of welfare services, the organisation of the food industry and retailing,
and domestic food production. Cultural factors include shared cultural norms
and classifications of food and eating in a country, as well as contemporary
public debates.
Kjærnes et. al: Introduction
1.4
43
Methodological design
1.4.1 A comparative design
Since food habits are embedded in the tacit, often self-evident routines of everyday life, the specific manifestations and adjustments of eating in a modern
society can be difficult to identify. We have chosen a comparative approach,
selecting countries based on the “most similar design” (Przeworski and Teune
1970, Tranøy 1993). The existence of many similarities, culturally as well as
politically, yields the possibility of isolating critical differences in terms of
food and eating.
Although at a very exploratory stage, market research data have indicated that
contemporary food cultures to a high degree follow national borders – in spite
of globalised media, increased food trade and more travelling (Askegaard and
Madsen 1995, Mennell 1985). One cannot ignore the possibility that the various contributions (most of them are not comparative) are strongly influenced
by the classifications and ideas that are prominent in their own country, perhaps also affected by conceptions typical of the class or social group that the
researcher belongs to.7 This means that caution must be exercised when trying
to draw conclusions on the structure of eating in modern everyday life.
Denmark, Finland, Sweden, and Norway are all relatively prosperous societies, with homogenous populations, ethnically and socio-economically. There
are strong historical and, in the case of Denmark, Norway and Sweden, language connections. All four countries are characterised by well-established
social democratic welfare policies with a heavy emphasis on public involvement. This is very evident in the field of family policy and public services.
The Nordic countries have been perceived as a political and cultural entity,
where freedom combined with solidarity has been a strong norm (Sørensen
and Stråth 1997).
Notwithstanding a number of similarities, there are evident differences between the countries that may influence food habits and eating practices, concerning both food industries and food markets, and the specific designs of
welfare policies, as well as cultural characteristics. Moreover, despite their
7
One might suspect that the focus on fashion and volatile lifestyles, where individualised public snacking is perceived as the up and coming trend, is inspired by the environment of the
trend analysts themselves, where people are young, urban, single and well off, with irregular
work patterns (Bugge and Døving 2000).
44
Eating Patterns
generally homogeneous populations, there are social and cultural variations
within the Nordic countries that can be used as a basis for discussions of varying degrees and manifestations of the “modernisation” of eating.
1.4.2 A quantitative survey
Quantitative data collection and representative samples. Our questions are
related to the extent of different modes of eating, i.e,. the occurrence of individualised, unpredictable eating practices vs. traditional meals vs. new patterns and structures. Quantitative surveys have mostly dealt with the frequency of the consumption of different types of foods or meals (“usually”,
during the past week, the past month, etc.). This means that interpretations
directed towards eating and meals as social events cannot be made. Foods and
dishes cannot be connected to when, where and with whom the food was consumed.
With the aim of overcoming these limitations, we have tried to record the
various aspects of eating events within the framework of a quantitative,
(mostly) pre-coded survey. Our point of departure has been the 24-hour recall
method of dietary research, but it has been modified considerably in order to
collect social data, rather than nutritional information.
A qualitative study will usually concentrate on particular groups or situations,
such as young families, teenagers or lunch at work. We wanted to be able to
make assumptions about the general population. Thus it has been important to
obtain samples that are representative with regard to the population in each
country, and to gather sufficient background data to be able to analyse variations between different groups. Representativity and background data are discussed in separate sections (for further details, see separate publication).
The samples were set at a size (1200 in each country) that would allow this
type of analysis, although not at a very detailed level.
1.4.3 The questionnaire
The questionnaires were, with a few exceptions, identical in all the countries.
We used one language only in each of the four countries. The translations
were made by the researchers themselves.
Kjærnes et. al: Introduction
45
Recording eating events. As mentioned above, the main structure of the questionnaire is based on a 24-hour recall. When asked chronologically and concretely, most people are able to make a reliable recall of what they did (and
ate) during the day before. Moreover, this approach allows a circumstantial
description of each event (when, what, where and with whom). The idea was
to record the succession of yesterday’s eating events, starting with the first
one, then the next, etc. Because of the type of data collection that we had chosen, we had to limit the number of eating events to a maximum of 10. As it
turned out, however, this limitation represented no problem, as no respondent
reached this maximum (see chapter 2).
The questions included in the record of an eating event were developed from
the model and operationalisations that have been described in section 1.4 and
Figure 1-3. Again, for practical reasons, we had to make certain simplifications or limitations. We did not want to make any association to particular
meals or types of meals, but we made a basic distinction between “hot” and
“cold” eating events. The former were supposed to be somewhat more elaborate that the latter ones, i.e., involving more cooking, and were therefore recorded in more detail with regard both to the foods and dishes and to the
preparation of the meal. The cold eating events included a range of different
types of eating events: events that are usually conceived as “meals”, even as
“proper meals”, as well as events organised around beverages (such as afternoon coffee or tea), and very simple snacks. The record of each additional
eating event therefore opened with the question: “Was the food hot or cold?”8
Pilot interviews revealed that the respondents had no problems understanding
this distinction between hot and cold, and it was therefore not explained unless the question was raised.
8
To avoid confusion, the interviewers were instructed that only the food was relevant here, not
the beverages.
Eating Patterns
46
Hour
hot food
cold food
main course
centre: meat fish veg. other
preparation
what (list)
supplements: potato veg. other
preparation
name of dish
first course
dessert
who prepared
beverages
where
with whom
Figure 1-3 The record of eating events (x 10)
The first eating event of the day (usually at “breakfast” time) proved an exception here, where pilot interviews showed that the hot/cold distinction did
not work. Hot and cold dishes were therefore recorded in a similar manner for
this event. Yet, because most of these first eating events were rather simple
and very uniform, involving very little cooking and few dishes, it was possible
to use a “cold event” format, expanded only with certain hot dishes, such as
porridge or eggs.
The time and format limitations of a telephone survey implied that we had to
define a lower limit of the eating events to be recorded. Events were omitted
when only a beverage was consumed (such as a cup of coffee) - or only one
small item of food, such as a piece of chocolate or an apple. Events where a
cup of coffee was taken together with a piece of chocolate were, however,
included. This limitation means that snacking has been somewhat underreported in our study. For further discussion of the number and types of eating
events, see chapter 2.
Kjærnes et. al: Introduction
47
Eating habits. There was no space for questions about the frequency of consuming certain types of food in our study. Simple questions were included
about whether yesterday was a normal or an unusual day in terms of food and
eating, and whether yesterday’s eating practices were influenced by particular
considerations (health, slimming, vegetarianism etc.). These questions were
meant mainly as checks in a discussion of generalisations based on our findings.
Moreover, in order to be able to compare with a British survey of eating out,
their questions about the frequency of meals eaten away from home (apart
from meals at work/school) were replicated (Warde and Martens 2000).
Attitudes. Our analytical focus has been directed mainly towards the association between eating practices, the organisation of everyday life, and social
structure. Yet, we also included a series of attitude questions in the questionnaire, directed towards norms and classifications of eating and meals, on the
one hand, and trust and governmental policies on the other. This report includes an analysis of the questions on trust and government policies (chapter
8).
Demographic and socio-economic background. As already mentioned, much
emphasis has been put on obtaining sufficient and comparable background
data. The following data were included:
- year of birth
- sex
- composition of household, number and age of children
- marital status
- education (respondent, respondent’s parents)
- occupational status (full-time, part-time, school, pensioner etc.)
- occupational type and level
- social class (based on respondent’s occupation; if no present occupation, on former – or partner’s occupation)
- annual gross income of the household
- place of residence.
The intention was, as in the rest of the questionnaire, to use identical questions in all four countries. Unfortunately, we found no former Nordic surveys
that included identical questions and response categories on sociodemographic background. Therefore, using mainly questions from the Swedish Statistical Bureau and a leading Norwegian market research and survey
institute (Markeds- og mediainstituttet) as templates, we had to formulate our
own questions and categories.
48
Eating Patterns
In most cases we succeeded in formulating identical, pre-coded questions. For
some types of questions, however, it was very hard to find pre-coded categories that were valid and applicable in all four countries. In some cases, we
formulated separate categories for one country (e.g., educational level in
Denmark). Income categories had to be related to the currency and income
distribution of each country. In other cases, we used open questions that were
coded by the bureau and recoded by the researchers afterwards (e.g., social
class based on occupation, region and urbanisation based on municipality or
postal code).
The background for and processes of coding and recoding background data
have been presented in detail in a separate publication (Kjærnes et al. 2000).
1.4.4 Data collection
The general design. We used Computer-Assisted Telephone Interviews
(CATI). Telephone interviews were chosen as a less expensive and less timeconsuming form of data collection, compared to personal interviews. The
CATI method also had some particular advantages in our case, where a postal
survey, and to some extent even personal interviews, would be unsatisfactory.
These advantages more than balance the shortcomings of this type of data collection, such as time limitations and consequences for the sample (you have to
be available by telephone).
First, as dietary practices imply weekly and seasonal variation, our recording
of the previous day meant that we had to control the exact day and date of the
interview. Interviews were to be evenly distributed throughout the week. This
could be accomplished most straightforwardly through telephone interviews.
Second, the aim of describing, in principle, one day of eating meant that the
interview period was to be the same in all four countries, and as well-defined
and brief as possible. In practice, two weeks were needed. Again, this could
be obtained most practically through telephone interviews. Third, the computer-based questionnaire was very flexible, allowing different “paths” and
“leaps”, restricted options and filters, as well as large numbers of response
alternatives when recording the foods eaten etc.. These options were utilised
extensively. Computer-assisted personal interviews would offer the same features, while this would be difficult with a self-administered questionnaire.
The duration of each interview was approximately 15 to 20 minutes, a maximum for this type of interview.
Kjærnes et. al: Introduction
49
The interviews were to be performed during the last weeks of April 1997.
This was a relatively “neutral” period, without particular events, feasts etc. in
any of the countries.9 With regard to the consumption of fruits and vegetables,
however, interpretations have to consider possible seasonal variations. Fortunately, also, no big crises emerged on the public agenda in any of the countries during this period with relevance to food issues. Such incidents might
have directed the respondents’ attention (in serious cases even consumption)
in special directions.
Conducting the interviews. A leading Norwegian market research institute
(MMI) was commissioned to conduct the interviews, under the management
of SIFO. Data collection was organised and co-ordinated by the MMI , and it
was carried out by them and their partners in the other countries.10 But
throughout the process, the researchers in each of the countries kept close
contact with the respective market research institutes, from the time the questionnaire was formulated, translated and tested, through the interviewing, and
to the final stages of coding open questions and transformation to the data
files that were to be delivered.
Instructions for interviewers were co-ordinated. This being a large study, interviewers were to be introduced to the study, its ideas and design, in advance.
Telephone interviews also allowed direct monitoring of the interviews. In addition, the researchers formulated a few written instructions related directly to
the questionnaire. These instructions included a brief introduction of the study
to the respondents, as well as interpretations and more precise explanations of
certain questions that proved unclear during the pilot interviews. This was
particularly relevant for interpretations of what should be recorded as eating
events, and what was to be omitted (see also discussion above).
After the interviews had been conducted and the open questions had been
coded (based on instructions from the researchers), data collected from the
participating countries was to be transformed into identical SPSS files.
In spite of all these efforts, there were inconsistencies in the final, computerbased questionnaires that have caused problems for our analyses. Being organised from Norway, the Norwegian design and data file were supposed to
9
Unfortunately, the Swedish data collection was delayed for a couple of weeks. But it was
emphasised that days with special national feasts or celebrations were to be omitted. Moreover,
due to cold weather that spring, the picnic and grill season had not yet started.
10
These are all private market research institutes: Gfk in Denmark and in Sweden, Taloustutkimus Oy in Finland.
50
Eating Patterns
constitute a norm to which the other countries were to adhere. Inconsistencies
include a few cases where response alternatives were not completely identical
or where filters did not work the way they were intended to. This was particularly problematic for the Finnish data set, the most serious consequence being
that Finland had to be omitted from some of the analyses (see chapter 4).
Problems with the Danish data set were first of all related to incompatibilities
in the final data files, causing a lot of work with re-coding. The main challenge concerning Sweden was, as mentioned above, the delay in the actual
data collection.
As mentioned, interviews were to be evenly distributed throughout all days of
the week. However, in Finland and Norway interviewing could not take place
on Sundays, and therefore Saturdays are only represented in the data from
Denmark and Sweden.
1.4.5 The sample
As already mentioned, the samples were to be representative for the population of the respective countries. The size of the sample was to be 1200 from
each country, recruited from the population 15 years of age and above (omnibus sample). The samples were drawn by a stratum lottery method, where the
strata consisted of persons listed in the telephone directories of the various
regions and other adult persons from the selected households.
If the selected person was not available, a new attempt was made when he/she
was expected to be at home. A minimum of 3 calls (often more) were made
for each person before the attempt to establish contact was given up.
The number of interviews to be conducted within each stratum was set in advance. The number was fixed so that the distribution of the total net sample
within the different strata was proportional to the population size 15 years and
above in each stratum (i.e., the region covered by the particular telephone directory).
The initial number of phone numbers drawn was higher than the designated
final number. Interviews were made randomly from this list. Calls within each
stratum were concluded when the designated number of interviews was
reached. Unused phone numbers on the list were not included in the gross
sample. This means that unsuccessful calls were replaced by new, randomly
selected phone numbers from the list (within each stratum).
Kjærnes et. al: Introduction
51
Different types of drop-outs were to be recorded. Table 1-3 shows that a substantial share of the calls were unsuccessful. Denmark represents an exception, particularly because of fewer refusals. The polling company has reported
that this is a normal response rate in Denmark in surveys carried out for research or governmental institutions. One reason for the high number of refusals in the other countries may be the long duration of the interview. 15-20
minutes may have seemed too long for some people. Another explanation may
be the focus on food and meals. On the one hand food is very trivial and ordinary, but, on the other hand, it may touch on very personal, sensitive matters.
With the strong moral elements of Nordic food discourse, some people may
have declined to “reveal” their concrete eating practices.
Table 1-3 Samples and response rates in the survey
11
Gross sample 1
12
Gross sample 2
13
Net sample
Final number of
respondents
Net response rate
Denmark
1900
1737
1494
Finland
3774
2388
2274
Norway
3480
3187
2347
Sweden
3569
3242
2413
Total
12723
10554
8528
1202
80%
1200
53%
1177
50%
1244
52%
4823
57%
1.4.6 Representativity
As the unit of analysis in our study in many cases is a country, the question of
representativity is important. In multi-variate analyses including sociodemographic variables, however, possible biases do not have the same implications, as weighting is part of the statistical technique in for example regression analysis.
Some discrepancies must be expected in this type of data collection. A separate publication (Kjærnes et al. 2000) has compared all background data from
our study with available statistical data. Because the categories are not quite
in accordance in all cases, reliable comparisons cannot always be made. Our
question about household members and our final recoded household variable
have no direct parallel in official statistics. Moreover, for comparative reasons
we used identical questions about educational levels (with some modifications
for Denmark). As government statistics reflect distinct national educational
11
The total number of used phone numbers.
Used phone numbers minus no contact (after repeated calls).
13
Contact has been made within target groups.
12
52
Eating Patterns
systems, such data are not immediately comparable – with other countries and
with our data. With the exception of gender, the divergences are not unacceptable.
The interviewers have reported that recruitment of male respondents was particularly difficult. This is not surprising, considering what we know about
gender differences with regard to interest in food issues. The recruitment
problem was reflected as a gender bias even in the final samples. Weighting
with regard to gender was needed in Norway and Sweden. However, a lower
response rate among men may also have resulted in a somewhat biased sample of men. As fewer women refused to participate, both women who are interested and those who are disinterested in food topics are probably included
in the sample. Among men, however, more among the disinterested may have
declined, thus giving an overrepresentation of men that are more interested in
food, compared to the overall male population.
There is a bias in the Finnish sample with regard to age. Due to a misunderstanding, an upper age limit of 75 was introduced. Only a few from the highest age groups have been included. However, as these age groups represent
only a very small part of the population, the bias almost disappears after recoding of the birth year variable into discrete age groups.
Typically, there are too few respondents from the younger age groups, while
people with a high educational level are over-represented. This is also the case
for two-parent families with children. These results are probably the outcome
of a combination of processes, by including people who are more easily available at home by phone and, on the other hand, well-educated and resourceful
people who are more receptive and open to this type of inquiry. However, as
far as comparisons with official statistics are possible, few discrepancies are
so great that weighting might be needed.
1.4.7 Data analyses
After the data files had been delivered, the initial work of recoding and making corrections for all four countries was performed at the National Institute
for Consumer Research in Norway. The aim was to produce identical files
that could form the basis for various types of analyses.
Background variables as well as the records of eating events were subject to a
considerable amount of recoding in this initial phase. The recoding of eating
Kjærnes et. al: Introduction
53
events has been described in paragraph 1.2 of this chapter, while background
variables have been described in a separate publication (Kjærnes et al. 2000).
A substantial task in this study has been to “reconstruct” the elaborate records
of eating, each including a large number of variables, into variables and categories that were theoretically intelligible and applicable in further analyses.
The ideas and operationalisation have already been described; the following
chapters also include more details on how this was accomplished.
Analyses of eating events are made in two, basically different ways. First, information about each eating event can be counted directly and added up
throughout the day, producing frequency variables, e.g., the number of eating
events that took place together with the family or with colleagues, the number
of eating events that took place at home, or the number of eating events consisting of sandwiches – or “proper” hot meals. The eating event is the analytical unit, and various types of information about each event can be combined
through bi-variate and correlation analyses.
Second, eating events can be categorised according to time intervals. This is a
much more laborious type of recoding and analysis, where time filters have to
be applied for all eating events for each respondent. However, as we are interested in meals as complex social events, as part of the everyday schedule,
much of the desired information can only be found in this way. Only in this
way can associations and interpretations be made with regard to commonly
used meal concepts. Moreover, as the individual respondent is the analytical
unit here, this is the type of recoding required for further analyses including
background variables.
As the initial number of variables in our study is very large, the aim being to
capture the structure and organisation of modern eating, statistical analyses
are to a large degree limited to comparisons of dependent variables between
the countries. These comparisons have been subject to significance testing
(student t test, χ2 – test). However, as mentioned, other bi-variate combinations, as well as correlation analyses have also been accounted for.
For important dependent variables, multivariate techniques have also been
applied, in order to capture variations between social groups within the total
sample and in each country. In most cases, multivariate regression analyses
(linear and logit regression) have been preferred. The concrete procedures of
these analyses will be described in more detail in each of the following chapters.
54
1.5
Eating Patterns
The Nordic countries: similarities and differences
Above we have identified various theses about the structure of eating and contemporary meal patterns. Very few of these contributions have given particular consideration to variations between countries with regard to traditions and
recent developments, social structure and cultural conventions. The comparative perspective has been important in the development of our conceptual
framework and the analytical model. We believe it is also important to provide some information on certain key characteristics that are similar and different in the selected countries. In the following, we will describe some important features with regard to social structure and welfare policies, food markets, food consumption levels, and nutrition policies. While not directly included in the analytical model, this context can be used as a frame of reference for the patterns and connections that are identified.
1.5.1 Social structure and welfare policies in the Nordic countries
With a total combined population of 24 million people, each of the Nordic
countries is relatively small. Sweden has the biggest population (8.8 million),
Norway has half that number (4.4 million), and Denmark and Finland have
5.3 and 5.1 million inhabitants, respectively. While Finland and Norway are
sparsely populated and population density is relatively low in Sweden as well,
the density is almost ten times higher in Denmark. Moreover, a larger proportion of the Danish population lives in big cities, in particular compared to
Finland and Norway.
The Nordic peoples are quite similar with regard to a number of demographic
and socio-economic indicators.14 Life expectancy is high and the elderly constitute a relatively large proportion of the populations. Nordic families comprise about two people on average. Around 30% of the households include
children – the highest proportion in Denmark, the lowest in Sweden. Birth
rates are relatively stable in Denmark and Norway (around 1.3 live births as
per cent of total population). In Sweden and Finland, the percentage of live
births is lower (1.0 and 1.2, respectively, in 1997) and declining.
Most immigrants in the Nordic countries come from one of the other Nordic
countries. The number of immigrants from Asia and Africa varies considerably, with the highest proportion living in Sweden, while the proportion in the
14
If not stated otherwise, the source in this paragraph is Nordic Statistical Yearbook (Feldbæk
1998).
Kjærnes et. al: Introduction
55
Finnish population is considerably lower. There are a considerable number of
Swedish-speaking Finns (about 7 per cent), while, the Sami people constitute
a small ethnic minority in Finland, Norway and Sweden. However, the general impression is that the Nordic countries constitute an area with relatively
high ethnic, linguistic and cultural homogeneity, although with a tendency
towards more diversity in recent decades.
Educational levels are high, but there is still considerable variation within the
region. Almost one-fourth of the people aged 25-64 have 13 years or more of
education (Sweden highest, Finland lowest). The percentage having 0-9 years
of education among men aged 25-64 years ranges from 34 in Denmark and
Finland to 28 in Sweden and 17 per cent in Norway. The proportions are quite
similar for women.
There are no direct accounts of the number of poor people in the various Nordic countries. The extent of absolute poverty is probably rather small in these
countries, perhaps somewhat higher in Finland. Relative poverty and economic inequalities have increased during the past decade, but the existence of
poor, marginalised groups is probably lower in the Nordic countries, compared to other European countries.
Figure 1-1 shows that the proportions of men and women aged 16-64 years
that are not in labour force is generally small, but there is some variation.
Denmark has the lowest proportion among men, Finland the highest. Among
women, the proportions are quite similar: about one-fourth of the total number
in these age groups. In Sweden the proportion is increasing among both men
and women, while the proportions are declining somewhat in all the other
countries. The unemployment rate is much higher in Finland compared to the
other countries (14 per cent vs. 4-8 per cent of men and 11 per cent vs. 3-5 per
cent of women aged 16-64 years). This means that employment (full-time or
part-time) is the norm even for most women with children. Only about onefifth of women aged 25-34 are not in the labour force, even less among
women aged 35-44 (12-15%).
Eating Patterns
56
100 %
90 %
80 %
70 %
60 %
50 %
40 %
30 %
20 %
10 %
0%
m en
wom en
Denm ark
m en
wom en
Finland
Em ployed
m en
wom en
Norway
Unem ployed
m en
wom en
Sweden
Not in labour force
Figure 1-4 Employment status among men and women in the Nordic countries (1996)
A relatively high proportion (one-third) are characteristically employed in
public services, while few are employed in primary industries. One-fourth are
employed in manufacturing and construction.
Universal access to social assistance and pensions constitutes an important
feature of the Nordic welfare states. The proportions depending on social
benefits as their sole income range from 11 per cent of the total population in
Finland to 5 per cent in Denmark.
Still, the ratio between public social services and cash benefits is higher in the
Nordic countries, compared to other Western countries. In particular, nonhealth and non-education services have expanded much more in the Nordic
countries, including childcare and home-help services (Esping-Andersen
1999). The number of children in day-care demonstrates this. There is some
variation, with Sweden topping the list and Finland at the bottom. While almost half of the Danish (48%) and 39% of the Swedish children under the age
of 2 are in day-care, the numbers are considerably lower in Norway and
Finland, 23 and 22 per cent, respectively. For pre-school children, the
percentages in day-care are much higher, ranging from 82 per cent in
Denmark to 63 per cent in Finland. For school children up to the age of 9-10,
special care programmes after school hours have been established.
Kjærnes et. al: Introduction
57
The Nordic welfare state characteristics have been discussed mainly in relation to their influence on the labour market, for example the extent and types
of female employment. However, it can be anticipated that they also have an
impact on consumption patterns, not only because catering can be included as
part of all public services, but also because they imply a certain institutionalisation of everyday life. Work, as well as day-care for preschool and school
children and services for the elderly, often means eating away from home during the day – frequently in an “institutionalised” setting. Eating within such
institutionalised settings receive variable levels of subsidies (from public authorities or employers).
1.5.2 The food markets
Nordic food distribution systems have changed rather dramatically over the
past two decades, first of all in terms of much higher concentration at the retail level (Dulsrud 1996). While discount stores have had a large impact, the
total outcome is probably that a larger variety of foods are offered. On the
other hand, as the supermarkets have a large degree of co-ordination of their
purchases, the selection has become much more uniform nationally. Still, the
Nordic food markets differ in several respects. In particular, the supply and
distribution of organic food varies considerably between these countries, in
terms of access, price levels and demand. While these types of foods have a
considerable market share and to some degree also represent alternative distribution channels in Denmark, this market is almost negligible in Norway.
An important aspect for the development of the respective food market and
market regulation is the country’s status with regard to EU membership.
While Denmark has been a member since 1972, Finland and Sweden joined
the EU in 1994. Norway is not a member, but the EEA agreement ensures that
the country is part of the internal market and has adopted the corresponding
market regulations.
Market structure influences supply in terms of availability, the product range,
price levels and quality. Food prices will have an impact on consumption patterns, both the absolute levels and relative to other foods. However, a comparison of price levels between countries is no simple task. Values of the different currencies, income levels in each country, price measurement procedures, and the costs of other important goods and services (such as housing)
may all influence the outcome.
Eating Patterns
58
While the relative prices of food have developed more or less parallel to the
total consumer price indices in Denmark and Norway, there was a clear drop
in food prices in Sweden and Finland in 1994/95. EU membership is an obvious reason for this, including a reduction of VAT on food. Comparison of
food prices in the four capitals (and Malmö) in 1994 and 1996 reveals a
somewhat different picture (Jacobsson et al 1996). On both occasions Oslo
clearly had the most expensive food. This is a prevailing feature, reflected for
all food groups. While the food in 1994 was considerably cheaper in Copenhagen, compared to all the other cities, this position had been taken over by
Helsinki in 1996. Malmö had a position in between Stockholm and Copenhagen.
A third approach to food prices is the proportion of household expenditure
spent on food. Again, the order is different, with Norwegians spending less on
food and Danes having the highest figures in 1995 and 96.
Table 1-4 Proportion of household expenditure spent on food (per cent)15
1990
Denmark
Finland
Norway
Sweden
1995
14.8
16.2
15.3
16.1
1996
14.4
14.2
13.6
14.4
14.0
13.4
13.0
13.4
The three indices point in different directions. Interpreted with some caution,
they seem to indicate that food purchase expenditure is influenced by various
factors, such as price levels, absolute and relative to other goods and services,
in relation to income levels – and to cultural features. The Norwegian position, with high and increasing food prices and a low proportion of expenditure
spent on food, can be interpreted both as a reflection of high incomes and of
low priority given to food. Steeply increasing housing expenditure must, however, also be considered. Similarly, the Danish position, with relatively low
and stable food prices and a larger proportion spent on food, compared to the
other countries, can be taken as an indication that higher priority is given to
food. The trend in Finland and Sweden is more directly in accordance with
decreasing food prices.
15
Sources: Denmark, Finland and Sweden: Eurostat Annual Report 1998/99. Norwegian data:
Consumption surveys, Statistics Norway.
Kjærnes et. al: Introduction
59
1.5.3 Food consumption
Data at the wholesale level (Table 1-5) show that food consumption patterns
vary considerably among the Nordic countries. While Danes favour meat,
eggs and vegetables, beer and wine, Norwegians prefer potatoes, fish and
fruits. However, Norwegians also top the list for soft drinks, salty snacks and
chocolate. The Finns have the highest consumption of milk and coffee – and
butter, while they seem to esteem neither wine nor soft drinks and snacks.
Swedish consumption levels take, in most cases, a position in the middle. The
only exceptions on this list are tomatoes and cheese (of which Swedes eat the
most) and potatoes (of which Swedes eat the least).
Table 1-5 Annual consumption of foods and beverages (wholesale level) in the Nordic
countries (1997) (kg per capita per year). Main source: FAO Food Balance Sheets
Denmark
Vegetables
- Tomatoes
Potatoes
Fruit excl. wine
- Apples excl. cider
Milk excl. butter
- Cheese
Butter
Eggs
Fish, seafood
Fish, freshwater
Meat
Poultry meat
Coffee
Beer, barley
Wine
16
Soft drinks (l/cap 1998)
17
Chocolate (bars/confectionary)
Snacks (chips, peanuts etc.)
18
1998
16
97.9
16.0
72.1
103.1
31.7
226.0
13.9
1.8
15.9
24.7
2.6
97.4
18.4
9.1
120.8
31.4
84
7.5
3.1
Finland
70.5
14.0
69.0
70.4
14.6
341.6
14.4
4.4
9.7
35.9
8.5
65.2
10.3
11.4
83.7
6.0
50
4.0
1.7
Norway
64.3
12.6
77.5
118.6
19.5
270.0
14.9
3.5
10.1
50.7
11.9
58.7
7.6
9.3
54.0
9.7
114
8.5
5.2
Sweden
71.9
17.8
64.9
99.8
23.0
339.0
16.0
4.3
11.6
27.5
3.1
67.9
9.0
10.0
63.8
12.2
67
5.1
3.0
Source: Bryggeriforeningen (Norway). Tall og fakta 2000 – Brus (sweetened soft drinks)
Source: CAOBISCO (the organisation for the European chocolate, confectionary and biscuit
industry)
18
Source: Estrella AB, Gothenburg (estimates made by Nilsen)
17
60
Eating Patterns
The absolute figures of per capita consumption at this level should, however,
be interpreted with care. Even though food balance sheets have been developed over the course of several decades, what is included in the categories can
be somewhat unclear, and the reliability of data may differ. Moreover, wholesale measurements should be distinguished from studies at the household or
individual level of consumption. Therefore, it is first of all the relative levels
that are of interest here.
A comparison of dietary studies from 15 European countries has revealed that
socio-demographic disparities in food habits19 also exist within the Nordic
region (Roos and Prättälä 1999). Among the Nordic countries, gender differences in food consumption were marked and quite similar. Women generally
ate more fruits and vegetables, while they ate less meat and added fat and they
drank less milk (except in Denmark). Cheese consumption patterns were less
unambiguous. There were disparities even with regard to the educational
level. Generally, the strongest effects of educational level were found in
Finland, where people with higher education ate more fruits, vegetables and
cheese, while they consumed less milk, meat and animal fat/butter. To the
degree that differences were found according to education in the other Nordic
countries, they pointed in a similar direction. The only exception was margarine/oils, where a positive association with high education was found in Finland, a negative in Denmark and no association in Norway (data lacking for
Sweden).
1.5.4 Food and nutrition policies in the Nordic countries
For a very long time the Nordic states have made efforts to influence eating
and food consumption patterns in accordance with norms about healthy eating. These efforts vary considerably among the countries, in terms of policy
goals and measures and also the emphasis and political will that has been put
into this sector of public policy. The perceptions of nutritional problems seem
to have been quite similar in the Nordic countries, but the formulation and
implementation of nutrition policy has differed considerably (Jul 1988,
Kjærnes 1985, Bolin 1993, Murcott and Prättälä 1993). The following section
will give a brief outline of the policies.20
19
It must be noted that food habits here mean individual habits, which should be clearly distinguished from the wholesale data in the table presented above.
20
This description draws mainly on (Kjærnes 1997).
Kjærnes et. al: Introduction
61
The elder generations of today were raised in an era when nutritional experts
were concerned about vitamins and energy. Educational campaigns told about
the value of protective foods, first of all milk. But many understood nutritional challenges also as a problem of poverty. This was particularly evident
in Norway and Sweden, where improved nutrition was associated with the
development of the welfare state and therefore an important task for public
policy in the 1930s and 1940s. But the orientations were quite dissimilar.
While attention was directed towards market regulation and coordination with
agricultural policy in Norway, the Swedish policy put much more emphasis
on the development of social services, in particular school lunches. There
seem to have been no ambitions to integrate welfare ideas into market policy
in Denmark and Finland. Thus the national nutrition policies were all characterised by educational strategies, but they were to varying degrees associated
with other measures and policies: market regulations in Norway, public service provision in Sweden, a strong association with agriculture in Finland, and
solely education as a rather weak policy in Denmark. These differences implied varying influence on people’s diets. But the institutions that were established also had an impact on later developments.
Nordic nutrition scientists acted on the newly discovered association between
dietary fat and cardiovascular disease (CHD) early on. New Nordic recommendations were published in 1968. Full-fat milk and butter were redefined
from healthy to unhealthy. However, while CHD was a major cause of early
death in the Nordic countries, a reformulation of nutritional advice was no
easy task. The tension between health and welfare on the one hand, and Nordic agriculture (in particular the dairy industry) on the other became evident.
Again, however, the problems and conflicts were handled quite differently.
Swedish authorities directed their efforts towards consumer education, and
quite early (1971) they launched a large-scale campaign on diet and physical
activity. Gradually, even catering menus were changed according to new dietary guidelines. Nutrition policy being an integrated part of agricultural policy
in Finland, new guidelines were much more difficult to establish. Parallel to
the struggles within the agricultural sector, however, new arrangements for
school lunches and preventive measures within primary health care were established. The strong association between welfare policy goals and agricultural policies in Norway also caused problems in this process of redefinition
and reorientation. A compromise was reached in 1975, however, when the
dietary guidelines were moderated and the main emphasis was put on market
regulation within the established system, giving a rather ambiguous impression of nutritional goals. It took another decade before the animal fat issue
emerged on the agenda in Denmark, where a nutrition policy was approved in
1984. A new nutrition unit was established at the Danish food authorities. The
62
Eating Patterns
main strategy was consumer education, but reference was made also to regulatory measures and catering.
Thus, the differences with regard to legitimacy and institutionalisation were
reflected even during this phase of redefining national goals. What has happened afterwards is that the emphasis on education and preventive measures
within primary health care seems to have increased in all the Nordic countries.
Towards the end of the 1990s, knowledge about diet and health had become
very widespread, even though the priority given to nutritional concerns may
vary both between and within the countries. In all the Nordic countries, however, changes had been observed in dietary habits that could be connected to
nutritional advice. The ratios of fat as a percentage of total energy intake have
become more similar in the Nordic countries over the past decade. The ratio
has been relatively stable in Norway and Sweden, in 1996 at 36.8 and 37.6 per
cent, respectively. The percentage has increased in Finland (38.6% in 1996)
and decreased from a very high level in Denmark (42.8% in 1996) (The European Commission 2000).
In recent years questions about food and health have been extended to include
a variety of issues, and the demands for political solutions have increased.
Structural changes in the food market, liberalisation and international integration have probably contributed to this mobilisation and a shift of focus towards regulatory measures. While educational campaigns are still important,
legal measures that can ensure sufficient access to product information receive
much more attention than before. Contrary to former episodes of mobilisation
on nutritional issues, these conflicts between consumers and the food industry
seem to have been more significant in Denmark. Still, as regulatory measures
within this field are developed by the European Union, the regulatory situation is becoming increasingly harmonised in these respects.
Welfare-related conflicts continue to vary among the Nordic countries. While
issues related to income distribution and social support are important in
Finland, connected to problems of poverty, this is not a major concern in the
other countries. Following the economic crisis in the early 1990s, there was
mobilisation connected to cutbacks in public services in Sweden, including
reduced subsidies of work lunches. In Norway, on the other hand, nutritional
issues have emerged linked to the development or improvement of public services, such as food in schools (there is no institution of school lunches; children bring sandwiches from home). But this has not caused any political conflict. In Denmark, where catering systems (and public involvement in them)
are quite similar to those in Norway, these issues seem to receive even less
attention.
Kjærnes et. al: Introduction
1.6
63
The organisation of the report
This report will describe the various aspects of eating events, the patterns,
formats and social contexts. The foods eaten are presented in Chapter 2,
where Mäkelä and Kjærnes have looked for the contents of hot and cold eating events. The typical national features are focused. Chapter 3 deals with the
daily time schedules of eating. Gronow and Jääskeläinen discuss the rhythm
of eating, when and how many hot and cold eating events during the day.
Chapter 4 again turns to the food eaten, where Mäkelä looks at the meal format in a more abstract manner. The chapter includes an analysis of the role of
the ‘proper meal’ format. In Chapter 5, Holm gives a broad presentation of
the social context of eating, including where eating takes place, with whom
and what goes on while eating. Individual vs. social eating is combined with
public vs. private eating. In Chapter 6, Holm continues with a special focus
on the role of the family meal, that is a hot meal where all household members
take part. Meals are not only about eating, but also cooking. In Chapter 7,
Ekström and L’orange Fürst has analysed the division of this work between
the genders, focussing particularly on what might be taken as conflicting
views on what cooking is about. Contemporary debates about food do not
only deal with the fate of meals, but also uncertainty related to food safety.
This is discussed by Berg and Kjærnes in Chapter 8, where the roles of public
and market institutions in relation to consumer trust and distrust are analysed.
In Chapter 9, Kjærnes tries to bring findings about all these various dimensions of eating together, in order discuss the overall questions posed in this
introductory chapter.
64
Eating Patterns
2
What Did They Eat?
Johanna Mäkelä, Unni Kjærnes
and Marianne Pipping Ekström
2.1
Introduction
The aim of this chapter is to describe the contents of cold and hot eating
events as recorded during the telephone interviews. This will provide the basis
for the analyses presented in other chapters that further develop our questions
of the structuration of modern eating. Yet, the contents of cold and hot eating
events interestingly depicts the grass-roots level of people’s everyday food
habits. It tells us what Scandinavian people are typically eating nowadays. We
will identify characteristics shared by all or a group of countries as well as
nation-specific features. The variation between weekdays and Sundays will
also be explored. In our system of eating (defined in section 1.4.1), the contents of eating events is part of the meal format, which will be analysed in
terms of the structure and the complexity of eating events in chapter 4.
2.2
Data and recoding
In the interviews each loop of recording started with a question about whether
the informant had eaten hot or cold food. The first eating event was always
seen as a breakfast, without any distinction between hot and cold. Accordingly, the contents of three types of eating events will be presented in this
chapter: breakfast, cold and hot eating events. A detailed description of data
collection and the questionnaire is given in the introduction in section 1.5.
Eating Patterns
66
The original variables were recoded in order to diminish the number of categories. Table 2-1 and Table 2-2 present the recoding of breakfast and cold
eating events. We wanted to provide a rather thorough description; consequently the reduction of categories is not very drastic. It is based partly on
logical combinations, partly on the frequencies of the original variables.
Table 2-1 The recoding of foods at breakfast
New category
Components
Bread
Breakfast cereals
Egg
Cold cuts
Fruit
Porridge
Sweet pastries
Yoghurt
Other
Bread alone, sandwiches, baguette, pies, crackers
Cold cuts, fish, bacon, cheese
Fruits, berries
Sweet pastries, cakes, sweet biscuits
Sweets, salty snacks, ice cream, vegetables, soup
Table 2-2 The recoding of cold eating events
New category
Components
Bread
Cold cuts, cheese
Fruits, berries
Salads, vegetables
Sweet pastries
Sweets, salty snacks
Other
Bread alone, sandwiches, baguette, pies, crackers
Cold cuts, cheese, bacon, fish
Sweet pastries, cakes, sweet biscuits
Sweets, salty snacks, ice cream
Breakfast cereals, eggs, other
The contents of hot eating events are classified into six categories: five food
categories and one for beverages (Table 2-3, Table 2-4). The main dish is
taken directly from the original question. The main dish could be meat, fish,
vegetables, or other. Staple, vegetables, trimmings, and bread are variables
constructed from one question about trimmings where all these types of foods
could be mentioned. A staple can be potatoes, rice, pasta, or beans and lentils.
Pasta was originally divided into two categories. “Spaghetti and macaroni”
presented familiar names or types of pasta and “tortellini, penne, ravioli, etc.”
new and modern trends. However, due to very low frequencies, these had to
be combined into the single category “‘pasta”’. The vegetable category here
represents vegetables served as side dishes, either cold (“salads”) or cooked.
Trimmings includes sauces, pickles and other condiments. Bread is bread.
(For a more detailed description see chapter 4, The Meal Format.)
Mäkelä, Kjærnes & Ekström: What Did They Eat?
67
Table 2-3 The recoding of hot meal components
Category
Main dish
Staple
Vegetables
Trimmings
Bread
Components
meat, fish, vegetables, other
potatoes, rice, pasta, beans and lentils
cabbages = broccoli, Brussels sprouts, cabbage, cauliflower, Chinese cabbage
carrot
cucumber
lettuces = ordinary and special lettuces
onion = onion, leek
pulses = green beans, peas, beans and lentils
sweet pepper
tomato
other = artichoke, asparagus, aubergine, avocado, beet root, celery, corn,
fennel, parsley root, Jerusalem artichokes, mushrooms, parsnip, root celery,
salad of grated root crops, spinach, rutabaga, turnip, vegetable mix, zucchini,
and other
hot and cold sauces = cold sauces, hot sauces, melted butter, salad dressing,
pickles, preserves, condiments = ketchup, soy sauce, pickles, jams and jellies
other
bread
The original list of vegetables had 36 different vegetables. It was possible to
choose more than one item at each eating event. The vegetables were coded
into nine new categories. The recoded list includes specific vegetables with
high frequencies (the most common ones such as carrots or tomatoes) and
classes of vegetables, such as lettuces or root crops. The idea was to use different types of vegetables as an indicator of traditional or modern food habits.
Therefore, the list included, on the one hand, traditional vegetables like root
crops, and, on the other hand, vegetables like aubergine or avocado, which as
novelties could be treated as indicators of modern food habits. We will therefore also return to the original list in some of the analyses.
Table 2-4 The recoding of beverages at hot and cold eating events
New categories
milk
water
juice and soft drinks
hot drinks
alcohol
other
nothing
Components
fresh milk and other milk products
tap water, mineral water
juice, lemonade, carbonated soft drinks
coffee, tea, hot chocolate
wine, beer, fortified wine, spirits
not specified
In addition we have information on beverages at each hot and cold eating
event. The original list of possible beverages had 17 options. Here as well the
alternatives were differentiated into modern and traditional ones. We wanted
Eating Patterns
68
to be able to distinguish drinkers of regular coffee from those who drink cappuccino or espresso, which are newcomers in the well-established Scandinavian coffee culture. However, due to low frequencies, these distinctions could
not be retained in this main classification. A new classification with seven
categories was created (Table 2-4).
Figure 2-1 summarises and illustrates the classification of the meal components: the main dish (here meat), the staple (here mashed potatoes), vegetables
(here peas), trimmings (here gravy), and a beverage (here a glass of wine).
Bread is of course an additional option.
Beverage
Trimmings
Staple
Vegetables
Main dish
Figure 2-1 A visualisation of the main components of a hot meal
It should be noted that the basis for the following calculations has been the
number of eating events and not the number of informants as the unit of
analysis. This is due to the structure of the questionnaire, where it proved difficult to take into account various types of information about each informant
and about each eating event at the same time. In most tables describing the
contents of eating events, N is the number of all eating events of that type
(breakfast, cold, hot) eaten during the day for the sake of consistency, most
often divided between weekdays and weekends.
Mäkelä, Kjærnes & Ekström: What Did They Eat?
2.3
69
Results
2.3.1 Breakfast
The first eating event has been automatically defined as a breakfast. On
weekdays over 90 per cent of the informants had breakfast before 11 a.m. On
Sundays the time of breakfast was a bit later, especially in Norway and
Finland. (Data not shown here. For more details see figure 3-3 in chapter 3.
Table 2-5 The number of components at breakfast on weekdays and Sundays. Per cent
Denmark
Finland
Norway
Sweden
Weekday Sunday Weekday Sunday Weekday Sunday Weekday Sunday
Only beverage(s)
6
6
12
10
8
4
5
4
One component
75
62
61
60
83
72
60
54
Two components
16
23
21
25
7
18
24
31
Three components
Four components
2
7
4
5
1
6
10
10
Mean
Total (N)
1
1
1
0
0
1
1
2
1.17
1.34
1.21
1.25
1.00
1.25
1.42
1.52
100 (735) 99 (257) 99 (976) 100 (211) 99 (961) 101 (221) 100 (902) 101 (168)
Differences between countries are statistically significant at 5 per cent risk level on weekdays
The form and contents of the first eating event varied considerably among the
countries. Table 2-5 shows that many respondents in Sweden had had a varied
breakfast with more than one type of food, while very few had eaten more
than one type in Norway. On the other hand, a somewhat greater percentage
had had only a beverage at the beginning of the day in Finland. The number of
components was somewhat higher on Sundays than on weekdays in all countries.
Figure 2-2 shows the types of foods eaten at the first eating event on weekdays. The relative lengths of the columns reflect that many Swedes mentioned
more than one food item, while very few did so in Norway. The figure shows
that in all four countries bread was the main component of a breakfast. It
should be noted that ‘bread’ in the Nordic countries at breakfast time in almost all cases means an open-faced sandwich: a slice of white or brown bread
Eating Patterns
70
with butter/margarine and jam/cheese/salami/liver paste etc.21 The addition of
eggs or foods like ham or bacon was relatively rare.
The consumption of foods eaten with a spoon, like breakfast cereals (muesli,
cornflakes etc), porridge or yoghurt, varied considerably. While the Swedish
respondents favoured all these types of dishes, porridge was most common in
Finland. At the other end, few Norwegian respondents had eaten breakfast
cereals, and porridge or yoghurt was rarely reported. While also having breakfast with few components, the Danes still had had more breakfast cereals and
yoghurt compared to the Norwegians. Few respondents had fruit, sweet pastries or other foods for weekday breakfast.
Sunday breakfasts were slightly more elaborate. The exception was Finland,
where the first eating event was very similar throughout the week. In Denmark, Norway and Sweden a few more respondents had had dishes like eggs
and bacon (data not shown). More Danes also favoured sweet pastries for
breakfast on Sundays (8 per cent on weekdays, 25 per cent on Sundays).
Denmark
68
Finland
62
9
3 4
73
Sweden
0
20
4
40
60
12
10
16
6
78
Norway
3
16
3
6
80
8
6
6
4
8
19
13
100
17
120
6
140
Per cent
Bread
Egg
Cold cuts, cheese
Breakfast cereals
Porridge
Yoghurt
Fruit
Sweet pastries
Other
Figure 2-2 Types of foods at breakfast on weekdays. Per cent22. N: Denmark 737, Finland
975, Norway 958, Sweden 904. More than one choice possible
21
Many Norwegian respondents had answered “bread alone” rather than “open sandwiches”.
Still, we have categorised this together with sandwiches. The term “bread” designates sandwiches in Norwegian and other studies do not indicate that bread alone is frequently eaten.
22
Frequencies below 3 per cent have been left out.
Mäkelä, Kjærnes & Ekström: What Did They Eat?
71
What did people drink for breakfast? The three most common beverages in
each country were the same on weekdays and Sundays (Table 2-6). In Denmark, Norway, and Sweden, the top three beverages were hot beverages (i.e.,
coffee or tea), milk and juice. In Finland, hot beverages were first on the list,
and the second most popular beverage was juice, with milk in the third place.
Similarly, in Sweden almost two-thirds included a hot beverage and onefourth drank milk. The Danish breakfasts were even more concentrated on
coffee and tea. On the other hand, only half of the Norwegian breakfasts included a hot beverage, and milk represented a similar proportion. Even though
the questionnaire allowed several alternatives to be chosen at each eating
event, a majority had only one type of beverage. Between 11 and 26 per cent
of the breakfasts included two beverages, the lowest proportion in Finland.
Table 2-6 Beverages at breakfast on weekdays and Sundays. Per cent. More than one
choice was possible23
Denmark
Finland
Norway
Sweden
Weekday Sunday Weekday Sunday Weekday Sunday Weekday Sunday
Alcoholic
beverages
Hot beverages
(coffee, tea) *
Juice *
0
3
0
0
0
0
0
0
70
75
75
70
49
52
63
75
9
17
17
21
12
19
16
21
Milk, buttermilk *
17
21
14
12
48
51
27
22
Water
9
9
11
9
11
7
9
7
Other beverages
1
0
1
0
0
0
1
0
No beverage
N
0
0
3
1
3
2
5
3
736
257
975
209
957
221
905
168
* Differences between countries are statistically significant at 5 per cent risk level on weekdays
In all countries, juice appeared more often and water less on Sundays than on
weekdays. On weekdays, no one reported drinking alcoholic beverages at
breakfast. However, on Sundays there were some Danish breakfasts with alcohol. In all other countries but Finland, hot beverages were consumed more
often on Sundays.
These findings indicate that breakfast is a simple, but distinct eating event in
the Nordic countries. The number and types of components of breakfasts varied somewhat among the countries. The pattern within each country, however,
was quite homogeneous. While open-faced sandwiches were the most wide23
The responses include those who had no, one or more than one beverage at each event.
Summing up percentages will therefore give little meaning.
Eating Patterns
72
spread type of food for breakfast, the frequencies of foods eaten with a spoon
varied considerably. Foods eaten with knife and fork, like fried bacon, were
rarely mentioned and less substantial foods, such as fruit or sweets, were uncommon at the first eating event of the day. Moreover, almost all of these eating events included a beverage, the most common being coffee or tea, but
with milk as a frequent addition or alternative.
2.3.2 Cold eating events
The three most popular foods at cold eating events in Denmark, Finland, and
Sweden were identical on weekdays (Figure 2-3). Sandwiches were the most
popular type of food in every country. The proportion of other food items was
considerably lower in Norway, compared to the other countries. The second
most common group consisted of sweet pastries, followed by fruits and berries. The proportion of sweet pastries and cakes was highest in Finland and
Sweden, with only half that frequency in the Norwegian sample. Similarly,
fruits and berries were included in cold eating events twice as often in Sweden
and Finland, compared to Denmark and Norway. Other types of foods constituted a minor proportion of the cold eating events. Snacks and sweets were
included in only 4-8 per cent of the events.
Denmark
65
5
4
21
11
7
4
2.3.3 Cold eating events
Finland
57
3 5
25
19
4
6
3
6
4
The cold eating events presented here include a range of events – from the very structured
‘proper’ cold meals to minor, arbitrary – or even random - snacks. Moreover, they include all
events after the first one. The distribution of various types of cold eating events throughout the
11
12
5
6
dayNorway
is described in chapter 3. 70
Sweden
49
0
20
4 3
40
22
24
60
80
8
100
120
140
Per cent
Bread
Cold cuts, cheese
Salad/veg.
Sweet pastries
Fruits, berries
Snacks
Yoghurt
Other
Figure 2-3 Types of foods at cold eating events on weekdays. Per cent. N: Denmark 1462,
Finland 1540, Norway 1666, Sweden 1412. More than one choice was possible
Mäkelä, Kjærnes & Ekström: What Did They Eat?
73
The picture was quite similar on Sundays (Figure 2-4), concerning both the
types of foods and the differences between the countries. However, sweet pastry and cakes appeared more frequently, sandwiches and fruit somewhat less
often.
D e nm a rk
61
F inla n d
7
53
N o rw a y
4
4
6
0
20
B re a d
C old c u t s , c he e s e
9
24
46
3
40
S a lad / ve g .
5
13
29
60
S w e e t pa s t rie s
9
16
29
61
Sw ede n
8
29
80
F ruit s , b errie s
8
9
7
9
3
8
12
5
100
S na c k s
8
120
Y o g hu rt
Figure 2-4 Types of foods at cold eating events on Sundays. Per cent. N: Denmark 417, Finland 313, Norway 275, Sweden 239. More than one choice was possible
Similar to the first eating event, hot beverages were the most frequently consumed type of beverage at cold eating events, but the levels were somewhat
lower later in the day (Figure 2-5). Milk, which was so dominant in the morning, was also less prominent. Juice and soft drinks constituted a varying proportion, the highest figures appearing in the Danish sample, the lowest in
Sweden. The frequency was somewhat higher on Sundays in Denmark,
Finland and Norway. A minor proportion of the events included alcoholic
beverages. But beer, wine or liquor was somewhat more common at cold
events later in the day than was observed for breakfast.
14 0
O t he r
Eating Patterns
Denmark
Finland
Norway
Sweden
74
Sunday
10
W eekday
5
Sunday
4
15
47
12
41
Sunday
5
W eekday
4
19
W eekday
19
0
Coffee, tea
10
20
Soft drinks
40
Milk
6
16
60
80
W ater (tap and mineral)
18
16
14
57
20
7
19
53
14
14
14
11
66
Sunday
6
24
13
10
13
27
52
16
16
10
64
W eekday
Alcohol
7
58
12
10
19
13
13
100
Other
120
140
Nothing
Figure 2-5 Beverages at cold eating events on weekdays and Sundays. Per cent. N: Denmark 1462/417, Finland 1540/313, Norway 1666/275, Sweden 1412/239. More than one
choice was possible
The contents of cold eating events reflect that several types of events were
included, big and small, at different times of the day. Most of the events included only one item, so the various food items mentioned seem to characterise the event. The respondents had either sandwiches, a cake, an apple or
sweets. The range of items was relatively limited and homogeneous within
each country. As mentioned in chapter 1, sweets, peanuts etc are probably
underreported due to the methodology chosen.
2.3.4 Hot meals
Similar to cold eating events, hot meals could also reflect several types of
events, ranging from a hotdog to elaborate multi-course meals. The description is, however, organised according to the most typical hot meal known
from earlier research; including one plateful with several components. The
description of the contents of hot meals is presented in accordance with the
five components outlined earlier: main dish, staple, vegetables, trimmings and
beverages (see chapter 4, The meal format). In addition, possible starters and
desserts are described.
Mäkelä, Kjærnes & Ekström: What Did They Eat?
75
Main ingredients
Table 2-7 shows the proportions of the various main ingredients of hot meals.
Meat was clearly the most common main ingredient, even more so on Sundays than on weekdays. The proportion of meat dishes was highest in Denmark, lowest in Norway. The difference between weekdays and Sundays was
especially evident in Norway, where half of the hot meals contained meat on
weekdays, three-fourths on Sundays. Fish, on the other hand, appeared much
more often on the everyday menu in Norway, particularly compared to Denmark. However, on Sundays very few hot meals contained fish in Norway.
A relatively small proportion of hot meals had vegetables as the main ingredient, the Finnish sample having the highest figure. The heterogeneous “other”
category constituted between 15 and 22 per cent of the hot meals, the highest
figure appearing in Sweden, the lowest in Denmark. The Finnish figures were
in the middle in most cases, again with insignificant differences between
weekday and Sunday meals.
The category “other” included dishes that might be difficult to classify
straightforwardly as meat, fish, or vegetables (e.g. omelette, soup or pizza).
The responses included a large variety of dishes, both traditional and modern.
“Other” main dishes were second most common in all countries but Norway
(Table 2-7). In Sweden, other main dishes were more common than in the
other Nordic countries. No particular dishes were especially prominent in this
category. The most frequent other dishes were pizza, pasta dishes, soups and
unspecified other dishes. Even when recoded into broader categories, however, the highest frequency was 6 per cent (pizza in Sweden). Data are therefore not shown.
Table 2-7 The main dish of hot meals on weekdays and Sundays. Per cent
Denmark
Finland
Norway
Sweden
Weekday Sunday Weekday Sunday Weekday Sunday Weekday Sunday
Fish1
Meat1 2
Vegetable1
Other
9
8
13
13
27
7
15
10
68
74
58
62
51
75
58
66
9
10
14
8
3
2
7
4
15
8
16
17
19
18
22
19
Total (N) 101 (686)100 (237)101 (1376)100 (260)100 (976)102 (222)102 (1222)99 (202)
Differences between countries are statistically significant at 0 per cent risk level
on Sundays
1
on weekdays
2
76
Eating Patterns
Table 2-8 gives an overview of how the main ingredients had been prepared.
While meat was most often served as whole meat dishes (such as steaks,
chops and roasted joints) in Denmark, dishes made from minced meat (hamburgers, meat balls, sausages etc) were more frequent on weekdays in the
other countries. This shifted on Sundays. Dishes where meat is mixed with
other ingredients, for example casseroles and stews, appeared less often, but
somewhat more frequently in Finland. The meat was most often pan fried,
with the highest figure in Sweden. Danes and Norwegians tended to boil the
meat more often (like boiled ham), while Finns were more in favour of oven
cooking. Oven cooked dishes were in general more popular on Sundays.
Other preparation methods, including microwave oven, deep-frying and wok,
appeared quite infrequently in all the countries, but with Sweden having the
highest figures.
Fish was most often served in the form of whole fish or filets. In Norway,
where fish dishes were more common on the weekday menu, minced fish
dishes like fish pudding and fish cakes were quite popular. In Finland, on the
other hand, fish was more often mixed with other ingredients. The fish was
most often fresh in Denmark and Finland, while frozen fish was more frequent in Norway and Sweden. Other types, like smoked or marinated fish,
were rare, again with Sweden as a small exception. Fish preparation methods
were quite varied, particularly on Sundays. Pan-frying was most commonly
used, particularly in Denmark, while the Finns favoured oven dishes and the
Norwegians more often preferred poached fish. The Swedish responses included all these methods, as well as other types.
Vegetables as the main ingredient were usually eaten mixed with other ingredients. The most common method of preparation was boiling. However, vegetables were often prepared by other methods than boiling, frying, or baking in
the oven (data not shown). But since the figures are very low (see Table 2-7),
it is difficult to make any conclusive statements.
Mäkelä, Kjærnes & Ekström: What Did They Eat?
77
Table 2-8 Meat and fish preparation methods. Per cent
Denmark
Finland
Norway
Sweden
Weekday Sunday Weekday Sunday Weekday Sunday Weekday Sunday
a) Types of meat as main dish
Steak, roast, joint
47
63
36
48
31
52
39
54
Minced meat
38
29
45
35
54
39
51
33
Mixed with other
ingredients
Total (N)
14
9
19
17
15
10
9
12
100(469) 100(176) 100(795) 100(162) 100(498) 100(166)
99(707)
99(135)
b) Meat preparation methods
Fried
53
40
40
35
53
45
61
54
Boiled
Baked
24
20
22
32
18
30
13
42
25
13
26
23
11
18
10
25
Microwave oven
1
1
6
7
2
1
3
0
Other method
2
4
4
1
6
5
7
10
0
1
3
1
1
1
0
0
100(401) 100(161) 100(643) 100(134) 100(421) 101(149) 100(644)
99(119)
No info
Total (N)
c) Types of fish as main
dish
Whole fish, fillet
84
79
63
57
54
87
73
75
Minced fish
14
11
13
29
35
13
13
15
2
11
24
14
10
0
12
10
100(19) 100(173)
100(35)
99(268)
100(15)
99(180)
100(20)
Mixed with other
ingredients
Total (N)
100(64)
d) Fresh and frozen fish
Fresh fish
56
60
72
70
39
38
38
13
Frozen fish
43
33
23
30
54
62
52
75
Other
0
7
5
0
3
0
7
6
No info
2
0
1
0
4
0
3
6
100(21) 100(147)
100(13)
99(133)
100(16)
Total (N)
100(54)
100(15) 100(109)
e) Fish preparation methods
Fried
62
35
31
33
38
29
45
42
Poached
21
18
15
7
48
50
25
21
Baked
17
12
41
50
8
14
19
16
Microwave oven
0
18
2
0
2
7
2
5
Other method
0
12
4
7
4
0
10
16
6
7
3
0
0
0
0
100(14) 100(157)
100(19)
No info
Total (N)
0
100(63)
100(17) 100(131)
100(30) 100(240)
Differences between countries are statistically significant at 5 per cent risk level on weekdays (at
second and third eating events)
Eating Patterns
78
Staples
Although often not perceived as the characterising centre of a hot meal, the
staple is usually a basic and less variable component than the main ingredient.
A staple could be potatoes, rice, pasta, or pulses. Bread is also included as a
staple. The different types of staples accompanying the main ingredient are
presented in Table 2-9. It should be pointed out that it was possible to choose
more than one item, for example potatoes and bread.
Table 2-9 Types of staples at hot meals on weekdays and Sundays. Per cent. More than
one choice was possible24
Denmark
Finland
Norway
Sweden
Weekday Sunday Weekday Sunday Weekday Sunday Weekday Sunday
Potatoes1
51
47
34
41
52
51
43
47
Bread1 2
11
15
38
33
10
9
24
25
Rice
8
9
11
11
5
11
9
9
Pasta1 2
7
6
6
6
4
5
11
12
Beans,
lentils1
N
0
2
0
0
1
1
1
0
686
237
1376
260
976
222
1222
202
Differences between the countries are statistically significant at 5 per cent probability level 1 on
weekdays, 2 on Sundays
There were some clear differences between the countries. In Denmark, Norway, and Sweden potato is the most common staple. On weekdays the incidence of potatoes was highest in Denmark, where half of the meals included
potatoes. Interestingly, in Finland, the meals included bread almost as often as
potatoes. This might be due to the fact that a main dish of mixed ingredients
was frequently used in Finland (see Table 2-8 a and c). These casserole types
of dishes can include potatoes and other root crops, and thus bread might be a
more appropriate accompaniment to them than potatoes. In Denmark and
Norway rice was third on the list, while in Sweden pasta came before rice.
Pulses were not often eaten.
On Sundays the order of different types of staples was the same in every
country but the share of each staple was a bit different. In Denmark and Norway, people reported having eaten potatoes less frequently than on weekdays.
In Denmark, more bread was eaten on Sundays, while in Finland the opposite
was the case.
24
Summing up percentages will therefore give little meaning.
Mäkelä, Kjærnes & Ekström: What Did They Eat?
79
Vegetables as side dishes
Table 2-10 shows the selection of vegetables included as side dishes in the
Nordic hot meals. The overall most popular vegetable was carrot, followed by
tomatoes, cucumber, lettuce and cabbages. There were, however, significant
differences between the countries. The Danes and the Norwegians had a lot of
carrots. This was particularly the case in Norway, where almost a third of all
hot meals included carrots. Other vegetable groups worth mentioning in
Denmark and Norway were cabbages and root vegetables. Many Danish hot
meals also included lettuce.
Table 2-10 Vegetables as side dishes at hot meals on weekdays and Sundays. Per cent.
More than one choice was possible25
Denmark
Finland
Norway
Sweden
Weekday Sunday Weekday Sunday Weekday Sunday Weekday Sunday
Beans, peas
10
11
4
3
8
15
5
Cabbages
12
14
5
6
15
18
6
5
Carrots*
18
21
11
10
29
27
11
15
5
5
18
22
5
8
16
20
Lettuces
11
13
16
20
7
9
16
19
Onions
11
7
3
5
5
5
5
8
Sweet pepper*
2
2
4
5
2
3
4
7
Tomatoes*
8
9
15
20
5
9
22
30
Other
vegetables
N
14
13
10
10
14
20
12
17
686
237
1376
260
976
222
1222
202
Cucumber*
11
* Differences between countries are statistically significant at 5 per cent risk level
The Finns and the Swedes preferred cucumber and tomato. The table also indicates that even other vegetables used in salads were mentioned more often
in these countries. In both Sweden and Finland, various lettuces were popular,
especially on Sundays when one meal in five included lettuce. It seems that
most vegetables were used somewhat more frequently on Sundays, and there
were no obvious “weekday vegetables”.
The method of preparation was divided into two categories: either raw or
cooked (Table 2-11). The Danes and the Norwegians seem to favour cooked
vegetables. One third of the meals in Denmark, and even more in Norway,
included vegetables that were cooked. In Finland and Sweden, the situation
25
Summing up percentages will therefore give little meaning.
Eating Patterns
80
was reversed: a third of the meals included raw vegetables. Cooked vegetables
were more frequently included in Sunday hot meals in all countries except in
Denmark, where the share of raw vegetables increased somewhat on Sundays.
Taken together, about half of all hot meals included a vegetable side dish, the
highest figures appearing in Norway, the lowest in Finland. Vegetables were
used somewhat more often on Sundays than on weekdays. It should be mentioned, however, that vegetable main dishes were slightly more frequent in
Finland, and that vegetables can also be included when the meat or fish is
served mixed with other ingredients (for example a mutton and cabbage casserole – “fårikål”).
Table 2-11 Preparation of vegetables as side dishes on weekdays and Sundays. Per cent
Denmark
Finland
Norway
Sweden
Weekday Sunday Weekday Sunday Weekday Sunday Weekday Sunday
Cooked vegetables
Raw vegetables
31
30
10
14
37
47
15
22
14
17
31
31
14
11
32
34
No vegetable
side dish
Total (N)
55
53
59
55
49
42
53
44
100
100
100
100
100
100
(686)
(237)
(1376)
(260)
(976)
(222)
Differences between countries are statistically significant at 5 per cent risk level
100
(1222)
100
(202)
Table 2-12 shows that the range of vegetables ordinarily used as side dishes at
hot meals was rather limited, particularly in Norway, where, as mentioned,
almost one-third of all hot meals included carrots. Again we find a dominance
of “salad” vegetables in Finland and Sweden. Some vegetables, such as onions, may have been used more frequently than these figures indicate, being a
common ingredient in mixed dishes. The lists of vegetables that were mentioned by only one or no respondents at all are long in all four countries, including traditional vegetables of domestic origin (such as parsnip and turnip)
as well as novel vegetables that are usually imported (such as artichoke and
aubergine). The use of traditional vegetables may, however, have been influenced by the time of the year, April being a low season for many domestic
vegetables.
Mäkelä, Kjærnes & Ekström: What Did They Eat?
81
Table 2-12 The top five and bottom of the vegetable choice list. Per cent of total number
of hot meals
The five
most
popular
vegetables
(per cent)
Listed
vegetables
not mentioned or
mentioned
by only
one respondent
Denmark
carrots
17%
ordinary lettuce 9%
tomatoes 8%
peas 8%
onion 7%
Finland
cucumber 23%
tomatoes19%
ordinary lettuce 6%
carrots 14%
sweet pepper 5%
Norway
carrots 29%
peas
8%
broccoli 7%
cucumber 5%
tomatoes 5%
Sweden
tomatoes 27%
cucumber 20%
ordinary lettuce 19%
carrots 14%
onion 6%
artichoke
aubergine
fennel
parsley root
Jerusalem artichoke
parsnip
grated root crops
spinach
rutabaga
turnip
artichoke
asparagus
aubergine
avocado
fennel
green beans
parsley root
Jerusalem artichoke
root celery
spinach
turnip
artichoke
fennel
green beans
Jerusalem artichoke
parsnip
turnip
zucchini
artichoke
aubergine
avocado
celery
Chinese cabbage
fennel
parsley root
Jerusalem artichoke
rutabaga
turnip
Trimmings
A considerable proportion of the hot meals included some kind of extra trimming, the highest frequency appearing in Sweden, the lowest in Finland. While trimmings were included more often in Sunday hot meals in Norway and
Sweden, there was a slight reverse tendency in Denmark and Finland. Table
2-13 shows that in all the countries, hot and cold sauces were the most common type of extra trimming.
Table 2-13 Trimmings at hot meals on weekdays and Sundays. Per cent. More than one
choice was possible26
Denmark
Finland
Norway
Sweden
Weekday Sunday Weekday Sunday Weekday Sunday Weekday Sunday
Hot and cold
sauces 1 2
Pickles, preserves, condiments 1
Other trimmings1
31
27
12
14
20
28
29
40
8
5
10
8
5
9
15
14
9
11
6
4
11
10
4
6
N
686
237
1376
260
976
222
1222
Differences between countries are statistically significant at 5 per cent risk level 1 on weekdays
on Sundays
26
Summing up percentages will therefore give little meaning.
2
202
82
Eating Patterns
But while a third of weekday hot meals included a sauce in Sweden, the frequency was only 12 per cent in Finland. There seems to be a slight “eastern”
bias in the popularity of pickles and preserves, but the figures are quite low.
Quite a few answers have fallen into the category “other” for unspecified
trimmings. Unfortunately, we do not have more specific information on this
category. There might also be some degree of underreporting of these kinds of
meal components. Very ordinary condiments like ketchup and mustard might
easily be forgotten. Moreover, a sauce might also be added to the main dish
before being served.
Beverages
shows the use of beverages at hot meals. It is difficult to detect general features, as the priorities were very dissimilar in the various countries. There was
a clear Danish and Norwegian pattern of using (tap) water and juice/soft
drinks at hot meals. This latter category includes even the popular lemonades
and fruit/berry syrups mixed with water. However, while alcoholic beverages
were quite frequent in Denmark, particularly on Sundays, this was much more
rare in Norway. The Finnish use of beverages at hot meals was very different,
as almost half of the hot meals were accompanied by milk, followed by water
and juice/soft drinks. Included in the milk category is also cultured milk,
which is very common in Finland (“piimä”) and Sweden (“filmjölk”). The
Swedish pattern was the most diverse, including all the various types of beverages.
While the Finnish pattern of beverage consumption shows little distinction
between weekdays and Sundays, there was a tendency towards more alcoholic
drinks and less water on Sundays in the other countries, in Sweden also less
milk. The proportion of juice/soft drinks was somewhat higher on Sundays in
Norway and Sweden. This probably reflects more frequent use of carbonated
soft drinks.
Denmark
Finland
Norway
Sweden
Mäkelä, Kjærnes & Ekström: What Did They Eat?
13
26
S un d ay
12
17
W e ek d ay
S un d ay
4
W e ek d ay
5
5
A lc o h ol
29
49
14
13
20
40
S oft d rin k s
23
19
24
C o ffee , te a
30
23
24
0
41
47
41
W e ek d ay
33
10
14
S un d ay
32
7
32
6
25
29
38
10
7
W e ek d ay
16
26
20
9
12
S un d ay
83
30
60
M ilk
W a te r
80
10 0
1 20
O the r
Figure 2-6 Beverages at hot meals on weekdays and Sundays. Per cent. N: Denmark
686/237, Finland 1376/260, Norway 976/222, Sweden 1222/202. More than one choice was
possible
2.3.5 Meals with several courses
A meal could include even a starter and/or a dessert (more courses than that
being very rare in the Nordic food cultures). Figure 2-6 shows that more hot
meals included a dessert than a starter. However, the majority of hot meals in
the Nordic countries included only one (main) course on all days of the week.
Starters were so rare that it is impossible to draw any conclusions with regard
to variation between the countries or between weekdays and Sundays. With
the exception of Finland, slightly more hot meals included a dessert on Sundays than on weekdays.
Table 2-14 The percentage of starters and dessert on weekdays and Sundays. Per cent of
all hot meals
Denmark
Weekday
Finland
Sunday
Norway
Weekday
Sunday
Weekday
Sweden
Sunday
Weekday Sunday
Starter
4
8
6
8
3
2
4
7
Dessert
21
31
25
23
27
36
22
30
686
237
1376
260
976
222
1222
202
N
Starters on weekdays p < 0.000; starters on Sundays p < 0.06; desserts on weekdays p < 0.000;
desserts on Sundays p < 0.006.
Eating Patterns
84
Table 2-14 shows that the two most popular desserts were berries and fruits
and ice cream, followed by other desserts, like cakes and puddings. Cheese
after the main course was very rare.
Table 2-15 Desserts at hot meals on weekdays and Sundays. Per cent
Denmark
Finland
Norway
Sweden
Weekday Sunday Weekday Sunday Weekday Sunday Weekday Sunday
Berries and fruits
(fresh/preserved)*
Ice cream and
frozen desserts
Other desserts*
No dessert
Total (N)
9
8
8
7
8
13
10
14
5
11
2
5
5
11
4
13
8
13
8
12
14
15
9
78
68
82
76
73
61
77
100
100
100
100
100
100
100
(686)
(237)
(1376)
(260)
(976)
(222)
(1222)
* Differences between countries are statistically significant at 5 per cent risk level on weekdays
2.4
8
65
100
(202)
Summary of findings and some interpretations
The main aim of this chapter was to describe the contents of cold and hot eating events. The findings are summarised in Table 2-16, where the typical features of different eating events are depicted. On the one hand, we explore cultural differences and similarities between countries and also the heterogeneity
within each country. On the other hand, we look at variations between weekdays and Sundays.
Cold eating events commonly consisted of sandwiches plus a hot beverage.
Breakfast (which in most cases was a cold eating event) could typically also
include food items such as breakfast cereals and yoghurt. The exception from
the ”cold” breakfast pattern was the use of hot porridge in Finland and Sweden. Later in the day, cold eating events included even a pastry taken with
coffee, a banana with a soft drink, or an ice cream. Very few cold eating
events included several food items and/or several beverages.
Hot meals were by definition “cooked” and therefore somewhat more complex. Yet, many hot meals were quite simple, the main components being
meat (or fish in Norway) and potatoes as a staple, frequently but not always
accompanied by vegetables.
Mäkelä, Kjærnes & Ekström: What Did They Eat?
85
As an illustration of common dishes, we will refer lists of the most frequent
dishes presented by marketing research companies.27 They reflect that simple
dishes from minced meat are very popular, and also that this popularity has
changed very little over the past decade. In Finland, the three most popular
dishes have for 15 years been meatballs, minced meat in sauce, and meat and
macaroni casserole. However, some changed have taken place. In 1985, pizza
ranked 15th, but it has gradually risen to fifth place (Food Facts 1999). In
Denmark, meatballs held first place among the most popular dishes in both
1985 and 1995. Pizza did not appear in this top ten list (Fagt et al 1999, 63).
In Sweden, falukorv (Swedish sausage), minced meat in sauce, and pizza rank
as the three most popular dinner dishes. This ranking has changed very little
over the past years (KF 2000, personal communication). Our own data show
that the Norwegian everyday menu is not very different from this. However,
fish being considerably more popular, dishes from minced fish, such as fish
cakes, are quite common.
The selection of vegetables was rather limited, both within each country and
in the Nordic region as a whole. The range of vegetables offered in the shops
indicates a larger variety, but our findings suggest that they are not a dominant
feature of the ordinary, everyday menu. Several lists of options included a
variety of traditional as well as modern food items, dishes and beverages. Invariably, the response rates for “novel” food items and dishes were low. None
or very few had had either cappuccino, kebab or avocado. The same goes for
preparation methods like wok or deep frying. The use of microwave oven as a
method of preparation was quite rare according to our results. However, according to another recent Finnish study the use of microwave ovens is more
frequent. They are used for both cooking and heating (Food Facts 1999).
We have described one day of eating, thus providing a picture of ordinary,
everyday eating patterns. Within this context it is clear that most novelties or
specialties still hold a relatively marginal position for most people. This does
not mean that these items or dishes are not consumed. Sales figures would tell
us differently. The main point is that they do not seem to have attained a major position in the everyday Nordic food cultures. For a large majority, eating
events seem to be structured around very well known food and drink items.
The selection of foods and dishes is limited and also relatively homogeneous
within each country. Due to the low frequencies of most modern/novel foods
and dishes it has been difficult to analyse particular modern/traditional or “so27
The origins of these lists are rather obscure, and the classifications are very heterogeneous.
Some dishes are very specific, while others are much more general, thus influencing their ‘ratings’. In Norway, no list of this kind has recently been produced.
Eating Patterns
86
phisticated” styles of eating. This will be discussed further in chapters 4 and
8.
Table 2-16 The typical features of different eating events in four Nordic countries. Items
used in 15 per cent or more of the respective types of eating events
Breakfast
Cold eating
events
Hot meals
Denmark
open sandwiches
breakfast cereals
(sweet pastry on
Sunday)
Finland
open sandwiches
porridge
Norway
open sandwiches
(eggs on Sundays)
Sweden
open sandwiches
breakfast cereals
yoghurt
coffee or tea
milk
(juice on Sunday)
coffee or tea
juice, milk or water
coffee or tea
milk, juice or water
coffee or tea
milk or juice
the simplest one
the most varied
one
sandwiches
sweet pastries
fruits
sandwiches
sweet pastries
sandwiches
sweet pastries
fruits
sandwiches
(cakes on Sundays)
coffee or tea
beer or wine
soft drinks
coffee or tea
water, (soft drinks
on Sundays)
coffee or tea
coffee or tea
milk, (soft drinks on
Sundays)
meal
meat (steak/
minced, fried),
other dishes
+ potatoes
+ cooked vegetables (carrots)
(+) sauce
snack
meat (minced/
steak, baked/fried),
other dishes)
+ bread, potatoes
+ uncooked vegetables (cucumber)
meal
meat (minced,
fried), fish
(poached) - not on
Sundays
+ potatoes
+ cooked vegetables (carrots)
(+) sauce
snack
meat (minced,
fried), fish (fried),
other dishes
+ potatoes, bread
+ uncooked vegetables (tomatoes)
+ sauce
(+) pickles
water, beer/wine,
soft drinks, milk
milk, water
water, soft drinks
water, milk, softdrinks, beer/wine
the lowest number
of components
the highest number
of components
Food items which were not easily categorised vanish into the category
“other”28. In spite of this, it seems that our categories have worked pretty well
in the most important questions. Yet, in questions about different trimmings a
28
Only in relation to the question of the main ingredient of the dish category “other” included a
list with 22 alternatives.
Mäkelä, Kjærnes & Ekström: What Did They Eat?
87
lion’s share of the answers fell into category “other”, which indicates that the
alternatives were too few.
It is possible to find differences in the food habits of Scandinavians. First, it
seems that Denmark and Norway, on the one hand, and Finland and Sweden,
on the other hand, have certain features in common:
a) The different traditions of hot and cold lunches should be pointed out (see
also chapter 4). As Finland and Sweden have the tradition of a hot lunch, the
more snack type of cold eating events are quite common. In Denmark and
Norway, many eat sandwiches for their lunch. Therefore, especially in Denmark, the list of food items at cold eating events is somewhat more diverse.
Some informants in Finland and Sweden have reported nothing to drink in
relation to cold eating events. This might reflect the difference in notions of
cold eating events in the respective countries. In Denmark and Norway a cold
eating event would often be defined as a cold meal, while in Finland and
Sweden, cold eating events are more likely to be treated as snacks.
b) This is also reflected in the use of beverages. In Norway, milk was a beverage of cold eating events. In Finland and Sweden, milk was reported in conjunction with hot meals. Norwegians drink a lot of milk at breakfast but far
less yoghurt than in the other three countries.
c) At breakfast, dishes eaten with a spoon appeared more often in Finland
and Sweden, compared to Denmark and Norway, most often represented by
porridge in Finland, breakfast cereals, cultured milk and yoghurt in Sweden.
d) Most features of hot meals did not seem to follow the east-west division
line. One exception is the common practices with regard to vegetables, where
vegetables used for cooking dominate in Denmark and Norway, vegetables
used raw in salads in Finland and Sweden. A “salad” including various raw
vegetables is a fairly new addition to hot meals, but in Finland and Sweden it
has established its place rapidly through the well-organised mass catering system and nutrition education promoting healthy eating habits (see e.g., Mäkelä
1996b). Hot lunches with a salad at school or in work canteens have provided
a model for the home meals as well. Actually, this salad seldom includes lettuce but is in general made out of raw vegetables such as grated carrots. In
Denmark and Norway, the same carrots would be served cooked.
88
Eating Patterns
e) There was also a slight difference in the use of additional bread at hot
meals, which occurred mainly in Finland, to some degree also in Sweden, but
very rarely in Denmark and Norway.
Second, there were a number of variations that seem to express particular national customs:
a) Fish appeared much more often on the menu in Norway, compared to the
other countries. The use pattern for fish was also distinct, with more minced
dishes and more frozen fish. Whole fish was usually poached. In the other
countries, minced dishes of fish were more uncommon, and the fish was usually fried. The frequent use of frozen fish was somewhat surprising. It might
be that fresh whole fish or fish filets were as common as in the other countries, but that the other (simple, everyday) dishes were used in addition. This
strong everyday position of fish in Norway was confirmed by the large difference between weekdays and Sundays in the relative proportion of fish as main
dish, a difference that was much larger than in the other countries.
Another typical feature of the Norwegian menu was the very standardised
breakfasts - a relatively solid, but simple breakfast with open-faced sandwiches, milk and/or a hot beverage. This standardisation in Norway was to
some degree evident even in the contents of cold eating events later in the
day. In many ways, this is in accordance with the so-called “Oslo breakfast”, a
type of school meal that was introduced in the 1920s and strongly promoted
for several decades. The patterns were somewhat more varied in the other
countries, in Sweden also at the individual level.
b) Oven dishes seem to be more popular in Finland, with meat or fish constituting the main ingredient. As these ingredients were often used in “mixed
dishes”, we can anticipate that they reflect the very common dish “laatikko”, a
casserole. Many also have pea soup on Thursdays. Oven baked dishes appeared even in the other countries, but then mainly with meat and mostly on
Sundays. This was probably a different type of dish, such as roasted meat or
loin.
While Finns share the tradition of hot lunches with the Swedes, the composition of their hot meals was quite different (apart from the frequency of salads).
The Finnish hot meals were less complex in terms of the number of components. Dishes containing several ingredients were quite common (meat/fish
mixed with vegetables and/or potatoes), rather than served as separate dishes.
Mäkelä, Kjærnes & Ekström: What Did They Eat?
89
Certain elements of the Finnish eating pattern may be interpreted as signs of
traditionalism, such as the use of porridge at breakfast, the generally frequent
use of milk (and cultured milk in particular), and bread as parts of a hot meal.
Porridge and cultured milk were very important elements of the traditional
rural Nordic diet.
c) Even though they were the most popular main ingredient of hot meals in
all four countries, meat dishes had a particularly solid position on the menu in
Denmark. The strong role of meat was confirmed by the more frequent use of
steaks, roasted meat and other dishes where meat had not been minced or
mixed with other ingredients.
The frequent use of alcoholic beverages also seems to be a significant feature
of the Danish eating patterns. Apart from weekday breakfasts, alcoholic beverages appeared quite often in all types of eating events: Sunday breakfasts,
cold eating events and hot meals. However, most of the alcohol consumed
was beer or wine, not liquor. Moreover, on Sundays even a considerable proportion of the Swedish hot meals included alcoholic beverages.
d) Variation seems to be the most characteristic feature of the Swedish eating events compared to the other Nordic countries. There was more variation
at two levels, both in the sense of larger heterogeneity in the Swedish sample,
and more diversity or complexity at the individual level. Breakfasts as well as
hot meals had more components with regard to food items as well as the number of beverages. Moreover, preparation methods for meat and fish dishes
were more diverse, as was the selection of vegetables.
Some general differences between weekdays and Sundays can be pointed out
if we look at both the complexity and contents of the various types of eating
events. Breakfast seems to be a little richer on Sundays than on weekdays,
with the addition of eggs in Norway and sweet pastries in Denmark. Sunday
cold eating events included more sweetened items, such as cakes and pastries
and soft drinks. The hot meals on Sundays were also somewhat more elaborate, with more components. Roasts and steaks were served more often. More
eating events also included beer or wine. However, the differences were not
striking, and in Finland they were only marginal and rarely significant.
Meat had a stronger position on Sundays. Could this indicate that Sunday
meals still have a special role in family life, as suggested by e.g. Charles and
Kerr (1988) and Lupton (1996)? It seems that in the Nordic countries there are
still signs of a tradition of Sundays roasts. In addition, Sunday meals were
accompanied by more vegetable side dishes than weekday meals. Interest-
90
Eating Patterns
ingly, the ideas of Sunday food were not similar in each country, especially
with regard to fish. In Finland, dishes like fish pudding were a Sunday food
whereas in Norway, fish and fish cakes (“fiskekaker”) appeared much more
frequently on the everyday menu.
3
The Daily Rhythm of Eating
Jukka Gronow and Anni Jääskeläinen
3.1
Is eating becoming more irregular?
The main aim of this study is to analyse whether irregular eating patterns have
taken the place of regular meals in the Nordic countries, as presumed by
many. Are we, as some sociologists studying food culture in Europe and
North America have suggested, in the era of 'gastroanomy’ where regular and
socially shared meals have become increasingly rare and irregular and 'asocial' snacks have taken their place, expressed in the concept of grazing
(Fischler 1990). The worries are crystallised in the debate on whether traditional family meal patterns and meal formats are being disrupted (Murcott
1982, Murcott 1997). Many arguments more or less explicitly presume that
there are normative expectations concerning the kind of food we should eat,
which are not, in reality, met in everyday practices. The sociological discussion was influenced by the structural concept of a proper meal introduced by
Douglas and Nicod in 1974 (Douglas & Nicod 1974, Douglas 1975).
The question of the irregularity of eating habits can refer to several different
social aspects of eating. It can refer to the rhythm and number of daily eating
events. (How many times, at which hours and in what order is eating repeated
during each day?) It can concern the meal format (the composition of the
main course and the sequence of the whole meal) and, finally, the social organisation of eating is also relevant (Where and with whom are people eating,
who did the cooking, etc.?) In this paper the daily rhythm of eating in the Nordic countries is analysed. Socially shared eating times are one of the major
aspects of the sociability of eating.
Eating Patterns
92
3.2
Studying the Nordic eating day
The data of the study "Eating and Modern Everyday Life: A Comparative
Survey of the Nordic countries" were collected in April 1997 with computer
assisted telephone interviews from a representative sample of approximately
1200 respondents from Denmark, Finland, Norway and Sweden each.29 The
method used is similar to the 24-hour recall often used in dietary surveys, but
we were not interested in the quantities of food consumed but in food combinations and the social context of eating. In order to avoid any predefined meal
concepts and to overcome the problem of the cultural variation in the naming
of the different meals and eating occasions, the analytical starting point of the
questionnaire was the eating event. By eating events we mean any occasion
when something is eaten. (Except for breakfast, just a drink or only eating
chewing gum or sweets, etc. were, however, excluded for practical reasons of
reporting.) The analysis of the empirical data thus involves a reconstruction of
what could be called meals, dinners, lunches, snacks, etc. (Even though the
Nordic countries resemble each other in many respects we could not simply
use concepts like 'dinner', 'lunch' or 'main meal' implicitly meaning, say, hot
'real' meals in our questionnaire since they would not necessarily imply the
same thing in each country and to all the respondents.)
This reconstruction is based on a model of what can be called the eating system. The model distinguishes three dimensions of the eating system: eating
pattern (or rhythm), meal format, and the social organisation of eating
(Mäkelä et al. 1999, Gronow et al. 1998). In the analysis of the eating pattern
presented here, we are interested in the schedule of eating events over a period
of 24 hours, namely when and how many times people ate something during
one day. The first eating event of the day was recorded as breakfast. In the
questionnaire breakfast could be more or less elaborate, cold, hot or both.
Therefore, in the following it is counted neither as a hot nor as a cold eating
event. The recording of all other eating events started with registering whether
they were cold or hot eating events, and the following recording was then designed differently according to this basic distinction. The informants were
asked to report on their eating during a single day, the day before the interview. It would have been ideal to record them all on the very same day in all
four countries, but for practical reasons the interviews had to be distributed
over a two-week period. Thus data were collected on all weekdays.
29
For a more detailed description of the method and data collection, see chapter 1 and
Kjærnes et al (2000).
Gronow & Jääskeläinen: The Daily Rhythm of Eating
3.3
93
The daily rhythm of eating
Figure 3-1 and Figure 3-2 describe the dispersion of all the recorded eating
events during the hours of the day in all four countries both on weekdays and
on Sundays. In all four countries there are clearly some peak hours during the
day when a large proportion of the population was eating something. These
peaks were most prominent in Denmark, where eating took place socially
more regularly than in any other Nordic Country. Such peaks occurred in all
four countries early in the morning at about 7 to 8 o'clock, at noon and late in
the afternoon or early in the evening. In addition there was a minor peak later
in the evening (either 7-8 p.m. or 8-9 p.m.). The early morning and noon
peaks coincided in all four countries, whereas the afternoon and evening
peaks occurred at slightly different hours in the different countries. On Sundays, eating occasions were spread out more evenly all through the day. In
Denmark eating was more regular also on Sundays, but even there the peaks
were clearly lower on Sundays than on weekdays.
Hot eating event
0
0
1st eating event
10
10
30
20
10
0
30
20
10
0
time (24 hours)
40
40
time (24 hours)
Sweden weekdays
time (24 hours)
Finland weekdays
Figure 3-1 The distribution of eating events during the day on weekdays in the Nordic countries, 1 hour intervals (per cent)
Cold eating event
50
50
Norway weekdays
60
20
20
60
30
30
time (24 hours)
40
40
60
50
Denmark weekdays
50
60
10
10
0
Sweden Sundays
Finland Sundays
Figure 3-2 The distribution of eating events during the day on Sundays in the Nordic countries, 1 hour intervals (per cent)
0
hot
15
15
cold
20
20
1st eating event
25
25
5
30
30
5
35
35
45
0
5
10
15
20
25
30
35
40
45
40
Norway Sundays
D e n m a rk S u n d a y s
40
45
0
5
10
15
20
25
30
35
40
45
Eating Patterns
96
Earlier comparable studies are available in Norway (Figure -3-3). There the
curves describing the frequencies of the eating times have preserved their
form to an astonishing degree ever since the early 70's, even though in the
90’s eating has become slightly more spread out during the day (Statistics
Norway 1992).
18
16
14
Percent
12
10
8
6
4
2
0
23
22
21
20
19
18
17
16
15
14
:0
:0
:0
:0
:0
:0
:0
:0
:0
:0
:0
:0
:0
:0
:0
:0
:0
:0
:0
:0
:0
:0
:0
:0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
19 8 0-81
13
12
10
11
09
08
07
06
05
04
03
02
01
00
1 97 1 -7 2
1 99 0 -9 1
Figure -3-3 Eating activities during the day in Norway on Mondays to Thursdays, 197172, 1980-81, and 1990-91 (per cent)
Similar studies were performed in Finland as well. Even though not directly
comparable, when we look at data from 1973 and our figures meal peaks seem
to be remarkably stable (Gallup Finland, unpublished data). Eating times were
quite similar. However, the peaks were somewhat higher thirty years ago (i.e.,
more people adhered to a shared national pattern). The second main meal
(dinner) took place somewhat later according to our data.
The results of our study show two different factors at work. On the one hand,
in each country there are quite evidently generally shared times of eating. E.g.
taken together, on weekdays two thirds of the whole Nordic population had
eaten something between 11 a.m. and 1 p.m. in all four countries. (A very
small proportion had eaten twice during this interval.) These percentages were
about the same in every country: Denmark 71%, Finland 61%, Norway 65%
and Sweden 68%. On the other hand, eating continued almost uninterruptedly
Gronow & Jääskeläinen: The Daily Rhythm of Eating
97
from early in the morning until late at night. From 6 a.m. to 11 p.m. at least
almost fifteen percent of the whole population were eating something each
hour (Figure 3-1). Thus there were no time periods, except in the very early
morning hours or late at night, when practically no one was eating anything.
This means that even though there are obviously commonly shared national
eating times, which differed slightly from one country to another - in each
country there is a specific daily rhythm of eating - there is also a lot of variation in the national eating times.
3.3.1 The number of eating events on weekdays and Sundays
The mean number of daily eating events on weekdays was the same in Denmark, Finland and Sweden (3.9), whereas in Norway it was slightly less (3.7)
(Table 3-1). The great majority had eaten at least three times (i.e. breakfast
plus two other occasions) and very few more than six times. No one had eaten
more often than nine times. Differences between the countries were not great.
As the lower average number indicates, in Norway there were significantly
fewer people who reported that they had eaten more often than four times during the previous day (19% in Norway vs. 27 - 29% in the three other countries). On Sundays eating took place less often, Norwegians again eaten least
frequently. (In Norway an average of 3.3, in Denmark, Sweden and Finland
3.6.) Only six per cent had eaten more than four times in Norway on Sundays.
In this respect Norway seems to be the most ascetic of all the Nordic countries.
If the cold and hot eating events - one of the basic distinctions in the questionnaire - are considered separately the Nordic countries fall into two distinct
groups (Table 3-2, Table 3-3). In Denmark and Norway food is eaten cold
relatively more frequently than in Finland and Sweden, which are 'hot food'
cultures. The average number of cold eating events (with the exclusion of the
day's first eating event, breakfast, which was registered in such a way that the
distinction between cold and hot eating events was not made) was highest in
Denmark (2.0), and second highest in Norway (1.7). It should be remembered
that there were less eating occasions in general in Norway) and less in Finland
and Sweden (1.6) (Table 3-2). The average number of hot eating events in
both Sweden and Finland was 1.4 and in both Denmark and Norway 1.0
(Table 3-1). About four-fifths of all Norwegians and Danes ate hot food once.
One-tenth had not eaten hot food at all during the previous day. In these two
countries there were rather few people who had eaten hot food more than once
(6% in Norway and 12% in Denmark), whereas the respective percentages in
Eating Patterns
98
Finland were as high as 43% and in Sweden 40%. In no country did anyone
eat hot food more than three times. As will be shown in more detail later on,
in Denmark and Norway there is a general custom of eating cold food at
around noon (what could be called the lunch hour), whereas Finns and Swedes are more likely to eat hot food at noon.
In the main, these differences between the countries remain the same on Sundays even though, on the whole, people ate fewer times on Sundays in all the
countries.
Table 3-1 The number of eating events in the Nordic countries (per cent). Weekdays*)
and Sundays*)
Number of
eating events
Denmark
Finland
Norway
Sweden
Weekday Sunday Weekday Sunday Weekday Sunday Weekday Sunday
1
1
1
1
2
1
1
0
1
2
7
11
8
13
8
17
6
12
3
29
38
26
30
31
46
31
35
4
36
34
36
34
41
31
35
33
5
21
14
21
15
16
5
19
16
6+
8
2
7
6
3
1
8
4
Total (N)
100 (736) 100 (257) 100 (977)100 (211) 100 (958) 100 (221) 100 (903) 100 (168)
Mean
3.94
3.55
3.93
3.63
3.74
3.26
3.93
3.63
0.98
1.15
1.19
0.98
0.82
1.06
1.04
Std.dev
0.98
Differences between the countries are statistically significant: *) p<0.05
Table 3-2 The number of cold eating events in the Nordic countries (per cent). Breakfast
not included. Weekdays*) and Sundays*)
Number of
cold
eating events
0
1
2
3
4+
Total (N)
Mean
Std.dev.
Denmark
Finland
Norway
Sweden
Weekday Sunday Weekday Sunday Weekday Sunday Weekday Sunday
7
13
15
18
10
21
17
20
28
35
36
37
32
40
34
36
35
34
31
30
40
33
29
29
21
14
14
13
15
6
15
14
8
4
4
3
4
1
5
2
100 (736) 100 (257) 100 (977) 100 (211) 100 (957) 100 (221) 100 (903) 100 (168)
1.99
1.62
1.58
1.48
1.73
1.25
1.56
1.43
1.10
1.05
1.11
1.02
1.11
1.04
0.99
0.87
Differences between the countries are statistically significant: *) p<0.05
Gronow & Jääskeläinen: The Daily Rhythm of Eating
99
Table 3-3 The number of hot eating events in the Nordic countries (per cent). Weekdays*)
and Sundays*)
Number of hot
eating events
Denmark
Finland
Norway
Sweden
Weekday Sunday Weekday Sunday Weekday Sunday Weekday Sunday
0
11
12
5
7
10
9
5
5
1
84
83
52
66
78
82
55
70
2
6
5
39
25
11
10
38
24
3
0
0
4
3
1
0
2
1
Total (N)
100 (736) 100 (257) 100 (977) 100 (211) 100 (957) 100 (221) 100 (903) 100 (168)
Mean
0.95
0.93
1.44
1.24
1.02
1.01
1.37
1.20
0.41
0.66
0.63
0.49
0.43
0.62
0.53
Std.dev
0.40
Differences between the countries are statistically significant: *) p<0.05
As has already been mentioned, almost all informants had eaten at least three
times during the day on weekdays. 6 - 8% had eaten only twice and hardly
anyone only once. On the other hand, no one had eaten more than nine times
or had eaten more or less uninterruptedly all day long. In this sense one could
say that the 'hypothesis' of increased grazing was not confirmed by this study.
One should, however, keep in mind that minor eating events (sweets, chewing
gum, apples etc.) were not reported at all in our study because it would have
been practically impossible with the technique at hand since in the case of
small snacks like candies or lozenges it would obviously be difficult to distinguish one such minor eating event from another. Therefore, on the basis of
this study alone one cannot say anything about the frequency and general
occurrence of very small 'in between' snacks. What can be concluded, on the
other hand, is that no respondent ate only such minor snacks. It should also be
kept in mind that in these kinds of surveys small groups with their peculiar
eating habits (e.g. some religious or ethnic minorities or people living in
closed institutions) do not show up, neither do people with strongly 'deviant'
eating habits, such as those suffering from eating disorders. Nationwide surveys give information only about the majorities.)
There are several previous studies of daily eating in the Nordic countries.
However, one has to be careful in comparing these results. Firstly, samples
and populations studied are often differently defined and more limited. Secondly, respondents were some times asked about their usual habits of eating
("How many times/meals do you usually eat every day or week?") and not
about the day before. Studies of the usual daily behaviour seem to report more
frequent eating than the method adopted here. Obviously, in such studies people wish to appear more conformist. This is the case, in particular, if the habit
100
Eating Patterns
in question is thought to be the normal or the expected. Compared to these
findings, one may assume that the number of eating events counted in the
Nordic study is rather reliable and representative.
According to a study using data from a dietary survey in Finland in 1992 with
1861 adult informants 47% of all men ate three daily meals (breakfast plus
two others) and almost as many only twice (Roos and Prättälä 1997). Women
ate a little less frequently: 44% ate three meals and 51% two meals. Differences between age groups were rather small but it was somewhat more common among the younger respondents to eat only twice a day. In this study a
meal was defined either as hot food or - in the case of the first eating occasion
- even cold food if it had been consumed before ten in the morning. These
figures can be compared with the number of those in the present study who
had reported having eaten hot food at least once or twice a day in Finland.
Since 52% ate once and 43% at least twice, in addition to the first eating
event, there is a rather good correspondence between the results of these two
studies.
In Denmark in an earlier study much higher frequencies of daily eating events
were reported (Haraldsdottir et al. 1987). The daily average was as high as
5.4. 62% said that they ate one hot main meal every day and of this number
82% usually ate it in the evening. 40% said they ate something (often a snack)
after breakfast and before lunch (these amounts included even pure drinking
occasions e.g. a cup of coffee). This study adopted the dietary history method,
however, which could account for the higher reported number of meals.
According to another recent nationwide study of the dietary habits of Danes in
1995 about half ate five meals daily of which three were on average 'main'
meals and 'two' minor, 'in between' meals or snacks. In this study, the respondents were asked to keep a pre-classified dietary record for seven days (Fagt,
Groth, and Andersen 2000, 42).
According to a study by Wold et al in 1985 almost all or 98% of Norwegians
ate breakfast, 84% ate a 'matpakke' (a lunch pack containing a couple of slices
of bread) for lunch and, in addition, 65% had eaten something or drunk coffee
in the afternoon. 71% had eaten some kind of evening meal (mostly bread).
As many as nine-tenths of the respondents reported that they eat dinner regularly seven times a week. In the beginning of the 90's, as many as 30-40% of
Norwegians had skipped dinner once or more during the week under study.
80% of all Norwegian dinners were eaten at home (Wold 1985, Wandel,
Bugge, and Ramm 1995).
Gronow & Jääskeläinen: The Daily Rhythm of Eating
101
Becker compiled data concerning standards of nutrition and daily meals from
practically all studies conducted in Sweden since the 1950's. The populations
in questions as well as study methods varied greatly (24-hour recall was,
however, often used) and so do the results (Becker, Enghardt, and Robertson
1994). According to Becker’s research the daily average number of eating
events was anything between 4.6 (adults in Southern Sweden in 1990) to as
much as 7.2 (women aged 25-44 in Stockholm). The average both in Sweden
and Denmark in the present study was almost four per day, and if one allows
for one extra small snack or drink per day (not necessarily recorded in this
study) it comes very close to the numbers reported in many earlier Swedish
studies.
Generally, however, previous studies conducted both in Denmark and Sweden
tend to give higher average figures than the present study. One should, however, be careful in comparing such results and drawing conclusions from them
since they obviously depend on the definition and criteria of an eating occasion/meal and the method of data collection used in them.
According to the results of a 'Nationwide' survey using 24-hour dietary recall
conducted in the USA in 1987-8 with 4114 households living in 48 states,
respondents ate on average 3.5 times, which was very close to the Nordic averages in the present study (Longnecker, Harper, and Kim 1997). In this
study, the number of calories consumed determined whether the eating occasion was counted as a 'real' meal. After subtracting all such minor eating
events during which less that 70 kcal were consumed the average was 3.1; if
150 kcal was used as the criteria the average was lower, namely 2.8.
3.3.2 The number of different types of cold food eating events
As was to be expected, in Denmark and Norway the mean number of cold
eating events was higher (2.0 and 1.7 respectively) than in the two other Nordic countries (see Table 3-2). Despite the greater share of hot eating events the
mean number of cold eating events in Finland and Sweden was, however, almost as high as in Norway (1.6 in Finland and Sweden). This was due to the
fact that the general number of eating events was significantly higher in both
Finland and Sweden compared to Norway. In Denmark and Norway, there
were less people who had not eaten cold food at all during the day before the
interview and more who had eaten it twice or even more times. In Denmark,
more respondents reported having eaten cold food on three (21%) and even
four times (8%) than in all the other countries in which the respective shares
were almost equal to one another (14-15% and 3-5%).
102
Eating Patterns
All the cold eating events were further divided into three categories. The first
category, snacks, consisted of those cases during which only fruits, ice cream,
sweets, nuts or chips were eaten. The second category was called 'sweet pastries'. This included all those cases where only cakes or other sweet pastries
were eaten. The third and numerically largest category, called cold food, consisted of all the remaining eating occasions. It consisted mainly of bread with
or without filling. All the cases can include drink(s). It should be noticed, anyway, that those eating events that have been classified in this third group may
also include sweet bread and fruits or other small snacks as their additional
components. This means that eating cakes, sweet pastry and fruits or other
snacks, etc. is more frequent in the Nordic countries than could be deduced on
the basis of these figures alone (see Johanna Mäkelä's more detailed analyses
of the contents of the eating events.) Furthermore, in the first category only
such events have been classified which consist only of 'snacks'. If both sweet
pastry and snacks were eaten then the event was classified as a 'sweet pastry'
event.
Table 3-4 The mean number of different cold eating events in the Nordic countries on
weekdays
Nature of eating
Denmark
Finland
Norway
events
All cold
2.0
1.6
Snack
0.3
0.2
Sweet pastry
0.3
0.4
Cold food
1.4
1.1
N
736
977
Differences between countries are statistically significant, p<0.05
Sweden
1.7
0.2
0.2
1.4
957
1.6
0.3
0.3
1.0
904
There were no marked differences between the four countries as far as the
frequency of the 'sweet pastry' and 'snack' events were concerned (Table 3-4).
In Norway, on weekdays there were slightly less sweet pastry eating events,
however, than in the other three countries, which further emphasised the ascetic nature of their eating. The overall number of cold eating events, and consequently of 'cold food eating events', was, however, higher in Denmark and
Norway than in Finland and Sweden. On Sundays, when eating all kinds of
food and cold food in particular took place less frequently in all the countries
the Norwegians differed less from the other countries in this respect. The
smaller total number of cold eating events in Finland and Sweden was, thus,
due to the less frequent number of 'cold food eating events'. 'Sweet pastry' and
'snacks' were eaten on Sundays as often as on weekdays. In general, on Sundays, the four countries resembled each other more than on weekdays: the
differences were small and statistically not significant.
Gronow & Jääskeläinen: The Daily Rhythm of Eating
103
Both on weekdays and Sundays most people had not eaten 'sweet pastry' (7080%) or 'snacks' (about 80%) at all in all the countries. Very few people had
eaten 'sweet pastry' more than once (of these almost all only twice) on weekdays in any country. It was most common in Sweden (6%) and least common,
as was to be expected, in Norway (1%). In Norway only 17 (1.8%) persons
had eaten 'snacks' more than once, Denmark led with 27 people (4%). On
Sundays, Denmark was the leading sweet pastry eating country (6% had eaten
more than once) and Norway was again in last place with only 3% who had
eaten 'a sweet pastry meal' twice. But again the differences between countries
were not great.
Both on Sundays and weekdays women ate sweet pastry and cold food
slightly more frequently in every country with the exception of Finland, where
men more often ate cakes and sweet bread on Sundays than women (Table
3-5).
Table 3-5 The mean number of different kinds of cold eating events eaten on weekdays
and Sundays by men and women in all four countries
Type of cold eating
event
Weekdays
Sundays
women
men
women
Snack
0.3
0.3
0.2
Sweet pastry
0.3
0.2
0.4
Cold food
1.3
1.2
1.0
N
1823
1752
446
Differences between men and women are statistically significant, p<0.05.
men
0.2
0.3
0.9
411
In all the countries other than Norway, there were two time periods during
which sweet pastry was eaten more often than at other times, first, in the afternoon and, second, later in the evening. Yet, only about five percent of all
respondents ate sweet pastry even at these 'peak' hours. In Denmark, these
time periods were a little later than in Finland and Sweden. Snacks were eaten
most often in the evenings in all the countries (Figure 3-4).
10
10
0
Sweden weekdays
Finland weekdays
Figure 3-4 The distribution of various cold eating events during the day on weekdays. 1 hour intervals (per cent)
0
snack
15
15
cake
20
20
other cold
25
25
5
30
30
5
35
35
45
0
5
10
15
20
25
30
35
40
45
40
Norway weekdays
Denmark weekdays
40
45
0
5
10
15
20
25
30
35
40
45
Gronow & Jääskeläinen: The Daily Rhythm of Eating
105
3.3.3 Frequencies of eating in different population groups
It is reasonable to presume that there are systematic differences in the daily
rhythm of eating in different social groups or segments of the population. It
might be thought, for instance, that older people would be more traditional in
their eating habits and would follow more regular eating patterns. One could
also presume that the daily lives of those employed as well as those living in
families are more regulated than the lives of retired or unemployed single
people. Therefore, in the following, variation both in the daily number of eating events and in their chronological occurrence is analysed in relation to different types of background variables.
The first variables describe such characteristics of the individuals that are irreversible: once ascribed they cannot normally be changed. They consist of
age, gender, and education. The second type of variables are reversible. At
least in principle they can be influenced by the individual him/herself even at
a later stage in life. They are income, status of employment (occupationally
active vs. non-active), household type and size and the size of the municipality where the respondent is currently living. Social status, in itself influenced
by several factors, is in between these two groups since it is determined,
among other things, both by the level of education and by the type of occupation.
In Denmark, Sweden and Finland, the mean number of all eating events on
weekdays was, as we have seen, almost the same (3.9), whereas in Norway it
was slightly less (3.7). In all countries, there were some significant variations
between population groups, but the variations were not the same. Neither
were the differences very great (See Table 3-12 to Table 3-17 in the attachment to this chapter). As far as the daily rhythm of eating and the average
number of different eating events were concerned, Norway and Finland differed quite noticeably from each other. In some other respects, however, they
resembled each other. In both countries 1) women ate more frequently than
men, 2) those who lived together with others ate more often than those living
alone and 3) couples with children ate more often than couples without children, who ate more times than single parents with children, who again ate
more times than singles without children. In both the countries the difference
in the mean number between the two extreme types of households was 0.4
times. In Finland the inhabitants of bigger cities (over 100,000 inhabitants) ate
more often than others. This was, however, not the case in Norway.
106
Eating Patterns
Denmark was the most homogeneous country. The only significant differences were found between age groups (those who were under 35 ate less frequently than older generations) and between municipalities of different size:
unlike in Finland, people living in big cities (over 100,000 inhabitants) ate
less often than others.
In Sweden, just like in Norway and Finland, women ate more often (4.0) than
men (3.8). (This was the only difference that was found in as many as three
countries.) The differences between age groups in Sweden were similar to
those found in Denmark (the middle-aged ate more frequently than others)
with the exception of the oldest age group, which in Sweden is bigger than in
other Nordic countries. Those in Sweden who were 65 or over ate almost as
rarely as the youngest age groups. Only in Sweden were there systematic differences between groups of different income level: on average the higher the
income, the more often did one eat on the average. Only in Sweden did the
occupationally active population eat more often than those at home (4.0 vs.
3.8).
Geographical regions differ rather little from each other in Denmark, Norway
and Sweden. The only country where there were rather clear and statistically
significant differences between the main geographical areas was Finland: In
Finland the most urban region, Uusimaa, on the one hand, and the least
densely populated Northern region, on the other, followed their own patterns.
In the county of Uusimaa in which the Finnish capital Helsinki is situated
people ate less often on average than in other parts of the country (Table 3-6).
The average number of hot eating events differed even more markedly: they
were highest in Northern and Eastern Finland, the most rural and least densely
populated regions, and lowest again in Uusimaa, the capital area. This was
caused by the fact that in Uusimaa there were remarkably more people who
had not eaten hot food at all during the day, whereas Northern Finland led in
the number of people who had eaten hot food more than once. Only one
fourth had eaten hot food at the most popular time period between 4 and 5
p.m. in Uusimaa, whereas half of the people had done so in the Northern region (Table 3-7). As could be expected, eating cold food was more rare in the
Northern region but it was not much more frequent in Uusimaa either. The
other three regions of the country all had higher averages (Table 3-6). On the
basis of these figures one could draw the conclusion that eating hot food more
than once has been the more traditional pattern in Finland. Those who have
started to eat hot food less often have decreased the total number of their daily
eating events rather than increased the number of cold eating events.
Gronow & Jääskeläinen: The Daily Rhythm of Eating
107
Table 3-6 The mean number of different types of eating events in Finnish regions. All
days
Region
Uusimaa
Southern F
Eastern F
Middle F
Northern F
All eating events
Hot eating events
Cold eating events
3.7
4.0
3.9
3.9
3.9
1.3
1.4
1.5
1.4
1.5
1.5
1.6
1.6
1.6
1.4
Total
3.9
1.4
N
1200
1200
Differences between regions statistically significant, p<0.05.
1.6
1200
Table 3-7 The percentage of those persons who had eaten hot food at dinner time (4-6
p.m.) in Finnish regions
REGION
Uusimaa
Southern Finland
Eastern Finland
Middle Finland
Northern Finland
Whole Finland
N
Percentage
24
30
43
39
50
34
333
As the above discussion has shown, differences between social groups were
often rather small and the significance of different background variables varied greatly between the countries. Furthermore, many background variables
did not influence in any way the mean number of eating occasions in the Nordic countries. For instance, there were no statistically significant differences
between groups with different levels of education or social status or between
those who had a job or studied and those who stayed at home (with the exception of Sweden).
Not very surprisingly then, as far as the average number of eating occasions is
concerned, the Nordic countries are internally rather homogeneous on the
whole. The differences that were noticed were, furthermore, rather difficult to
interpret in many cases in particular since they did not point in the same direction in all the countries. It is, for instance, rather difficult to say why on average middle-aged people eat more times both in Sweden and Denmark but not
in Finland and Norway. (The differences were not statistically significant in
Finland.) It was easier to understand why people living together or married
couples (and married couples with children in particular) eat more times than
single individuals. One would suppose there would be more 'order' in the lives
of couples and people living together with others in other respects, too. This
Eating Patterns
108
was the case in all the four countries, but the differences were not statistically
significant in all of them. Women ate more often than men, but again the difference was not statistically significant in Denmark.
Some of these differences can be better interpreted if other background variables are standardised or held constant. In Table 3-8 the results of a linear regression analysis are presented. If all the countries are taken together,
1) the differences in the number of eating events between the countries remain the same even if other variables are standardised. In Norway people
ate less frequently than in the other countries;
2) older age groups ate more frequently than younger ones, the two oldest
ones about 40% more than the youngest (under 25 years);
3) women ate more often than men;
4) the occupationally active ate more often than those who were not occupationally active;
5) those living with someone ate more often than those living alone.
The explanatory value of this model was, however, low, which was to be expected because the original differences were small.
Table 3-8 The number of eating events in the Nordic countries on weekdays. Linear regression analysis. A cleaned model**
Variable
Category
COUNTRY
(ref. Norway)
Denmark
Finland
Sweden
0.21*
0.19*
0.23*
AGE
(ref. 15-24 years)
25 - 34
35 - 44
45 - 54
55 - 65
65 +
0.19*
0.30*
0.31*
0.41*
0.39*
GENDER (ref. male)
Female
0.19*
OCCUPATIONAL STATUS
(ref. non-active)
Active
0.15*
LIVING (ref. alone)
Together with other(s)
0.21*
Constant
Adj. R2
N
Standardised B-value*
2.91
3%
3 570
* statistically significant, p<0.05
** the variables education and social status are not statistically significant and are not included
Gronow & Jääskeläinen: The Daily Rhythm of Eating
3.4
109
Eating events throughout the day
3.4.1 The first eating event: similarity between the countries
Since the questionnaire was designed in such a way that the first eating event
of the day could be either cold or hot it will be analysed separately. In constructing the questionnaire it was thought that the first eating event (which
might in most cases reasonably be called the breakfast) would mostly consist
of cold food such as bread with or without cheese or ham, etc. plus cereals,
yoghurt and/or hot porridge. This assumption, which was confirmed by our
findings, was based on the authors' everyday knowledge of the respective
food cultures. Bread was overwhelmingly the most common item in the
menus of these first eating events; in Denmark and Finland almost two-thirds
had eaten bread, whereas in Norway and Sweden the figures were even higher
(about three fourths had eaten bread). In other respects the Scandinavian
'breakfasts' differed from each other to some extent. Finns and Swedes also
ate porridge (about 15%), Danes and Swedes cereals (corn flakes etc. 16% vs.
19%). Yoghurt was rather common in all countries except Norway (most
common in Sweden, 17%). Danes and Finns might also eat some cakes and
other sweet bread in the morning. Once again the diet of the first eating occasion was most monotonous in Norway. The other three countries had more
varied diets but even with them bread was the only item that was eaten by the
majority of the people.30
As was to be expected, the first eating event took place in all the countries
rather early in the morning, on weekdays between 6 a.m. and 10 a.m., on Sundays a little later from 7 a.m. till 11 a.m.
In Finland the eating day started on average a little earlier and in Norway a
little later than in the other countries, both on weekdays and on Sundays.
These differences were, however, not great. Both in Finland and Sweden 78% ate something for the first time before 6 a.m. on weekdays (in Denmark
and Norway somewhat less.) The 'peak hour' was on weekdays from 7-8 a.m.,
whereas on Sundays it was from nine to ten in the morning. In Denmark the
peak occurred one hour earlier on Sunday (Figure 3-1, Figure 3-2).
30
See in more detail, chapter 2.
Eating Patterns
110
3.4.2 The daily eating pattern
The order of the different kinds of an individual's total daily eating events
could be called the eating pattern in a narrow sense. The first eating event,
which in most cases could reasonably be called a breakfast (B), could be followed by any number of eating events alternating between hot (H) and cold
(C). Depending on the daily amount and the internal order of such occasions a
great number of widely different patterns were possible, starting from either
B, BC or BH and ending up in combinations like BCCHCHHCC or
BCHCHCCHC etc. However, since, on the one hand, very few people had
eaten only once or twice or, on the other hand, more than six times the most
common combinations were to be found among those consisting of three to
six eating events.
Table 3-9 The relative proportion of the most common daily patterns of eating events in
the Nordic countries on weekdays and on Sundays (per cent)
Weekdays
Pattern
Total (N)
Sundays
Pattern
Total (N)
Denmark
BCHC
14
BCCHC
14
BCH
13
BCCH
10
BCCCHC
9
Other
40
100(736)
Finland
BHC
BHCC
BHCHC
BHHC
BHH
Other
BCH
BCHC
BCCHC
BCCH
BHC
Other
BHCC
17
BHC
15
BHCH
9
BHCCC
6
BCH/BHHC 5/5
Other
38
100(211)
17
14
10
9
8
42
100(257)
10
9
8
8
6
59
100(977)
Norway
BCHC
26
BCH
13
BHC
11
BCCHC
7
BH/BCHCC 5/5
Other
38
100 (957)
Sweden
BHC
BCHC
BHH
BHHC
BHCC
Other
BHC
BHCH
BH
BHCC
BCH
Other
BHC
BH
BHCC
BCH
BCHC
Other
25
14
12
12
8
19
100(221)
10
10
10
7
6
57
100(904)
13
10
9
9
8
51
100(168)
On weekdays, the five most common patterns covered about 60% of all the
cases in Denmark and Norway and only about 40% in Finland and Sweden
(Table 3-9). This means that the two countries in which cold food predominates at lunch hour (the so-called cold food cultures) would seem to be, rather
surprisingly, in this respect internally more unified than the other two. On
Sundays larger parts of the population adhered to the five most popular patterns, and in this sense eating was more uniform in all the countries except for
Denmark. On Sundays, the most popular patterns covered as many as 81% of
all the cases in Norway. These patterns were in general, however, not the
same ones on Sundays as on weekdays and their internal order of 'popularity'
was also different. Only in Norway did as many as about one-fourth of the
Gronow & Jääskeläinen: The Daily Rhythm of Eating
111
respondents follow the most common pattern both on weekdays and Sundays.
These patterns were, however, not the same on weekdays and on Sundays.
The most frequent 'Sunday pattern' was simpler (BHC) than the 'weekday pattern' (BCHC). In the other countries there was more variation and dispersion.
The greater concentration of events in fewer patterns in Norway followed, at
least partly, from the smaller average number of eating events in this country.
3.4.3 The two cultures: hot vs. cold food at lunch hour
The analysis of the daily rhythms of eating has revealed that there are certain
peak hours during which a large part of the population reported that they had
eaten something (Figure 3-1). Generally, on Sundays the peaks were not as
clearly discernible as on weekdays, and eating was more evenly spread out, in
particular, throughout the afternoon and early evening (Figure 3-2). The only
times of eating which were generally shared in all the Nordic countries were,
in fact, early in the morning ('breakfast time') and at noon ('lunch time'). Later
in the afternoon and evening there was more variation between the countries.
The Nordic countries can clearly be divided into two distinct groups. The first
includes Denmark and Norway, in which cold food is often eaten at noon.
Finland and Sweden belong to the second: here hot food is more common at
that time (Table 3-10).
Table 3-10 The number of persons who had eaten cold and hot food at lunch hour (from
11 a.m. to 1 p.m.) on weekdays in the Nordic countries (per cent)
Country
Cold food
Hot food
N
Denmark
55
13
Finland
17
45
Norway
52
7
Sweden
22
42
Differences between countries are statistically significant, p<0.05.
736
977
958
903
Despite some similarities, once again the differences between population
groups were generally not similar, but instead varied from country to country.
(See tables in the attachment to this chapter). For instance, in both Norway
and Denmark the age groups between 35 and 44 were the ones who had eaten
cold food during 11 a.m. and 1 p.m. (60% and 62% of the total, respectively).
Only in Norway did the oldest age group eat remarkably less than all the others (39%). In Norway, there was a big difference between those with only basic education and those with higher education (39% vs. 62%), whereas in
112
Eating Patterns
Denmark the less educated ate cold food at noon slightly more often. In general, the differences between population groups were much smaller in Denmark than in Norway.
In Finland, where a cold 'lunch' was much more rare (in the social groups analysed its occurrence varied from about one-tenth to one-fifth of the total) women and lower white collar workers ate cold 'lunch' more often than other
groups. In Sweden, where it was a little more common, the only significant
differences were found between different age groups: the two youngest groups
(under 35 years) were least likely to eat cold food at noon.
There were rather few people in Denmark and in Norway who reported that
they ate hot food at around noon. On the other hand, in Finland and Sweden
almost half the population ate hot food between 11 a.m. and 1 p.m.. There
were very few significant differences between population groups: in Finland
both the youngest and the oldest age groups ate it most frequently, in Sweden
those who were occupationally active ate it more often than those who did not
go out to work (45% vs. 34%).
A logistic regression analysis with six background variables makes it possible
to study the influence of each separate background variable when the others
are standardised. As the results show, in Norway (see attachment Table 3-15)
eating cold food at lunch hour is more common with people of higher the educational status. It is also noticeably more frequent among those who are occupationally active and among those who live together with others. The oldest
age group (over 65) resembles the other groups more once the differences in
their social and occupational statuses are taken into account.
Denmark (see attachment Table 3-15) was more homogeneous as far as eating
cold food at noon was concerned, but some differences were accentuated once
other factors were controlled for. The influence of age becomes more marked
(eating a cold food is more common in the middle aged groups). The result
that lower and middle white-collar workers ate cold 'lunch' more often than
both manual workers and higher white-collar workers remained the same even
if the influence of other factors was controlled for.
In Sweden and Finland, where eating 'hot' food is more common, the original
difference between age groups diminished once the influence of social status
was controlled for. In the end, there were no statistically significant differences among the social groups in Finland (see attachment Table 3-16) concerning eating hot food at noon. In Sweden (Table 3-16), both those with
white-collar status and those who were occupationally active (as opposed to
Gronow & Jääskeläinen: The Daily Rhythm of Eating
113
those not employed) preserved the higher frequency of eating 'hot' at noon in
the same model. (The higher the status, the more often hot food was eaten.)
In Finland and Sweden, where eating cold food at noon was much rarer than
in the other two countries, very few clear differences could be found among
the social groups in this respect. In Finland (Table 3-15) women, middle
white-collar workers and the age group between 55 and 65 ate 'cold' lunches
significantly more often than others. When other factors are controlled for,
however, the only remaining difference was between men and women:
women ate cold food at lunch hour much more often than men. This could
probably be explained by the tendency of educated women to be more likely
to eat 'lighter' vegetarian food and salads (Holm and Møhl 1994, Lien,
Bjørkum, and Bye 1998, 60-63). The differences between the age groups also
remained but in this model they were not statistically significant.
In Sweden (Table 3-15), the only noticeable differences were between people
of different ages. The two youngest age groups ate 'cold' lunches at noon more
seldom than others. These differences remained even after controlling for the
possible influence of other factors. (This might have something to do with
organised institutional lunches for younger generations; however this concerned even those as old as 25-34 years.)
In Denmark and Norway, eating hot food at lunch hour was rare, and the
small number of these cases decreased the power of the statistical analysis.
Some remarks can, however, be made. In Denmark (Table 3-16) the oldest
age group were more likely to eat hot food more often at noon (27% ate hot
lunches in this group), whereas those with an academic education, those who
were occupationally active and those living with someone else ate it less often
than other groups. When the effect of other variables is controlled for only
older people differed significantly from others. Eating hot food at lunch has
earlier been traditional among people living in the countryside.
Also in Norway (Table 3-16), where even fewer people ate hot food at noon,
the oldest age group (over 65) ate it more often than others, but even among
them only 13% did so. Those with further education ate it significantly less
often (but on the other hand the group with higher academic education did not
differ significantly from the least educated.) These differences remained even
though other factors were controlled for. In both Denmark and Norway, eating
hot food at noon seemed to be more common among the oldest age group and
it did not depend on their work or life situation (whether they were retired,
stayed at home, lived alone, etc.).
114
Eating Patterns
As has already been pointed out, even though there were clear differences in
the frequencies of eating cold and hot food at around noon these results are
difficult to interpret. Firstly, they were not systematic in the sense that they
would point in the same direction in all the countries. Quite the contrary, they
seem mostly to represent 'national peculiarities'. Secondly, the differences are
not easy to explain. In most cases there is no self-evident, 'natural' reason why
things should be as they were. For instance, why should one expect people
with higher social status to eat remarkably more often 'cold' at noon in Norway when at the same time in Denmark both highest and lowest status groups
are leading hot food eaters? Or why should higher social status groups as well
as those who are employed eat hot more often in Sweden but not in Finland?
There are different institutional arrangements for eating during the regular
working hours in the Nordic countries. Both in Sweden and Finland it is more
common to provide hot food at canteens at the work place or at an educational
institution than in Denmark and Norway. Publicly subsidised and collectively
organised eating at one's work place or at school is common in Finland and
Sweden. It could therefore be expected, for instance, that people who are either employed or attending a school would more frequently than others fit into
their 'national eating pattern' (e.g. eating 'hot' at noon in Finland and Sweden.)
This was, however, the case only in Sweden with those who were occupationally active.
3.4.4 The eating of hot food at dinnertime
In all the Nordic countries, there was a third peak consisting mainly of hot
food eating events. It was more marked in Denmark and Norway than in
Finland and Sweden, i.e. in countries that share the tradition of eating cold
food at noon. The peak occurred, however, at slightly different times in the
different countries. Finland and Norway were the earliest (4-6 p.m.), followed
by Denmark (6-8 p.m.) and Sweden (5-7 p.m.) one or two hours later (Figure
3-1, Figure 3-2).
Table 3-11 shows the percentages of people who had eaten hot and cold food,
respectively, at these peak hours in the four countries.
Gronow & Jääskeläinen: The Daily Rhythm of Eating
115
Table 3-11 The percentage of people who had eaten hot and cold food in Finland and
Norway at 4-6 p.m., Denmark at 6-8 p.m. and Sweden at 5-7 p.m. on weekdays
Hour
Denmark
Finland
Norway
Sweden
6-8 p.m.
4-6 p.m.
4-6 p.m.
5-7 p.m.
Hot
Cold
58
34
44
37
N
13
18
12
20
736
976
960
904
In Norway (see attachment Table 3-17), it was more common to eat hot food
at 4 to 6 p.m. among those who had a higher education, had the highest social
status, were employed and lived with someone else. Age had the opposite effect: in the oldest groups it was rarer to eat hot food at dinnertime. When the
effect of living alone or with someone else is controlled for, the differences
between age groups were not, however, statistically significant any more.
Denmark (Table 3-17) resembled Norway in other respects than the influence
of education: differences between educational groups were not statistically
significant as such. In the model with six independent variables only age and
education turned out to have an effect on eating hot food at dinner time: the
eldest did so less frequently, those with further education more.
In Finland (Table 3-17) the effect of education was the opposite of that in
Denmark and Norway: those with higher education ate significantly less frequently hot food at the peak hours than others. Living with others increased
the probability of having eaten hot food at dinnertime too. It was also higher
in the 'middle aged' groups (35-54 years). These differences remained almost
the same even after the influence of other factors was controlled for.
In Sweden (Table 3-17) only 'living with someone else' significantly increased
the probability of having eaten hot food between 5 and 7 p.m.. When the influence of other factors was taken into account this difference remained and
the higher frequency of eating hot food in the second oldest age group (55-65)
was also accentuated and became statistically significant.
Living with someone else was the only factor that increased the probability of
eating hot food at dinnertime in all the four countries. (In Denmark it was,
however, not statistically significant when the effect of other factors was controlled for.) Age had some influence in all countries too but not in the same
way. Education and social status had the opposite effect in Norway compared
to Finland. Sweden and Denmark were more homogeneous in this respect
than the other two countries. Again these results are rather difficult to inter-
Eating Patterns
116
pret except for the fact that when one lives together with other adults and
children one is more apt to eat one's food hot in the later afternoon or early
evening. It could be presumed that more effort is directed towards more social
than individual eating occasions.
3.4.5 Nordic eating later in the evening
After the peak of hot eating events during late afternoon or early evening
hours, eating went on until late in the evening (11 p.m.) almost uninterrupted:
during each hour there were quite a large number of people who ate something. A small 'peak' occurred, however, later in the evening between 8-9 p.m.
(In Denmark one hour later between 9 and 10 p.m.). At this time of the evening more eaters ate cold food (30-40% of all) in all the countries and only a
few hot (under 10% of all). Since there were rather many 'cold food eaters' at
dinnertime too it is on this basis alone difficult to say when dining is transformed into an evening meal or a snack in these countries. After 11 p.m. very
few - and after 12 p.m. hardly any – people ate anything at all.
3.5
Conclusion
The main results of this analysis can be summarised as follows:
a) The biggest differences were to be found between the countries or
groups of countries. In many respects Finland and Sweden, on the one hand,
and Denmark and Norway, on the other hand, resembled one another. Denmark and Norway are cold lunch cultures, Finns and Swedes eat their lunches
more often hot. Therefore one could reasonably speak about distinctly different cultural traditions of eating in the Nordic countries.
b)There are some commonly shared eating times and patterns in each
country. At the same time there is a lot of internal variation in the population
as a whole and among different social groups. Hardly anyone eats at night
when most people are asleep. The normal eight hour work day and school
hours also have a clear influence on the daily rhythm of eating extending outside the occupationally active population.
c) Despite some systematic variation between different social groups
or population groups these differences are often not very big and they differed
from country to country. The noticed variation could result from such systematic differences between some population groups that were not, however, recorded and could not be analysed in this study. It could also be individual by
its nature and mean one - or both - of two things. Either each individual has
Gronow & Jääskeläinen: The Daily Rhythm of Eating
117
his or her own routine of daily eating which he or she usually follows more or
less strictly but these routines vary in a random way between individuals. Or
each individual, while often following some daily routines, also makes a lot of
exceptions to them (for instance, skipping lunch or dinner or adding an inbetween snack every now and then). In both cases, modern Nordic eating habits - as far as their timetable is concerned - are a good example of such flexible social habits or conventions which, obviously, are not subject to any
strong normative regulation nor to any strong structural constraints.
Many studies have presumed that such eating habits, like a family meal taking
place at a regular hour every day, have been more binding or compulsory in
earlier times in Europe. As Anne Murcott has rightfully pointed out, there is,
in fact, very little systematic historical evidence of the real occurrence of such
normative or empirical patterns of eating (Murcott1997). Some recent historical studies have, however, suggested that a five-meal pattern seemed to have
been the typical one in pre-industrial Europe at the turn of the century both in
commercial cities and among peasants in the countryside (Sjögren deBeauchaine 1988). During the last century a daily pattern of three meals, typically
found in our study too, slowly evolved from it (Rotenberg 1981, Prättälä and
Helminen 1990). Presuming that older age groups or, say, people with lower
educational or social status would be more traditional in their behaviour one
would have expected them to eat more often and at more regular hours than
younger and better educated, urban people. This was, however, generally not
the case in our study.
One explanation for the existence of such socially shared routines of which
regular eating hours are a good example is the need for social co-ordination. It
would be natural to presume that, for instance, people living or working together and sharing many other daily routines would be more likely to follow
the same eating pattern too. Even though there was some evidence supporting
such a conclusion it was not very systematic or convincing. In all countries
living together with others seemed to be the one factor which more than anything else increased the probability of following socially shared eating patterns.
After all, the daily rhythm of eating would seem to be an ideal example of
such flexible social routines or habits which have been pondered by the sociological classics (Weber 1968, 29-31): for some reason, people often follow
some common routines or conventions even when there are no strong constraints that would compel them to do so.31
31
See (Gronow & Warde 2001)
Eating Patterns
118
Table 3-12 Mean numbers of all eating events in the Nordic countries in different population groups. Weekdays
All eating
events
Sign. p<0.00
Gender
Mean
St.dev.
N
Men
Women
Sign.
Occupatio- Non-active
nal status
Active
Sign
Living
Alone
With someone
Sign.
Age
15-24
25-34
35-44
45-54
55-64
65+
Sign.
Education
Mandatory
Further
Academic
Sign.
Income
Low
level
Middle
High
Sig.
Social
Worker
status
Low white collar
Middle white collar
High white collar
Sig.
Household Couple with children
type
Couple without children
(children
Single with children
<20 years)
Single living alone
Sig.
Size of
> 24 999
muni25 000 - 99 999
cipality
100 000 >
Sig.
Sig. *** p<0,001, ** p<0,01, * p<0,05
Denmark
3.94
1.08
Finland
3.93
1.15
Norway
3.74
0.98
Sweden
3.93
1.06
3.89
3.99
n.s.
4.03
3.90
n.s.
3.83
3.98
n.s.
3.60
3.85
4.12
3.98
4.04
4.01
**
3.89
3.94
4.00
n.s.
3.91
4.05
3.98
n.s.
4.03
3.89
3.99
3.90
n.s.
4.13
3.99
3.97
3.83
n.s.
3.98
4.08
3.77
*
3.87
4.08
***
3.87
3.96
n.s.
3.76
4.03
***
3.84
3.97
4.02
4.05
4.05
3.87
n.s.
3.91
3.94
3.92
n.s.
3.99
4.09
4.02
n.s.
3.97
3.99
4.04
3.89
n.s.
4.08
3.95
3.90
3.73
**
3.89
3.84
4.08
*
3.60
3.88
***
3.71
3.75
n.s.
3.57
3.80
***
3.60
3.71
3.78
3.73
3.84
3.80
n.s.
3.64
3.71
3.83
n.s.
3.64
3.76
3.79
n.s.
3.64
3.85
3.75
3.78
n.s.
3.95
3.75
3.66
3.57
***
3.77
3.71
3.71
n.s.
3.88
3.98
0
3.82
3.97
*
3.83
3.99
*
3.62
3.87
4.04
4.14
4.14
3.91
***
3.98
3.97
3.88
n.s.
3.79
3.93
4.10
***
3.93
3.94
4.01
4.02
n.s.
4.05
3.99
3.89
3.82
n.s.
4.03
3.94
3.86
n.s.
Gronow & Jääskeläinen: The Daily Rhythm of Eating
119
Table 3-13 Mean numbers of cold eating events in the Nordic countries in different population groups.
Cold eating
events
Sign.
p< 0.00
Gender
Mean
St.dev.
N
Men
Women
Sig.
Occupatio- Non-active
nal status
Active
Sig.
Living
Alone
With someone
Sign.
Age
15-24
25-34
35-44
45-54
55-64
65+
Sig.
Education
Mandatory
Further
Academic
Sig.
Income
Low
level
Middle
High
Sig.
Social
Worker
status
Low white collar
Middle white collar
High white collar
Sig.
Household Couple with children
type
Couple without children
(children
Single with children
<20 years)
Single living alone
Sig.
Size of
> 24 999
muni25 000 - 99 999
cipality
100 000 >
Sig.
Sig. *** p<0,001, ** p<0,01, * p<0,05
Denmark
1.99
1.11
Finland
1.58
1.11
Norway
1.73
0.99
Sweden
1.56
1.10
1.92
2,06
n.s.
2.10
1.94
n.s.
1.90
2.02
n.s.
1.52
1.92
2.18
2.04
2.10
2.08
***
1.93
1.99
2.08
n.s.
1.98
2.09
2.04
n.s.
2.06
1.91
2.10
2.04
n.s.
2.12
2.05
2.13
1.90
n.s.
2.04
2.13
1.81
*
1.43
1.72
***
1.54
1.60
n.s.
1.47
1.60
n.s.
1.32
1.62
1.57
1.68
1.71
1.47
*
1.49
1.61
1.63
n.s.
1.59
1.73
1.61
n.s.
1.60
1.62
1.63
1.56
n.s.
1.64
1.67
1.42
1.48
n.s.
1.53
1.52
1.74
*
1.55
1.89
***
1.78
1.71
n.s.
1.69
1.74
***
1.54
1.67
1.76
1.71
1.87
1.84
n.s.
1.69
1.70
1.80
n.s.
1.74
1.73
1.69
n.s.
1.64
1.86
1.72
1.71
n.s.
1.83
1.76
1.60
1.68
n.s.
1.73
1.75
1.70
n.s.
1.43
1.69
***
1.56
1.57
n.s.
1.56
1.56
n.s.
1.14
1.38
1.66
1.84
1.79
1.69
***
1.68
1.61
1.48
n.s.
1.49
1.55
1.66
n.s.
1.60
1.60
1.64
1.62
n.s.
1.57
1.64
1.39
1.56
n.s.
1.64
1.56
1.51
n.s.
Eating Patterns
120
Table 3-14 Mean numbers of hot eating events in the Nordic countries in different population groups.
Hot eating
events
Sig.p< 0.00
Gender
Mean
St.dev.
N
Men
Women
Sig.
Occupatio- Non-active
nal status
Active
Sig.
Living
Alone
With someone
Sig.
Age
15-24
25-34
35-44
45-54
55-64
65+
Sig.
Education
Mandatory
Further
Academic
Sig.
Income
Low
level
Middle
High
Sig.
Social
Worker
status
Low white collar
Middle white collar
High white collar
Sig.
Household Couple with children
type
Couple without children
(children
Single with children
<20 years)
Single living alone
Sig.
Size of
> 24 999
muni25 000 - 99 999
cipality
100 000 >
Sig.
Sig. *** p<0,001, ** p<0,01, * p<0,05
Denmark
0.95
0.40
Finland
1.43
0.66
Norway
1.02
0.49
Sweden
1.37
0.62
0.97
0.93
n.s.
0.94
0.96
n.s.
0.94
0.96
n.s.
1.07
0.92
0.94
0.94
0.94
0.93
n.s.
0.96
0.97
0.92
n.s.
0.93
0.96
0.94
n.s.
0.97
0.98
0.89
0.87
n.s.
1.00
0.95
0.84
0.94
n.s.
0.94
0.95
0.96
n.s.
1.50
1.38
***
1.41
1.45
n.s.
1.31
1.46
**
1.55
1.39
1.50
1.39
1.36
1.41
n.s.
1.48
1.42
1.40
n.s.
1.41
1.40
1.44
n.s.
1.41
1.39
1.43
1.40
n.s.
1.53
1.37
1.54
1.31
***
1.45
1.43
1.41
n.s.
1.05
0.99
n.s.
0.96
1.04
*
0.90
1.06
***
1.06
1.04
1.03
1.02
0.99
0.97
n.s.
0.99
1.02
1.04
n.s.
0.91
1.03
1.10
***
1.00
1.00
1.03
1.08
n.s.
1.12
1.00
1.08
0.90
***
1.05
0.98
1.00
n.s.
1.46
1.29
***
1.26
1.42
***
1.27
1.42
***
1.48
1.49
1.37
1.30
1.36
1.23
***
1.30
1.36
1.41
n.s.
1.30
1.39
1.44
*
1.34
1.33
1.37
1.40
n.s.
1.49
1.35
1.50
1.27
***
1.39
1.38
1.35
n.s.
Gronow & Jääskeläinen: The Daily Rhythm of Eating
121
Table 3-15 An explanatory analysis of cold lunch eating (11am-1pm) in the Nordic countries on weekdays. Logistic regression of all countries together and each separately
All countries
Denmark
Finland
Norway
Sweden
Variable
Category
COUNTRY
ref.
Norway
Denmark
Finland
Sweden
Separate Odds
influence ratio
of variable
1.09
1.23*
0.18*
0.19*
0.26*
0.29*
AGE
ref.
15 – 24
25 – 34
35 - 44
45 – 54
55 – 65
65+
1.31*
1.67*
1.59*
1.60*
1.31*
1.18
1.61*
1.64*
1.78*
1.46*
n.s.
+
+
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
+
+
+
GENDER
ref.male
Female
1.13
1.26*
n.s.
+
n.s.
n.s.
1.15
1.33*
1.05
1.16
n.s.
n.s.
n.s.
n.s.
+
+
n.s.
n.s.
EDUCATION Further
ref.
Academic
mandatory
Significant variation (p<0.05)
(+ positive, - negative, n.s. not significant)
SOCIAL
STATUS
ref. worker
Lower white 0.91
Middle white 1.24*
Higher white 0.77*
0.78*
1.07
0.91
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
OCCUP.
STATUS
ref. nonactive
Active,
student
1.24*
1.21
n.s.
n.s.
+
n.s.
LIVING
ref. alone
With
someone
1.13*
1.35*
n.s.
n.s.
+
n.s.
N
Sig. of the
model
3 576
3 241
680
857
874
830
0.000
0.000
0.041
0.234
0.000
0.100
Eating Patterns
122
Table 3-16 A description of hot lunch eating (11 a.m. –1 p.m.) in Nordic countries on
weekdays according to different background variables. Logistic regression of all countries
together and each separately
All countries
Denmark
Finland
Norway
Sweden
Variable
Category
COUNTRY
ref.
Norway
Denmark
Finland
Sweden
Separate Odds
influence ratio
of variable
2.10*
1.85*
11.82*
11.52*
10.38
9.53*
AGE
ref.
15 – 24
25 – 34
35 - 44
45 – 54
55 – 65
65+
0.69*
0.74*
0.60*
0.82
0.94
0.96
0.94
0.70*
1.06
1.74*
n.s.
n.s.
n.s.
n.s.
+
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
GENDER
ref.male
Female
0.94
0.88
n.s.
n.s.
n.s.
n.s.
0.69
0.83
0.80*
0.93
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
EDUCATION Further
ref.
Academic
mandatory
Significant variation (p<0.05)
(+ positive, - negative, n.s. not significant)
SOCIAL
STATUS
ref. worker
Lower white 1.07
Middle white 1.05
Higher white 1.75*
1.25
1.25
1.34*
+
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
+
OCCUP.
STATUS
ref. nonactive
Active,
student
0.98
1.36*
n.s.
n.s.
n.s.
+
LIVING
ref. alone
With
someone
1.08
1.08
n.s.
n.s.
n.s.
n.s.
N
Sig. of the
model
3576
3241
680
857
874
830
0.000
0.000
0.000
0.71
0.069
0.042
Gronow & Jääskeläinen: The Daily Rhythm of Eating
123
Table 3-17 An explanatory analysis of hot dinner eating in the Nordic countries. Relative
hot dinner eating in the Nordic countries on weekdays (odds ratio. Logistic regression of
all countries together and each separately
All countries
Denmark
Finland
Norway
Sweden
Variable
Category
COUNTRY
ref.
Norway
Denmark
Finland
Sweden
Separate Odds
influence ratio
of variable
1.73*
1.88*
0.65*
0.64*
0.74*
0.81*
AGGE
ref.
15 – 24
25 – 34
35 - 44
45 – 54
55 – 65
65+
1.23
1.24
1.28*
1.02
0.77*
1.36*
1.31
1.38*
1.21
0.95
n.s.
n.s.
n.s.
n.s.
-
+
+
+
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
+
n.s.
GENDER
ref.male
Female
0.88
0.92
n.s.
n.s.
n.s.
n.s.
1.34*
1.31*
1.25*
1.17
+
n.s.
n.s.
-
+
+
n.s.
n.s.
EDUCATION Further
ref.
Academic
mandatory
Significant variation (p<0.05)
(+ positive, - negative, n.s. not significant)
SOCIAL
STATUS
ref. worker
Lower white 0.98
Middle white 0.99
Higher white 0.95
0.93
0.95
0.96
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
+
OCCUP.
STATUS
ref. nonactive
Active,
student
1.39*
1.06
n.s.
n.s.
n.s.
+
LIVING
ref. alone
With
someone
1.64*
1.60*
n.s.
+
+
+
N
Sig. of the
model
3576
3241
680
857
874
830
0.000
0.000
0.007
0.000
0.000
0.011
124
Eating Patterns
4
The Meal Format
Johanna Mäkelä
4.1
Introduction
Studying food can be motivated from several different perspectives. We could
study the health effects of different types of foods or determine the price of an
average ‘food basket’; or, as we have chosen to do here, we could regard food
and eating habits as a perspective on everyday life and an indicator of current
cultural aspects of everyday life. Without going into a detailed discussion of
the meaning of food, eating and meals, the basic standpoint we have adopted
is to approach food and meal habits as products of culture (see e.g. LéviStrauss 1966). In this sense, eating habits tell us something essential about the
present state and nature of Nordic cultures. The general aim of our project
‘Eating Patterns. A Day in the Lives of Nordic Peoples’, was not to explore
food and eating per se but to use it as a keyhole to modern Nordic living. (See
Chapter 1.) The theme, which runs through all chapters of this book, is the
structuration of eating or rather the destructuration of eating as epitomised in
concepts like ‘grazing’ (Caplan 1997) and ‘gastroanomy’ (Fischler 1990). We
wanted to find out whether traditional regular meals are being substituted by
modern irregular eating patterns and at the same time describe the content and
structure of different types of meals. Specifically, we wanted to know whether
32
Acknowledgments. I want to express my gratitude to Anni Jääskeläinen for her assistance
and expertise in processing and analysing data. Thanks also to Jukka Gronow, Lotte Holm,
Unni Kjærnes, Seppo Pöntinen, and Juhana Vartiainen for their valuable comments and contributions.
126
Eating Patterns
it is possible to find any empirical evidence that could verify the discussion
about grazing, destructuration of eating, and rapid change in food habits in the
Nordic countries. We approached the state of Nordic eating habits by asking
one apparently simple question: are people eating meals? Furthermore, we
were interested in the shared patterns of eating and, on the other hand, in the
variation, even individualism, of eating. At the same time as we compare four
Nordic countries and their social structure in relation to eating, we can pinpoint the general shared features of these (food) cultures. Conversely, the idea
of comparative research allows us to depict the particular characteristics of
each nation’s food habits.
In this chapter, the dimension of the meal format is applied to throw light on
these research questions. Focusing on the content and the structure of hot eating events, we want to analyse the complexity of such eating events in terms
of the number and type of components included. Furthermore, we want to see
to what extent there are common formats in each country and in the Nordic
countries as a whole. Finally, we want to examine the occurrence of meals
that in the British literature are labelled as ‘proper’ ones vs. other types of
meal formats.
During the planning stage of our study we had many intriguing discussions
about the definition and essence of a meal within the Nordic setting. Even
though Denmark, Finland, Norway and Sweden in many ways share a common Nordic cultural background, we soon learned that there are significant
differences in food habits as well as in the vocabulary for eating. We realised
that it would not be possible to operationalise our research question into simple items concerning dinners, lunches or main meals. While similar concepts
are used in each of the four languages, we were not at all sure that they would
be understood in the same way in every country.
In Denmark and Norway, lunch could be a sandwich. The Norwegian variant
of two slices of dark bread with goat cheese eaten with your fingers is more
modest than its Danish cousin ‘smørrebrød’ – an open sandwich which often
appears with many different fillings and often eaten with fork and knife. In
Finland and Sweden, defining the main meal is complicated by the fact that
many people eat hot food both at lunch and dinner. One of these meals could
be a sparse one. A soup alone can be a meal in Finland. Therefore, we decided
that for our questionnaire, we had to deconstruct these everyday concepts in
order to get rid of their culture-specific definitions and reconstruct them in the
process of data analysis.
Mäkelä: The Meal Format
127
What emerged is the model of the eating system, which distinguishes between
three dimensions of the eating system: the eating pattern, the meal format, and
the social organisation of eating. (A more detailed description of the eating
system is given in Chapter 1, see also Mäkelä et al. 1999.) Together, these
three dimensions in our model describe the content, the structure, the schedule, the place and the social context and organisation of eating. These dimensions do not form a hierarchy. In principle each of them could be explored
separately, as we have done in this report. As part of the model, each dimension has a new angle to offer. However, from a sociological point of view, the
most interesting aspect certainly lies in combinations of different dimensions.
We can provide a better answer to the question of how organised or unorganised eating is if we take into account both the eating pattern and the meal format. The sociability of eating will remain outside the analysis if we do not
focus on the social organisation of eating. A thorough analysis of what has
been considered to be a ‘proper’ or ideal meal needs to concentrate both on
the content, i.e., the meal format, and the social context and organisation of
eating.
This chapter begins with an introduction to the theoretical background of the
concept of meal format and then proceed to the methodological design of the
study, focussing on the definitions and recodings relevant to the analyses of
the meal format. Next, the results depict the complexity of hot meals. In addition, there is a discussion of the similarities and differences of meal formats
between countries as well as variation within each country. Here three different classifications of hot meals are presented. Finally, the main findings are
discussed in relation to the main topics of this study: the (de)structuration of
meal patterns and eating as part of modern everyday life.
4.2
Theoretical background
One of the main concerns of sociologically oriented research on food and eating has been with the vivisection of meals. Since the 1970s, British scholars in
particular have been both active and influential in the analysis of the meal
structure (Marshall 2000, for a review of cultural definitions of a meal, see
Mäkelä 2000). However, the whole idea of connecting meals to a wider social
system in which a meal is one ordered system related to other ordered systems
(Douglas 1975) has been the dominant paradigm of sociology of food for the
past 20 years.
128
Eating Patterns
The approach according to which the order of food is connected to the social
order has been eminently clear in Mary Douglas’s work. During the two past
decades no sociological study on food has neglected to refer to Douglas, who
together with Nicod analysed the structure of British working class meals
(Douglas and Nicod 1974). Their criteria for the classification of meals are
complexity, copiousness, ceremoniousness and the structural and sensory
qualities embedded in the binary oppositions savoury/sweet, hot/cold, liquid/dry. To Douglas, meals and drinks are two important opposite categories,
as in the relation between solids and liquids. A meal has both solid and liquid
components and it has to have a dimension of bland – sweet – sour. The
meaning of a meal is based on a system of repeated analogies (Douglas 1975).
A meal is not, however, the only type of eating. Douglas and Nicod have presented four terms to describe different types of eating. 1. A ‘food event’ is an
occasion when food is eaten. 2. A ‘structured event’ is a social occasion organised by rules concerning time, place and sequence of action. 3. Food eaten
as a part of a structured event is a ‘meal’. A meal is connected to the rules of
combination and sequence. 4. A ‘snack’ is an unstructured food event without
any rules of combination and sequence.
The meal system proposed by Douglas and Nicod consists of three types of
meals: (A) a major meal/the main meal, (B) a minor meal/the second meal,
and (C) an even less significant meal/the third meal (biscuits and a hot drink).
(Douglas 1983, 83, Douglas and Gross 1981, 6–7, Douglas and Nicod 1974).
Thus, the model consists of complementary classifications. The first course of
meal (A) has, both on Sundays and weekdays, the same basic structure based
on a staple (potato), a centre (meat, fish or egg), trimmings (vegetables) and
dressing (gravy). Everything is savoury and hot. The second course has the
same structure except that everything is sweet. The staple is cereal, the centre
is fruit and the dressing is liquid custard or cream. Meal (B) follows the structure of meal (A), but the staple is cereal (bread) and not potato.
A few years, later Murcott (1982, 1983 and 1986) studied people’s concept of
‘a proper meal’ in South Wales. In this qualitative study, a proper meal always proved to be a cooked dinner of a certain type. The creation of a proper
meal requires transforming ingredients by cooking and combining them into a
meal. A proper meal consists of one course only, which is always a variation
on meat and two vegetables. ‘Meat’ must be (fresh) meat; sausages or offal do
not qualify. Poultry will do as a last resort, but fish is definitely ruled out. Potatoes and other vegetables are necessary and at least one of them must be
green. Finally, the gravy gives the finishing touch to this plateful and combines all the ingredients into a proper meal.
Mäkelä: The Meal Format
129
In the late 1980s, Charles and Kerr (1988, 19–20) still identified a similar
proper meal as Murcott as well as Douglas and Nicod: a cooked meal with
meat (fish), potatoes and vegetables as opposed to a snack, which is bread
based and merely prepared instead of cooked. Murcott and Charles and Kerr
extended Douglas and Nicod’s ideas by involving the aspect of gender and
food preparation as an important part of proper meals (see Chapter 7).
Interestingly, a recent study on changes in food choice and eating habits during the transition from single to married or cohabiting in Scotland suggests
that the notions of a proper British meal have changed in the 1990s (Kemmer,
et al. 1998). Even though the idea of a proper meal as ‘meat and two veg’ was
self-evident to young couples, this traditional format was seen as a little boring. Influences from non-British food cultures have changed the food habits of
the younger generation. To them, a salad, pasta, chilli or curry meets the criteria of a proper meal (see also Marshall 2000).
Research on meals in the Nordic countries has been inspired by the British
discussion. In a quantitative Swedish study, Ekström (1990) identified four
components in a cooked meal. The main dish gives the name to the course and
is usually animal protein. The first trimming is the starchy base of the meal,
the second trimming consists of vegetables. Extra trimmings are vegetables or
different condiments. Bugge and Døving (2000, 47) used qualitative data to
define a Norwegian proper meal as a cooked one with vegetable trimmings. In
Denmark similar concepts of proper meals have been found in qualitative
studies, and surveys of dietary practices show that the most typical hot meal
consists of meat, potatoes, vegetables, and sauce (Haraldsdottir et al. 1987),
39–40). A qualitative study on eating in Finland concluded that a proper meal
could be defined as ‘a hot dish’ (i.e. cooked food) accompanied by ‘a salad’
(i.e. uncooked vegetables) (Mäkelä 1996). The most important change in the
basic structure of a Finnish meal over the past few decades has been the addition of a salad. However, the transition from rural to urban eating patterns diminished the number of daily hot meals (Prättälä and Helminen 1990). At the
same time meat and fish have taken over in the central role from potatoes,
bread, and porridge (Prättälä et al. 1993).
However, the idea of a proper meal as a plateful is not the convention in all
European countries. In Italy, a traditional proper meal has four or five courses
(see e.g. Fischler 1990, 160). In France, the proper or real meal (un vrai repas)
with a three-course structure is well established in the everyday routines of the
Parisian bourgeoisie (see e.g. Sjögrende Beauchaine 1988). However, Fischler
Eating Patterns
130
(1990) has voiced some concern over the state of French eating during the era
of gastro-anomie.
In order to determine the categories of food, Douglas (1975) classified meals,
courses, helpings, and mouthfuls from the standpoint of a linguistic analogy.
Elaborating on the idea of meals as a language with a grammar, it can be suggested that each meal follows both syntagmatic and paradigmatic rules. The
order of dishes is determined by the syntagmatic rules. The three partite basic
structure of a starter, a main course, and a dessert is known in Scandinavia,
although this structure obviously varies from country to country, from culture
to culture. The paradigmatic rules, then, define what kind of dishes can be
eaten in each category. For example, a dessert can be berry pie or crème
caramel, but fish soup would never be accepted as a dessert.
Interestingly, both of these dimensions seem to be present in the discussion of
contemporary meal formats and in the debate on the existence of proper
meals. Fischler (1990) claims that the traditional order is replaced by disorder
and unpredictability, while the findings by Kemmer et al. (1998) indicate the
emergence of new types of order. It should, however, be noted that most earlier studies have been mainly concerned with characterising typical or even
ideal (proper) meals, rather than with describing the heterogeneity of everyday eating of modern times – as we intend to do.
4.3
Data
4.3.1 Operationalisation
This study is grounded mainly in qualitative and theoretical research on
meals. However, since we were concerned specifically with the extent and
distribution of different types of eating, we wanted to do a quantitative study.
Although we were inspired by the 24-hour recall interview method, the focus
was on the combinations of food and the social context of eating rather than
on the quantities of food consumed or the frequencies of eating. The challenge
was to formulate questions that could reveal the construction of meals within
this quantitative and comparative setting. Our solution was to first deconstruct
eating into number of questions and then to reconstruct the answers into
snacks or meals.
Mäkelä: The Meal Format
131
We needed a concept that would capture the different aspects of the eating
system. In the deconstruction process we decided to concentrate on hot and
cold eating events, i.e. on every occasion of eating something (just a beverage
or only eating chewing gum or sweets, etc, were excluded). As a concept, the
‘eating event’ is inspired by the four different types of eating identified by
Douglas and Nicod (1974): food event, a structured event, a meal, and a
snack. Our analysis of the empirical data would then involve a reconstruction
of eating events into events with varying degrees of complexity and structuration. In everyday parlance these ‘eating events’ could be called ‘dinners’,
‘lunches’ or other kinds of eating events like ‘snacks’.
A description of data collection33 and other details of the methodological design and the questionnaire is given in Chapter 1. Furthermore, some methodological problems of the study are discussed in an earlier article by the project
(Mäkelä et al. 1999). To sum up, the questionnaire consists of loops where
questions on the structure, the content, and the context of eating events were
presented in a certain order according to a number of filters. The eating events
could occur at any time during the 24 hours of recording. The questionnaire
allowed ten rounds of reporting, but in practice the hot eating events were reported from the second eating event to the eighth eating event. The first eating
event was always defined as ‘breakfast’ with no distinction between hot and
cold food (see Chapter 2). The analyses in this chapter are based on reports of
hot eating events.
The questionnaire was so constructed that every informant had to answer the
question concerning the main ingredient of each hot eating event. Indeed, everyone who was asked the question answered it. This question was followed by
a battery of items concerning other ingredients (staple, vegetables, trimmings), ways of preparation, and beverages. Content has already been described in Chapter 2, together with the occurrence and types of starters and
desserts. The entire questionnaire is presented in Appendix 1 to the report.
4.3.2 Components of a meal
In order to reconstruct the hot eating events into ‘hot meals’, we created five
new variables: centre, staple, vegetables, trimmings, and bread, which, ac33
The data were collected in April 1997 with computer assisted telephone interviews (CATI).
A total of 4,800 (1 200 in each country) individuals above the age of 15 were interviewed in
Denmark, Finland (where the age range is from 15 to 77), Sweden and Norway.
132
Eating Patterns
cording to our theoretical perspective, represent the essential components
within the meal structure. One (main) course could then include one to five of
these components. The reconstruction process is described in detail in Chapter
2. The main ingredient of a hot dish is its ‘centre’, the heart that often gives
the dish its name. Meatballs, fish casserole, pork sausages, lamb chops or
vegetable soup tell a lot about the food eaten. ‘Staple’ and ‘vegetables’ are the
side dishes, and together with the ‘trimmings’ they form the main course of a
hot meal. The starter is the first course and the dessert is the third or the last
course of a meal.
While this reconstruction is an extended version of the ideas of proper meals
by British scholars, it does not necessarily imply a traditional (British) proper
meal format of ‘meat and two veg’. Our application differs from its predecessor in two respects. Firstly, in addition to meat, the ‘centre’ could be either
fish or vegetables. We know about the traditionally strong role of fish, and we
also know how vegetables have established themselves in the Nordic diets as
more modern new arrivals. We even added a category ‘other’ to capture the
variety of dishes typical of different Nordic countries. This option included
traditional dishes such as a soup, pancakes, or porridge, but also more recently
introduced dishes such as pasta, pizza, risotto, and taco. Secondly, bread is
lifted up in our application. Douglas and Nicod (1974) placed cereal staple
(bread and pastries) as a part of minor meals (snacks). Although the British
concept is used as a methodological tool, other types of hot eating events
could be recorded within the structure of our questionnaire.
Staple, vegetables, trimmings, and bread are variables constructed from the
question of trimmings. ‘Staple’ can be potatoes, rice, pasta, or pulses. The
variable ‘vegetables’ contains vegetables served as side dishes either
uncooked (‘salads’) or cooked. In the variable ‘extra’ we included other trimmings like sauces, pickles, and other condiments. Bread is bread. The vegetable categories (originally 36) were coded into nine new variables. In addition,
we have information on beverages for each eating event. A new variable of
beverages with seven classes of the original 17 alternatives were created.
Table 4-1 illustrates this recoding process (see also Chapter 2).
Mäkelä: The Meal Format
133
Table 4-1 Meal components
Component Centre
Categories meat
fish
vegetables
other
Staple
potatoes
rice
pasta
beans
and lentils
Vegetables
cabbages
carrot
cucumber
pulses
lettuces
onion
tomato
sweet pepper
other
Trimmings
hot and cold
sauces
pickles, preserves, condiments
other
Bread Beverages
bread alcohol
beverages
hot beverages
juice
milk, buttermilk
water
other
no beverages
These components were used as a basis for classifying different meal combinations. In principle, there are 31 possible combinations with one to five components. Our questionnaire logically ruled out some of these variations, since
the question on the main ingredient acts as a filter. As a result, only 16 combinations were found in the data (Table 4-2). These combinations were then
used to categorise all hot eating events and to describe the complexity of hot
meals.
Table 4-2 clearly illustrates some similarities and differences between countries. Interestingly, there seems to be a clear concentration on certain combinations in all countries. The combination centre + staple + vegetables appeared most frequently in all countries except Finland, where it was the second most common combination. Nonetheless, this combination covered a
relatively minor proportion of all hot eating events. The highest figure, 27 per
cent, was recorded in Norway. The most common Finnish combination was
centre + bread, but even this had a rather low frequency of 15 per cent. In
Finland and Norway, a solitary centre was the second most common combination, whereas in Denmark and Sweden it was centre + staple. Thus, the less
complex formats included an additional staple (or bread) rather than vegetables or trimmings. In Sweden, the combination centre + staple + vegetables +
trimmings was as popular as centre + staple + vegetables. In Finland and
Sweden, hot meals spread across all 16 combinations, whereas hot meals in
Denmark and Norway were more concentrated on certain combinations. Some
combinations were not found at all, in particular those alternatives including
bread.
Eating Patterns
134
Table 4-2 Hot meal combinations on weekdays. Per cent of all hot meals
Combination
Denmark
Finland
Norway
Sweden
Weekdays Sundays Weekdays Sundays Weekdays Sundays Weekdays Sundays
C
10
11
13
12
15
9
8
6
CS
15
15
12
8
11
7
14
11
CV
8
7
9
6
6
8
6
5
CT
8
6
7
7
8
8
8
7
CB
CSV ■
5
6
15
10
6
4
7
8
18
17
13
19
27
29
15
17
CVT
2
3
1
2
2
1
2
1
CST
15
13
4
5
7
9
9
11
0
CSB
1
2
3
6
0
0
1
CVB ■
2
3
5
2
1
1
1
3
CTB
1
1
3
2
1
3
4
2
CSVT ■
14
13
5
8
13
19
15
19
CVTB ■
1
2
2
1
1
0
2
2
CSTB
0
0
2
1
0
0
2
2
CSVB ■
0
1
5
8
0
0
2
2
CSVTB
■
%
0
0
3
2
0
0
4
7
100
100
100
100
100
100
100
100
N1
700
239
1388
260
982
224
1240
199
C = centre (meat, fish, vegetables, other), S = staple (potatoes, rice, pasta, pulses), B = bread,
V = vegetables (as a side dish), T = trimmings (sauces, preserves, condiments)
■ proper meals
1
the total number of hot meals
4.3.3 The proper meal format
The content, structure and complexity of hot meals are described in Chapter 2.
The next step was to find out to what extent Nordic populations have been
eating hot food in a format that could be defined as a meal, or alternatively,
whether the habit of ‘grazing’ is so well established that hot eating events
have come to look more like snacks than meals. We also wanted to know how
the occurrence of different formats varies between countries and, say, between
age groups or between families and single person households.
We started out from the idea of proper meals that was used during the planning stage of our questionnaire. The construction of ‘proper meals’ followed
the general structure indicated by Douglas and Nicod and Murcott. We did
not, however, apply these British ideas of proper meals as such, but created a
Mäkelä: The Meal Format
135
Scandinavian application adjusted to our quantitative frame and to our existing knowledge about Nordic meal habits.
A variable for ‘proper meals’ was constructed. It divides the data set into two
groups: those who had eaten at least one proper hot meal and those who had
eaten only other types of meals. ‘Proper meals’ were defined as containing a
centre, a staple or/and bread, and vegetables, with trimmings as an additional
option. The essential difference here is that a ‘proper meal’ must besides a
centre include both a staple and vegetables, whereas other types of hot meals
could include a centre and either a staple or vegetables or only a centre. The
other types of meals, i.e. ‘not proper’ meals obviously include many different
combinations. Furthermore, it should be emphasised that in our analyses the
properness of meals does not take into account any nutritional aspects of
eating. Neither did we inquire about informants’ cultural ideas about proper
meals in a normative sense. ‘A proper meal’ is a construction that is used as a
tool in our analyses of structure and complexity of hot eating events. Earlier
studies (see above) have established that the idea of proper meals does exist in
the Nordic countries. However, our concern here is not with the question of
how and to what degree these ideas are manifested. Instead, our aim is to
analyse actual modern eating practices by using the construct of ‘proper
meals’ embedded in earlier theoretical discussions and empirical findings.
‘Proper meals’ are depicted in Table 4-2. We will illustrate these four combinations as follows. A CSV could be a fried fillet of fish with rice and green
beans, or it could be a meat and vegetable casserole with boiled potatoes and a
tomato and onion salad. A CVB could be sausages, a salad and a piece of
bread. A CSVT could be oven-baked chicken with chips, a salad of grated
root crops, and pickles. A CSVB could be a vegetable stew with spaghetti, a
cucumber salad, and bread. A CVTB could be oven baked fish with steamed
sugar peas and remoulade sauce and a piece of bread. A CSVTB could be
roast beef with baked potatoes, boiled broccoli, a salad, gravy, and bread – a
Sunday dinner par excellence.
The other types of meals consist of all other combinations than those indicated
in Table 4-2. A solitary C could be porridge or pizza. A CS could be stir-fried
vegetables with rice – or a hot dog. A CVT could be grilled meatballs with
green salad and brown sauce or ‘kumle’ (Norwegian steamed potato balls
with pork or fish) with boiled carrots and melted butter. A CB could be chilli
con carne with bread. A CT could be macaroni casserole with ketchup. A CST
could be fish cakes with mashed potatoes and a sauce.
Eating Patterns
136
The aim is to depict the structure and the content of a meal and, as a result,
portray the state of patterned eating. Table 4-2 describes all these aspects.
Firstly, the number of components describes the structure of a meal. Secondly,
these components are of a certain kind, i.e. they tell us something about the
content of a meal. Furthermore, the number and type of meal components together depict the complexity of a hot meal.
4.4
Meal combinations: a description of the findings
Figure 4-1 summarises the proportions of different components included in all
hot meals in each country. All hot meals contained by definition a centre. A
beverage was almost always reported as part of the hot meals. Interestingly, a
staple was present in two thirds of hot meals in all countries except Finland.
However, the relatively high proportion of bread in Finland may explain this
difference. Another difference is found when we look at trimmings. In Sweden, hot meals included trimmings twice as often as in Finland. Vegetables as
side dishes were most common in Norway, where over half of all hot meals
included vegetables as side dishes.
100 100 100 100
96 97 95 97
100
90
80
70
65
61
64
60
53
49
49
47
50
42
47
43
40
37
34
27
30
20
24
13
10
10
0
Centre
Denmark
Staple
Finland
Vegetables
(side dish)
Norway
Sweden
Trimmings
Figure 4-1 The different components of all hot meals. Per cent
Bread
Beverage
Mäkelä: The Meal Format
137
4.4.1 Number of components
The number of components provides a simple measure of the complexity of
hot meals. Table 4-3 shows small but significant differences between countries. In every country, the most typical hot meal had two or three components. Meals with one component were more common in Norway, whereas in
Sweden there were more meals with four or five components. In Norway and
Denmark, no meals with five components were reported.
While a hot meal could include one to five components, we see from Table
4-3 that the mean on weekdays was as low as 2.6. The mean number of components on weekdays was the lowest in Finland (2.5) and the highest in Sweden (2.8). In Denmark, there was no significant difference between the mean
number of components on weekdays and on Sundays. In the other three countries the mean was higher on Sundays. On Sundays, the mean number of
components was still the highest in Sweden (3.0) and the lowest in Denmark
(2.6).
Table 4-3 Mean and number of components in hot meals on weekdays and Sundays. Per
cent of all hot meals
Number of components
One component
Denmark
Weekdays
10
Finland
Sundays
10
Weekdays
13
Norway
Sundays
13
Weekdays
15
Sweden
Sundays
11
Weekdays
8
Sundays
6
Two components
35
29
44
32
32
35
36
31
Three components
39
43
26
34
39
36
30
33
Four components
15
19
14
19
14
21
22
24
Five components
0
0
3
2
0
0
4
7
Total
100
100
100
100
100
100
100
100
Mean
2.5
2.6
2.5
2.7
2.6
2.8
2.8
3.0
Std. Dev.
0.86
0.88
0.85
0.88
0.86
0.85
0.88
0.97
N
700
221
1388
254
982
229
1240
200
Differences between countries are statistically significant at 5 per cent risk level on weekdays (at
second, third and fourth eating events)
4.4.2 Types of staples
The number and types of staples at hot meals illustrate both cultural differences and the structure of a meal. While cold sandwiches appear as significant
and structured meals in Denmark and Norway, they are more likely to be clas-
Eating Patterns
138
sified as snacks or minor meals in Finland and Sweden, where a meal is considered to include hot food. Bread was frequently included as a component of
hot meals in Finland and Sweden, but not in Denmark and Norway (see Chapter 2).
Despite the obvious traditional importance of bread, its position is different
from all other components. While other components or ingredients of a meal
are on a plate and eaten with utensils, bread is eaten with fingers and has a
separate plate, or can occasionally even be placed on the table.
All hot meals were classified according to the number of staples. This produced three groups: hot meals with both staple and bread (double staple),
those with either staple or bread (one staple), and those with no staple or
bread at all (no staple). Double staple means that a meal includes both potatoes, rice, or pasta and bread. Meals with one staple include either potatoes,
rice, pasta, or bread. The third class is hot meals without any type of staple or
bread. In principle, double staple could also be potatoes and rice, for example.
Typically, however, it seems that a double staple includes bread and rarely
contains other staple combinations without bread. It should be noted that this
classification does not distinguish one-pot dishes, which might include meat
and potatoes from centres with mixed ingredients (see Chapter 2). The category ‘other centre’ includes several dishes identified through the staple, such
as pasta and porridge, and where a staple constitutes a major part, such as the
dough of pizza.
Table 4-4 Number of staples in hot meals on weekdays and Sundays. Per cent of all hot
meals
Number of
staples
Double staple
One staple
No staple
Denmark
Weekdays
2
Finland
Sun- Weekdays
days
3
12
71
27
69
28
Total
100
N
700
Norway
Sund- Weekays
days
18
1
Sweden
Sundays
1
Weekdays
10
Sundays
10
67
23
71
19
58
30
55
28
68
31
73
26
100
100
100
100
100
100
100
239
1388
260
979
223
1239
201
Differences between countries are statistically significant at 5 per cent risk level on weekdays (at
second, third and fourth eating events)
Table 4-4 shows that in all four countries, a clear majority of hot meals included at least one staple in addition to the centre. Hot meals with double staple were much more common in Finland and Sweden, where one in ten hot
Mäkelä: The Meal Format
139
meals had staple and bread. In Denmark and Norway double staple meals
were very rare. To have no (separate) staple at all was more common in Norway, where almost one third had had such a meal. In Sweden about one quarter of all meals had no staple. On Sundays, the number of meals with no staple
was higher in Denmark. In Finland, the share of meals with double staple was
higher, accounting for almost one fifth of all meals.
4.4.3 Frequencies of proper hot meals
The following presents our findings concerning the ‘proper meal format’. A
‘proper meal’ was defined as consisting of a centre, a staple and/or bread, a
side dish of vegetables and optional trimmings. We will explore how well this
construction based on British food culture worked in relation to Nordic eating
habits and how relevant it is.
The unit of analysis in the tables above was hot eating events. In order to be
able to link the data on eating to background variables, we need to change our
unit of analysis and focus instead on individual respondents. Each individual
may have had no, one, or more than one hot meal on the day before. Eating
two hot meals was quite common in both Finland and Sweden, where about
two in five had had a hot meal twice on weekdays (see Chapter 3). In Table
4-5 we see that the proportion of people who had no hot meals at all was
slightly higher in Denmark and in Norway (11–12%) than in Finland and in
Sweden (5–6%). In order to analyse meal formats at the individual level, we
needed to link one meal to one individual. We concentrated on certain time
periods during which it was unlikely for people to have had more than one hot
meal. Rather wide time frames were applied in dividing the data set in order
to include as many cases as possible. All hours with at least 5 per cent of hot
eating events in one country were included. Bearing in mind the earlier analyses of ‘lunch’ and ‘dinner’, we divided the day into two parts, from 11 a.m. to
3 p.m. and from 3 p.m. to 9 p.m. These frames cover 93 per cent of all hot
meals during the day (data not shown).
Table 4-5 Eating hot meals on weekdays. Per cent of all informants
Denmark Finland
Norway Sweden
Total
No hot meals
12
5
11
6
8
At least one hot meal
88
95
89
94
(proper or not)
92
Total ( N)
100 (735) 100 (976) 100 (962) 100 (903) 100 (3576)
Differences between countries are statistically significant at 0.1 per cent risk level
Eating Patterns
140
Table 4-6 describes the relative frequency of eating one hot meal within the
time frames between 11 a.m. and 3 p.m. and between 3 p.m. and 9 p.m. The
main tendency is the same as in the tables presented in Chapter 3. More people in Finland and Sweden than in Denmark and Norway had two hot meals
per day: one early in the day and a second one later.
Table 4-6 Informants who had eaten hot meals between 11 a.m. and 3 p.m. and between 3
p.m. and 9 p.m. on weekdays. Per cent of all informants
.
Denmark
Finland
Norway
Sweden
Total
11 a.m.–3 p.m.
At least one hot meal
15
59
17
59
39
No hot meals
85
41
83
41
61
100 (735)
100 (976)
100 (961)
100 (903)
100 (3575)
77
59
74
67
69
Total (N)
3 p.m.–9 p.m.
At least one hot meal
No hot meals
Total (N)
23
41
26
33
31
100 (735)
100 (976)
100 (961)
100 (903)
100 (3575)
Differences between countries are statistically significant at 0 per cent risk level
We found that eating ‘proper meals’ was quite common in all countries. In
Table 4-7 we see that more respondents in Sweden had at least one ‘proper
meal’ during the day before compared to the other countries, 6 per cent even
had two. The Norwegians also had a relatively high proportion of ‘proper
meals’, but very few had more than one. The proportions were lower in Denmark and Finland, but there were some Finns who had two such meals. These
results show that even if people had two hot meals during the day, both of
them were rarely proper ones.
Table 4-7 Informants who had eaten proper meals on weekdays. Per cent of informants
who had eaten hot meals
Number of proper meals
Denmark Finland
Norway
Sweden
No proper meals
63
57
53
50
One proper meal
37
38
46
43
More than one proper meal
1
5
1
6
Total (N)
101 (658) 100 (927) 100 (863) 99 (856)
Differences between countries are statistically significant at 0.1 per cent risk level
In sum then, the highest proportion of hot meals not categorised as ‘proper
meals’ was recorded for Denmark. Going back to Table 4-2, we can see that
more Danish hot meals consisted of a centre with a staple or a centre with staple and trimmings. Table 4-2 also shows that the ‘proper meal’ combination
Mäkelä: The Meal Format
141
of centre, staple and vegetables (CSV) was the most common combination on
weekdays in all countries except Finland, where it took the second place.
4.4.4 Type of centre in proper meals
Murcott (1983) argues that the centre of a British proper meal always is meat.
We consciously broadened this definition by adding fish, vegetables and other
(especially one pot dishes) as suitable centres. Table 4-8 (which is across
meals) reveals interesting variations in centres between countries and types of
meals. In all four countries, meat was the most common centre of all types of
hot meals. The highest figure was recorded in Denmark. However, meat was
even more popular as a centre when the hot meal in question was a ‘proper
meal’. This was true for all countries except Norway. In Denmark, 85 per cent
of ‘proper meals’ had a meat centre. In Norway, fish seemed to be almost as
popular as meat, but in other countries the share of fish centres with ‘proper
meals’ was clearly lower. Vegetables were more often the centre of other
types of meals than ‘proper meals’. In Finland, however, 10 per cent of meals
with a vegetable centre were ‘proper meals’. Further, ‘other’ types of centres
(see Chapter 2) were quite common in Finland and Sweden in both types of
hot meals.
Table 4-8 Types of centres in hot meals on weekdays. Per cent of proper and other types
of meals
Denmark
Proper meals
Other meals
Finland
Proper meals
Other meals
Norway
Proper meals
Other meals
Sweden
Proper meals
Other meals
Meat
Fish
Vegetables
Other
Total (N)
85
56
9
9
3
22
3
22
100 (264)
100 (436)
63
52
15
10
12
20
12
20
100 (632)
99 (756)
50
51
42
14
7
30
7
30
100 (463)
100 (516)
67
47
18
11
13
32
13
32
101 638)
101(598)
Differences between countries are statistically significant at 5 per cent risk level (at
second, third and fourth eating events)
Eating Patterns
142
4.5
Eating proper meals
Logistic regression models were used to examine the social patterning of
‘proper meals’. The event we were interested in modelling was the eating of a
proper meal.34 The odds ratio (OR) for a predictor indicates the relative
amount by which the odds of the outcome increases (OR greater than 1.0) or
decreases (OR less than one) when the value of the predictor is increased by 1
unit. Therefore, the dependent variable is the occurrence of ‘proper meals’,
the probability of which we analysed by using a logistic regression. The
dependent variable was recoded into an indicator variable with value 1 for
those who had a ‘proper meal’ during the relevant time period.
In these analyses we excluded those individuals who had not eaten any hot
meal at all (proper or not). Thus, our logistic regression concerns the conditional probability of eating a proper meal, given that the individual had eaten
some kind of hot meal (proper or not). The relative frequency of not eating a
hot meal was about 8 per cent of the entire sample (Table 4-5) and the logistic
regression models are rather representative of the population as a whole.
We fitted several models sequentially, starting with one regressor (or one
group of regressors) only and then increasing the number of regressors. The
explanatory variables were introduced in the following order: country, age,
sex, education, occupational status, social status, living alone or with
someone, size of municipality, day of the week. The variable social status is
based on actual occupation, earlier occupation, and spouse’s occupation (a
description of background variables is given in Kjærnes et al. 2000).
The main idea of our logistic regression models is to compare countries.
Therefore the same set of background variables was used in all analyses, even
though not all the same variables are significant for each country. Another
possibility would have been to fit separate models for each country with significant variables only, but we thought that fitting the same model for all the
countries would be more interesting for comparative purposes. However, the
number of observations in the different specifications diminishes as more
34
Denote by p the probability of having eaten a proper meal. The logistic regression model can
then be written
ln(p/(1-p)) = ΣβiXi
where the Xi:s are the regressors (i.e. explanatory variables) and the βi :s are the (logit) coefficients.
The odds of having eaten a proper meal is estimated as
odds = p/1-p.
Mäkelä: The Meal Format
143
variables are taken in. The effect of adding a variable to a model is tested by
the Wald statistic. The model chi-square tests are used to test the null hypothesis that the coefficients for all of the terms in the current model are 0. It
should be noted that the tests of the significance of the estimates are concerned with the difference of each group vis-à-vis the reference group. Therefore, the use of different reference groups yields different significance patterns. The logistic regression results are presented as odds ratios (OR). The
statistical analyses were run with SPSS (v. 8 and 10).
We chose to report the full models as well as models with one group of
regressors only. Therefore, the table for each estimation consists of three
columns. The first column depicts the relative frequency of eating ‘proper
meals’. The second column shows the odds ratios for eating ‘proper meals’
for each separate variable, the third and last column presents the odds ratios
for the full model. In our tables of results, the reference group always gets the
odds ratio of unity. In these tables, N is the number of informants who had
eaten hot meals (not the number of hot meals eaten, as in the tables describing
the number of components and staples).
The first series of analyses concern the occurrence of ‘proper meal’ eating
within the time frames defined above. In the first three columns of Table 4-9,
estimation results for the probability of eating a ‘proper meal’ between 11
a.m. and 3 p.m. are presented for all four countries. The first rows of the second column indicate that ‘proper meals’ were somewhat more likely to occur
in Norway, although the only statistically significant difference was between
Norway and Finland. When all the regressors were included (see the third
column), this difference is no longer statistically significant. Looking at the
other odds ratios, we conclude that only age factors seem to affect the likelihood of ‘proper meals’. In all countries taken together, the youngest age group
were less likely to eat ‘proper meals’ than the other age groups. The social
groups did not differ from one another in other respects.
The last three columns of Table 4-9 depict similar estimates for the relative
likelihood of eating a ‘proper meal’ between 3 p.m. and 9 p.m. (‘dinner’). In
this specification, the inter-country differentials remained even in the full
model specification. Again, Norway had the highest odds ratios and Finland
the lowest. However, the pattern of the age effects comes very close to that of
the first three (‘lunch’) columns. The main difference between the two estimations has to do with the effect of education and gender. Higher education and
female gender increased the probability of eating ‘proper meals’. The effect
was in the opposite direction for social status, with the middle white-collar
group having had ‘proper meals’ in the afternoon/evening less often.
Eating Patterns
144
Table 4-9 Description (%), relative eating (odds ratio) and full model (odds ratio, logistic
regression) of eating a proper meal between 11 a.m. and 3 p.m. and between 3 p.m. and 9
p.m. of those who have eaten a hot meal. Weekdays
Proper
meal
%
11 a.m.– 3 p.m.
Relative
Model
eating
OR
OR
Proper
meal
%
3 p.m.–9 p.m.
Relative
Model
eating
OR
OR
Country
Norway
Denmark
Finland
Sweden
48
37
38
42
1.00
0.66
0.67*
0.78
1.00
0.68
0.71
0.85
46
35
28
40
1.00
0.66*
0.46*
0.80*
1.00
0.67*
0.45*
0.75*
Age
15–24
25–34
35–44
45–54
55–64
65+
30
39
42
45
42
44
1.00
1.43
1.64*
1.87*
1.61*
1.76*
1.00
1.75*
1.96*
2.20*
2.22*
2.64*
26
38
43
40
37
40
1.00
1.74*
2.19*
1.88*
1.70*
1.90*
1.00
1.86*
2.42*
2.30*
2.00*
2.28*
Sex
Male
Female
41
40
1.00
0.92
1.00
0.91
34
42
1.00
1.41*
1.00
1.44*
Education
Primary
Secondary
University
41
39
42
1.00
0.91
1.03
1.00
0.84
0.87
31
37
44
1.00
1.33*
1.75*
1.00
1.33*
1.62*
Occup. status
Not active, retired
Active, student
48
55
1.00
1.12
1.00
1.39
38
38
1.00
1.01
1.00
0.95
39
41
44
42
1.00
1.05
1.17
1.11
1.00
1.09
1.10
0.98
37
40
39
42
1.00
1.12
1.08
1.24
1.00
0.93
0.73*
0.98
41
40
1.00
0.91
1.00
1.06
37
38
1.00
1.01
1.00
1.13
37
40
46
1.00
1.12
1.41*
1.00
1.07
1.34
37
41
33
1.00
1.18
0.84
1.00
1.08
0.81
Social status
Worker
Lower white collar
Middle white
collar
Higher white
collar
Living
Alone
With someone
Size municipality
< 25 000
25 000–99 999
> 99 999
0.083
1 152
0.000
2 111
Significance
N
* The probability of eating proper meal differs from the reference group at a significance level of 5
per cent
Mäkelä: The Meal Format
145
4.5.1 Proper meals in different social groups
Next, we repeated the analysis for individual countries. We analysed the
probability of eating ‘proper meals’ between 11 a.m. and 3 p.m. separately for
Finland and Sweden. Due to the small number of observations of hot
‘lunches’, this could not be done in any reliable way for Denmark and
Norway.
Table 4-10 shows that in Sweden, age made a difference (as was the case for
the pooled model for the four Nordic countries), whereas for Finland, these
effects were not statistically significant. The second column of Table 4-10
indicates that in Finland, economically active and students ate ‘proper meals’
significantly more often than the economically non-active population. Similarly, people living in big cities were more likely to eat proper meals. However, only the latter effect remained statistically significant after the effect of
other factors was controlled for. Looking again at the results for separate variables, it seems that elderly and economically non-active Finnish persons were
less likely to eat ‘proper meals’. This difference was not significant in the full
model. Technically this is probably due to multicollinearity: elderly people
are retired and their occupational status seems to be the background for the
age effect when analysed separately. These separate country analyses suggest
that the variation between age groups appeared mainly in Sweden and the effect of the size of municipality in Finland.
As to eating a ‘proper meal’ (‘dinner’) between 3 p.m. and 9 p.m., there were
enough observations in Denmark and Norway as well to analyse the probability of eating ‘proper meals’ separately in all four countries (Table 4-11). We
see that the effect of gender and age on the probability of eating proper evening meals varied from country to country. In Denmark and Finland, women
were relatively more likely to eat ‘proper meals’ than men, whereas in Norway and Sweden no such difference could be found. In Norway and Sweden,
the youngest age group (15-24 years) differed from all other age groups by
eating ‘proper meals’ relatively less frequently. In Denmark this was true of
the two youngest age groups (that is up to the age 34). In Finland, differences
between age groups were less clear. The effect of social status in Table 4-9
was (as we can see in Table 4-11) limited to Finland, where middle white collar employees were less likely to eat ‘proper meals’ during this time period
than both workers and higher white collar people. In Sweden – and to some
extent in Denmark too – education mattered greatly: the more educated they
were, the more likely they were to have eaten ‘proper meals’ between 3 p.m.
and 9 p.m. Otherwise, the four countries were internally rather homogeneous.
Eating Patterns
146
Table 4-10 Description (%), relative eating (odds ratio) and full model (odds ratio, logistic
regression) of eating a proper meal between 11 a.m. and 3 p.m. of those who have eaten a
hot meal. Weekdays. Finland and Sweden
Proper
meal
%
Finland
Relative
eating
OR
Model
OR
Proper
meal
%
Sweden
Relative
eating
OR
Model
OR
Age
15–24
25–34
35–44
45–54
55–64
65+
35
35
41
52
33
28
1.00
1.02
1.29
1.99*
0.91
0.72
1.00
1.05
1.23
1.90
1.20
1.23
29
41
53
47
39
43
1.00
1.71
2.67*
2.06*
1.52
1.80*
1.00
2.41*
3.37*
2.68*
2.33*
3.87*
Sex
Male
Female
37
39
1.00
1.07
1.00
1.11
46
38
1.00
0.72
1.00
0.70
Education
Primary
Secondary
University
33
38
46
1.00
1.22
1.67
1.00
1.05
1.22
42
38
44
1.00
0.83
1.07
1.00
0.66
1.00
Occupational
status
Non-active,retired
Active, student
28
43
1.00
1.96*
1.00
1.70
37
44
1.00
1.32
1.00
1.71
Social status
Worker
Lower white
Middle white
Higher white
33
39
46
34
1.00
1.32
1.78*
1.07
1.00
1.27
1.50
0.85
43
41
41
49
1.00
0.92
0.91
1.25
1.00
0.96
0.79
0.87
Living
Alone
With someone
36
38
1.00
1.07
1.00
1.07
41
42
1.00
1.08
1.00
1.18
Size of municip.
< 25 000
25 000–99 999
> 99 999
34
34
50
1.00
1.00
1.94*
1.00
1.02
1.68*
46
41
40
1.00
0.81
0.80
1.00
0.79
0.84
Significance
0.048
0.117
N
461
452
* The probability of eating proper meal differs from the reference group at a significance level of 5
per cent
1.00
1.16
1.90*
2.04*
1.86
1.67
1.00
2.26*
1.00
1.46*
1.43
1.00
1.06
1.00
0.91
1.15
0.58
1.00
1.27
1.00
1.36
0.90
26
29
40
41
39
37
27
45
29
38
37
35
36
37
35
40
26
31
37
35
42
33
DENMARK
Relative
eatingOR
1.00
1.43
0.91
1.00
1.22
1.00
0.75
0.90
0.59
1.00
1.67
1.00
1.55
1.47
1.00
2.53*
1.00
1.11
1.82
2.42*
2.41*
2.91*
Model
OR
27
30
27
23
28
30
32
19
38
23
30
24
29
30
23
33
21
23
38
31
24
16
Proper
meal %
1.00
1.16
1.03
1.00
1.34
1.00
1.10
0.54*
1.42
1.00
1.37
1.00
1.30
1.36
1.00
1.63*
1.00
1.06
2.25*
1.63
1.17
0.68
FINLAND
Relative
eatingOR
1.00
1.17
1.04
1.00
1.32
1.00
0.77
0.56*
1.43
1.00
1.35
1.00
0.98
1.07
1.00
2.15*
1.00
0.99
1.95
1.67
1.19
0.59
Model
OR
48
46
40
49
45
44
48
47
50
52
44
43
43
51
44
47
23
47
46
51
52
52
Proper
meal %
1.00
0.90
0.71
1.00
0.85
1.00
1.16
1.12
1.25
1.00
0.71
1.00
0.98
1.34
1.00
1.13
1.00
2.93*
2.89*
3.51*
3.69*
3.59*
NORWAY
Relative
eatingOR
Significance
0.000
0.001
N
522
437
* The probability of eating proper meal differs from the reference group at a significance level of 5 per cent
Age
15–24
25–34
35–44
45–54
55–64
65+
Sex
Male
Female
Education
Primary
Secondary
University
Occ. status
Not active
Active
Social stat.
Worker
Lower
Middle
Higher
Living
Alone
With some
Size munic.
< 25 000
25–99 999
> 99 999
Proper
meal %
0.000
623
1.00
0.97
0.73
1.00
1.12
1.00
1.15
0.91
0.99
1.00
0.73
1.00
0.95
1.16
1.00
1.16
1.00
3.24*
3.25*
4.13*
3.90*
3.28*
Model
OR
41
42
32
40
40
37
42
40
47
44
38
30
38
45
39
41
31
45
50
34
34
47
Proper
meal %
1.00
1.08
0.69
1.00
1.02
1.00
1.26
1.16
1.54
1.00
0.79
1.00
1.45
1.88*
1.00
1.06
1.00
1.79*
2.22*
1.13
1.13
1.96*
SWEDEN
Relative
eatingOR
0.001
502
1.00
1.01
0.60
1.00
1.16
1.00
1.29
0.93
1.18
1.00
0.57
1.00
1.62
2.50*
1.00
0.91
1.00
2.35*
2.75*
1.59
1.51
2.21*
Model
OR
Table 4-11 Description (%), relative eating (odds ratio) and full model (odds ratio, logistic regression) of eating a proper meal between 3
p.m. and 9 p.m. of those who have eaten a hot meal. Weekdays. Denmark, Finland, Norway, and Sweden
Eating Patterns
148
Table 4-12 Description (%), relative eating (odds ratio) and full model (odds ratio, logistic
regression) of eating at least one proper meal during the day of those who have eaten hot
meal(s)
Proper meal %
Relative eating OR
Model OR
Country
Norway
Denmark
Finland
Sweden
49
40
44
51
1.00
0.69*
0.84*
1.11
1.00
0.65*
0.82*
1.08
Sex
Male
Female
44
48
1.00
1.19*
1.00
1.24*
Age
15–24
25–34
35–44
45–54
55–64
65+
39
43
52
49
47
47
1.00
1.18
1.73*
1.50*
1.40*
1.38*
1.00
1.33*
1.97*
1.77*
1.82*
2.12*
Education
Primary
Secondary
University
42
46
50
1.00
1.16*
1.40*
1.00
1.19
1.29*
Occupational status
Not active, retired
Active, student
44
47
1.00
1.11
1.00
1.17
Social status
Worker
Lower white collar
Middle white collar
Higher white collar
45
46
48
50
1.00
1.04
1.09
1.20
1.00
0.94
0.87
0.94
Living
Alone
With someone
43
47
1.00
1.18*
1.00
1.30*
Size of municipality
< 25 000
25 000–99 999
> 99 999
45
49
44
1.00
1.16*
0.96
1.00
1.04
0.93
Day
Weekday
Weekend
45
50
1.00
1.24*
1.00
1.30*
Significance
N
0.000
3921
* The probability of eating proper meal differs from the reference group at a significance level of
5 per cent
Mäkelä: The Meal Format
149
Using a more encompassing variable, we now move on to analyse the event of having
eaten at least one ‘proper meal’ during the entire day, regardless of when this occurred. Thus, in the analyses reported in Table 4-12 and Table 4-13, the dependent
variable is the likelihood of the event ‘having eaten at least one proper meal during
the day’. These two analyses show more statistically significant differences.
The third column in Table 4-12 reveals the estimates for the full model specification.
Being a Dane or a Finn lowered the likelihood of eating a proper meal, relative to
Sweden and Norway. In the Nordic countries as a whole, women were relatively more
likely to have a ‘proper meal’ than men, as were people with middle or higher
education. The youngest age group was less likely to eat ‘proper meals’ even in this
analysis. People living alone were less likely to eat ‘proper meals’ than people living
in larger households. Furthermore, ‘proper meals’ were more common during
weekends than on weekdays.
We also did an analysis of probability of not eating hot meals, but the results
did not change the overall picture. It seems that eating or not eating hot meals
was not clearly socially differentiated. Furthermore, living alone or with
someone did not seem to be of significance. Possible connections between the
structure and social aspects of hot meals will be further discussed in
Chapter 9.
4.5.2 Proper meals among men and women
Since Nordic women were more likely to have ‘proper meals’ than men, we
decided to fit a second series of logistic regression models, this time separately for men and women. The fitting of the model was identical to the earlier
analyses with one exception: the variable describing whether or not the informant lived alone was replaced with a variable related to the type of household.
This variable includes information on children living in the same household
with the informant. This information should be relevant in an analysis of gender differences in eating ‘proper meals’. These analyses are reported in Table
4-13. The results are interesting. Age effects seem stronger for men than
women, and this was the case for university education and gainful employment. Older, highly educated and economically active men were more likely
to enjoy a ‘proper meal’ during the day compared to young, uneducated and
economically non-active men. For women, this difference was much less
marked. Men and women living with children were both more likely to have a
proper meal.
Eating Patterns
150
Table 4-13 Description (%), relative eating (odds ratio) and full model (odds ratio, logistic
regression) of eating one or more proper meal during the day of those who have eaten hot
meal(s). Separate analyses for men and women
Country
Norway
Denmark
Finland
Sweden
Age
15–24
25–34
35–44
45–54
55–64
65+
Education
Primary
Secondary
University
Occupational
status
Nonactive, retired
Active, student
Social status
Worker
Lower white
Middle white
Higher white
Household type
Couple +children
Couple, no childr.
Single +children
Single, no childr.
Size municipality
< 25 000
25 000–99 999
> 99 999
Day
Weekday
Weekend
Significance
N
Proper
meal
%
Men
Relative
eating
OR
Model
OR
Proper
meal
%
Women
Relative
eating
OR
Model
OR
46
34
44
51
1.00
0.60*
0.90
1.19
1.00
0.64*
1.02
1.32
51
45
45
52
1.00
0.79
0.77*
1.04
1.00
0.72*
0.73*
0.96
35
43
49
48
42
44
1.00
1.36*
1.74*
1.71*
1.30*
1.45*
1.00
1.56*
1.87*
2.21*
2.31*
3.49*
42
43
56
49
51
48
1.00
1.01
1.71*
1.30
1.43*
1.28
1.00
0.98
1.64*
1.42
1.76*
1.64*
38
42
50
1.00
1.20
1.67*
1.00
1.28
1.69*
45
49
50
1.00
1.14
1.19
1.00
1.19
1.18
39
45
1.00
1.30*
1.00
1.53*
48
48
1.00
1.02
1.00
0.96
41
41
47
50
1.00
0.99
1.25*
1.40*
1.00
1.04
0.90
0.93
52
48
48
50
1.00
0.88
0.86
0.93
1.00
0.79
0.72
0.77
48
43
40
41
1.00
0.83
0.71
0.74*
1.00
0.69*
0.76
0.63*
50
50
46
45
1.00
0.98
0.82
0.81
1.00
0.80
0.84
0.66*
44
46
41
1.00
1.06
0.86
1.00
0.89
0.82
46
52
47
1.00
1.27*
1.07
1.00
1.25
1.09
42
48
1.00
1.28*
1.00
1.28*
47
52
1.00
1.20
1.00
1.19
0.000
1666
0.001
1 939
* The probability of eating proper meal differs from the reference group at a
significance level of 5 per cent
We then again divided the day into the two periods of 11 a.m. to 3 p.m. and 3
p.m. to 9 p.m. and carried out separate analyses for the two sexes. Table 4-14
and Table 4-15 show that the overall picture remains more or less the same as
Mäkelä: The Meal Format
151
that suggested by Table 4-13. Women were a more homogeneous group,
whereas men differed from each other due to age, education and occupational
status. Country differences were, however, not significant.
Table 4-14 Eating one proper meal between 11 a.m. and 3 p.m. of those who have eaten a
hot meal. Weekdays. Description (%), relative eating (odds ratio) and full model (odds
ratio, logistic regression). Men
Country
Norway
Denmark
Finland
Sweden
Age
15–24
25–34
35–44
45–54
55–64
65+
Education
Primary
Secondary
University
Occupational status
Not active, retired
Active, student
Social status
Worker
Lower white collar
Middle white collar
Higher white collar
Household type
Couple with children
Couple, no children
Single with children
Single, no children
Size of municipality
< 25 000
25 000–99 999
> 99 999
Men
Proper meal %
Men
Relative eating OR
Men
Model OR
46
37
37
46
1.00
0.69
0.70
1.01
1.00
0.93
0.74
1.23
31
46
40
50
41
42
1.00
1.88*
1.47
2.27*
1.59
1.65
1.00
3.66*
2.84*
4.92*
5.71*
7.64*
39
38
48
1.00
0.99
1.50*
1.00
0.91
1.28
34
44
1.00
1.54*
1.00
2.12*
39
38
46
46
1.00
0.97
1.29
1.30
1.00
0.87
1.02
0.90
44
42
33
39
1.00
0.92
0.63
0.83
1.00
0.62
0.87
0.61
38
43
45
1.00
1.26
1.33
1.00
1.12
1.20
Significance
0.008
N
481
*The probability of eating proper meal differs from the reference group at a significance level of 5
per cent
Eating Patterns
152
Table 4-15 Eating one proper meal between 3 p.m. and 9 p.m. of those who have eaten a
hot meal. Weekdays. Description (%), relative eating (odds ratio) and full model (odds
ratio, logistic regression). Separate analyses for men and women
Country
Norway
Denmark
Finland
Sweden
Age
15–24
25–34
35–44
45–54
55–64
65+
Education
Primary
Secondary
University
Occupational
status
Nonactive, retired
Active, student
Social status
Worker
Lower white
Middle white
Higher white
Household type
Couple +children
Couple, no childr.
Single +children
Single, no childr.
Size municipality
< 25 000
25 000–99 999
> 99 999
Proper
meal
%
Men
Relative
eating
OR
Model
OR
Proper
meal
%
Women
Relative
eating
OR
Model
OR
43
27
24
39
1.00
0.47*
0.42*
0.85
1.00
0.51*
0.47*
0.88
47
45
33
41
1.00
0.92
0.56*
0.77
1.00
0.95
0.53*
0.66*
22
35
37
39
34
37
1.00
1.92*
2.09*
2.26*
1.84*
2.03*
1.00
1.78
1.92*
2.83*
2.55*
3.04*
31
41
50
43
42
43
1.00
1.57*
2.24*
1.66*
1.58
1.66*
1.00
1.25
1.75
1.80*
1.69
1.88
27
33
41
1.00
1.29
1.85*
1.00
1.32
1.79*
35
43
47
1.00
1.41*
1.63*
1.00
1.29
1.53
31
35
1.00
1.22
1.00
1.18
43
42
1.00
0.93
1.00
0.84
33
31
36
40
1.00
0.91
1.14
1.36
1.00
0.92
0.67*
0.92
45
44
41
48
1.00
0.98
0.87
1.15
1.00
0.88
0.68
1.04
34
35
33
35
1.00
1.02
0.95
1.01
1.00
0.76
1.18
0.78
47
39
43
41
1.00
0.72*
0.83
0.79
1.00
0.54*
0.90
0.61*
33
38
31
1.00
1.23
0.90
1.00
1.08
0.82
43
45
37
1.00
1.12
0.79
1.00
1.17
0.79
Significance
0.000
0.004
N
876
1 013
* The probability of eating proper meal differs from the reference group at a significance level
of 5 %.
For men, age increased the probability of eating ‘proper meals’ in both time
intervals. The coefficients were rather high. This might be a technical problem
due to the sample size. A closer look at the head counts revealed that there
was a slight shortage of cases in the older age groups. This was the case especially in Finland (see Kjærnes et al. 2000). In the middle of the day, those who
were economically active tended to eat more ‘proper meals’, whereas in the
Mäkelä: The Meal Format
153
afternoon and evening education had a similar impact on ‘proper meal’ eating
among men.
In Table 4-15 we see that, during the time period between 3 p.m. and 9 p.m.,
women living with children were more likely to have ‘proper meals’ than
women without children. There were similar effects among men but they were
not statistically significant. There were some country differences. Age and
gender were relevant factors in most analyses, as was education. It seems,
however, that the relative effects were not similar in all countries. While
women in general ate ‘proper meals’ relatively more often, men were more
heterogeneous in this respect. The probability of eating ‘proper meals’ increased among men who were older, better educated and/or economically active and who lived in families. The (Finnish) effect of social status was further
limited to men, while there was no significant effect for women.
4.6
Discussion
Hoping to contribute to the ongoing discussion on the possible destructuration
of eating in modern everyday life, the main task of this chapter was to analyse
the structure and the complexity of hot meals. A key concept in this
discussion is that of ‘grazing’, implying a destructuration with respect to time,
social context and meal formats. Within the confines of our study it is not
possible to establish what changes have taken place over time, but in this
chapter we are interested in the variation that has occurred with respect to
meal formats – both between socio-economic groups and between the four
countries.
In order to grasp the structure and variety of hot eating events, we constructed
a variable called ‘proper meal’. This is heavily inspired by the work on proper
meals in Britain by Douglas and Nicod as well as Murcott. Even though our
multivariate analyses showed some interesting differences in the probability
of having eaten ‘proper meals’, these were not very clear and systematic.
However, in all countries younger people were less likely to have ‘proper
meals’. There was not much difference in this regard between social groups,
but it does seem that education increases the probability of eating ‘proper
meals’. In general, people were more likely to have ‘proper meals’ during
weekends than on weekdays.
We also found some nation-specific characteristics. Norwegians were more
likely to have ‘proper meals’ than other Nordic nations, whereas Danes ate
154
Eating Patterns
less ‘proper meals’ than people elsewhere in Nordic countries. The impact of
gender was intriguing. It seems that women, especially in Denmark and
Finland, had more ‘proper meals’ than women in Norway and Sweden. These
results were all the more interesting if we bear in mind that in many other respects, Denmark and Norway on the one hand and Finland and Sweden, on
the other, shared similar patterns of eating. However, an analysis done separately for men and women revealed that while women in general ate more
‘proper meals’, there were only few differences between different sociodemographic groups of women. Men constituted a more heterogeneous group
in that age, education, occupational status, and family situation all affected
men’s probability of having ‘proper meals’.
In Finland, middle white-collar employees were less likely to have proper
dinners (i.e. ‘proper meals’ between 3 p.m. and 9 p.m.) than other groups, but
they were more likely than others to do so during typical lunch hours. One
possible, albeit somewhat speculative interpretation is that for this group,
‘proper meals’ are located in their professional lives. The group includes bank
clerks and other office workers who often have a canteen lunch that is
subsidised by the employer. Therefore we may assume that these people do
not have the same kind of need for a ‘proper meal’ in the evening as other
groups; they have already had one.
It is difficult to offer any exhaustive interpretation of these results. They could
mean one of two things. First, the habit of eating ‘proper meals’ is not very
systematically differentiated within the Nordic population. Second, one
should bear in mind that in all these countries other types of hot meals are as
common as or even more common than these ‘proper meals’. This could also
mean that ‘proper meals’, thus defined, are not very good criteria for the
‘properness’ – in the sense of being typical of a meal in the Nordic countries.
In particular, the differences between men and women and the different
dispersion of eating ‘proper meals’ within these two groups taken separately
make quite sensible reading. It has been established in a number of studies
that attitudes and practices around food are gender specific (see e.g. Charles
and Kerr1988). Even though the situation is gradually changing, it has
traditionally been an integral part of a woman’s role to be interested and
involved in matters of food and home cooking. Therefore, it is understandable
that women are in general likely to have ‘proper meals’, not only continuing
traditions, but also accepting new vogues of (healthy) eating (Roos 1998).
In all four countries, people reported eating events which show both complexity and variation. Our findings do not support the idea of ‘grazing’. People do
Mäkelä: The Meal Format
155
still have meals with a structure, but it should be noted that they also eat simple meals with only a few components, and some informants did not eat any
hot food during the day. The average number of components on weekdays
was 2.6, which indicates that there were many hot meals with fewer than the
three components required of a ‘proper meal’ (see above).
The bread eating of Finns and Swedes created an anomaly within our
theoretical frame that needed to be taken into account. To this end we
produced a classification of hot meals by focussing on the number of staples.
Finns and Swedes stood apart from the other Nordic nations on account of
their use of so-called double staple: it was quite common in Finland and
Sweden to have both bread and potatoes or pasta at the same time. In Finland,
bread could be the only staple. This may be explained by the fact that Finns
eat a lot of pot dishes, where potatoes may be mixed in with other ingredients.
However, variation in the number of staples might suggest some deeper
cultural explanations. Double staple may indicate that bread is associated with
hot meals in Finland and Sweden because there is no tradition of sandwiches
as (cold) meals, as is the case in Denmark and Norway. Finns and Swedes do
have sandwiches as snacks, but their place in the eating system is different
from that in Denmark and Norway. This is merely a question of definition, as
it is not the content of the sandwiches that labels them as a snack or a meal.
Norwegian and Finnish sandwiches are usually both quite modest. In Finland,
one could also speculate that the widespread consumption of bread might be a
remnant of an agrarian way of life where grain and potatoes were used as a
supplement to meagre portions of meat and fish (Prättälä et al. 1993). The
growth of wellbeing has manifested itself in rising consumption of meat, fish
and vegetables other than root crops (Maula 1995, 47). The number of staples
may also indicate increasing complexity, ranging from no staple at all to two
staples.
It is reasonable briefly to discuss deconstruction and reconstruction of ‘proper
meals’ as a methodological tool. It was clearly useful to deconstruct or break
down hot eating events or meals into questions according to the idea of a hot
meal consisting of a centre, a staple, and trimmings. The interviewees had no
difficulty answering our questions. We continued by reconstructing these
components into a separate variable called ‘proper meals’. As it turned out,
quite a large number of the reported hot eating events fitted into this category.
However, many hot meals did not. ‘Proper meals’ are only one type of diverse
hot Nordic meals. Here, the focus of our analyses was on hot meals, even
though a part of Danish and Norwegian cold eating events are traditionally
understood as meals and are quite common. Another focus in our analyses has
156
Eating Patterns
been a plateful, i.e. the hot main course. Even though people seldom reported
any other courses, one could think about a concept or a construct of ‘elaborate
meals’, which in addition to a main course would include a starter and a dessert.
However, our results also raise the question as to what is to be considered
traditional or modern eating. In Chapter 2, we showed that those vegetables
we thought were novelties in Nordic food consumption were hardly ever
reported. Looking at these data it is possible to trace trends of quite modern
diets in Finland, where the number of centres consisting of vegetables is
somewhat higher than in the other countries. Yet, it is evident that meat and
fish are the typical centres of a hot meal. It seems that the Finnish diet is an
intriguing combination of old traditions and modern novelties.
The types of ‘proper meals’ found in British studies are neither the most
common nor traditional types of hot meals in the Nordic countries. In Finland
and Norway, in particular, various one-pot dishes, in which all ingredients are
either cooked or baked in the oven, are quite common. One could say that
these dishes are the most traditional ones. Therefore, a ‘proper meal’ as a
combination including both a staple and vegetables as separate components,
might be relatively modern per se in the Nordic countries. Statistics on food
consumption in the Nordic countries reveal quite consistent trends for the
consumption of vegetables, which has increased markedly since the 1960s
(Nordisk Ministerråd 2001). Interestingly, it seems that the notions of a British proper meal are changing in the opposite direction, as dishes with mixed
ingredients, like curry or chilli, nowadays qualify as proper meals (Kemmer et
al. 1998).
It should be pointed out, however, that our analyses of ‘proper meals’ did not
have a normative dimension; we did not ask our informants what they thought
was the proper meal par excellence. We are therefore not in a position to
make any assumptions about the normative dimension of eating.
Nonetheles, it might be difficult to completely sever our construct of ‘proper
meals’ from its cultural background. Our construct is based on a British concept, and ideas of what is considered a proper meal are culturally differentiated and not constant. In Finland, for a meal to be called ‘proper’, it is enough
that the food is hot and accompanied by a salad (Mäkelä 1996). As has repeatedly been pointed out here, in Denmark and Norway a cold sandwich definitely
goes for a ‘proper’ lunch. Interestingly, the variation in eating ‘proper meals’
does not follow the boundaries of cold and hot lunch cultures (see Chapter 3).
In Denmark and Norway, people eat hot lunches less often than in Finland and
Mäkelä: The Meal Format
157
Sweden. Even so, Danes and Norwegians differ clearly in their habits of eating ‘proper meals’, Danes ranking lowest and Norwegians highest among the
Nordic countries.
Our results do not seem to support the idea of increasingly unstructured eating
habits. One could speculate whether the debate on grazing has grown up in
part out of fears of an Americanisation of culture. This is crystallised in the
idea of the McDonaldization of society, a thesis put forward by Ritzer (1993)
to highlight the rationalisation of especially food service provision (see also
Chapter 5). It should, however, be kept in mind that the method applied here
forces eating events into pre-coded categories. We were also unable to record
small eating events like eating a piece of chocolate or chewing gum, i.e.
putting something edible into the mouth perhaps without the frames which
characterise eating meals. Still, it can be said that the majority of people do on
a regular basis eat meals with a certain structure. In addition, most people also
had one or a few less structured eating events during the day.
4.7
Conclusion
The model and method used in this study are unique in allowing us to take
into account different levels and dimensions of eating with the aim of
providing an exhaustive description of eating patterns. In our vivisectionist
approach, deconstruction and reconstruction served as a theoretically based
means of empirical analysis as we sought to test the concept of proper meals
within a quantitative data set. The focus in this chapter has been on the
structuration of hot eating events. In addition, we have explored the meal
formats and the social structure of eating events in the Nordic countries.
The original aim of our study was to use food and eating habits as a window
on everyday life and an indicator of current cultural aspects of modern life. It
is a recurring argument that that eating patterns in modern society are
changing (see e.g. Bell and Valentine 1997, 82). Be that as it may, our main
findings in this chapter indicate that there is both structure and variation in
Nordic meals. A significant proportion of hot meals followed a structure of
‘proper meals’. On the one hand, a hot eating event including fish fingers,
mashed potatoes, a salad and a cold dressing has been classified as a ‘proper
meal’. On the other hand, nothing but lasagne is a good example of other
types of meals, in which the centre also includes other ingredients.
158
Eating Patterns
Socio-economic differences in food use have been reduced in Western countries during the past 100 years (Beardsworth and Keil 1997, Levenstein 1988,
Mennell 1985, Roos 1998). In terms of nutrient intake there remain only marginal differences in how different social groups eat nowadays. However, there
are still some differences at the food item level and in practices related to eating. Our results show some interesting variation both between and within
countries in the choice and number of meal components, but less variation
between ‘proper meals’ and other meal formats. However, there seem to be
some differences between countries and social groups, even though they are
not especially clear or systematic. Firstly, there is variation between different
countries as Norwegians are most and Danes least into having ‘proper meals’.
Secondly, younger age groups are less likely to eat ’proper meals’ than older
age groups. Thirdly, our results indicate that men and women differ in relation
to ’proper meals’. Women seem more likely to have ’proper meals’ in general,
while there is more variation among men. Finally, people who do not work
outside the home and who are less educated seemed less prone to have ‘proper
meals’. We conclude that the position of ‘proper meal’ eating tends to fluctuate within the life cycle. The style of one’s eating varies according to age,
gender, occupational status, and education.
5
The Social Context of Eating
Lotte Holm
5.1
Introduction
Meals are sometimes described as constituting the very foundation of social
life. In his ‘Sociology of the Meal’, Simmel describes the meal as a mediator
of socialization: “That we must eat is such a primitive and lowly fact in the
development of our life values, that each individual unquestionably has it in
common with every other individual. This is precisely what makes the coming
together for a shared meal possible, and the socialization mediated thereby
promotes the overcoming of the sheer naturalism of eating.” (Simmel 1994).
The various ways in which meals are defined in the social scientific literature
depend upon which aspects of food or of the eating process are in focus
(Mäkelä 1991, Mäkelä 2000). However, most contributions to the sociology
of food and eating habits emphasize that meals are social events. This is underlined by the etymology of the word community. Its parts, com and munes,
refer to shared munitions, that is, shared provisions or supplies. Similarly,
companion or compagnon means “celui qui mange son pain avec” (Otnes
1991, 101). It would seem that the idea of the social world is conceptually
related to the world of food and eating.
Meals can be understood as the embodiment of a group - they signify unity
and sharing. “The meal briefly is the original impersonal mediator, our primary school of mediation”, writes Per Otnes in his discussion of ‘What do
meals do?’ (Otnes 1991, 105). Meals are media for bringing people together.
Otnes quotes Simmel as saying that: “Precisely the exclusive selfishness of
160
Eating Patterns
eating makes for a frequency of congregation, ties it to a habituation of being
together”. Meals unite, they assemble groups, most often groups of the kind
known as households or families, and they tend to do so at more or less regular intervals. Thereby meals form basic occasions for groups to discover
themselves as groups. “A meal may signify unity and sharing; but it certainly
does more,” writes Otnes (1991, 105). “It enacts or effects this sharing of the
not (at each individual end) really shareable - all the more important exactly
for that reason. It not only says “we share!”, it operates the actual sharing”
(Otnes 1991, 105). Therefore, Otnes suggests that one might even go so far as
to say that the meal is the group, that there could be no human groups were
there no sharing of meals.
In the discussion of the ways in which modernity influences eating habits it
has been suggested that the social context of eating is radically changing.
Modern eating is thought to increasingly take place in isolated circumstances
and less often within the social context of the family (Whit 1995, 147). This in
turn suggests that the very foundation of social life may be undergoing profound changes. Learning more about what actually occurs in regard to the social organization and contexts of eating events is therefore a topic of major
interest - not only to those who are specifically interested in food or eating
habits as such, but also to those who are interested in changing patterns of
social interaction and in everyday social life as influenced by modernity.
The focus in this chapter is upon the social contexts of eating. We will introduce some of the hypotheses put forward in the scientific literature about the
ways in which modernity influences the social contexts of everyday eating
and we will present data from our survey which describe the social contexts of
eating in four Nordic countries. The more recent discussion about modern
eating habits and how they have changed is based in many cases on empirical
studies undertaken in the USA. American trends are seen as indicating likely
developments of eating habits in the Nordic countries and in Europe more
generally. Considerable attention has been given to the idea that the locations
in which eating takes place have changed, such that the private sphere of the
home no longer constitutes the social context in which eating predominantly
takes place. It has been found that eating increasingly tends to take place outside the home - in public places such as restaurants, cafes, fast food outlets,
etc. (Dumangan and Hackett 1995, Mogelonsky 1998, Dinkins 1992). It has
also been suggested that modern eating is increasingly becoming an individual
matter, in so far as it takes place in social isolation (Falk 1994). Finally, it has
been suggested that the meal is losing its significance as an event in its own
right, since eating events increasingly take place simultaneously with other
activities (such as walking down a street, working at a desk, watching televi-
Holm: The Social Context of Eating
161
sion, etc.) (Senauer, Asp, and Kinsey 1991, Andersen 1997, Bugge and Døving 2000).
In order to address these ideas our analysis aims to present a descriptive picture of eating patterns that are widespread today among the Nordic populations. Our first and primary level of analysis therefore concerns the population
as a whole in each country. To what extent are common patterns identified in
these four countries and in which respects do these countries differ from each
other? Secondly, we identify the extent to which eating patterns within each
of these countries are homogenous or differentiated. This is accomplished by
describing the social variations in eating patterns among sociodemographically defined population groups. The central question for analysis
in regard to the latter issue is whether we find the same kind of social variations in each country.
A major focus of the analysis is upon where eating takes place and with
whom. We ask to what extent the home predominates as the place in which
eating takes place and to what extent eating takes place outside the home? We
analyse this by identifying how many of the individual’s eating events occur
inside the home and outside it, which locations are reported most frequently
and to what extent specific locations vary according to the time of the day.
Similarly, we ask whether eating events today are predominantly social
events, or whether they are individual activities that are not shared with others? We analyse this by identifying how many of the individual’s eating
events occur alone and how many occur in the company of others, with whom
respondents eat and to what extent the social contexts of eating events vary
according to the time of day.
In addition, we approach the issue of the relative status of eating as an event
in its own right by looking at the extent to which other activities take place
while eating, whether or not individuals sit down to eat and for how long they
eat. The overall aim of this chapter is to provide a broad empirical description
of how eating is socially organized in modern everyday life in the Nordic
countries.
5.2
Earlier studies of the social context of eating
The ways in which meals are located, timed and socially organized can be
understood as reflecting the demand of societal production as well as those of
the private sphere. Structural trends as regards the organization of production
162
Eating Patterns
contribute to shaping the patterns of eating practices. This patterning is experienced in everyday life as a set of conventions, which define the appropriateness, as well as the practicalities of eating.
The timing of meals, and their relative size is a topic of interest to nutritionists, since the rhythm of energy and nutrient intake is thought to influence
bodily performance and well-being (Gibney and Wolever 1997). Interestingly,
dietary recommendations in regard to the rhythm of eating vary between
countries (see chapter 1.6). This suggests that an understanding of the physiological needs of humans tends to be filtered through cultural conventions and
practices regarding the rhythms of everyday life. The discussion of dietary
recommendations, however, tends to focus more on the physiological needs of
humans than on the societal structuring and patterning of eating habits. One of
the very few studies to analyse meal patterns in their societal context was undertaken by Robert Rotenberg, and concerns the impact of industrialization on
meal patterns in Vienna, Austria (Rotenberg 1981). This study was based on
participant observation and household interviews as well as documentary materials. Rotenberg re-traced the process whereby a meal system with five daily
meals, which was commonly found in commercial cities at the turn of the last
century, was gradually replaced by the three-meal system, which is commonly
found in industrialized societies today.
The five-meal system was one that allowed daily meals to be consumed in
three different locations and social contexts: at home with family members
(morning, midday and evening), at the workplace with colleagues (midmorning) and at a café with friends and old school-mates (afternoon - the latter being reported as a common habit among middle-class men). It was a system that reflected a form of societal organization in which adults tended to
work within or close by the family residence. This made it possible for the
family to share meals in their home in the course of the workday, and it allowed domestic activities to be accommodated within the work schedule.
Women were therefore able to prepare a complex, hot meal in the middle of
the day, at the same time as they were either occupied in the home or in the
family business.
According to Rotenberg, this pattern gradually changed into a three-meal system in the course of the process whereby Vienna became industrialized. The
number of meal-breaks during the workday decreased, and their duration became shorter. This new system, which Rotenberg describes as the typical meal
pattern in industrialized societies allows three daily meals at two different locations and in two social contexts: at home with the family (morning and evening) and at the workplace with colleagues (midday). Snacks are eaten be-
Holm: The Social Context of Eating
163
tween meals. During work hours, this consumption is not organized in the
form of a meal (i.e. as a social occasion, as defined by Rotenberg), except on
special occasions. A snack is often eaten when the individual worker can take
a short break, and it is frequently eaten at the work desk or workbench. In the
same way, eating with friends or schoolmates lost the status of being a daily
activity and became an optional event that tends to take place in the course of
weekends rather than on ordinary workdays.
Rotenberg argues that the change in eating patterns and their social contexts is
strongly influenced by the industrial re-organization of everyday life in Vienna. The daily routines of the urban household became redirected in a manner that stems from changes in the organization of activities in the workplace.
Taylorism and scientific management on the one hand and trades unions on
the other became engaged in a struggle and the ensuing negotiations eventually led to a strict definition of every minute of the worker’s day. The midmorning and the mid-afternoon meals of earlier times were lost in this process. Another influential force was provided by changes in the relationship between workplace and household. By 1971, Rotenberg notes, less than one percent of employed persons worked in family firms. All others worked under
systems of management that were free from family control and did not need to
accommodate domestic activities. Furthermore, the new industries were
commonly located at the periphery of the city, far removed from traditional,
multi-functional neighbourhoods. The present pattern of the convenience
lunch can therefore be seen as an adaptation to the lack of adequate time to
prepare meals in the course of the working day and to the increase in commuting time between the home and the workplace. Consequently, the lunch,
which had played a prominent role in the family life of commercial Vienna,
declined in importance.
Rotenberg’s analysis suggests that patterns whereby the home and the workplace became the primary locations in which meals are consumed, family
members and work colleagues constituting the social contexts of such meals,
can be understood as by-products of the process of industrialization. In the
course of this process personal friends became excluded from participation in
the ordinary meals of everyday life.
A similarly profound change in meal patterns has been described as having
taken place in the course of the development from a rural to an industrial society in Finland during the period following World War II. Urbanization and
industrialization are seen as having led to the provision and consumption of
fewer meals, especially fewer hot meals (Prättälä and Helminen 1990, Prättälä
et al. 1993). It is also suggested that in this process the home has been losing
164
Eating Patterns
some of its significance as the major organizer of Finnish meals: “When parents lunched at their workplaces, and children at kindergartens or schools,
dinners began to lose their role as the main meal of the day” (Prättälä and
Helminen1990).
The question is, whether other recent changes in modern everyday life have
given rise to further changes in the social contexts of eating? In each of the
four Nordic countries examined in our survey, household size has decreased
since World War II. Families have become smaller and more individuals live
alone. It must be presumed that these structural changes have influenced the
social organization of eating events. But it is less clear whether they have led
to an overall decrease in eating events as social events.
All four countries are characterized by having a large proportion of women in
the labour force. To a varying extent they also provide public welfare services
which cater for the care of pre-school children, school children during leisure
hours and elderly citizens, thus enabling workers to work full-time for longer
periods of their lives. The question is whether these initiatives have led to a
decrease in the role of the home and the family in regard to eating patterns?
As women spend larger parts of their time outside the home, does the provision and consumption of meals also tend to occur increasingly outside the
home?
The character of the workplace has also changed in all four countries. The
larger part of the workforce is now employed in the service sector, and relatively fewer people are employed in large industrial plants with their negotiated and strictly scheduled working day. How does this influence eating habits
during the workday? Is eating during working hours still organized in structured breaks during which colleagues eat together in canteens or does eating
take place as an isolated activity at irregular times of the day, perhaps even
while working?
Income has risen in all four countries, yielding more money to spend on food
and food preparation. Has this led to an increase in more expensive eating i.e. eating outside the home? In the four countries - as in other countries - the
number of restaurants, cafes, take-aways and fast food chains has increased
dramatically during recent decades. Is this reflected in the overall eating patterns of the population on an ordinary day? To what extent are more meals
eaten in these places and when they are, do they constitute social events in the
form of shared meals or individual eating events? Is eating losing its social
significance? Is it becoming a de-structured activity with little or no ceremony
Holm: The Social Context of Eating
165
- a busy matter that takes place while other activities are accomplished at the
same time?
In the following a descriptive analysis of the social context of eating in the
Nordic countries will be presented. The description is based on data from a
cross section picture of one day of eating. It does not therefore lend itself to
conclusions about developmental trends, but the questions raised above lay
behind the interest that has formed the survey.
5.3
The present analysis of the social context of eating
The aim of this chapter is to present data that we hope will serve as one step
in the process of providing an empirical basis for the ongoing discussion of
changing eating patterns in contemporary society. We will do this by describing where Nordic people eat, with whom they eat, whether or not they sit
while eating, for how long they eat and what other activities they engage in
while eating.
Social context - operationalisation
In this chapter we operationalise ‘social context’ in the following manner:
¾ Firstly, we focus on the location of eating, by summarizing our data into
three main categories: eating at home, eating at a workplace or educational institution, and eating at other places. The category ‘other places’
includes eating while visiting other people’s homes, eating in cafes, restaurants or fast food venues and eating on a picnic or out of doors.
¾ Secondly, we analyse social context with reference to the company in
which eating takes place, focusing on four categories: eating in the company of family members, eating with work-mates, colleagues or business
associates, eating with friends or others, and eating alone.
¾ Thirdly, we analyse the character of the eating event by identifying: (a)
whether or not other activities are undertaken while eating at home - i.e.
whether individuals watch television, listen to the radio or read while eating, (b) the duration of eating events, and (c) where eating events take
place in the home, e.g. seated at a table, sitting on a sofa, etc., and where
they take place at work or in school, e.g. in a canteen, at a work desk, etc.
Eating Patterns
166
The questions related to the theme of this chapter were posed during the interviews with regard to every eating event in which more than one item had been
eaten throughout the day, starting with breakfast. The instructions to interviewers were designed to filter out events in which only one item was eaten or
one drink was consumed, such that further questions concerning social context would not be posed in regard to such events. Unfortunately, this filtering
process did not proceed as planned in Finland. The Finnish version of the
questionnaire filtered out further questions in regard to all eating events in
which cold food or drink had been consumed. This deviation has made it necessary to exclude the Finnish data from significant parts of the following
analysis.
In presenting our results we focus first upon the general patterns in each of the
countries included in the analysis, followed by an analysis of how patterns
vary between different groups in the respective populations.
5.4
Time and social context
In the following, we will describe how the social context of eating varied according to the time of day in three Nordic countries.
First, we focus on the location of eating events. The response categories for
the question Where did you eat? were: in my home, in the home of other people, in a café or restaurant, in a fast food outlet or kiosk, on an outing (in the
woods, at the beach, in the boat etc), out of doors (on the street, in a car, etc.),
at work in a canteen, in a room devoted to work-breaks, at work while working, at school, at other places. As mentioned above, we have summarized the
responses into three main categories: at home, at work or in school, at other
places.
Figure 5-1 shows that most eating took place at home on weekdays. At almost
all times of the day, more people had eaten at home than in any other place. In
each country large peaks of home eating occurred in the morning. There were
also peaks indicating home eating in the middle of the day, but they were
typically smaller. In the evening there were large and broader peaks again.
Peaks in the middle of the day indicate that many people had eaten at their
workplace or at school. These peaks appeared approximately one hour earlier
than those indicating home eating during the middle of the day.
Holm: The Social Context of Eating
167
50
D e n m a rk
n =1 1 8 7
45
40
35
Percent
30
25
20
15
10
5
3
2-
-1
3
-2
24
20
22
-2
1
9
-1
7
18
14
16
-1
-1
5
3
12
-1
1
9
8-
-1
7
6-
10
5
4-
0
T im e
50
N o rw a y
45
n=1178
40
35
Per cent
30
25
20
15
10
5
3
2-
-1
24
22
-2
3
1
20
-2
9
18
-1
7
16
-1
5
-1
14
12
-1
3
1
9
8-
-1
7
6-
10
5
4-
0
T im e
50
S w ed en
n = 12 44
45
40
35
Per cent
30
25
20
15
10
5
3
2-
-1
24
3
-2
22
1
20
-2
9
18
-1
7
-1
16
5
-1
14
3
-1
12
1
9
8-
-1
7
6-
10
5
4-
0
Tim e
H om e
W ork/school
O ther
Figure 5-1 Time and place of eating. The distribution during the day on weekdays. Percentages of populations eating at home, at work/in school or at other places.
On weekdays, eating at the workplace or at school exceeded domestic eating
only in a few cases. This occurred most clearly in Norway in the middle of the
168
Eating Patterns
day at the typical lunch hour. Here, more individuals had eaten at work or in
school than at home. In the other countries, this pattern was not quite so clear.
In the mid-morning, there were peaks indicating eating events at the workplace, clearly so in Sweden, to some extent in Denmark and less clearly so in
Norway. There were no corresponding peaks in domestic eating during the
mid-morning. In the afternoon, there was a small peak in Sweden indicating
eating events at the workplace, but not in the other countries. The number of
peaks of events that took place at work varied between the countries. Denmark and Norway had two such peaks, while Sweden had three. The pattern
therefore indicates that in Sweden there are more institutionalised breaks in
which eating can take place during the working day than in the other countries, suggesting that such breaks in Sweden are probably designed for the
consumption of meals or snacks.
Relatively few people in each country had eaten at ‘other places’ (whether
visiting others or dining out in restaurants, cafes, fast food outlets, etc.). There
was a fairly even distribution of these eating events throughout the day in
Denmark and Norway, whereas there seems to be a small peak in Sweden
around the middle of the day. This may reflect a Swedish convention of having workday lunches at restaurants near one’s workplace. For some years during the 1970s this practice was subsidized by the government. This is no
longer the case, but the practice seems to prevail.
On Sundays, eating mainly took place at home.
Figure 5-2 shows, that at all times of the day many more individuals had eaten
at home than at any other places. Also, eating at ‘other places’ was more frequent than eating at the workplace on Sundays. In Denmark and Norway,
these eating events were fairly evenly distributed throughout the afternoon,
with only small peaks during the evening, while in Sweden eating events at
other places were concentrated between the hours of 13.00 -16.00.
Data concerning eating patterns on Saturdays were only available for Denmark and Sweden (data not shown). In both countries eating on Saturdays also
took place first and foremost in the home, while eating at ‘other places’ came
second. Neither on Saturdays or Sundays were there any peaks indicating
mid-morning eating events. This confirms that these events are typically connected to the working day and the workplace, and do not constitute a part of
typical eating patterns outside the work-sphere.
Holm: The Social Context of Eating
169
80
Denmark
n=1187
70
60
Percent
50
40
30
20
10
0
5
480
6-
7
8-
9
10
-1
1
12
-1
3
5
7
-1
16
Time
Norway
n=1178
14
70
-1
18
-1
9
20
-2
1
22
-2
3
24
-1
2-
3
60
Percent
50
40
30
20
10
0
4-
5
6-
7
8-
1
-1
10
9
-1
12
3
14
-1
5
7
-1
16
Time
18
-1
9
20
-2
1
22
3
-2
24
-1
2-
3
80
S w eden
n=12 44
70
60
Percent
50
40
30
20
10
3
2-
-1
24
3
-2
22
1
-2
20
9
-1
18
7
-1
16
5
-1
14
3
-1
12
1
-1
9
8-
10
7
6-
4-
5
0
T im e
At hom e
W ork/school
O ther
Figure 5-2 Time and place of eating. The distribution during the day on Sundays. Percentage of populations eating at home, at work/in school or at other places.
170
Eating Patterns
So far, the analysis has shown that the home was the absolutely dominant location in which eating took place - even more so during weekends than on
weekdays. The analysis also shows that eating events tended to be patterned
by the location in which they took place: eating at the workplace or school
during the middle of the day occurred somewhat earlier than equivalent events
at home, and mid-morning breaks in which eating occurred at places of work
seem to follow a distinct rhythm, which differed from the pattern of eating
events at home. This was most clearly the case in Sweden.
But what about the social company in which eating events take place? A
question on this issue was posed immediately following that concerning the
location in which the eating event had taken place. The wording of the question was: Did you eat alone or with others? This was followed up by asking:
With whom? in the event that a response indicated that ‘others’ were present.
The response categories were: alone, with partner/spouse, with mother, with
father, with child(ren), with siblings, with other relatives, with boyfriend/girlfriend, with housekeeper or au pair, with colleagues, with business
associates, with friends, with others. We have summarized the responses into
four main categories: Alone, with family members (‘family’), with colleagues,
schoolmates or business contacts (‘colleagues’), with friends, boyfriend/girlfriend or others (‘friends/others’). The latter three categories are
sometimes further conflated in the following analysis to the single category of
‘social eating’, in contrast to ‘individual eating’.
Figure 5-3 shows that the proportion of the population who had eaten alone
during the day was almost the same or a bit larger than the proportion that had
eaten in the company of family members. In the late afternoon and early hours
of the evening, however, more people had eaten with family members than
had eaten alone. Both individual and social eating fluctuated throughout the
day, and largely the peaks that occurred took place during the same periods of
time. However, peaks indicating social eating events were generally larger
than those indicating individual events. This suggests that the more ‘important’ eating events, i.e. events which took place at the most popular/conventional eating hours in each country, were more often social events.
Individual eating events tended to be somewhat more evenly distributed
throughout the day, but it is also clear that many individual eating events also
followed the conventional eating rhythm. The figure indicates a pattern
whereby eating in the earlier part of the morning was more typically an individual affair, whereas eating events in the evening were typically a time for
family members to meet in the household.
Holm: The Social Context of Eating
171
Denmark
50
n=1187
45
40
35
Percent
30
25
20
15
10
5
3
2-
-1
24
3
-2
22
1
-2
9
18
20
-1
16
-1
7
5
-1
3
-1
14
50
12
1
9
8-
-1
7
6-
10
5
4-
0
Time
Norway
45
n=1178
40
35
Percent
30
25
20
15
10
5
23
24
-1
22
-2
3
20
-2
1
18
-1
9
16
-1
7
14
-1
5
12
-1
3
50
10
-1
1
89
67
45
0
Time
Sweden
n=1244
45
40
35
Percent
30
25
20
15
10
5
23
24
-1
22
-2
3
20
-2
1
18
-1
9
16
-1
7
14
-1
5
12
-1
3
10
-1
1
89
67
45
0
Time
Alone
Family
Colleagues
Friends/others
Figure 5-3 Time and social context of eating. The distribution during the day on weekdays. Percentage of populations eating alone, with family, with colleagues/business contacts or
with others.
172
Eating Patterns
In each country, eating events in the company of colleagues or business associates started later in the day, typically in the mid-morning, and stopped in the
course of the afternoon. The figure shows that the mid-morning eating event
at the workplace, which was discussed earlier, was often a social event. This
was most clearly the case in Sweden. In each country, however, the main eating event with colleagues occurred during the hours that correspond to the
typical lunch break.
Eating with friends and others played a role that is barely perceptible in
Figure 5-3. It can be added that in Denmark, these events were spread rather
evenly throughout the day, in Norway they were mainly concentrated in the
middle of the day and after 18.00 hours, while in Sweden they tended to take
place during the afternoon and evening.
It emerges from Figure 5-4 that on Sundays eating with the family was much
more common that on workdays. It can be seen that family eating events exceeded individual eating events in each country and at all times during the
day, except for the very early morning hours. In Denmark, individual and
family eating events tended to follow the same time pattern throughout the
day, whereas in Norway and Sweden this was only the case at the beginning
of the day. During the later parts of the day, individual as compared to social
eating events, were more evenly distributed throughout the hours of the day in
Norway and Sweden. Relatively few people had eaten in the company of colleagues or business associates on Sundays, while relatively more had eaten in
the company of friends or partners. Eating in the company of friends tended to
follow the same time pattern as eating in the company of family members.
These events were concentrated during the afternoon and early evening.
Data from Denmark and Sweden concerning the social context of eating
events on Saturdays (data not shown) show a similar pattern: More eating
events had taken place in the company of family members than alone.
Holm: The Social Context of Eating
173
50
Denmark
45
n=1187
40
35
Percent
30
25
20
15
10
5
23
24
-
1
3
22
-2
1
20
-2
9
18
-1
7
16
-1
5
14
-1
3
12
-1
1
10
-1
89
67
45
0
Time
50
Norw ay
45
n=1178
40
35
Percent
30
25
20
15
10
5
3
2-
3
-1
24
20
22
-2
1
-2
9
18
-1
16
-1
7
5
14
-1
3
-1
12
1
9
8-
-1
7
6-
10
5
4-
0
Tim e
50
Sweden
45
n=1244
40
35
Percent
30
25
20
15
10
5
3
2-
-1
24
3
-2
22
1
-2
20
9
-1
18
7
-1
16
5
-1
14
3
-1
12
1
-1
10
9
8-
7
6-
4-
5
0
Time
Alone
With family
Colleagues
Friends/others
Figure 5-4 Time and social context of eating. The distribution during the day on Sundays.
Percentage of populations eating alone, with family, with colleagues/business contacts, friends
and others.
174
Eating Patterns
The analysis has indicated three social contexts in which eating events primarily take place: Individual eating events, eating in the company of family
members and eating in the company of colleagues, schoolmates or business
contacts. Among these, individual eating and eating with the family constituted the social contexts that predominated. Family eating dominated during
weekends, at all hours on weekdays and in the evenings. Individual eating was
more frequent on weekdays and during the day than on weekends and during
the evening. Eating in the company of colleagues occurred on weekdays during the day. Eating in the company of friends or others appeared to play a very
small part in the general picture, indicating that such events were not a frequent habit in the general population. However, the tendency to do so was
somewhat more common during the weekend than on weekdays.
So far, this analysis has shown that people in Denmark, Norway and Sweden
did most of their eating at home. There were fluctuations reflecting the fact
that many people went to work on weekdays and that they therefore also
tended to eat at their place of work during working hours. Nevertheless, the
overall patterns clearly indicate that the home was the absolutely dominant
location in which eating took place. Our data show that everyday eating
events among Nordic populations currently take place at fewer locations than
were identified by Rotenberg as common among Viennese people at the turn
of the last century: the restaurant or café does not seem to be a frequent place
for eating today as revealed by general eating patterns among the populations
at issue.
It would appear that the family still represents the dominant social context in
which eating takes place. Whether or not this implies that eating events include all family members is an issue examined in chapter 6. However, many
individuals had also eaten alone. This theme was not discussed in Rotenberg's
analysis, partly because it was focused upon social eating events. We have no
indication as to how common individual eating events were in Vienna at the
turn of the 19th century, and historical data on trends in this regard would
seem to be very sparse.
5.5 Location and company: two dimensions of social
context
The social contexts of eating events differ in regard to location as well as
company, factors that may be combined in several ways: Eating at home can
take place alone, with family members or with friends, eating at the workplace
Holm: The Social Context of Eating
175
can take place alone or with colleagues, and so on. Two major distinctions are
drawn in Figure 5-5: whether eating events take place at home or outside the
home, and whether they take place as individual events or as social events in
the company of others, yielding four types of social context.
The findings confirm once again the familiar pattern that eating at home was
more common than that outside the home. It emerges from Figure 5-5 that in
Norway and Sweden, when eating took place at home, it was undertaken by
slightly more people as an individual event in which people had eaten alone
than as a social event in the company of others. The opposite pattern was obtained in Denmark where eating events at home more often tended to be social
events undertaken in the company of others. It also emerges that eating outside the home was much more often a social event than an individual one, and
this was the case in each of the countries at issue.
50
45
43
40
36
36
39
38
37
35
Per cent
30
25
21
20
21
18
15
10
5
5
3
4
0
D e n m a rk
A t h o m e /so cia l
A t h o m e /ind ivid u a l
N o rw a y
O u tsid e h o m e /so cia l
Sw eden
O u ts id e h o m e /in d ivid u a l
Figure 5-5 Social context and location of eating. Per cent of total number of eating events.
(N= Denmark: 3266, Norway: 3363, Sweden: 3920).
5.6
Eating out
It has been clear from the analysis presented thus far that eating at ‘other
places’, which include restaurants, cafes and the like, was not a very common
pattern among the Nordic peoples. In the more general patterns of eating
events, those that took place at ‘other places’ were barely perceptible. In fact
an average of 4-6% of eating events took place at ‘other places’. It transpired
that eating while visiting other people’s homes was the most frequently used
Eating Patterns
176
response category among ‘other places’ in all countries, while eating events in
cafes or restaurants was the second most frequently used category (data not
shown).
100 %
9
13
90 %
7
7
80 %
37
46
70 %
45
54
60 %
50 %
40 %
33
30 %
20 %
32
35
29
23
10 %
11
12
F in la n d
N o rw a y
8
0 %
D e n m a rk
W e e k ly
M o n th ly
1 -1 1 /y e a r
Sw eden
N e ve r
Figure 5-6 Eating out. Per cent of populations with different frequencies of eating out.
N= Denmark: 1187, Finland: 1200, Norway: 1178, Sweden: 1244.
Our questionnaire also included a more general question about the frequency
of eating out at restaurants, cafes etc. Finland is included in the analysis with
regard to this question. The responses to this question demonstrate that there
were some variations between the countries (Figure 5-6). The results demonstrate that more Swedes eat out on a weekly basis than any of the other populations, while relatively more Danes eat out less frequently than once a month.
Among those who had eaten out in a restaurant or café on the day preceding
the interview, more Danes had eaten in the company of family or friends than
was the case in the other countries, while relatively fewer had dined with colleagues or business contacts (data not shown). Our findings suggest that in
Denmark eating out constitutes a leisure-time activity to a greater extent than
in the other Nordic countries, where it appears to be an activity that is anchored to a greater extent in working life (Bjørkum et al. 1998). Furthermore,
among those who had eaten at a restaurant or café on the day preceding the
interview, 48% of Swedes and 55% of Finns considered that ‘yesterday’ had
been an ‘ordinary day’, whereas only 35% and 38% did so respectively in
Denmark and Norway (Bjørkum et al. 1998). This pattern suggests that eating
out is a relatively more integrated aspect of ordinary everyday life in Sweden
and Finland than is the case in Denmark or Norway. The questions regarding
Holm: The Social Context of Eating
177
the place of specific meals showed that when individuals in the three Nordic
countries had eaten out, relatively more of them had eaten in restaurants or
cafes than in fast food venues (Finland excluded again). In Denmark and
Norway, 6% of the population reported that they had visited a restaurant or a
café on the day before the interview, while in Sweden 14% had done so. In
Denmark, less than 1% reported that they had visited a fast food venue, while
2% and 3% had done so in Norway and Sweden respectively.
Thus far, the units of analysis have been national populations. The question
remains to what extent the patterns we have identified are uniform patterns in
each of the populations under examination, or whether they are the result of
an aggregate pattern that rests upon a variety of differences among specific
population groups. We will pursue this issue in the following analysis while
maintaining cross-national comparisons, since differences between countries
can imply that living conditions influence eating patterns in culturally specific
ways.
5.7
Variations in the social context of eating
A large majority in all three countries had eaten more than twice at home on
the day preceding the interview, averages ranging between 2.5 and 2.7 of eating events. Only 3-4 percent of the respective populations had not eaten at
home at all during that day. Table 5-1 shows how the average number of eating events at home varied between population groups. It also shows the proportion of each population group that had eaten at home less than twice on the
day preceding the interview.
It emerges from Table 5-1 that the largest variations with respect to eating at
home in all three countries are found according to age. The number of eating
events at home increased with age. On average, older age groups had one extra eating event or somewhat more as compared to the youngest age group.
Only 4-8% of the oldest age group had eaten at home less than twice, while
approximately one third of the youngest age group had done so. The overall
significance of age was also reflected indirectly in some of the other variations that can be identified. People with shorter education and couples with no
children are over-represented among the older age groups, and the tendency to
eat more at home also emerged in relation to these population groups. Among
households with children, couples ate more frequently at home than did single
parents.
Eating Patterns
178
Table 5-1 Eating at home. Average number of eating events at home in population groups
and proportion of population who had less than 2 eating events at home on the previous
day. 35
Denmark
Norway
Sweden
% with
% with
% with
Mean s.d.
<236 Mean s.d.
<2
Mean s.d.
<2
Country
(N)
Total population37
(1187)
2,7
1,3
19
(1178)
2,5
1,2
20
(1244)
2,6
1,2
16
Gender
(N=
country)
Men
Women
sig
2,5
2,8
***
1,2
1,3
20
18
***
2,4
2,6
***
1,1
1,2
23
17
**
2,5
2,8
**
1,2
1,2
18
14
0
Age
(N=
country)
15-24
25-34
35-44
45-54
55-64
65+
sig
1,8
2,3
2,6
2,7
3,1
3,2
***
1,0
1,3
1,3
1,2
1,1
1,2
38
30
21
13
6
8
***
2,2
2,0
2,4
2,5
3,0
3,2
***
1,1
1,2
1,0
1,1
1,1
1,0
27
33
19
18
9
4
***
2,2
2,3
2,4
2,6
2,9
3,4
***
1,0
1,2
1,0
1,1
1,3
1,1
26
23
17
14
11
5
***
3,2
2,4
***
1,2
1,2
7
24
***
3,1
2,3
***
1,0
1,1
8
23
***
3,2
2,4
***
1,2
1,1
7
20
***
(1107)
2,6
2,9
2,4
1,6
**
1,3
1,3
1,2
1,3
20
14
19
19
NS
(1102)
2,5
2,8
2,2
2,4
***
1,1
1,2
1,1
1,2
17
15
33
24
***
(1169)
2,6
2,9
2,3
2,5
***
1,0
1,2
1,0
1,3
13
10
20
23
***
Education (N)
Mandatory
Further education
Academic
sig
(1187)
2,8
2,6
2,6
*
1,2
1,3
1,3
15
21
19
**
(1174)
2,8
2,5
2,4
***
1,2
1,1
1,6
14
20
22
**
(1244)
3,0
2,7
2,5
***
1,3
1,2
1,2
11
14
19
***
Urbanisa- (N)
tion
<25.000
25-99.999
100.000+
sig
(1177)
2,8
2,6
2,5
*
1,3
1,3
1,2
17
21
21
NS
(1178)
2,4
2,6
2,6
NS*
1,1
1,2
1,2
21
18
19
NS
(1153)
2,7
2,7
2,5
NS
1,2
1,2
1,3
16
14
20
NS
Occ.
Not active
Active
status
(N=countr sig
y
Household
type
35
(N)
2+ adults w/children
2+ adults no childr.
1 adult with children
1 adult
sig
Statistics: Differences between means: ANOVA. Chi-square test for differences in pct of
population groups who had varying number of eating events at home is based on the full distribution between 0,1,2, and more eating events at home.
36
Percent of population who had less than 2 eating events at home on the previous day
37
Differences between means in countries are statistically different, p<0,000
Holm: The Social Context of Eating
179
This pattern indicates that the arranging of eating events in the home depended more on the presence of adults (and children) than on that of children
only. In each country, women tended to eat at home a little more frequently
than men. In Denmark, people living in large cities ate somewhat less frequently at home than those living in smaller urban areas. This trend did not
emerge in Norway or Sweden.
For a considerable proportion of the population in each country, all eating
events had taken place at home on the day preceding the interview. In Denmark more than half of the populations had only eaten at home (52%), while
in Norway and Sweden somewhat less than half (41% and 43% respectively)
had done so. Older people (aged 45 or more) and those who were not occupationally active were more likely to be included among those who had eaten
exclusively at home (data not shown).
Eating at the workplace or in school was a common pattern. In Denmark, 27%
had done so, while among the occupationally active, the figure rose to 40%. In
Sweden, the equivalent figures were 34% of all and 46% of the occupationally
active; while in Norway 41% of all and 53% of the occupationally active had
eaten at their place of work or in school. Among those who had eaten at their
place of work, the average number of such eating events was 1,3 in Norway,
1,6 in Denmark and 1,7 in Sweden. Unsurprisingly, the largest social variations in regard to eating at a place of work or in school were found in relation
to occupational status. Hardly any of those who were not occupationally active had eaten at either such location.
Small but significant variations according to gender were identified in each
country. Men tended to have had relatively more eating events at work or in
school than women (on average, 0.5 versus 0.3 in Denmark, 0.6 versus 0.5 in
Norway and Sweden). These differences reflect the fact that relatively more
men than women are occupationally active. When occupational status was
controlled, no significant difference between men and women in regard to
eating patterns at work or in school remained (data not shown, statistics: chisquare test).
Eating at ‘other places’ was not a very common feature. Only 25% had done
so in Denmark and Norway and 32% in Sweden (averages ranging between
0.4 and 0.5 of eating events on the day preceding the interview). No significant differences between men and women on this point emerged in any of the
countries. In Norway and Sweden, eating at ‘other places’ occurred to a
greater extent among younger age groups. There were no significant differences between individuals living in different types of households. Perhaps
180
Eating Patterns
contrary to expectations, our data indicate that individuals who live alone do
not tend to eat at ‘other places’ (visiting the homes of other people, restaurants, cafes, etc.) to a greater extent than those whose households comprise
two or more persons (data not shown).
Eating with family members was earlier shown to be the most typical social
context in which eating took place. It should be noted at this stage of the
analysis that the category ‘family’ here refers to an eating event that took
place in the company of one or more family members. It does not necessarily
imply that all family members were present. To what extent such events were
instances of what has been discussed in the scientific literature, as the ‘family
meal’ is an issue that will be considered in more detail in chapter 6.
Table 5-2 documents that relatively more Danes had eaten in the company of
family members than had Norwegians or Swedes. 71% of the Danish sample
had had such eating events on the day preceding the interview, while 67% in
Norway and 63% in Sweden had done so. On average, Danes had participated
in 2.6 eating events in the company of family members, while Norwegians
and Swedes had participated in 2.1 and 2.0 respectively. This table shows
clearly that eating with family members varied greatly according to household
type. As might be expected, very few of those living in one-person households
had eaten with family members (not part of their own household), whereas the
large majority of people living in all other types of households had done so.
Individuals from households, which comprised two or more adults as well as
children, had eaten in the company of family members more often than any
other type of household.
In Norway, relatively more women than men had eaten in the company of
family members. The same tendency was seen in Denmark and Sweden, but
here the differences were not significant. Age was significant in each country.
Relatively more people in the age group 35-44 years had eaten in the company of family members than either younger or older age groups. Urbanization was a significant factor in Denmark only, where it emerged that more
people living in smaller cities or towns had eaten with family members than
those resident in larger urban areas.
Holm: The Social Context of Eating
181
Table 5-2 Eating with family members. Average number of eating events in population
groups and proportion of population who had no eating events with family members on
the previous day.38
Denmark
Norway
Sweden
% with
% with
% with
Mean s.d.
039
Mean s.d.
0
Mean s.d.
0
Country
(N)
Total population40
(1187)
2,6
2,5
29
(1178)
2,1
2,1
33
(1244)
2,0
2,1
37
Gender
Men
Women
sig
2,4
2,7
NS
2,4
2,6
30
28
NS
1,8
2,3
***
1,9
2,2
36
30
**
2,0
2,1
NS
2,1
2,2
37
37
NS
Age
15-24
25-34
35-44
45-54
55-64
65+
sig
2,1
2,9
3,5
2,7
2,3
1,9
***
2,3
2,7
2,9
2,3
1,9
2,1
35
24
20
21
25
47
***
2,0
1,9
2,6
1,9
2,0
1,9
***
2,0
2,3
2,1
1,8
2,0
1,9
36
36
21
30
32
40
***
1,5
2,0
2,5
2,1
2,1
1,9
**
2,0
2,5
2,3
2,0
2,0
2,1
50
40
24
28
27
46
***
2,2
2,8
***
2,2
2,6
40
24
***
1,9
2,1
NS
1,9
2,1
37
31
*
2,0
2,0
NS
2,1
2,1
39
36
*
2,7
1,9
10
8
(1102)
3,2
2,4
2,2
1,7
9
13
(1169)
3,5
2,8
2,3
1,9
11
11
2,3
1,4
15
80
***
1,6
0,3
***
1,5
0,9
33
85
***
1,8
0,3
***
1,7
0,8
25
82
***
Occupa- Not active
tional
Active
status
sig
Household
type
(N)
(1107)
2+ adults w/children
4,1
2+ adults no children 3,0
1 adult with children
2,6
1 adult
0,5
sig
***
Education (N)
Mandatory
Further education
Academic
Sig
(1187)
2,5
2,6
2,7
NS
2,3
2,6
2,4
30
29
27
NS
(1174)
2,1
2,0
2,2
NS
1,9
2,0
2,2
31
34
31
NS
(1244)
2,0
2,1
1,9
NS
2,1
2,2
2,1
36
33
40
NS
Urbanisa- (N)
tion
<25.000
(1177)
2,9
2,3
2,1
***
2,7
2,3
2,2
26
32
33
**
(1178)
2,1
2,1
1,9
NS
2,1
2,1
1,9
31
32
37
NS
(1153)
2,3
2,1
1,8
NS
2,2
2,1
2,2
33
35
41
NS
1
38
25-99.999
100.000+
sig
Statistics: Differences between means: ANOVA. Chi-square test for differences in pct of
population groups who had varying number of eating events at home is based on the full distribution between 0,1,2, and more eating events with family members
39
Percent of population who had no eating events with family members on the previous day.
40 Differences between means in countries are statistically significant
, p<0,000
182
Eating Patterns
Table 5-3 documents that somewhat more Swedes and Norwegians had eaten
alone on the day preceding the interview than had Danes. A total of 67% of
Swedes had eaten alone at least once, while 66% of Norwegians and 58% of
Danes had done so. On average, Swedes had eaten alone 1.4 times, Danes 1.3
and Norwegians 1.2 times (remember that the total number of events was
lower in Norway). Not surprisingly, household type was the significant factor
in accounting for variations on this point. Individuals living in one-person
households had eaten alone three times more often than had others. Among
members of multi-person households, those who belonged to households
comprising only one adult and children had eaten alone to a greater extent
than had members of other household types. Age and occupational status also
appeared to have impacts on the tendency to eat alone and, taken together,
they indicate that the older generations had more often eaten alone than
younger generations. Variations according to educational background were
small, but point in the same direction. People with the minimum mandatory
education had more individual eating events than had people with further education, including academic education. There were no significant variations
according to gender on this point in any of the countries. In Sweden and Norway, the average number of individual eating events was largest among people resident in big cities, but only significantly so in Norway.
The proportion of people in the three countries who had eaten in the company
of colleagues or business contacts was 23-33%. Differences between population groups were large and varied primarily according to occupational status,
hardly any of those who were not occupationally active having eaten in such
company (averages: 0.02-0.05 such events), whereas between 58% and 68%
of the occupationally active had done so, with an average between 1.0 and 1.2
such events. Individuals in the age group 25-44 years had participated in more
eating events with colleagues than other age groups, and in each country small
but significant gender differences are found. Men had participated in more
eating events in working or professional contexts than had women.
Eating with friends was less frequent than eating in any other kind of company. Only 18-24% of individuals had participated in such eating events. In
Denmark and Norway the average for such events was 0.3, while in Sweden it
was 0.4. The youngest age group (15-24 years) was the population group that
had eaten with friends to a significantly greater extent than any other age
group. Approximately half of the people belonging to this age group had done
so. Single adults had eaten with friends to a greater extent than had other
adults whether or not they lived alone or in households with children.
Holm: The Social Context of Eating
183
Table 5-3 Eating alone. Average number of individual eating events in population groups
and proportion of population who had no individual eating events a day. 41
Denmark
Norway
Sweden
% with
% with
% with
Mean s.d.
0
Mean s.d.
0
Mean s.d.
042
Country
(N)
Total population43
(1187)
1,2
1,3
42
(1178)
1,2
1,2
34
(1244)
1,4
1,3
33
Gender
Men
Women
sig
1,1
1,2
NS
1,3
1,3
41
43
NS
1,3
1,2
NS
1,2
1,2
32
37
NS
1,3
1,4
NS
1,2
1,4
34
32
NS
Age
15-24
25-34
35-44
45-54
55-64
65+
sig
0,9
0,9
0,9
1,1
1,3
1,7
***
0,9
1,1
1,0
1,3
1,4
1,6
40
44
47
43
39
27
***
1,2
1,1
1,0
1,3
1,4
1,6
***
1,1
1,1
1,1
1,1
1,2
1,5
30
36
39
30
33
36
***
1,3
1,2
1,1
1,2
1,4
1,8
***
1,1
1,1
1,1
1,3
1,4
1,7
21
37
39
35
33
35
***
1,6
1,0
***
1,6
1,1
37
44
***
1,5
1,1
***
1,5
1,1
34
34
***
1,6
1,2
***
1,5
1,2
35
32
***
1,0
1,0
54
59
(1102)
0,8
0,9
0,9
1,0
44
44
(1169)
0,8
0,8
1,0
1,0
43
54
0,9
1,4
33
8
***
1,1
2,2
***
1,1
1,2
34
11
***
1,0
2,3
***
0,9
1,3
31
8
***
Occupa- Not active
tional
Active
status
sig
Household
type
(N)
(1107)
2+ adults w/children
0,7
2+ adults no children 0,7
1 adult with children
0,9
1 adult
2,4
sig
***
Education (N)
Mandatory
Further education
Academic
sig
(1187)
1,3
1,2
1,0
*
1,3
1,4
1,3
38
42
46
NS
(1174)
1,4
1,3
1,1
**
1,3
1,2
1,1
34
33
36
*
(1244)
1,5
1,3
1,3
NS
1,5
1,3
1,2
36
36
30
**
Urbanisa- (N)
tion
<25.000
(1177)
1,1
1,3
1,3
NS
1,3
1,4
1,3
45
41
36
**
(1178)
1,2
1,3
1,4
*
1,1
1,2
1,3
35
36
31
*
(1153)
1,2
1,3
1,4
NS
1,2
1,3
1,3
35
36
28
NS
25-99.999
100.000+
sig
41
Statistics: Differences between means: ANOVA. Chi-square test for differences in pct of
population groups who had varying number of eating events at home is based on the full distribution between 0,1,2, and more eating events at home.
42
Percent of population who had less than 2 eating events at home on the previous day
43
Differences between means in countries are statistically different, p<0,000
184
Eating Patterns
Occupationally active people and those with higher academic education had
also participated in more eating events with friends than had others, while this
pattern was also found among women as compared to men in Norway. No
gender differences were found in Denmark or Sweden.
The analysis in this section has demonstrated that the location and company in
which eating events take place varies between population groups. Some of
these variations follow patterns that might be expected, given the varying living conditions among population groups. Thus, those who were living alone
tended to have eaten alone to a greater extent than others, while those who
were employed or following a course of education had to a greater extent than
others eaten at these locations and in the company of colleagues or business
contacts, etc. These variations appear to be straightforward reflections of variations in material living conditions. However, there were also a variety of
practices among and between the members of population groups. Those who
were living alone did not always eat alone, those living with family members
did not always eat with family members, and not all of those who were employed had eaten in the company of colleagues or professional contacts. Even
though socio-demographic characteristics and living conditions influence eating practices, these data demonstrate that there are also variations that cannot
be ascribed to living conditions as such.
Furthermore, there were other variations that were less straightforward but
nevertheless constituted systematic differences: women had more eating
events at home and with family members than had men; men had more eating
events outside the home and with colleagues. Younger age groups had to a
greater extent eaten at other places and with friends; older age groups had to a
greater extent eaten at home. The occupationally active had not only eaten at a
place of work to a greater extent than had others, they had also eaten to a
greater extent at other places as well as in the company of family members,
colleagues and friends. Those who were not active on the other hand had more
often eaten at home and alone. People living in households with one or more
other adults had to a greater extent eaten at home, with family members and
with colleagues; singles had to a greater extent eaten at the workplace, with
colleagues and with friends; those with lower education had eaten at home to
a greater extent, whereas those with further or academic education had more
often eaten at other places and with friends.
Taken together, these results suggest that in spite of female participation in
the workforce women are more tied up with the duties of home and family
than are men. They also suggest that those who are occupationally active have
Holm: The Social Context of Eating
185
more social contacts, not only with people they meet through working activities but also with family members and friends. However, contact with friends
appears to be much more widespread among the younger age groups, suggesting that such activities in so far as they involve eating events occur more frequently among those who have not yet formed families of their own. The
older age groups seem to be characterized by eating to a lesser extent outside
the home and by a greater tendency to eat alone. The presence of children in
the household is related to a greater tendency to eat in the company of family
members, but this tendency is even more pronounced among households that
include children as well as more than one adult. This difference between single parents and couples with children may reflect that joint eating is more
likely to occur when more than one adult participates in the event. On the
other hand, it may reflect a common Nordic pattern whereby children of separated or divorced couples live alternatively at the residence of their mother
and father, often changing residence on a weekly basis. This pattern would
also imply that single parents regularly live alone on certain weekdays or
weeks.
5.8
The character and course of eating events
Respondents were also asked in more detail about the character and course of
their eating events. These questions focused upon the specific place of eating
events in the home and at the workplace, respectively, the duration of eating
events and which other kinds of other activity, if any, took place while eating.
For eating events at home, questions were asked about whether and where
informants sat down while eating - at a dining table, at a coffee table (while
seated on a sofa or armchair), somewhere else or not at all.
Figure 5-7 shows how eating events at home were distributed between different seating arrangements. It emerges that the majority of these eating events
took place at a dining or kitchen table in all three countries, but the proportion
was somewhat smaller in Norway than in Denmark and Sweden. Correspondingly, more Norwegian eating events took place at a coffee table. Only 1-2 per
cent of eating events took place without the eating person being seated. Data
about how these seating arrangements were practiced by individuals throughout the day (data not shown) indicate that a large majority of people (80-90%)
had eaten at such a table, while relatively fewer had eaten at a coffee table. In
Norway, however, this latter practice was a little more widespread than in
Denmark or Sweden: 53% of Norwegians had eaten at a coffee table at some
point during the day preceding the interview, while 43 and 40% respectively
Eating Patterns
186
had done so in Denmark and Sweden. These results also indicate that eating
while walking around the house was not at all common. Only a small minority
had eaten at home without being seated (5.3% in Denmark, 4.6% in Norway
and 2.4% in Sweden). So, eating events at home would seem to imply that
people generally sit down to eat - quite often on a sofa at a coffee table, but
most frequently it seems at a proper dining table.
Per cent
Figure 5-7 Eating arrangements – where to sit. Percentage of eating events at home in dif100
90
80
70
60
50
40
30
20
10
0
76
73
65
30
23
19
2
2
2
Denmark
Norway
Kitchen/dining table Coffee table/sofa
4
2
1
Sweden
Elsewhere Didn't sit
ferent kinds of sitting arrangements. N= Denmark: 3160, Norway: 2945, Sweden: 3170).
100
90
80
70
79
75
68
Per cent
60
50
40
32
30
25
21
20
10
0
Denmark
Canteen
Norw ay
Sw eden
At work desk
Figure 5-8 Eating at the workplace. Percentage of eating events in different places. N= Denmark 437, Norway: 553, Sweden: 607).
Holm: The Social Context of Eating
187
At the workplace eating can take place in a canteen or other room designed
for such breaks or alternatively at the workbench or desk without leaving
work. Figure 5-8 shows how eating events at the workplace were distributed
between these two main kinds of arrangement.
The results presented in Figure 5-8 show that most eating at workplaces had
taken place in canteens or rooms designed for breaks at a remove from work
as such. However, between one fifth and one third of eating events had taken
place without the respondent leaving his or her work. Cross tabulations (not
shown here) revealed that most individuals who had eaten at their workplace
on the preceding day had done so both in rooms designed for this purpose and
while remaining at their work. A small group of them, between 5 and 7%,
however, had only eaten without leaving their work. No systematic sociodemographic differences with respect to this practice could be identified in
our data.
For all eating events in which more than one item had been consumed, respondents were asked to indicate the duration of the event (at 10-minute intervals, ranging from less than ten to more than 40 minutes).
50
47
45
39
40
37
35
32
29
30
Percent
27
27
25
22
23
20
20
19
16
15
15
12
9
10
8
5
2
8
7
3
0
D en m ark
<10 m in.
Fin lan d
10 -20 m in.
2 1-3 0 m in.
Norw ay
31-40 m in .
S w eden
>4 0 m in.
Figure 5-9 Eating arrangements – how long did the eating event take? Percentage of eating
events at home with different durations. N= Denmark: 4792, Finland: 5159, Norway: 4221,
Sweden: 5043).
The distribution of eating events by duration is presented in Figure 5-9. The
Finnish data are included in this part of the analysis since on this point the
questionnaire design was similar to that employed in the other Nordic coun-
188
Eating Patterns
tries. It transpires that in all four countries most eating events had lasted 10-20
minutes, and the least usual response had been 31-40 minutes. There were,
however, clear and striking national differences especially between Finland
and Denmark. In Finland, one third of all eating events had lasted less than 10
minutes and hardly any had lasted more than 40 minutes. In Denmark, in contrast, only 15% of eating events had lasted less than 10 minutes, while 20%
had lasted more than 40 minutes.
Our data indicate that in all four countries more than half of the population
had participated in eating events that lasted 11-20 minutes (data not shown).
In Finland, nearly 70% of the population had participated in very brief eating
events (less than 10 minutes), while in Denmark only 39% had done so. (The
equivalent figures in Norway and Sweden were 45 and 56%, respectively). In
Finland, only 11% of the population had participated in an eating event that
had lasted more than 40 minutes, while in Denmark 51% had done so (the
equivalent figures in Norway and Sweden being 24 and 36%).
These data indicate that individuals had typically participated in eating events
of variable duration during the day preceding the interview. But they also indicate apparent differences in eating habits as between the Nordic countries.
The question posed to respondents was formulated in a similar manner in each
of the four countries, as follows: For how long did the eating event last?
There may have been some variation in respondents’ understanding of this
question. Some may have reported for how long they had remained seated at
the table, while others may have reported the duration of actual eating as such.
Based on knowledge of our native cultures, however, we see no reason why
Finns or Danes should systematically understand this question in different
ways. We therefore confidently interpret these results as indicating genuine
cultural differences between these countries.
There were small but significant social variations with respect to the duration
of eating events (data not shown): Brief eating events that lasted less than 10
minutes were more frequent among the occupationally active, and also among
women in Finland, Norway and Sweden, but not in Denmark. Single parent
households had more such brief eating events and fewer events of long duration (lasting more than 40 minutes) than any other household type. Variations
according to age and educational level were unsystematic.
Holm: The Social Context of Eating
189
Denmark
100 %
80 %
60 %
40 %
20 %
0%
At home/social
At home/alone
Outside
home/social
Outside home/alone
Norway
100 %
80 %
60 %
40 %
20 %
0%
At home/social
At home/alone
Outside home/social Outside home/alone
Sweden
100 %
80 %
60 %
40 %
20 %
0%
At home/social
<10 min.
At home/alone
10-20 min.
21-30 min
Outside
home/social
31-40 min.
Outside
home/alone
>40 min.
Figure 5-10Types of eating arrangements and their duration. Percentage of eating events
in different contexts and locations by duration. N: Denmark 3257, Norway 3334, Sweden
3917.
Eating Patterns
190
The results presented in Figure 5-10 show that individual eating events were
more frequently of brief duration than were social eating events and that eating events of longer duration were more frequently social events. Furthermore, whether eating events were individual or social, those which took place
at home were more frequently of longer duration and less frequently brief than
eating events outside the home. A similar pattern with regard to the duration
of different types of eating events seems to be reflected in Denmark, Norway
and Sweden (again Finland had to be excluded). The national differences in
duration seem to be particularly pronounced for eating events at home.
When eating events had taken place in the home of the respondent, further
questions were asked about what other activities had been carried out while
eating (watching TV, listening to radio, reading, or none of these activities).
The data presented in Figure 5-11 indicate that in the three countries between
39 and 45% of eating events at home were events in which no other activity
had been carried out simultaneously. The remaining events (half of all eating
events at home or more than half) were thus events in which other activities
had taken place while eating. Some national differences seem to be at issue on
this point. In Norway, the practice of watching television while eating would
seem to be slightly more widespread than listening to radio. In Denmark and
Sweden preferences appeared to be the other way around. Reading while eating was more widespread in Sweden than it was in Denmark or Norway.
60
50
45
43
39
40
26
30
24
20
20
21 22
21
11
17
9
10
0
Denmark
TV
Radio
Norway
Reading
Sweden
Neither
Figure 5-11 Eating arrangements – what went on while eating? Percentage of eating events
at home with different kinds or no simultaneous activity. N= Denmark: 3268, Norway: 3028,
Sweden: 3323.
Approximately 40-50% of the population in each country had watched television while eating at some point during the day preceding the interview, 36-
Holm: The Social Context of Eating
191
44% had listened to radio, while 20- 40% had read while eating. However, 6167% had participated in one or more eating events on the previous day in
which no other such activities had taken place simultaneously (data not
shown). This indicates that individuals had typically participated in several
kinds of eating events during the particular day that preceded the interview.
There were some variations between population groups with respect to eating
while listening to radio (the practice being more common among the elderly)
and while reading (variations being unsystematic). The practice of eating
while watching television, however, seemed to be very evenly distributed between population groups in each country, only few and small differences being significant. In Sweden, the practice was somewhat more widespread
among people living alone, while this was the case in Norway among the
youngest and the oldest age groups.
Some people had not engaged in any other activities while eating at any point
of the previous day. This was the case for 21% in Sweden, 25% in Denmark
and 26% in Norway. This pattern varied first and foremost according to age,
being most widespread in the age group between 25 and 44 years, when all
other background factors were controlled for (logistic regression analysis,
odds ratio 1.5, reference group 45+ years).
100 %
31
29
33
80 %
46
60 %
16
19
45
49
27
9
12
6
40 %
24
29
20
20
21
28
25
24
22
24
20 %
20
21
I n d iv id u a l
S o c ia l
I n d iv id u a l
S o c ia l
I n d iv id u a l
S o c ia l
D e n m a rk
D e n m a rk
N o rw a y
N o rw a y
Sweden
Sweden
0 %
TV
R a d io
R e a d in g
N e ith e r
Figure 5-12 Types of eating arrangements at home: what went on while eating? Percentage of eating events in different contexts and location with simultaneous activities. N= Denmark: 2454, Norway: 2378, Sweden: 2694.
192
Eating Patterns
Whether or not other activities were carried out while eating depended on
whether the event was an individual or a social one. Figure 5-12 shows that
simultaneous activities occurred to a greater extent in conjunction with individual eating events than with social ones. Reading was markedly more common when eating alone, while listening to radio or watching television did not
seem to depend upon whether the eating event was an individual or social one.
To sum up, this analysis of the character and course of eating events indicates
that most eating events in the home had taken place while seated at a table - a
kitchen or dining table (the most frequent), or coffee table. There was little
evidence that eating events, whether in the home or the workplace, had taken
place while standing or walking around. Most of those who had eaten at their
place of work on the previous day had done so on at least one occasion in a
canteen or other room that was devoted to this purpose.
Many eating events were of brief duration, 10-20 minutes, but the typical duration of eating events also varied greatly between the four countries, a tendency to very brief events (less than 10 minutes) being a characteristic pattern
in Finland and a tendency to longer events being a characteristic pattern in
Denmark. There were smaller but significant differences between population
groups, the tendency to relatively brief eating events being characteristic of
those who were occupationally active and those living in single-parent households.
It was common to engage in other activities while eating. About one third of
the populations had done so in relation to all eating events, while the remainder had participated in at least one eating event in which no such activities had
occurred. There were also some differences between population groups with
regard to which activities had taken place. However, watching television while eating was relatively evenly distributed among different population groups,
and this practice as well as listening to radio had occurred independently of
whether the eating event was an individual or social one. There were no
indications that such practices might be increasing - that watching television,
for example, had been more widespread among younger age groups. On the
contrary, there were relatively more people among the younger age groups
who had not engaged in any such activities in conjunction with eating events.
Holm: The Social Context of Eating
5.9
193
Discussion
Quantitative studies of the locations and social contexts of eating are rare, and
we have not been able to identify any studies that have registered patterns of
behaviour with regard to where, with whom and how people in modern societies eat. We are therefore not in a position to compare most of our findings
with the results of similar surveys undertaken in other countries. Dietary surveys, which are undertaken in the Nordic countries on a more or less regular
basis, do not systematically register the locations or social contexts in which
the intake of nutrients occurs. Some few studies have focussed on the issue of
location, but in regard to more general patterns we are not able to compare our
present findings with earlier data. This entails that we are also unable to discuss to what extent these findings might represent new trends or gradual
changes and, if so, in what direction.
However, eating out is an issue that has been in focus in several studies. In the
American nationwide and representative food consumption survey CSFII (the
Continuing Survey of Food Intakes by Individuals, USDA 1998), eating out is
registered. In this survey, the question: Was this food ever in your home? was
posed, following each report of food that the respondent had consumed. On
this basis, domestic consumption was differentiated from ‘eating out’. This
survey documented that on any given day 57% of Americans would eat at
least one food away from home. This represented a 33% increase over a period of approximately two decades, the corresponding figure from a 1977-’78
survey having been 43%. Furthermore, the type of eating out had also
changed during that period. During the 1970s, eating out had first and foremost taken place in restaurants, while eating out in someone else’s home had
ranked second, and eating in fast-food venues third (USDA 1998). By the
1990s, one third of those eating out did so at fast-food venues and only 27%
in restaurants. The most frequent type of meal eaten out was lunch. On any
given day, 40% of men in the age group 12-59 years ate at a fast-food venue,
while one third of women in the same age group also did so.
Compared to these data, the results of our survey document that the Nordic
pattern of eating out is very far removed indeed from the American pattern,
which is nevertheless often adopted as the exemplar of ‘modern eating’ (cf.
(Senauer, Asp, and Kinsey1991). Our data indicate that eating out in fast-food
venues or restaurants is a pattern that is less widespread than eating out in
other people’s homes, while eating at home and at one’s place of work are the
most common locations in which eating events take place. It also emerges that
eating out in fast-food venues is less widespread in the Nordic countries than
visits to restaurants and cafes. Contrary to the USA, fast-food venues do not
194
Eating Patterns
seem to constitute an important element in ordinary everyday eating patterns
among the populations of the Nordic countries. In Sweden, Norway and
Finland eating out in restaurants, cafes or fast food venues does constitute a
feature of the lunch break among some population groups, while this is not
generally the case in Denmark. Furthermore, our data show that in Denmark
and Norway eating out seems to be considered as an event that is out of the
ordinary (Bjørkum et al 1998). Our survey documents that the dominant pattern of everyday eating in the Nordic countries is a home-based activity, while
eating out - other than at one’s place of work - would seem to be rather more a
leisure-time activity than an everyday routine. These patterns underline some
of the marked differences that appear to obtain between eating habits in the
USA and in the Nordic countries.
In Britain, the frequency of eating out was surveyed among 1000 people living in three cities (London, Bristol, Preston) (Warde and Martens 2000). According to this survey, 21% of inhabitants in these cities ate a main meal out
on a weekly basis, 44% did so monthly, 29% occasionally or annually, while
7% never did so. These frequencies are broadly similar to our findings among
Swedes only, and are considerably higher than the frequencies with which
people eat out in other Nordic countries. The proportions of the population
that never eats out are similar, however. It would seem that eating out is an
activity engaged in by approximately the same proportion of the population in
Denmark, Finland and Norway as well as in Britain, but only in Sweden does
the frequency of doing so reach the same level as that registered in some British cities.
5.10 Conclusions
This analysis has revealed that eating today in the Nordic countries takes
place first and foremost in the home. The majority of people eat the larger part
of their meals in their own home, and approximately half of the populations in
each country eat all of their meals at home. Those who are occupationally active, either because they have a job, are at school or otherwise pursuing a
course of education, also eat at their place of work or education during the
day. These general patterns are common to all three of the Nordic countries
that were included in this analysis.
Eating out in restaurants, cafes or fast food venues and similar outlets is only
undertaken by a minority, and very rarely as an individual eating event undertaken alone. In Sweden, the pattern of eating out, when seen in relation to
Holm: The Social Context of Eating
195
hours of the day, suggests that much of this activity is a part of professional
life. Eating out in restaurants or cafes at lunchtime on an ordinary working
day is relatively common, suggesting that this is a frequent pattern among
some groups of the population. This is not the case in Denmark where eating
out is less common and takes place at times of day that suggest eating out is
first and foremost a leisure-time activity.
The social contexts of eating confirm the relationship between location and
company. It follows from the home-based character of eating events that the
most widespread pattern among the Nordic peoples is that of eating at home
either in the company of one or more family members or alone. The second
most widespread pattern is that of eating with colleagues, schoolmates or
business contacts at the place of work or education. People more rarely eat in
the company of friends as such. This hardly occurs at all on weekdays, but is a
little more common during weekends. Young people (in the age group 15-24
years), in the period before they have dependent families, eat more commonly
with their friends than do any other groups of the national populations.
There are variations in these general patterns. Some of these reflect variations
in living conditions - those who do not work or go to school more commonly
eat at home, while those who live alone more commonly eat in social isolation
than others do. People who do go to work or to school are those who eat in the
company of colleagues or other professional contacts, but it also transpires
from this analysis that they are among the population groups who more commonly eat in the company of friends or other personal contacts. The older age
groups are those who most commonly eat at home and alone.
It does occur that people who do not live in one-person households eat alone,
but this eating pattern occurs far more commonly among those who do live in
such households. Adults who share their household with a child or children
more commonly eat in the company of family members, but in households
comprising two adults this pattern is even more common. The preparation of
shared meals at home therefore seems to depend not only on the presence of
children, but on the presence of children and two adults or more. The reasons
for this are unclear and may be due to one or more of several structural factors. It may reflect a familiar Nordic practice, according to which children of
divorced or separated parents alternate between the homes of each parent, residing in each home parts of the time. It may reflect a status difference between household members, such that the inclination to arrange and prepare
social meals is stronger when the meal is likely to include a ‘full’ family in
the traditional sense - i.e. two adults and children. Single partents may also
have less time to spend on organising meals.
196
Eating Patterns
Differences between men and women with regard to the social contexts of
their eating patterns are generally small. The differences that have been identified in this analysis indicate that women more commonly spend time at home
with their families than do men.
Our analysis has revealed that even when eating events are of relatively brief
duration, they commonly constitute events that are undertaken in their own
right rather than being one activity among others, which are undertaken simultaneously. In most cases, eating does not seem to be a very hasty event, typically taking more than 10 minutes, and most people in the Nordic countries
participate in eating events of variable duration in the course of a day. Moreover, eating most commonly takes place while seated at a kitchen or dining
table, but quite commonly too while seated at a coffee table. At work or educational institutions, most people eat in canteens or rooms designed for the
purpose of taking breaks from work, which suggests that people are also seated at a dining table in these contexts. However, the pattern of eating in the
actual location in which work is carried out is also relatively common, and a
small group of people - who do not have any distinguishing social characteristics - would appear to eat exclusively in this fashion when they eat at their
place of work or education.
More than half of the eating events which take place in the home are events in
which television or radio are playing or people are reading while eating.
These activities occur more commonly in conjunction with individual eating
events, that is, when people eat alone. It is also clear that these practices vary
from one eating event to another for many individuals. There are some variations between social groups in regard to the practices of reading or listening to
radio while eating. Watching television while eating, however, is a practice
that is relatively evenly distributed throughout the national populations. There
are no indications that this practice is more common in specific population
groups, such as younger age groups, which might ground an expectation that
television viewing while eating might increase in the future. On the contrary,
it transpired that more people in the younger than in the older age groups had
not watched television while eating at any time of the day at issue.
Many of the general features of eating patterns are found to be similar in the
countries included in our analysis. This is the case regarding the significance
of home-based eating patterns and the tendency to eat either in the company
of family members or alone, but some significantly different national variations have also been identified. In Denmark, the home and family tend to dominate the picture even more than in the other Nordic countries, and eating
Holm: The Social Context of Eating
197
events are typically of longer duration than in the other countries. In Finland,
eating events tend to be of remarkably shorter duration. In Norway, eating
while seated in a sofa and while watching television are practices that are
more widespread than in other Nordic countries. In Sweden, it is more common to eat out in a restaurant or cafe and to have more eating events in the
company of other people in the course of the working day.
These findings suggest that some similarities in the eating patterns of these
different populations are heavily dependent upon the organization of time,
space and social relationships in Nordic societies, which have broadly similar
features in these societies. These features include the scheduling of the working day, the physical location of places of work relative to homes, and the participation of women in the workforce and the proportion of small and singleperson households. Some other characteristics of eating patterns, however,
appear to reflect local or national traditions and habits, which do not depend
on structural features of this kind. These include negotiated welfare arrangements regarding breaks during working hours, which differ from one workplace and one society to another, as well as informal norms and cultural traditions with respect to eating and eating events, which may differ from one
household to another as well as between societies. The latter is exemplified in
differences with respect to the duration and character of eating events. Additionally, there are many variations in eating practices which are not related in
any systematic manner to national differences or differences between population groups. This indicates that the social contexts in which eating takes place
also reflect highly variable conditions of everyday life, which the present
study was not designed to delineate. It also indicates that the social norms
guiding everyday eating patterns are flexible in character.
These results do not confirm the views put forward in the scientific literature
to the effect that the sociality of eating practices has disappeared and that eating events are no longer significant events in their own right. On the other
hand, neither do they indicate that all members of the national populations at
issue have an eating pattern that is social in character, in the sense that eating
events include the participation of other people. Our analysis suggests that
some aspects of the picture drawn by Rotenberg’s analysis of eating patterns
during the process of industrialization (Rotenberg1981) still stand. On a daily
basis, modern individuals in the Nordic countries commonly eat in the company of family and (for some) colleagues. It would seem, however, that the
company of friends - even during weekends - is markedly less widespread
than that suggested by Rotenberg’s analysis, while it transpires that the practice of eating alone is widespread.
198
Eating Patterns
The variable character of contemporary eating events in the Nordic countries
at the individual level suggests that these events are also flexible in character.
Different social contexts and courses of events are reported as occurring on
the same day. Some eating events are undertaken alone, and some while reading, listening to radio or watching television. An eating event at work may
take place in a canteen with colleagues or alone without leaving one’s work.
While most people eat while seated, the dining table may be at issue or the
sofa. Some eating events still seem to be important events of longer duration,
although there are characteristic differences between the countries on this
point. These data also document nevertheless that when eating takes place in
social isolation, they seem to loose some of the features which are here viewed as signs of significance: such eating events are of shorter duration than
eating events that are shared in the company of others, and are more frequently undertaken while other activities are also occurring at the same time.
Some people exclusively reported individual eating events, but the larger parts
of the national populations reported a pattern that included both individual
and social eating events. For as long as modern individuals live in households
with other people, and for as long as they leave their homes to go to work or
to their place of education it would seem most likely that they will continue to
engage in eating together with the people around them - be it at the workplace
or in the home. The meal as a social event therefore forms and in all likelihood will continue to form a significant element in modern living.
6
Family Meals
Lotte Holm
6.1
Introduction
One of the important themes in the ongoing discussion about food and eating
in contemporary society was addressed in the previous chapter: the concern
that the meal as a social institution is on the decline. This discussion has been
pursued in the scientific literature with particular regard to the so-called family meal. Qualitative studies of family food habits which have been undertaken in recent decades reveal that family meals play a very important part in
maintaining the cohesiveness of the family unit. This is documented in studies
conducted in the United Kingdom (Murcott 1982, Murcott 1983, Charles and
Kerr 1988) Italy (Counihan 1988), Sweden (Jansson 1988, Ekström 1990,
Jansson 1993), Finland (Mäkelä 1996, Mäkelä 2000), Denmark (Haastrup
1992, Holm 1996, Iversen and Holm 1999) and Norway (Wandel, Bugge, and
Ramm 1995, Bugge and Døving 2000) as well as the USA (DeVault 1991).
It is a common theme in these studies that the meal is an important element in
what has been described as ‘the project of creating a family’ (Iversen and
Holm 1999). The idea that the whole family should be gathered together for a
daily meal is a central one in many households, and informants (women for
the most part) give numerous detailed reports of their endeavours to plan, organize, create and conduct such meals. However, several reports from both
the Nordic countries and the USA indicate that difficulties arise in regard to
putting the idea of the family meal into practice. Family members in modern
households typically spend the larger part of their day in separate social
spheres and often only meet in the home for recreation. It is only possible to
200
Eating Patterns
be together as a family when everybody is at home at the same time. Arranging a joint meal at this juncture of time is therefore seen as an important task
to be accomplished by the adult members of the family or by one of them. In
this sense, a common meal can be described as a medium for building the
family unit (DeVault 1991).
The qualitative sociological studies which have been conducted from an everyday life perspective leave no doubt that in the UK, Italy, the Nordic countries and the USA, most informants consider family meals as being vital to a
happy family life. However, it is not possible to make direct inferences from
this to the actual practices of modern households. The many extensive reports
in qualitative studies about the meaning and significance of family meals may
regard what is as a matter of fact an ordinary and important daily practice. But
these reports could also reflect an ongoing concern or worry about the fact
that families do not gather together for a joint meal as frequently as is thought
to be appropriate. The fact that informants in many of the qualitative studies
do report great difficulties in gathering the family together for joint meals
supports the latter interpretation.
These reported difficulties are also reflected in modern discourses about food
and nutrition as a concern that the meal as an institution is currently in a process of dissolution. Family meals, it is said, are disappearing, and eating is
thought to take place increasingly outside the home - in restaurants, cafés,
bars and in the streets (Fischler 1990, Senauer, Asp, and Kinsey 1991). As a
consequence, it is feared that the institution of the family meal is disintegrating, and that eating is becoming an individualized activity that lacks any social significance in so far as it becomes an isolated and mechanical activity.
This concern is not only embedded in modern discourses about food, but also
appears in debates about the broader issues of modern everyday life and its
degenerative features. One example is a recent Danish book entitled The Centrifuge of Everyday Life. Here the theme is conceptualized as ‘eating-on-thego’, a concept that is parallel to that of ‘grazing’ (Andersen 1997). The tendency to ‘eat-on-the-go’ is presented as an expression of the individualization
of modern family life. The busy family members do not sit down and enjoy
joint meals, but eat independently of each other. However, according to Andersen, a ritualized break with this busy pattern of eating does exist: the good
meal, the special meal, for which everybody gets together in order to enjoy
good home-cooked food. Both the special meal and the tendency to ‘eat-onthe-go’ are seen as undermining the traditional pattern, in which “one daily
meal, is eaten at fixed times and at which family members each have their
fixed seat around the dining table” (Andersen 1997, 30-31). The dissolution
of this tradition is viewed by Andersen as undermining the family.
Holm: Family Meals
201
Similar concerns have been expressed in other Nordic countries. Ekström
cites the gastronomic debate in Sweden as a case in point. The modern family
is seen as a family who will not sit down for breakfast: “Father on his way
out, briefcase in hand, is served coffee by mother who is in the midst of preparing luncheon sandwiches for little brother. Each of them is in a rush daughter concerned that her hair is in a mess. No one has time to sit down at
the dining table; the question is, whether there is a dining table as such” (Ekström 1990, 70). Jansson describes his expectations at a point in the 1980s
when he was about to start interviewing in the course of a qualitative study of
family food habits. Official reports and media coverage in Sweden had led
him to expect to meet “a bunch of fast-food eating people, who didn’t care to
cook on weekdays and who rarely took the time to gather around a laid table
...” (Jansson 1988, 44).
The same concern has also cropped up in the Norwegian food debate. The
National Union for Public Health published a pamphlet in 1998 under the
heading: “Protect the shared meal”. Its plea was made in the following terms:
“The National Union wants to protect the shared meal and warns against the
development of so-called “single-handed food”- fast food, which can be eaten
at any time and any place” (quoted from Bugge and Døving 2000, 23).
The Finnish nutritional recommendations were revised in 1998. Under the
heading ‘Special questions in nutrition’, regular meals were emphasised as an
important basis for balanced and moderate eating. The concept of a ‘meal’
designates more than an eating event for many Finns. It also designates a social event. In a qualitative study Finnish women are reported as defining and
understanding proper meals as those which have three characteristics: hot
food, a salad and company (Mäkelä 1996). Concern has been expressed that
Finnish norms in this regard are becoming more relaxed, there being fewer
family dinners and lunches (Varjonen 2000).
In Sweden and Denmark there are nationwide initiatives devoted to the aim of
restoring the good meal. A national network in Sweden, ‘The Day of the
Meal’, arranges an annual celebration day with activities across the nation
which aim to promote the meal as a good, social experience (Möller 2000). In
Denmark, an official report has been issued by the Ministry of Food, Agriculture and Fisheries marking the initiation of a new institution, the ‘House of
Meals’ (Ministry of Food 1999). These initiatives are official attempts to address what is understood to be the decline of the meal, including the family
meal.
202
Eating Patterns
The issue as to whether or not the tradition of sharing family meals on a daily
basis has really existed in the Nordic countries, and if so, how widespread it
has ever been, is discussed by Ekström (1990). The point is made that “the
shared meal was not, historically speaking, shared by everyone. Earlier, there
were several social categories who did not belong to the sharing group:
women, servants and children each had a different social status and there are
reasons to believe”, Ekström maintains, “that everybody did not eat from the
same pot” (Ekström 1990, 75). Murcott proposes a similar argument in her
discussion of how family forms and household structures have varied considerably from one social class to another. A study undertaken by Littlejohn in
1963 reported contrasts between family meals in working-class and middleclass households. The middle-class wife would act as a fellow diner during
family meals, whereas in working class households her role would more
closely resemble that of a waitress or a servant (Murcott 1997, 43). “It is often
forgotten”, Murcott continues, “that over the last 150-200 years the children
of upper-class, wealthy and aristocratic families never ate with the rest of the
family” (Murcott 1997, 43). The image of sharing a joint meal at the dining
table breaks down when gender, age and class are taken into account, and
Murcott therefore suggests that the ideal of the family which is expressed in
worry about the decline of the family meal is probably an “ideal-typical model
of the middle-class and (respectable) working-class family” (Murcott 1997,
44).
Murcott documents that the dissolution of the family meal is a recurrent theme
in the sociological literature regarding food. This dissolution is taken for
granted by several sociological writers (Murcott referring to Mintz, Fischler
and Falk as examples), and is commonly presented as reflecting an underlying
and general process of social and cultural dissolution. Murcott traces the empirical foundation for this conception and concludes that there is only meagre
and scattered historical evidence of the status of the family meal, rendering
any firm conclusion as to its currently changing status somewhat dubious. She
documents that all empirical references to the dissolution of the family meal
refer back to the study Middletown, by Lynd and Lynd, first published in
1929. In this study, informants were reported as complaining that family reunions at the dining table, which had been taken for granted by the previous
generation, were becoming increasingly difficult to establish. This would date
the dissolution of the traditional family meal to the period before World War
I. It could be the case, Murcott argues, that the dissolution of the family meal
dates back to this period. She also suggests, however, an alternative interpretation: “This alternative proposes that an anxiety lest the family meal is waning is also to be understood as if it were a standing item on the agenda of
twentieth-century public commentary on the nature of family life” (Murcott
Holm: Family Meals
203
1997, 46). The latter interpretation is consistent with the findings of qualitative studies to the effect that family meals are seen as symbols of the family.
The more recent studies of food in the context of everyday life document that
shared family meals are still seen as an important means of maintaining the
family as a unit. But these studies have not addressed the issue as to whether
this expressed ideal reflects a currently important practice or the bemoaning
of its disappearance. This issue still needs to be investigated empirically. In
the following we will present a descriptive analysis of the frequency of family
meals in the Nordic countries. We see this analysis as a point of departure for
the kind of empirical investigation that needs to be undertaken to-day.
6.2
Family Meals – operationalisations and expectations
It was shown in Chapter 5 that most eating takes place at home and that most
of those who share their household with others do report having shared one or
more eating events with family members on the day before the interview. It
would seem therefore that eating with the family is an important and frequent
mode of eating.
But ‘The Family Meal’ is more than a matter of sharing a meal with another
family member. As described in qualitative sociological studies, it is a meal
which gathers all members of the household together. It is often described as
being the main meal of the day and, in the Nordic studies at least, it is also
typically a hot meal.
Our data regarding the pattern of ‘eating with the family’ as thus far described
does not necessarily imply eating in the company of the assembled household.
Some members of the household may be absent, and even if everybody is at
home they may not sit down to eat at the same time. In order to operationalise
the concept of the ‘Family Meal’ we therefore decided to reserve this label for
eating events which fulfil the following criteria:
•
•
•
•
the meal is eaten by a person who lives in a multi-person household
it takes place in the home
it is eaten in the company of other family members
it is eaten by all household members at the same time.
As described in Chapter 5, an interview filter did not function as planned in
the interviews undertaken in Finland. For that reason we were obliged to ex-
Eating Patterns
204
clude the Finnish data from most of the analysis concerning the place and social context of eating. However, we have decided to include the Finnish data
in the analysis presented in this chapter. According to the qualitative studies
previously undertaken in Finland and other countries, the family meal is more
often a hot meal than a cold one. We therefore add an extra criterion:
• the food that is eaten is hot food.
The qualitative studies of family food habits indicate that family meals are
considered to be vital to the cohesiveness of the family unit. Informants report
how the task of building a family unit becomes important when children arrive in the household. We expect therefore that family meals will be more frequent among households with children. Furthermore the analysis presented in
Chapter 5 suggests that these meals will be more frequent among couples with
children than among singles with children.
It has been suggested in the literature that families encounter difficulties in
their attempts to arrange joint meals on a daily basis, and that family meals
are therefore arranged as special occasions when extra time is available. The
supposition is that this occurs regularly. If this is the case, we might expect
that family meals will be more frequent during weekends than on weekdays.
Family meals, by our definition, gather the entire household together. We do
not therefore expect to find gender differences in regard to participation in
such meals. In the light of the available qualitative studies it would also seem
that the meaning and significance attached to family meals are global, rather
than related to and reflecting differences in social background. We therefore
expect that socio-economic background will not account for variation in the
frequency of family meals. Qualitative studies of eating patterns in families
have been conducted in each of the Nordic countries with broadly similar results. We do not therefore expect to find national differences regarding the
frequency of family meals.
6.3
Results
Figure 6-1 shows that the tendency to have a family meal (or meals) on the
day preceding the interview varied between household types in the four Nordic countries. These data support some of our expectations, but fail to support
others. It is confirmed that family meals are more frequent among couples
with children than among singles with children. This is the case in all four
countries. But it also emerges - contrary to our expectations - that family
Holm: Family Meals
205
meals are less frequent in households with children than in multiple-person
households without children. The Danish data constitute an exception to this
pattern, which otherwise holds in Finland, Norway and Sweden. Older agegroups are typically over-represented among couples without children, and
these data suggest that family meals will vary with age. We will return to this
issue presently (see Figure 6-2).
80
60
67 66
60
70
63
60
51
66
55
57 55
46
40
20
0
D e nm a rk
F inlan d
C ou p le, n o c h ild (ren )
N o rw a y
S w e de n
C o u p le , w c h ild (re n )
S in g le, w c h ild (re n )
Figure 6-1Family meals and household types (N: Denmark 810, Finland 912, Noreway
811, Sweden 744).
It emerges from Figure 6-1 that there are national differences with respect to
family meals. This is contrary to our expectations. Family meals are less frequent in Finland first and foremost, but they are also less frequent in Sweden
as compared with Norway and Denmark. The variation between household
types is generally smaller in Denmark than in the other countries. It was
shown in Chapter 5 that eating at home and eating with other family members
is more frequent in Denmark than in the other countries. It emerges from the
data in Figure 6-1 that eating events which include all household members are
also characteristic of the eating pattern in Denmark.
Finland and Sweden are the two countries in which the pattern of eating hot
food at lunch-time is a common one. For this reason there are two plausible
interpretations of the results shown in Figure 6-1. It is possible that individuals in Finland and Sweden less frequently have meals - either hot or cold - in
which all household members participate. But it is also possible that the meals
in which all household members participate are relatively often not hot meals.
Unfortunately, due to the flaw in the data collection in Finland, we are not
able to pursue these alternative interpretations of the data any further. Our
Eating Patterns
206
data only allow us to draw conclusions about ‘family meals’ in so far as they
are hot meals. However, this particular criterion of the ‘family meal’ is also
applicable to the Finnish understanding of this term as it emerges from qualitative studies of food in everyday life (Mäkelä 1996). What remains indisputable is that Finnish and Swedish households do have family meals, as defined,
less frequently than households in Denmark and Norway.
100
90
80
70
Percent
60
50
40
30
20
10
0
1 5 -2 9
3 0 -4 4
4 5 -5 9
6 0+
A g e g ro u p
D e n m a rk
F in la n d
N o rw a y
S weden
Figure 6-2 Family meals and age groups (N: Denmark 890, Finland 993, Norway 888,
Sweden 820).
Figure 6-2 shows that family meals vary between age groups. The data document that the frequency of family meals increases with age in all four countries, and differences between age groups are marked in each country. This is
particularly so in Sweden, where the frequency of family meals among the
elderly (aged 60 years or more) is double that of the youngest age group (29
years or less). For this particular agegroup family meals are more frequent in
Sweden than in the other countries.
Figure 6-3 shows that the frequency of family meals varies according to the
day of the week. There are differences between weekdays and weekends, but
they are generally small. Furthermore, differences are not consistent as between the four countries. In Denmark, family meals are a little more frequent
on weekdays than during weekends. In Finland and Sweden they are more
frequent during weekends, while in Norway there are hardly any differences
at all. Our expectation that family meals would be more common during
Holm: Family Meals
207
weekends is therefore only confirmed in regard to the Finnish and Swedish
data.
70
65
61
60
61
61
63
66
59
53
50
40
30
20
10
0
Denm ark
W eekday
Finland
N orw ay
Sweden
W eekend
Figure 6-3 Family meals and day of the week (N: Denmark 890, Finland 996, Norway 888,
Sweden 820).
An overview of how family meals vary in relation socio-demographic background factors in each of the four countries is given in Table 6-1. It offers a
summary of the results of a series of logistic regression analyses for each
country as well as one logistic regression analysis of the assembled data from
all four countries. This analysis confirms that when all other factors are controlled, the differences between Finland and the other countries remain. The
pattern is that in all socio-demographic groups, with the exception of the age
group 45-59, Finns have less family meals than the other Nordic people, while
Danes have more. It also emerges that differences between days of the week
remain when other factors are controlled, but only significantly so in Finland.
Some differences between household types become less marked when other
factors are taken into account.
However, age differences, which were presented in Figure 6-2, remain when
all other factors are controlled, and it is confirmed that age is the background
factor that accounts for the greatest variation. Variation by occupational status
presents a consistent pattern as between the countries, but occupational status
is only statistically significant in Finland and Sweden. Further, there are large
and significant variations in Finland in regard to urbanization. This is not
found in the other countries. In Finland, higher white-collar workers appear to
be less likely to have family meals than other status groups. Similar trends are
not found in the other countries. Neither gender nor education show any significant differences in any of the four countries when other factors are controlled.
Eating Patterns
208
Table 6-1 Family meals.Percent who ate a hot meal at home with the whole household.
Relative family meal eating, odds ratio. (Persons living alone excluded).
Country
(N)
45
Total population
Gender
(N)
Men
Women
sig
Age
15-29
30-44
45-59
60+
sig
Education Mandatory
Further education
Academic
sig
Household 2+ adults w/childr
type
2+ adults no child.
1 adult + children
sig
Occupatio- Not active
nal
Active
status
sig
Social
Worker
status
Lower white collar
White collar
Higher white col
Not applicable
sig
Urbani<25.000
sation
25-99.999
100.000+
sig
Day of
Weekday
week
Weekend
Sig
sig/of whole model
Constant
44
Denmark
% of
Odds
pop ratio44
890
64
1,04
803
63
1,00
64
1,13
NS
47
1,00*
65
1,70*
66
2,06*
75
2,85*
***
71
1,00
60
0,72
63
0,86
*
66
1,00
67
0,67*
60
0,73
NS
76
1,00
60
0,69
***
68
1,00
59
0,72
61
0,69
66
0,81
65
0,87
NS
65
1,00
64
1,03
60
1,04
NS
65
1,00
61
0,82
NS
NS
0,001
2,738*
Finland
% of
Odds
pop
ratio
Norway
% of
Odds
pop
ratio
Sweden
% of Odds
pop radio
996
54
995
62
820
61
54
55
NS
44
52
57
72
***
59
53
50
NS
51
63
46
**
68
49
***
60
52
56
44
53
*
60
51
44
***
53
61
*
NS
0,67*
905
1,00
1,01
1,00
1,57*
1,67*
1,93*
1,00
1,04
1,15
1,00
1,27
0,86
1,00
0,57*
1,00
0,78
0,92
0,56*
1,13
1,00*
0,72
0,60*
1,00*
1,55*
0,000
1,457
61
62
NS
47
61
64
82
***
62
60
64
NS
60
70
55
**
74
59
***
62
61
63
67
49
NS
61
60
64
NS
61
63
NS
NS
1,00
811
1,00
1,01
1,00*
1,22
1,58
3,75*
1,00
1,13
1,49
1,00
0,83
0,89
1,00
0,58
1,00
0,89
0,99
0,90
0,68
1,00
0,80
0,89
1,00
1,03
0,000
2,038
62
60
NS
44
57
62
87
***
70
68
52
***
57
66
55
*
78
55
***
62
66
57
61
53
NS
65
64
57
NS
59
70
NS
NS
0,91
712
1,00
0,82
1,00*
2,03*
2,44*
5,45*
1,00
1,21
0,77
1,00
0,97
0,90
1,00
0,59*
1,00
1,54
0,88
1,10
2,07*
1,00
1,37
1,18
1,00
1,20
0,000
0,907
Logistic regression analyses done separately in each country (Model including all sociodemographic variables in table).
45
Chi-square test for differences in pct of population groups who had varying numbers of family meals. The test is based on the full distribution between 0, 1, 2 and more family meals.
Holm: Family Meals
209
We are now in a position to return to our initial expectations. The analysis has
shown that practices related to family meals are not the same in the four countries. This is contrary to our expectations. Finns have significantly less family
meals than do other Nordic people, and this is the case both for the population
as a whole and for almost all population groups. This national difference is
confirmed when other factors are controlled.
Our expectation that family meals were more frequent during weekends is
confirmed in Finland where there are significantly more family meals durings
weekends. There is a similar tendency in Sweden (and Norway) but not in
Denmark. The hypothesis to this effect was originally proposed in Denmark in
a book which debated modern everyday life (Andersen 1997). It is not supported, however, by the Danish data, which suggest that family meals occur
more often on weekdays.
We expected households with children to have more family meals than other
households. This expectation was not confirmed in any straightforward manner. Looking at the simple frequencies, households without children have
more family meals, but this picture changes when other factors are controlled.
It emerges that households with children (and more than one adult) do in fact
have more family meals than households without children. But Finland once
again constitutes an exception to this pattern. However, in all countries couples with children were more likely to have family meals than singles with
children. This underlines the pattern discussed earlier in chapter 5, that the
task of arranging joint family meals seems to be highly dependent upon the
presence of other adults and not only upon the presence of children in the
household.
Our expectation that there would be no gender differences with respect to
family meals was confirmed. The analysis showed that in all countries older
age-groups were more likely than others to have family meals. Employed persons (independently of age) were also somewhat more likely to do so - significantly so, only in Finland and Sweden. Age was the single factor in all
four countries which had the greatest impact upon the likelihood of having
family meals, and these meals were more frequent among older persons and
less frequent among younger ones in each country. This pattern may be due to
a generational effect, according to which the practice of having family meals
is more strongly embedded in the social norms of the older generation. In that
event, it might well be expected that family meals will decline in the future. It
is equally likely that it is due to structural effects, however, according to
which older individuals engage less in activities outside the home and are
therefore more likely to spend time at home with other members of their
Eating Patterns
210
households. Also, the households of older individuals tend to be smaller,
rarely including children. It is therefore a less complicated task to arrange
meals to be shared by all members of such households. In so far as structural
factors are decisive, there would be no reason to suppose that family meals
will decline in the future. Our descriptive data regard the patterns obtaining at
a particular point in time and do not provide support for choosing between
these two explanations.
According to our analysis, Finland differs from the other Nordic countries in
regard to practices with respect to family meals. These meals are less frequent
in Finland, and social and demographic variations in that country also differ to
some extent from the pattern in the other countries. The specific Finnish
variations show that family meals are less frequent in urban households and
among higher (white collar) professionals. It may be, therefore, that family
meals are associated with a more traditional and rural way of life in Finland,
and that further migration to cities will lead to a further decline in family
meals. There is no evidence of a similar pattern in the other Nordic countries.
6.4
Discussion
Our analysis demonstrated that more than half (between 54 and 64%) of those
living in multiple person households had a hot meal (or meals) on the day preceding the interview - a meal that was eaten by the assembled household.
There are not many other available data with which these results can be compared.
In the most recent nationwide dietary survey undertaken in Denmark informants were asked about family meals. This survey was based upon a representative sample of the Danish population and comprised 2000 adults
(Andersen et al. 1996). In this survey, 81% reported that they usually had dinner with their assembled household more than 5 times a week (Groth 1998).
The term ‘dinner’ in this context referred to the evening meal, whether hot or
cold, and cold evening meals are rather frequent in Denmark (Fagt, Groth, and
Andersen 2000). Still, the results of this dietary survey suggest that family
meals take place with a greater frequency than that indicated by our data.
However, the two surveys adopted different methodologies, which may explain the difference. Respondents in the dietary survey were asked to state
how many times a week the entire household ate an evening meal together. In
so far as there are strong norms in favor of family meals in Denmark and the
other Nordic countries, it might be expected that individuals will report a
Holm: Family Meals
211
higher frequency of family meals when asked to report about their usual habits than when being asked about actual behavior on a specific day, as in our
survey.
A survey undertaken by postal questionnaire in Norway, comprising 3,536
individuals aged 15 or more, found that more than 80% of individuals living
in multi-person households ate dinner at home in the company of others on
the last preceding weekday, but it did not seek to establish whether or not the
entire household ate together (Bugge and Døving 2000, 119). A representative
survey from the early 1980s, comprising 808 Norwegian households and employing a combination of self-administered dietary records and structured personal interviews, concluded that dinner is a social event, which gathers the
household together (Wold 1985). However, the data presented in the survey
report showed that household members in fact participate in such dinners with
varying frequency. Mothers had 6.22 dinners per week at home on average;
fathers, 6.17; daughters, 5.76, and sons, 5.93 dinners per week at home. This
suggests, that family meals according to our definition (including the entire
household) were likely to take place on average somewhat less than 6 times a
week. This survey also indicated a markedly higher frequency of shared family meals than we find in our present study. The Norwegian survey used a
methodology which is similar to the national Danish dietary survey - but, in
contrast to the Danish survey, the social context of meals was registered in the
diet record and was not a response to a question regarding habitual practices.
The methodology of the Norwegian survey more closely resembles that of the
present study in this respect, and its results are therefore more comparable to
ours than those of the Danish dietary survey. There is therefore good reason to
believe that family meals are less frequent today in Norway than they were in
the 1980s.
In her article concerning family meals, Murcott discusses the lack of empirical
studies of actual practices in regard to eating together as families (Murcott
1997). At the time of her article, few academic studies of this issue were
available. However, marketing researchers and mass market magazines had
collected some information about mealtime patterns. Murcott refers to two
British surveys and one German, which indicate a similar pattern: “Very approximately, the family meal could be described as making a half-time appearance” (Murcott 1997, 35). More recently, Warde and Martens reported
that 75% of the respondents in their study (which concerned dining out) had
eaten an evening meal at home, which included all household members, on
the day preceding the interview (Warde and Martens 2000, 107).
Eating Patterns
212
The trends Murcott reports from marketing studies appear to be consistent
with our findings, whereas Warde’s and Martens’ results indicate that family
meals are more frequently eaten in British households than in Nordic households. However, their results regarding home eating patterns are only mentioned in passing in their report and lie outside the focus of their analysis. We
are therefore unable to examine the matter in more detail.
6.5
Conclusion
The analysis has shown that family meals are still part of ordinary everyday
eating in the Nordic countries. This is the case both in households with children and in households without children. Family meals, i.e. hot meals shared
by all household members are more frequent among people in the older age
groups and are least frequent among the young generation. Earlier Norwegian
data provide some support for the view that family meals are less frequent
today than they were 20 years ago, but similar data are not available for the
other Nordic countries. The Finnish data suggest that family meals are less
common among urban and higher professional, white collar households. This
may be a sign of potential future decline. Aside from that, our data provide no
firm basis for drawing conclusions about the future prospects of family meals.
However, our study does provide sound quantitative support for the findings
of a number of qualitative studies of family food habits. Family meals do concern actual practices and not merely social norms in regard to ideals. According to the qualitative studies, family meals have great significance and meaning as symbols of family cohesiveness. Our study demonstrates that if family
cohesion is confirmed through shared meals, this does not occur on a daily
basis. The indications are that it may very well occur on an every-other-day
basis.
7
The Gendered Division of Cooking
Marianne Pipping Ekström and
Elisabeth L’orange Fürst
In this chapter we will explore what at first glimpse seems to be a very simple
question: "Who did the cooking?" This question was posed to the respondents
of the survey after they had answered questions about what hot food they had
eaten at home the day before (for a further description, see chapter 1 of this
report). What seemed to be so simple, however, turned out to be quite complicated. It has been somewhat ponderous to understand the rather confusing
figures regarding the division of labour by gender which appeared in the beginning of our analysis. In this presentation we will start by answering the
more basic question what is cooking by focusing on certain theoretical perspectives. Next we will describe the procedure of asking, as well as the answers to the question: who did the cooking? The central analysis, however,
will consider our findings concerning the gendered division in today’s everyday cooking. These will also be discussed in relation to aspects of ‘tradition’
and ‘modernity’. Finally, we will present our interpretations of the confusing
findings which we label a tendency of “over-reporting”. This includes a discussion of differences according to gender in the understanding and evaluation of cooking as part of housework.
7.1
Cooking: A many-dimensional kind of work
Cooking, or the making of food, is one of the basic activities in society, as it is
a necessity not only for our bodily existence, but also for our social well-
214
Eating Patterns
being. Food is characterised by ambiguity; it is both symbol and substance,
both individual and social (Fürst 1995). According to Claude Lévi-Strauss
cooking is a universal expression of the transformation of nature into culture
by the help of heat (Lévi-Strauss 1966). As a raw material food represents
nature, as a prepared dish it represents culture. Preparing the food before eating it represents a human specificity, distinguishing human beings from the
animal kingdom; according to Lévi-Strauss this signifies the threshold of civilisation. It was the control of fire that effectuated this. Fire is the medium in
the cultural transformation through which the raw (nature) is transformed into
the cooked (culture). The natural transformation, on the contrary, converts the
raw into the rotten. Lévi-Strauss (ibid.) speaks in terms of a Culinary Triangle
of the raw, the cooked and the rotten. The human's need for food exposes us,
on an everyday basis, to the ambiguity of being both nature and culture, an
ambiguity that creates tensions that the art of cooking attempts to resolve.
This is why cooking is a culinary46 activity that involves creativity. The raw
material is thus transformed into something beyond the mere edible, into
something palatable, pleasurable and enjoyable for nose, eyes and mouth.
In everyday life we may eat at home or at restaurants, cafes or cafeterias.
This chapter concerns eating at home, more specifically the prerequisite for
having a (hot) meal at home - that someone in the household has done the
cooking. While the activity of producing a meal for sale in the public sphere
has always been understood to be a kind of work, this has not been the case
for such production in private homes. There may be several reasons for this.
One may be that this production has traditionally been women's activity and
her responsibility as part of her role as a housewife. However, an important
task of women's studies since the middle of the 20th century has been to make
visible that which was formerly invisible: the fact that what is given and
served in the home - food, care, tidiness and cleanliness - are products of a
certain kind of work, even though they have not been counted in the GNPs
(Prokop 1978, Kaluzynska 1980, Åkerman 1983, Gisselberg 1985, Nyberg
1997). This production was labelled unpaid work (Leira 1976). Some of the
feminist discussion has concerned the possible consequences of starting to pay
the housewives for this work (Grønseth 1986). One important argument
against this was that this would bind women to housework in their private and
isolated kitchens, instead of giving them the opportunity to participate in the
public sphere of wage labour, which was emerging for women on a large scale
in the 1970s.
46
Culinary – connected to especially delicious food or to cooking. From Latin culinarius =
associated to the kitchen (National Encyclopedia 2000).
Ekström & Fürst: The Gendered Division of Cooking
215
In connection with the attempt to make housework visible, some feminists
argued that also meals served at home, were actually crystallised working
hours (Åkerman 1983, see also Harmaja 1928). By this, they referred to Karl
Marx' definition of commodities: crystallised working hours meant the
amount of time used to convert some raw material into a product. If one
started to count hours, women's housework would become expensive. But one
characteristic of such work according to Marxian interpretations is that it is
necessary labour in society, although not paid for, in opposition to wage labour for capitalist production. It must be done, it meets intrinsic human needs
for life and satisfaction. It is a production of use values.
Today we may state that cooking, like other housework, is still unpaid labour.
This implies that it is a kind of work, but not wage labour, and hence not a
commodity on the market. The production of use values in opposition to exchange values for sale and for profit at the market has been analysed as related
to a different kind of rationality than that which characterises capitalist production (Prokop 1978, Fürst 1981). This has been labelled a rationality of the
gift, underlining the giving aspects of work connected to family relations, including care for others (Fürst 1995, Fürst 1997, see also Theophano & Curtis
1991). Hence cooking may be understood as a form of production marked by
the logic of the gift, rather than of the commodity. What cooking may include
in addition, as compared to housework in general, is, in line with LéviStrauss, the creative and culinary aspects of food.
7.2
Cooking and gender
Defined as a kind of work, cooking as part of housework has been analysed in
relation to the gendered division of labour in society. While Marx considered
this to be a division by nature, feminist researchers have labelled it a product
of patriarchal society (Carlsson et al. 1983). When women, in large numbers
joined the public labour force from the 1970s onwards, feminists imagined
that this would imply a breakdown of the division of labour. They expected
men to correspondingly take part in the private sphere of housework. This,
however, did not happen in the 1970s and 1980s (Government Report 1979,
Wærness 1978, Haraldsen & Kitterød 1992). But what about the 1990s? What
will our study tell us about men's participation in everyday cooking at the end
of this decade?
Studies of the gendered division in cooking in homes have demonstrated that
approximately 80-85 per cent of this work has been done by women. This
216
Eating Patterns
holds for the Nordic countries (Glefjell 1985, Glefjell 1986, Ekström 1990,
Ahrne & Roman 1997, Bonke 1997) as well as England (Charles & Kerr
1988) and the USA (DeVault 1991).47
Studies of housework in general have shown that men and women differ in
their participation in household activities (Haavind 1987, Jakobsen & Karlsson 1993). Time studies show that men do less than women at home, also in
regard to cleaning, washing and childcare (O'Doherty Jensen & Holm 1997).
Has time use in households changed during the last decade? A fairly new
Swedish study focuses on the use of time in relation to gender, power and
equality (Ahrne & Roman 1997, Government Report 1998). The study presents and discusses the division of labour (concerning cooking, laundering
and washing up) between couples in households of four groups. About 13%
are what was termed Equal. Men and women share these three types of
household work equally. The next level is termed Semi-equal. Here we find
25% of the couples. Women do more, but men are doing some of the tasks.
For example: The couple share cleaning and cooking, but women are responsible for the laundry. A little more than one third is what they term Conventional. Women are doing most of the household work, but men are helping.
Finally among what they term Patriarchal, they found somewhat more than
25% of the couples. Women do most of the household work, and men are a
little helpful, doing some smaller tasks. This study indicates that in the 1990s
household work is still a women’s task, but men’s contribution to cooking
seems to be somewhat higher than in the two previous decades.
In our study we examine the participation of the respondents regarding cooking on an everyday basis by asking what the respondents did yesterday in relation to the preparation, cooking and serving of hot meals. This question illustrates the degree of invisibility of this kind of work. In market research there
is a tendency to conceive the consumer, and hence the one who is responsible
for the making of food, as only a customer. In this conceptualisation cooking
starts in the shop, and the buying of ingredients is a question of personal
choice, a choice the merchant tries to control. From a woman’s perspective
this is only part of the story. The planning of the evening meal ("dinner") may
start in the morning, before leaving home for work outside. It may start with a
glance in the refrigerator, the freezer or cupboard. Ekström (1990) has fo47
Studies on time use in households has shown that the time used in connection with cooking
in Sweden is 2 hours, 34 minutes a week for men and 5 h 41 for women (Statistics Sweden
1992), in Finland 1h 38 for men and 5 h 01 for women (Niemi & Pääkkönen 1989) and in
Denmark 2h 06 for men and 5h 43 for women (O’Doherty Jensen & Holm 1997). This is about
men and women irrespective of living single or as couple.
Ekström & Fürst: The Gendered Division of Cooking
217
cused on this planning process underlining that it is women's responsibility by
saying that they have the cupboard in their heads (ibid, 212). The questions
which may occur to the woman are manifold: What do I have at home that
needs to be eaten? What do I need to buy? What do the children like to have?
To ensure some variation in the week menus, which also is good for nutrition,
what would be wise? These and other questions, we will claim, are part of the
planning of a meal and a process beyond merely being a customer. It represents having the responsibility for the family meal.
It is important to underline that the planning of the meal as well as the choice
of raw materials, the treatment of them and the way to serve the meal, are embedded in cultural standards that differ in various countries and regions, as
well as in subgroups of different kinds, as for instance distinguished by economy, education, occupation and generation. The procedures of cooking also
follow national and regional traditions, typically transmitted from mother to
daughter. These procedures also include following recipes found in cookbooks, magazines and eating experiences abroad or recipes learned from
friends.
Our findings show that cooking activities seem to be done for a meal eaten
together with someone in the household. Thus eating in the Nordic countries
is still part of a family gathering. We may notice that only a very small part of
those respondents who reported that they lived with someone had eaten the
hot meal alone (5%). There was no difference between the countries or between men and women (No table). Thus the idea of an on-going disruption of
family meals as a sign of (post) modern everyday life does not seem to be
supported by our findings. (See Holm, this volume). 48
7.3
The procedure of asking "Who did the cooking?"
The questions which were posed to the informants concerning cooking related
to the preparation of hot food. When constructing the questionnaire, it was
decided that eating something cold, such as yoghurt and corn flakes, sandwiches, drinks, etc., should not fall under the concept of ‘cooking’. This, it
was argued, was more a question of "putting on the table". There are of course
borderline cases, such as a mixed salad, which therefore is not included. It
should be noticed that according to our findings these kinds of dishes were
48
According to Yearbook of Nordic Statistics 1996 (Mørch 1996) nearly half of all households
consisted of couples with or without children (in Denmark 45%, in Finland 44%, in Norway
45% and in Sweden 42% of all households).
218
Eating Patterns
very few. The question "who did the cooking" was consequently only raised
to respondents who had answered that they had been eating hot food. As also
stated above, the question was posed only to those living with someone. If the
respondent lived alone, it was assumed that s/he had made the food
her/himself.
The question "Who did the cooking" was posed at the end of every ‘block’ of
answers about the eating event (where the respondent reported on the composition and preparation of the meal). In a telephone interview the respondent
cannot see the fixed alternatives which are given as possible answers. The
procedure was as follows: When the respondent gave her/his answer, the interviewer marked the corresponding alternative in the questionnaire with one
tick or more, depending on the number of persons involved in the cooking
(see Table 7-1). It is worth while noting that the question "Who did the cooking?" was asked of each respondent one or several times, depending on how
many times s/he had eaten hot food at home the day before. In Norway and
Denmark only one hot meal is usually eaten each day, while especially in
Sweden and also in Finland it is usual to eat two hot meals each day. This implies a habit of eating hot food at lunch time, at the work place in a canteen, or
in a cafe or restaurant (see Gronow and Jääskeläinen, this volume).
7.4
Who did the cooking in the household and who didn't?
A first very rough summary of the findings (Table 7-1) shows that there were
only small differences between the four countries. The answer “I myself” had
a range of 57% (Sweden) to 60% (Denmark). In this table the answers are not
gender specific, as the question was put to people irrespective of sex. Hence
the answer “I myself”, could come from either a man or a woman. Mean 27
per cent answered that the partner, (husband or wife)49 did the cooking and
about 5 per cent answered that they did the cooking themselves, but together
with another person. Fewer than 10 per cent answered that the mother, father
or others (for example siblings) and, very rarely, a domestic help did the
cooking. “Mother” scored the highest in Finland, 9%, and lowest in Denmark,
5%.
49
In the sample there might be a small group of people living together with partners of same
sex (lesbian or gay). In the study presented here however, there is no data for analysing this
possibility. For research on the division of labour in households among homosexuals, see
(Dunne 1998).
Ekström & Fürst: The Gendered Division of Cooking
219
Table 7-1 Persons answering the question "Who did the cooking?" living together with
someone, and who have had one hot meal or more during the day. Per cent.
I myself
My partner
Mother
Father
Child/ren
Siblings
Other family members
Home servant, Au pair
Other person
Don’t know, don’t remember
Total
n (% of N)
N (total study sample)
50
I myself+another person
Denmark
Finland
Norway
Sweden
Mean
60
27
5
2
<1
<1
<1
<1
4
58
27
9
2
1
1
<1
<1
2
59
28
8
2
2
<1
<1
<1
1
57
27
8
1
1
<1
<1
5
59
27
7
2
1
<1
<1
<1
<1
3
98
784(66)
1187
100
895(75)
1200
100
900(76)
1178
99
804(65)
1244
99
3383(70)
4809
6
2
5
6
5
The percentage of child respondents in the study was quite low.51 Hence the
percentage of respondents who replied that mother or father did the cooking
was low. As seen in Table 7-1 the percentage of children, siblings or others,
who did the cooking is minimal. It is still the adults (the parents) in the Nordic
households who do the cooking. Statistically speaking, children do practically
nothing, nor were there any domestic servants. When the parents were reported to have done the cooking, this was more often the mother (7%) than
the father (2%). Most of the respondents seem to be aware of who actually did
the cooking. Only three per cent said they did not know or did not remember
who actually cooked the meal.
7.5
Answers according to gender
As stated above, gender studies have shown that women in Western society
have been doing a greater part of the household work, including cooking.
50
The percentage who had more than one answer to this question, indicating that cooking was
carried out jointly. In the table above, these are included in the category “I myself and one of
the others”.
51
Persons from the age of 15 were included in the study (6.4 per cent were between 15-18
years old (Kjærnes 2001)
220
Eating Patterns
However, as also mentioned, there are good reasons to assume that lately
there has been a change in this division of labour. This assumption is due to
the fact that women are increasingly taking part in the labour market on a
fulltime basis, as well as the observation of a strong ideology of gender equality in the Nordic countries.52 It has indeed been expected that the male partners thus will start to take more part in the home arenas of housework, including cooking. Such expectations of more equally shared work between men
and women may be considered to be an important aspect of the ongoing
world-wide modernisation processes.
Does our study find results that indicate less-traditional gendered divisions of
labour in the Nordic homes? When observing the distribution of answers to
the question "Who did the cooking?" we found that among the repondents
who answered "I myself", 1⁄3 were men and 2⁄3 were women.53 Among those
who answered that someone else did the cooking, 3⁄4 were men and 1⁄4 were
women. This is a very simple distribution of the answers, however, and one
must also consider how the answers as dependent variables are distributed
when tested in relation to gender as independent variable. In line with other
social background variables like age, education, and occupation gender here
has an impact on the activity of cooking. This has been done in Table 7-2.
In order to avoid double reporting of those who had eaten more than one hot
meal during the day (especially Swedes and Finns), we focus on only the hot
meal in the afternoon/evening.
Table 7-2 shows the percentage of men and women in the Nordic countries
who answer that they themselves did the cooking or that it was done by someone else (the partner, mother, father etc. combined to one variable). Looking
at the mean, 42% of the men said they did the cooking and 77% of the women
gave the same answer. We also observe that Norwegian women score the
highest (79%) and men in Finland the lowest (36%).
52
These countries in an international perspective have been in the forefront of developing gender equality, not least due to political decisions (Government Report1998, 6).
53
Male and female respondents were about 50/50 but the recruitment of male respondents was
particularly difficult. Weighing with regard to gender was needed in Norway and Sweden.
Ekström & Fürst: The Gendered Division of Cooking
221
Table 7-2 Who did the cooking? Persons living together with someone, and who have had
one hot meal in the afternoon (2-9 p.m.). Per cent.
Denmark
Men
I myself
Someone
else
42
58
Finland
Women
77
23
Men
36
64
Norway
Women
77
23
Men
49
51
Sweden
Women
79
21
Men
41
59
Mean
Women
73
27
Men
42
58
Women
77
23
Total
100
100
100
100
100
100
100
100
100
100
n (% of N)
702 (59)
588 (49)
690 (59)
596 (48)
2576 (54)
Level of significance: the difference between men and women are significant at p<.000 for each
country. (Chi 2 test)
If one adds the men’s answers and the women’s answers “I myself” did the
cooking, the sum for each country is more than 100 per cent (Table 7-2). This
is the complicated finding referred to in the introduction of this chapter. It
may be interpreted to represent an over-reporting from the respondents. The
percentage of the men who answered “I did the cooking” is especially high in
Norway (49%). In this country we also observe the highest percentage among
women who answered “I myself” (79%). This gives a sum of 126 per cent. In
Finland the percentage among men answering “I myself” is the lowest, 36%.
In this country 77 % of the women answered the same. This sums up to the
lowest "over-reporting" of 113 per cent.
7.6
Analysis in three steps
Interpreting the answers in Table 7-2 poses the above-mentioned problems,
which we will explore further. We will discuss the tendency of over-reporting
in relation to some different interpretations. Over-reporting one’s own contribution may be quite normal as a response to questionnaires and interviews in
general, especially in connection with positively valued activies. However, we
find it worth while to explore this question in some depths in relation to the
answers we have registered to the questions of who did the cooking in a gender perspective. We ask if over-reporting may be understood as a way of
“boasting” about valued activities, alternatively that it demonstrates different
ways of understanding the contribution to housework and cooking according
to gender.
In order to explain the somewhat confusing results, an analysis of men’s and
women’s reporting of who did the cooking was done first, in respect to other
social background variables (Table 7-3). The analysis was done in three steps,
which will be described below.
Eating Patterns
222
First, we will consider the differences between men’s and women’s answers
(Table 7-2, Table 7-3). The obvious impression, statistically speaking, when
we first looked at the results regarding the division of labour in cooking, was
the very large differences according to gender. In order to see if we would
find large differences also in respect to other social background variables,
such as age, education, income and type of household, we decided to further
examine our findings. This analysis is described below in the section “Observing the differences”.
Secondly, we want look more closely at the highest and lowest reported
scores (Table 7-4). Which groups would demonstrate top and bottom results
in relation to cooking activities? We have constructed a classification
“model”54 according to the gendered division of labour in cooking and labelled the findings respectively ‘traditional’, ‘average’ and ‘modern’, as further described in the paragraph “Traditional or modern division of cooking”.
Thirdly, we will analyse the differences in men’s and women’s answers in
reporting their own and their partner’s contribution. This entails a systematic
scrutinisation of the supposed over and underreporting, to discover any observable patterns (Table 7-5). This is described below in “Gender differences
in ways of reporting”.
7.7
Observing the differences
If one assumes that people in the Nordic countries today are egalitarian also in
respect to everyday life activities at home and not only according to public
ideo-logy, one might be surprised by the large (and highly significant) differences in the contribution to cooking activities of men and women living together in households. Here is a considerably higher percentage of women
(mean 77%) than of men, (42%) who report having done the cooking (Table
7-2). This difference is demonstrated for all four countries, irrespective of
weekdays or weekends, and irrespective of having children in the household
or not. The difference is also irrespective of education and occupation (Table
7-3).
54
Inspired of the labelling by Ahrne & Roman (1997) termed Equal, Semi-equal or Conventional.
Weekdays
Sunday
With someone
Couples without
children
Couples with
children
15-29
Age
30-44
45-59
60+
Education Mandatory
Further
Academic
Non-active
OccupaActive, student
tionally
Low
Income
Middle
High
Worker
OccupaLower white collar
tional
Middle white coll.
status
Higher white coll.
Household Worker
status
Lower white collar
Middle white coll.
Higher white coll.
74
57
84
75
88
75
79
75
85
74
70
90
80
78
73
86
57
78
75
87
53
47
36
54
40
26
35
45
42
36
43
31
37
43
38
43
52
49
38
44
49
40
46
34
42
36
**
***
***
***
***
***
***
***
***
*
***
***
***
***
***
NS
***
***
***
NS
***
Denmark
Wo- Sign.
men
82
***
72
***
77
***
79
***
158
241
187
116
168
321
213
154
548
46
100
429
139
170
96
47
240
227
138
65
311
446
134
702
297
(n)
32
46
28
39
29
41
32
41
35
36
36
38
34
42
46
30
35
37
44
37
38
36
36
36
36
Men
55
75
91
97
76
76
82
85
73
86
87
72
78
95
84
58
83
88
82
65
68
**
***
***
***
***
***
***
***
***
***
***
***
***
***
***
NS
***
***
***
*
***
Finland
Wo- Sign.
men
78
***
72
***
77
***
91
***
Levels of significance. p<.000***; p<.01**; p<0.05*; NS=not significant. Chi2-test
Day of
week
Living
Men
148
193
172
72
183
321
82
164
423
102
148
236
79
57
96
39
169
120
143
80
288
479
105
588
216
(n)
48
55
58
28
36
52
54
37
51
44
40
53
55
49
54
65
48
45
55
51
50
53
28
49
45
Men
63
81
85
95
77
81
79
90
77
75
83
80
77
82
78
73
82
85
79
72
79
*
***
***
***
***
***
***
***
***
NS
***
***
NS
***
***
NS
***
***
***
NS
***
Norway
Wo- Sign.
men
83
***
61
***
79
***
84
***
177
249
167
99
126
326
233
115
574
32
150
406
86
100
193
38
165
148
262
60
316
572
118
690
278
(n)
31
59
48
20
40
52
37
28
44
33
41
44
51
40
59
41
40
39
51
41
41
43
31
41
38
Men
Table 7-3 "I did the cooking." Men's and women's answers about who did the cooking in different social groups. Per cent.
47
78
84
81
72
82
68
79
71
75
79
77
67
78
78
70
76
79
74
69
70
**
**
***
***
***
***
***
***
***
**
***
***
NS
***
*
*
***
***
**
**
***
Sweden
Wo- Sign.
men
76
***
72
***
73
***
76
***
152
164
176
61
127
169
301
135
460
59
185
283
66
94
99
76
114
167
139
115
232
424
82
596
255
(n)
224
Eating Patterns
If one were to assume that participation might have changed lately, and compared these results with earlier studies (from the 1980s), one would be surprised to find that the per cent for women's contribution is not lower. Our
findings show that not very much seems to have changed during the last decade in regard to women's (reported) contribution to everyday cooking; the
picture must still be characterised as surprisingly traditional. If we look at
men's reported contributions, however, the picture indicates a notable
change.55 This will be considered further, according to the above mentioned
questions of interpretation.
Regarding women's reported contributions we might notice a certain difference in the Nordic countries, particularly between Norwegian (79%) and
Swedish women (73%) (Table 7-2). Sweden, according to this gender measure, seems to be the most egalitarian (or least unegalitarian) in Scandinavia.
We will also return to this suggestion.
All the background variables age, education, middle and high income, as well
as the occupational status have a statistically significant effect on men's and
women's reported contribution in cooking (Table 7-3). Actually, the gendered
difference is high in all countries except Sweden for people between ages 3044 and for all countries for people over this age. In relation to occupational
status, lower and middle white-collar show gender differences, (except for
middle white collar in Sweden). However, since these findings follow the tradition, (with an interesting exception in Sweden, which will be analysed
later), we may suggest that the most interesting findings will be those showing
the smallest differences. This is because these values indicate a more equally
shared contribution in cooking than those in the broad tradition. We find
smaller differences between men and women related to certain age groups; the
15-29 years old in all countries (15 – 23 percentage points) and also, as mentioned over, the 30-44 years old in Sweden. The same is true for two groups
of class: for workers, (in Norway 21 percentage points, and in Sweden 16 percentage points) and for higher white collar in Denmark (9 percentage points)
and Norway (8 percentage points) and middle white collar also in Sweden (19
percentage points).
To summarise, the youngest partners, and the workers, but also the upper
white collar partners show less difference in the gender division of cooking.
Another interesting finding is that Sweden seems to be the country with less
difference (for certain groups) between how men and women report who did
55
We here touch upon a methodological problem: studies of the gender division of housework
are traditionally based on interviews with women alone.
Ekström & Fürst: The Gendered Division of Cooking
225
the cooking in the households. These findings show that young couples and
higher middle class couples as well as certain Swedish groups may be classified as more equal than other groups.
7.8
Traditional or modern division of cooking
As mentioned above we have made a classification model of those answers
scoring the highest values and those scoring the lowest values in relation to
the contribution in household cooking. This was done by analysing the range
of the reported answers.
In Table 7-4 the top and bottom marks are presented. What we here call a
“traditional” division of labour implies that women more often and men more
seldom do the housework of cooking. A “modern” pattern implies a more
equally-shared contribution. The levels in the model are determined in the
following way. A traditional division of labour implies that most women do
the cooking (85 per cent or more), and only 30 per cent or less among men. At
a level of 20 per cent or less there are no groups represented, and hence the
level of 30 per cent was chosen in order to determine those groups with the
lowest contribution.
In order to indicate a “modern” or equal division of labour, the levels are defined as close as possible to 50 – 50. For men, we decided that it should be 50
per cent or more, and for women - in order to sort out certain groups – we felt
required to state 60 per cent or less. We observe that in the table almost no
groups of women scored less than 50 per cent.
The standard or “normal” division of labour in cooking for men is thus reported to be between 31 – 49 per cent (compared to total Mean 42%) and for
women 61 – 84 per cent (Total Mean 77%) (Table 7-4).
Also in Table 7-4 we see that the background variables age and occupational
status have effects on the contribution. A modern division of labour in cooking is found among the youngest and among the middle class. The only categories of women who are below the 60 per cent level are the very young in
Denmark, Finland and Sweden, as well as the Danish and Finnish women
with middle or higher occupational status. A traditional division of labour in
cooking is found mostly among ages over 45. Norwegian men seem to be the
most modern as they to a great extent score more than 50%, while some
groups of women in Norway and Finland seem to be the most traditional. This
226
Eating Patterns
is according to our interpretation of the gender division of labour as it is presented here by men and women in their self-reporting. This is the case if we
define “modern” according to the men's self-reporting: Norwegian men report
most cooking activities and Finnish men least. Finnish and Norwegian women
however are the most traditional according to our definitions so far, as their
reports score the highest.
The important question is whether this tells us about actual practices or
whether it is rather a statement about ideologies and self-representations. Are
both men and women over-reporting their contributions, and if they are, why?
Are men eager to tell how clever they are, according to those ideological expectations, which are part of egalitarian attitudes? Are women eager to tell
about how much work they do at home? Are they complaining about their
double burden, double work? Or do they like the idea of caring for the family,
and thereby giving them what they need and like? We must admit that it is a
bit puzzling when men are credited with the honour of representing modernity
while women represent the traditional. Norwegian men not only report highest
parttaking in cooking activities, but are also those, whose over-reporting
seems highest. Over-reporting one’s own activities also means underestimating the other’s contributions. Men’s over-reporting is seen, then, when the
results are put in relation to women’s answers, and vice versa. The Finns seem
to be most traditional, but are also those whose reporting is the most concordant if we look at the means for over-reporting (Table 7-5). Perhaps there is
less of a gender struggle in these matters in Finland as compared to Norway?
Weekdays
Sunday
With someone
Couples no ch.
Couples +ch.
15-29
Age
30-44
45-59
60+
Education Mandatory
Further
Academic
Occupa- Non-active
Active, stutionally
dent
Low
Income
Middle/
average
High
Worker
Occup.
status
Lower white c.
Middle white c
Higher white c
Worker
HouseLower white c.
hold
Middle white c
status
Higher white c
Day of
week
Living
52
54
57
53
59
51
55
51
57
54
65
53
55
51
52
52
54
51
59
55
58
50
58
55
47
"Modern"
Men ≥50%
Women≤60%
Dk Fin No Sw Dk Fin No Sw
53
49
38
44
49
40
43
38
44
31
37
35
37
44
37
38
34
42
46
36
36
48
45
49
44
40
41
41
40
39
40
44
33
41
78
75
80
78
73
70
82
65
83
84
72
78
79
72
80
77
82
78
73
82
75
83
77
67
78
78
70
76
79
74
69
75
79
30
87
86
90
88
95
86
87
85
"Normal"
"Traditional"
Men 31 - 49%
Women 61 - 84%
Men≤30%
Women≥85%
Dk Fin No Sw Dk Fin No Sw Dk Fin No Sw Dk Fin No Sw
46 36
43 82 78 83 76
34 36
31 72 72 61 72
28
42 36 49 41 77 77 79 73
36 36 45 38 79
84 76
91
47 38
41 74 68 79 70
36 32 48 31
63
46
84 75 81 78
40
48 75
84
28
91 85
39
81 26
28 20 88 97 95
35
36 40 75 76 77 72
29
45 41
79 76 81 82
42 32
37 75 82 79 68
36 41 37
79
28 85 85 90
43 35
44 74 73 77 71
Table 7-4 Division of labour according to cooking. Divided into the categories "Modern", "Normal" and "Traditional". Per cent.
Eating Patterns
228
7.9
Gender differences in ways of reporting
The somewhat confusing findings in Table 7-2 demonstrated that the sum of
answers as to who did the cooking by men and women for each country was
more than 100%. In the following discussion we want to suggest some further
interpretations. We will consider three meanings of over-reporting. The first
implies a tendency of simply over-reporting one’s own activities when interviewed. When this is considered to imply some kind of error, we will argue
that it is presupposed that men and women understand the concept and the
activity of cooking in the same way. The second meaning represents an alternative interpretation to this. As we have learned from various studies, the understanding of what housework is all about – including perceptions of what
needs to be done – differs according to gender (Haavind 1982, Haavind 1987,
Johansson 1996, Flood & Gråsjö 1997, Roman 1999). This is also the case
with cooking (Ekström 1990). The third meaning of over-reporting that we
will consider implies that the activities are “boasted of” because they have
positive connotations, they are valued or marked by political – or cultural –
correctness.
Concerning the second meaning, we have personally observed in connection
with interviews in households that women and men have different understandings of the division of labour and of their own contributions in the households'
labour. To give an example: On one occasion there were two interviewers visiting a family, one of them sat down with the husband, the other interviewed
the wife. The wife said to her interviewer: “ I do approximately 90 per cent of
all housework” and the husband said: “I do about 50 per cent…”A little later
all four were sitting around the kitchen table, discussing what has to be done
in the home. Now their differences in understanding became obvious. The
discussion ended up with the interviewers asking, with a little laughter, but
also with some seriousness, if it was possible to leave the couple alone, or if
the fight were to continue (Ekström et al. 1996). This example and others
gave us the idea to interpret the differences between men’s and women’s reporting of their contribution to cooking in the context we suggest here. They
report differently because they understand cooking in different ways.
The observations and findings in relation to the division of labour in cooking
are particularly confusing if we do not take into consideration the fact that
men and women represent different social groups, with differing experiences,
values and understandings in certain areas of life. The problem arises when
one expects the household to be homogenous and therefore expect women and
Ekström & Fürst: The Gendered Division of Cooking
229
men to have unanimous opinions about what has to be done in regard to
housework and cooking.
It is important to note that in our study the questions concerning the division
of labour are posed to individual women and men living as partners or couples, and not to the couples jointly. These individuals have given their own
opinions, not a shared answer. The special case in our study is that also men,
and not only women are asked, in contrast to most studies focusing on gender
and housework and cooking (see footnote 10).
The third level of statistical analysis, as mentioned, concerns the differences
in women's and men's reporting. Here we have calculated the differences between the men's and women's respective answers "I myself" and “someone
other(s) did it”56. As previously mentioned, over-reporting one’s own contribution implies a diminishing of the other’s efforts. Over-reporting is when the
answer “I did” from both men and women sums up to more than 100 per cent.
Underreporting is when “I did” sum up to less than 100 per cent. Underreporting could also be interpreted to mean that one appreciates the efforts of the
other member(s) of the household (Table 7-5).
In Table 7-5 we observe a relatively small difference or high correspondence
in answers about who did the cooking on Sundays, from both those with low
and academic education and also those with low income (especially in Denmark, but also in Sweden). We also observe such correspondence in high occupational status in Denmark and Sweden. It is worth noting that minus values implies an inverted relation, the one's "I myself" did the cooking scores
below the other's answers “someone else did it”, i.e. we here observe an under-reporting. Minus values exist among the youngest in all countries except
Norway, as well as in Finland among those with high occupational status.
This might be interpreted to indicate that it is socially acceptable to give credit
to “someone else” for his or her contribution to cooking. Alternatively, it may
indicate that cooking is not evaluated as a positive activity.
The largest differences between women's and men's reports on their own contributions to cooking, and hence a low correspondance, are found in Norway,
mean 27 percentage points (Table 7-5). The differences in the other countries
are smaller (Mean value: 16 - 17 percentage points). One interpretation of this
may be that small differences between the two genders' understanding about
who did the cooking, may indicate that the respondents agree about how this
56
If men and women regarded their contributions equally, then the men’s answers have to correspond to the women’s and vice versa.
Eating Patterns
230
work is supposed to be divided. That is why their reports correspond. Large
differences, on the contrary, may be interpreted to indicate the opposite, that
there are discrepancies in comprehension about the division of cooking.
Table 7-5 Levels of over- or underreporting or differences in men's and women's understanding of division of labour concerning cooking. Percentage points.
Weekdays
Sunday
With someone
Living
Couples without ch.
Couples with children
15-29
Age
30-44
45-59
60+
Mandatory
Education
Further
Academic
Occupationally Non-active
Active, student
Low
Income
Middle/average
High
Occupational Worker
Lower white collar
status
Middle white collar
Higher white collar
Worker
Household
Lower white collar
status
Middle white collar
Higher white collar
Mean
Day of week
Denmark
28
6
19
15
22
-8
38
16
14
10
23
17
22
16
1
27
23
15
17
38
6
16
19
35
-8
17
Finland
14
8
13
27
5
-13
21
19
36
6
17
14
26
8
22
24
10
12
37
30
-12
18
25
26
2
16
Norway
36
-11
28
29
28
12
36
44
23
13
33
33
27
29
19
23
33
32
31
32
39
29
30
34
23
27
Sweden
19
2
14
14
10
-22
36
32
1
12
38
5
8
16
8
20
21
18
18
37
11
17
18
25
10
16
Average
24
1
19
21
16
-8
33
28
19
10
28
17
21
17
13
24
22
19
26
34
11
20
23
30
7
19
These interpretations imply that the question concerns not so much the realities in the division of labour, but rather whether the two genders agree or disagree about how it is divided or shared in practice. When groups of Norwegian men over-report their contribution, it may – in line with the third meaning of this concept - be that they share the cultural ideals of more gender
equality in the kitchen and believe that they do take part in the cooking. Their
female partners, however, may experience this participation differently, what
the partner calls cooking, is perhaps not what she calls it. Perhaps she feels
that he contributes too little in the way that she wishes, his contribution is
considered to be more like a helping hand. Alternatively she is over-reporting,
because she wants to present herself as the one responsible for cooking, an
Ekström & Fürst: The Gendered Division of Cooking
231
activity which might be important for her feminine identity, including a rationality of the gift (Fürst 1997).
The interpretation of an existing ideal of gender equality in Norway, (in line
with a strong norm in this country of equality in all social relations), is supported by the finding that the discrepancy is especially large in the age group
45-59 years (44 percentage points). If we analyse age as a matter of generation, we may state that this group is what in Norway is called "the 68generation", a generation who has identified themselves with radical politics
of different kinds, including feminist claims for gender equality, also in
housework matters. These policies have, in Norway, been supported by the
government, which has formed strong opinions of "political correctness" in
questions regarding gender equality. Norwegian men who seem to over-report
their contribution may be special in our study in the way they want to present
themselves as ideologically acceptable persons, because of the very strong
ideological emphasis in Norway in these matters. It should be noted that our
findings of high scores differ from other studies of men's participation in
cooking, where this, however, is reported by female respondents.
7.10 Interest in cooking
One question was posed to the respondents about their interest in cooking.
Table 7-6 shows the relation between those who reported they did the cooking
the day before and their interest or lack of interest in cooking. 67 per cent of
those men who answered that they were not interested in cooking reported
that they had not done the cooking. 87 per cent of the women who were not
interested in cooking, however, reported that they cooked the day before.
Those men who answered that they were interested in cooking, had in a higher
degree also done it (60%). Thus we may state that women do the cooking regardless of their interest, while men seem to cook mainly because they are
interested.
Table 7-6 Relations between cooking and interest. Per cent.
Did the cooking myself
Someone else did the cooking
Total
N
Interested in cooking
Men
Women
60
92
40
8
100
100
670
917
Not interested in cooking
Men
Women
33
87
67
14
100
101
448
252
232
Eating Patterns
7.11 Conclusion
A summary of the results of this study so far shows that cooking is still to a
great degree women's work. Certain changes, however, may be observed indicating a more equal sharing of the work involved, most visibly among the
youngest and among those with middle and higher occupational status. It may
also be observed that the self-reporting by men and women differs most in
Norway. This is because among men, Norwegians report the largest contribution in cooking. According to this criteria they appear to be the most modern
of the men, followed by Swedes. Most traditional, according to the same measures, are the Finns, both men and women. When it comes to the most modern
women, they are to be found among the youngest in Denmark, Finland and
Sweden and among those with high occupational status in Denmark and Finland.
There is some uncertainty in this study in relation to the question of who answered in the telephone interview. It must be remembered that the answers
were not shared answers, given by actual couples who are living together. The
interviewed persons are women and men living as partners, but their specific
partners were not interviewed. Thus we know nothing about how the actual
couple share the labour of cooking.
A further problem with the sample is the difficulty of finding men who were
willing to take part in the study. Therefore, those men who actually took part
could be seen as being more interested in food matters than is the average
man. Thus the differences between men and women according to contribution
to cooking could actually be larger than our results show. This might explain
why so few groups of men scored less than 30 per cent as seen in Table 7-4.
The question ”who did the cooking” illustrates certain methodological problems in research concerning division of labour at home. Both men and women
have to be involved in the studies and asked for their opinions in order to gain
an adequate understanding of gender relations in households. The very simple
question “Who did the cooking?” revealed that this question concerns gender,
the social relation between men and women. Over- and underreporting can
take place only in relation to the expectations of the other partner.
Studies of the gender division of cooking have to a certain extent been based
on the notion that cooking, like housework, is a burdensome, repetitive Sisyphus activity which ought to be shared equally between man and woman, in
order to obtain equality and to support modern gender roles. Our findings,
however, indicate that cooking may be experienced somewhat differently.
Ekström & Fürst: The Gendered Division of Cooking
233
This is revealed by our interpretation of over-reporting. We have observed a
certain tendency to competition between men and women in their representations of "who did the cooking". Perhaps cooking, in line with Lévi-Strauss'
theories, is also experienced as a creative and culinary activity of transforming
nature into culture - implying a certain civilising aspect. Earlier studies indicate that cooking represents aspects of housework that women take part in
with pleasure (Murcott 1983, Fürst 1985, DeVault 1991). Cooking may also
be experienced as a fulfilling activity when it meets others' needs and satisfies
hunger. This, we have suggested, may imply a gift-relationship which may
be linked to feminine identity.
Such perspectives represent some reluctance to interpret the women who
score high in cooking activities as necessarily traditional in a negative sense,
or that modern woman do not cook. It also suggests an answer to the question
why the developments in gender equality seem to run so slowly in this regard.
Perhaps sameness is not the paramount value in every gender relation, it may
be that difference does not necessarily counter equality (see Braidotti 1990).
In this we may see a positive evaluation of women's traditional work, which
challenges the misrepresentation of it which lies in making it invisible. A wish
to redefine the cultural representations of women’s work and its value lies
behind some of the focus on the difference between men and women in gender research since the 1980s. In her book "In a different voice" Carol Gilligan
argues that even though women's reasoning and moral evaluations may be
different from men's, this does not mean that they are inferior (Gilligan 1982).
Other feminists have in line with this argued that women's orientations are
characterised by an ethic of care, and are a product of women's work and responsibilities in their everyday practice of reproduction and subsistence
(Hartsock 1983).
When men in our study seem to over-report their participation in this kind of
work it might be interpreted to mean that it has acquired a higher value in today’s society, perhaps as a consequence of feminist endeavours to visualise
it's worth. However, to the present women's activities have been culturally
ambiguous, representing in between processes in a borderland which ascribes
to women a secondary value in relation to men. If men, as we might interpret
our results, to a greater degree want to take part in this activity, there is every
good reason to welcome them. Hopefully this will help to upgrade the value
of this kind of necessary work in society. This is a way to understand gender
equality other than one which only underlines the duty of sharing burdens.
234
Eating Patterns
8 Consumers and Food Safety Policies in the
Nordic Countries
Lisbet Berg and Unni Kjærnes
8.1
Introduction
Current debates about food and eating do not only deal with eating patterns.
On the contrary, much attention is directed towards food safety as a question
of consumer protection. Consumer trust has emerged as a key issue that links
food policy and everyday life. As a supplement to the description of Nordic
eating patterns, this study therefore also included questions about attitudes
towards food safety policies.
The supply of sufficient and safe food for everybody is crucial for the overall
welfare level in a society. Events during the last decade have demonstrated
that distrust in the safety of the food system creates turbulence and uncertainty
– for the individual consumer, market actors as well as politicians (Grünert et
al 1996, Latouche et al 1998, Miller and Reilly 1995, Mitchell 1997). No matter whether the consumers are right or wrong in their worries, when consumers fear that food consumption might involve short-term or long-term health
risks, this must be dealt with by actors in the food system. In this chapter we
take the perspective of the consumer, and our main question is: How does distrust in the food system affect consumers’ attitudes towards food safety policy?
One main distinction is associated with the freedom of the individual consumer. Public protection is associated with delegation of control to politicians
and experts and thereby freedom from the need to take safety precautions by
236
Eating Patterns
the individual consumer. Individual responsibility, on the other hand, is based
on some sort of individual control and associated with independence and individual freedom to choose. In order to investigate preferences for these two
classical, and perhaps opposite, policy principles, we interpret a wish for
‘public regulation of the food market through prices and directives’ as a preference for protection, while stressing the need for ‘public information on
healthy food’ signal a preference for staying in control and taking individual
responsibility. We ask: Do consumers who distrust the food system, tend either to ask for more information or to ask for stronger public regulations or
both?
The food system is not a coherent system we either trust or distrust. Distrust is
usually directed towards specific actors or institutions operating in the food
system. Analytically we can only distinguish between trust in public interference and control on the one hand and trust in market mechanisms and its actors on the other hand. Does distrust in the market and distrust in public authorities have the same effect, if any, on the wish for information and the wish
for regulation?
Trust and distrust are of course, if not totally, at least notably, a result of how
the food system functions and operates in real life, viz.; on the one hand, to
what extent are public authorities worth trusting, and, on the other hand, to
what extent are harmful forces in the market kept in check. The Nordic countries have many similarities, but still there are differences in how the food systems operate and what kinds of situations consumers have experienced. We
therefore ask: Are there differences in preferences for public interference in
the food market between the Nordic countries? Finally, we ask whether such
differences between the Nordic countries, if there are any, can be explained by
different levels of distrust in public authorities and in the market, respectively.
8.2
How can we describe the food system of today?
Consumers’ preferences for protective measures reflect the present society
and the food system they are living in. We shall first discuss how the scientific development in general might affect consumers’ relation to trust and distrust in foods. The level of knowledge related to food production and consumption has increased considerably during the last decades. Available
knowledge about food, from production technologies to nutrition and future
health consequences, is enormous, and today only specialists in specific and
limited areas can keep their knowledge up do date. But as consumers we also
Berg & Kjærnes: Consumers and Food Safety Policies in the Nordic Countries
237
get the impression that the specialists do not agree on quite fundamental topics like risks connected to the use of vaccines, additives or biotechnology in
the food industry. As Giddens points out, the only thing we can be sure of today, is that we cannot be certain about the future consequences of such modern production techniques (Giddens 1994). Fischler also claims that today’s
sophisticated and large-scale food production and distribution systems result
in consumers who literally constantly know less about what they are eating
(Fischler 1988). In our everyday meals we need to rely on, not only the person
who prepares the food, but also those unknown and unseen persons involved
in the whole complex and opaque production, distribution and controlling
processes, from the farmer, hunter or fisherman, through the processing and
distribution channels to the dinner table. Like a paradox, rapid scientific development and increased available knowledge has resulted in greater uncertainty (Kjærnes 1999).
Beck explicitly connects the concept of trust, or rather distrust, to uncertainty,
risk and fear (Beck 1992). He argues that industrial societies have created potential dangers unknown in earlier times. What acknowledge and separate
these modern risks from earlier risks is that they are results of human activity
(new production forms) and not natural forces. One example of a modern risk
is long-term effects of food pollution. Beck’s major point is that living in what
he calls a risk society, generates worries and distrust, which in turn promote
more active, reflexive and critical individuals (Beck 1994).
As argued above, rapid scientific development seems to have resulted in
greater consumer uncertainty. Social complexity and increased opportunities
(or necessity) to choose is another dominant aspect of our post-modern society. We shall argue in the following that this aspect too can represent unpleasant challenges:
During the last decades we have witnessed that the number of individual consumption choices have increased considerably in the Western world. Expanding local markets initiated by technological innovations and globalisation of
the food market, together with higher standards of living with increased individual buying power, increase our everyday opportunities to choose. This
need not be only positive. When there are many options, choosing can be difficult, time-consuming and even unpleasant (Berg 1997). Even though information and knowledge is available, we seldom have a satisfactory overview
that will guarantee the best buy with regard to the whole range of attributes
that we want to consider. Often, it takes too much time to gather sufficient
information to be fully informed about a product, including information about
238
Eating Patterns
future consequences of eating it. This is called the information problem57
(Elster 1989). It may not be a big problem that we cannot taste all available
cheeses before making our decision on what to serve our guests. The information problem, though, feels more unpleasant if our trust in food is questioned
and we need to consider whether the food we buy and serve is safe or not.
Anyhow, the complex and extensive food market, with almost infinite possibilities of choice, can easily be perceived as overwhelming by the consumer.
We need strategies that can save us from the freedom of choices.
According to Luhmann, ‘trust’ is a classifying mechanism that reduces such
problematic social complexity. In modern societies with advanced division of
labour, we cannot be in control of everything, nor is it possible always to
make informed decisions (Luhmann 1979, chap.7). By trusting others, institutions or systems, we solve this problem. Luhmann treats trust (as well as distrust58) as a mechanism that helps us to simplify our lives.
To sum up: While Beck focuses on the necessity of reducing risk, Luhmann
focuses on the necessity of reducing complexity. According to Beck, distrust
complicates our lives by demanding a more active, reflexive and critical consumer attitude, and, according to Luhmann, trust (and distrust) simplifies our
lives by liberating us from complicated choice situations. It is easy to recognise and accept both Beck’s and Luhmann’s approaches and arguments,
which imply that modern societies are characterised by both new forms of
risks and a problematic complexity.
If we return to consumers’ food policy preferences, public regulation of the
complex food market could release consumers from time demanding personal
food safety precautions, while information could help the reflexive consumer
to choose what s/he considers to be the safest available food.
57
One of the problems that reduces our capacity to make rational choices.
Luhmann (chap.10) also regards “distrust” as a mechanism that reduces the social complexity. Distrust is then not only seen as opposite to trust, but also equivalent to trust. In our everyday purchasing of food we can choose between the two complexity reducing mechanisms, that
is to trust or distrust the product, the producer, the salesman, the market, the public control etc,
etc.
58
Berg & Kjærnes: Consumers and Food Safety Policies in the Nordic Countries
8.3
239
Methodology
We want to stress that “trust” is a complicated concept and also that “the food
system” is multifaceted and diffuse. It is therefore almost impossible, we
think, to find a perfectly valid assessment of “trust in the food system”. This
implies that the empirical results should be interpreted cautiously, and, unfortunately, that results based on other operationalisations might differ. Anyhow,
the following analysis is an incipient attempt to illuminate this complex, but
still important and interesting, topic.
Our material is based on the study ‘Eating and Modern Everyday Life. A
Comparative Survey of the Nordic Peoples’ (see chapter 1). In addition to a
record of one day of eating and its social context, there was a battery of questions concerning, partly, attitudes towards social norms in relation to meals
and eating, partly towards trust, food regulations etc.
The following multivariate analysis is based on responses to two statements
related to preferences for how the state should interfere in the food system
(dependent variables), four statements about trust and distrust directed towards the food system (intervening variables), three variables measuring individual characteristics; gender, age and education, and finally country of residence (independent variables).
With each statement the respondents could choose between agree, partly
agree, don’t know, partly disagree, and disagree. Even though these variables,
strictly speaking, do not represent continuous variables, which are preferable
in regression analyses, we still find it justifiable to use the variables in regression analysis, which has proven to be a rather robust technique (Labovitz
1967).
Since the material includes as many as 4800 respondents, even relatively
small effects will yield statistically significant estimates. In the interpretation
of a multivariate stepwise regression analysis such small effects will only be
confusing, and we therefore choose to exclude estimates smaller than beta =
.06, from the presentation. The material is analysed by SPSS.
8.4
Mapping
Initially, we will present how the consumers in the Nordic countries responded to the different statements that later will be included in the multivari-
240
Eating Patterns
ate analysis. Nordic consumers were asked to evaluate different statements
about the food system in their country. The following analysis is mainly based
on six such statements. Two statements refer to preferences, i.e. what kind of
food policy (regulation or information) the respondents think should be followed in order to promote food safety (which need not say anything about
how s/he evaluates the present situation). Four statements refer to the status
quo, i.e. how market and public institutions, in the opinion of the respondents,
affect food safety in their everyday lives.
In the questionnaire, some statements were formulated positively, some negatively. In order to get a more simple picture, we will present the results in
such a way that all the statements point in the same (in Figure 8-1- positive)
direction.
As shown in Figure 8-1, almost every Scandinavian agreed, or partly agreed,
that “[i]t is an important duty for the public authorities to inform the population about healthy food59” (Information). The majority also agreed, or partly
agreed, that “[p]ublic authorities should regulate the food market through
prices and directives in order to ensure safe and healthy food60”(Regulation).
It was expected that more respondents agreed on information than on regulation, since one can choose to use available information or not, while regulation is a much stronger measure affecting everybody. Anyhow, the distribution of answers on these statements signals that most people want a strong and
active food policy in order to promote food safety, including safety regulations and information. These statements do not, however, say anything about
whether the respondents are satisfied with the situation in their own country or
not.
The next two statements refer to trust/distrust in public authorities. One third
agreed, and another third partly agreed that “[p]ublic food inspections guarantee that the consumer does not get food with salmonella61”(Public authority
control). It was not that common to see the European Union as an institution
promoting safety in the food market. Only nine per cent agreed and another 24
per cent partly agreed that “European common market regulations guarantee
that the quality of the food on the market is satisfactory62”(EU regulations).
59
Det er en viktig oppgave for myndighetene å opplyse befolkningen om sunn kost
Myndighetene bør regulere matvaremarkedet gjennom priser og lovgivning for å sikre gode
og sunne matvarer.
61
Offentlig kontroll garanterer at forbrukeren ikke får mat med salmonella.
62
EUs regelverk garanterer at maten som selges holder god nok kvalitet.
60
4
International trade
7
12
3
34
32
Fully agree
9
10
3
24
20
7
Partly agree
31
62
30
36
40
Neutral
34
34
50
29
60
56
80
19
13
44
29
Partly disagree
3
5
70
31
6
12
14
1
Disagree
90
3
100
Figure 8-1 Wish for protective measures against food insecurity, and opinions on what institutions promote consumer trust among Scandinavians. (N= weighted frequencies due to sample size and number of habitants in Norway (4.4), Denmark (5.3), Sweden (8.8) and Finland
5.1)).
5
Food producers
EU regulations
State authority control
Trust in:
Regulation
Information
W ish for:
0
Eating Patterns
242
The last two statements refer to trust/distrust in market actors. Rather few
consumers think that “[f]ood producers are - not – more concerned with making money than with consumers’ health63”(inversely presented in the questionnaire) (Food producers), and “[i]ncreased international trade has - not created new kinds of food risks64” (inversely presented in the questionnaire)
(International trade). Based on the responses to these statements, the overall
pattern seems to be that the typical Nordic consumer trusts public authorities
more than market actors. In the field of food, s/he wants market mechanisms
to be regulated by public authorities. Nordic people are concerned about the
problematic sides of the market, and they want the state to regulate it and protect them against food scandals caused by the greed of some market actors
(i.e. Belgium, summer 1999).
8.5
Comparison of the Nordic countries
The Nordic countries are quite similar in many respects; culturally, with regard to the relation between state and market, as well as the forms of welfare
states (Esping-Andersen 1990). This is true also when it comes to the particular field of consumer protection with heavy emphasis on public intervention in
the market (Ilmonen and Stø 1997). Food safety regulations have been harmonised between the countries for quite a long time. Such common structural
and normative frameworks allow a more focussed discussion of concrete
variations between the countries (Przeworski and Teune 1970).
But, still, although less well documented, there are differences with regard to
the food market and food consumption. Moreover, public policy in this area is
organised differently, both with regard to decision-making procedures and the
concrete enforcement. And we do expect that general conditions will influence consumer attitudes with regard to food and nutrition policies, the overall
situation of political trust, party politics etc., as well as more situational factors, including food scandals and media debates. Denmark, with its export
oriented food industry, has been a member of the European Union since 1973.
Sweden and Finland, who both joined the EU in 1995, have a considerable,
but much less export oriented agricultural industry. Norway is still not a
member, but takes part in the common market (with harmonised regulations)
through the EES agreement. With a small, but strongly protected, domestic
agriculture the Norwegian food market is heavily dependent on imports. The
large fisheries, though, is an export-dominated industry.
63
64
Inverse of: Matprodusentene tenker mer på å tjene penger enn på forbrukernes helse.
Inverse of: Mer handle over landegrensene har skapt nye former for matrisiko.
Berg & Kjærnes: Consumers and Food Safety Policies in the Nordic Countries
243
The traditions of nutrition policy and education are quite diverse in the four
Nordic countries (Kjærnes 1997). Norwegians have been most concerned with
market regulations in order to obtain nutritional goals, and Swedes (and
Finns) with public catering and school meals. But they have all, including the
Danes, shared an interest in education through mass media campaigns etc.
The Nordic countries are generally regarded as high-trust societies, but there
are variations even here. Norwegians have a high and stable trust in political
institutions. While this used to be true in Sweden, too, the level of trust has
been reduced in later years, now being similar to the more moderate level of
Denmark. Of the four countries, Finns seem to have the lowest level of trust in
political institutions (Listhaug and Wiberg 1995, Miller and Listhaug 1998).
Finally, other parts of this Nordic study indicate that even the food cultures
differ considerably between the Nordic countries (Mäkelä et al 1999).
As exemplified above, the food systems in our four countries have both similarities and differences. The differences do, however, not seem to form a specific and reinforced pattern that could help us predict how the citizens in the
different countries would respond to our attitude statements.
Figure 8-2 shows what protective measures against food safety problems the
consumers in our four selected countries prefer, and also how they evaluate
the present situation when it comes to food safety. The variables that measure
distrust in the food market and public authorities, respectively, are constructed
by the four last variables in Figure 8-165.
The main impression from Figure 8-2 is that we find both similarities and differences: As expected, in all the countries, more consumers were in favour of
public information than public regulations. Based on our operationalisations
65
To measure distrust, we have chosen to use two index variables, each based on responses to
two statements. And in order to separate distrusting consumers from the others, we only count
those who show strong distrust towards the food system, and the indexes are then labelled
“strong distrust on both statements”(2) ‘strong distrust on one statement”(1) “no strong distrust”(0).:
i) Distrust Market: Totally agree in the statement “[f]ood producers are more concerned with
making money than with consumers’ health” + totally agree in the statement “[i]ncreased international trade has created new kinds of food risks” (International trade).
ii): Distrust Public Authorities: Totally disagree in the statement “public law enforcement
guarantees that the consumer would not get food with salmonella” + totally disagree in the
statement “[t]he European common market regulations guarantee that the quality of the food
on the market is good enough”.
Eating Patterns
244
of distrust, Figure 8-2 also shows that in all the countries it was more common
to express distrust towards the market than distrust towards public authorities’
regulation of the market.
0
10
20
30
40
50
60
70
80
90
100
64
37
Denmark
42
73
60
40
Finland
33
69
54
22
Sweden
38
63
76
43
Norway
33
79
Wish for information
Wish for regulation
Distrust public authorities capacity to secure safe food
Distrust food market
Figure 8-2 Preferences for protective measures against food insecurity (‘fully agree’) and
distrust with status quo (see footnote 12). Percentages “Fully agree”. (N=4805).
But we also find differences: Norwegian consumers were more often than the
other Nordic peoples in favour of both public information and public regulation of the food market. While 76 per cent of the Norwegian consumers were
very positive to public information and 43 per cent to regulation, this was the
case only for 54 and 22 per cent of the Swedes.
We also see that Denmark had the highest level of distrust directed towards
public authorities’ capacity to secure safe food, while Norway had the highest
level of consumers who were sceptical towards market actors.
There is no obvious and simple explanation to these differences. We expect
preferences for different food policy strategies to be influenced by several
factors, e.g. differences in food culture, agriculture, nutrition policies and the
real level of information and regulation. In other words, we expect several
Berg & Kjærnes: Consumers and Food Safety Policies in the Nordic Countries
245
explanations connected with the country of residence that affect the level of
wish for information and/or regulation. In the analysis to come, however, one
objective was to investigate whether preference for protective measures are
affected by the level of distrust. In other words, we asked whether variations
in the level of distrust in the food system could explain some of the variations
among the Nordic countries in preferences for protective measures.
Separate calculations show that it was not unusual to wish both regulation and
information. While the majority of those who wanted regulations also wanted
information, less than half of those who wanted information also wanted
regulation. This pattern recurred in all the countries.
8.6
The analytical model
In order to illuminate the questions raised in the beginning of this chapter, we
have based our analysis on the following model (Figure 8-3): It is a two-step
path analysis constructed by four regular linear regression models. The independent variables include country of residence, gender, educational level and
age. The intervening variables measure distrust in the food system, where we
distinguish between ‘distrust in market’ and ‘distrust in public authorities’66.
Finally, the dependent variables measure what protective measures against
food safety problems that the consumers prefer67.
The most important independent variable is country of residence, which was
included in the model as dummy variables. Even though the Nordic countries
have many similarities, we do expect that countries of residence also represent
different frames, not only in terms of policy regimes, but also different food
experiences for the consumers. For instance, the incidence of salmonellosis in
humans seems to be considerably lower in Norway than in Denmark68. We
expect such facts to affect the level of trust among the citizens. In the analyses, we used Norway as the basis for the comparison, and the model thereby
allows us to investigate whether Swedes, Danes and Finns distrusted the food
system to a larger or smaller extent than Norwegians do.
66
The intervening variables are labelled from 2(strongly disagree on two statements) 1(strongly
disagree on one statement) 0 (no strong disagreement) See also previous note.
67
The two statements used as dependent variables are labelled from 5(agree) to 1(disagree).
68
Sources: Trends and Sources of Zoonotic Agents in Animals, Feedingstuffs, Food and Man
1999 for Denmark (the Danish Zoonosis Centre) and Norway (the Norwegian Zoonosis Centre). It is emphasised that there is considerable uncertainty about the incidence and comparison
between countries can only be made in very general terms.
Eating Patterns
246
Country of Residence/
Individual Characteristics:
Distrust in
Food-system:
Wishes for
Protective
Measures:
Norway=0
Sweden=1
Public authorities should
regulate the
food market
through prices
and directives in
order to ensure
safe and healthy
food
Denmark=1
Finland=1
Man=1
High Ed.=1
Age (cont)
Distrust
Public Authorities
Distrust
Market
It is an important
duty for the
public authorities to inform
the population
about healthy
food
Figure 8-3 Model 1: The analytical model.
As seen in the model, we also included some individual background characteristics that we expected to have some influence on the level of trust. It is a
common finding that gender, age and educational level influence attitudes
related to food and health risks (Wandel 1994, Kearney et al 1997). Berg
(2000a) also found that female and older consumers tend to be more concerned about food related issues, while the impact of educational level was
less clear. In the following analysis, however, the individual variables are not
our main focus.
The intervening variable ‘distrust in the food system’, will, according to our
questions, be treated both as intervening and dependent variables. The food
system is quite complex, with actors and institutions with different interests.
And as consumers we hardly perceive the food system as one coherent institution we either trust or distrust. Different mechanisms of trust are expected
(Kjærnes 1999). We therefore found it convenient to distinguish between distrust directed towards the food market, on the one hand, and distrust directed
Berg & Kjærnes: Consumers and Food Safety Policies in the Nordic Countries
247
towards the role of the public authorities (at national and European levels) in
the food system on the other hand.
Finally, in the second step of the model we tried to investigate whether a consumer who feels that the food system is not quite safe and secure, preferred to
take personal control by asking for more information, or preferred to be protected by public regulations of the food market.
8.7
Empirical results
In the following, we present the results when using wish for information as the
dependent variable and then the results when using wish for regulation as the
dependent variable in two separate figures. (The first step in the analyses is of
course identical.) The main purpose of the model was to investigate whether,
and possibly how, distrust affects wishes for information and regulations, and
whether there are differences among the Nordic countries.
Based on the results presented in Figure 8-4 and Figure 8-5, we can illuminate
our main question: How does distrust in the food system affect consumers’
food policy preferences?
First, we notice that distrust in public authorities, on the one hand, and distrust
in the market, on the other hand, have opposite effects on what role the respondents want the state to play in the food system. Those who distrust the
market (otherwise all equal), tend to prefer both regulations (.14**) and information (.17**) more often than others. But those who distrust public authorities rather tend to prefer regulations (-.09**) and information69 more seldom than other. In other words: Distrust in how the food market operates contributes to a wish for public intervention, while this is not at all the case when
the consumer distrusts the way the state and/or the European Union operates.
We did not expect to find these negative effects from distrust in public authorities on wishes for protective measures. Rather, we expected that distrust
in public authorities would reflect consumers who thought that the public authorities’ capacity to ensure safe food was low due to too little intervention
and too few regulations, and that these consumers accordingly would want
69
The estimated effect from ’distrust public authorities on ’information’ is -.04**,
which is smaller than -.06, but still included in the model in a parenthesis because
this effect is essential for our understanding.
Eating Patterns
248
stronger public regulations and more information. Instead, our results indicate
that those who distrust public authorities’ capacity to ensure safe food think
that public intervention is not trustworthy and accordingly want neither public
regulations nor information.
Norway=0
-.17**
Sweden=1
-.08**.
Denmark=1
-.14**
Finland=1
-.10**
Male=1
-.09**
High.Ed.=1
-.20**
Age
Public authorities should
regulate the
food market
through prices
and directives in
order to ensure
safe and
healthy food (15)
.11**
-.06**
.07**
-.10**
Distrust Public
Authorities
(0-2)
.14**
.13**
Distrust in
Market
(0-2)
**Sig. p<0.01
Figure 8-4 Effects from distrust in public authorities and distrust in the market on the
wish for public authority regulations on the food market. Smaller effects than .06** are
not included in the model. Standardised regression coefficients. (Beta) (N=4784)
Berg & Kjærnes: Consumers and Food Safety Policies in the Nordic Countries
249
Norway=0
-.17**
Sweden=1
It is an important duty for
the public
authorities to
inform the
population
about healthy
food
(1-5)
-.10**
Denmark=1
-.10**
Finland=1
Male=1
High.Ed.=1
-.20**
.11**
-09**
(-.04**)
Age
.
Distrust
-.06**
Public
.07**
Authorities
-.10** (0-2)
.13**
.14**
Distrust in
Market
(0-2)
**Sig. p<0.01
Figure 8-5 Effects from distrust in public authorities and distrust in the market on the
wish for public information about healthy food. . Smaller effects than .06** are not included in the model. Standardised regression coefficients. (Beta) (N=4784)
We also asked whether distrust in the food system seemed to result in a wish
for either more public information or regulations, or if consumers followed
both strategies simultaneously. As shown in the models, distrust in the food
market seems to activate both a wish for information and for regulation, but
the models do not tell whether it is the same individuals who wish public information and regulations. Separate calculations show, however, that among
those with strong distrust towards the market, about 40 per cent strongly desire both information and regulations and about 40 per cent strongly desire
either information or regulation70. Based on this, we will claim that there need
not be any contradiction between preferences for public protection and a wish
70
Since the majority of our respondents are in favour of both public information and state regulations, several of those who do not strongly wish for both information and regulation, still have
preferences for both (Figure 8-1).
250
Eating Patterns
to keep personal control and take individual responsibility when it comes to
food consumption.
The results could probably also be interpreted in a second way: Both dependent variables - wish for public information and wish for public regulations can implicitly be understood as indications of trust in public authorities. From
this point of view, we might interpret the results as follows: If we cannot trust
the food market, Nordic consumers want to trust public authorities; that is,
establish stronger trust in public interference in order to compensate for reduced trust in the market. The authorities will thus act as an independent third
party that can reduce the pressure on the consumers’ individual choice. We
tend to react in such a way, because, in Luhmann’s term; we need to trust.
If we cannot trust the way public authorities interfere in the food system, the
distrust has more serious consequences. If that is the case, our model does not
cover how the consumer will solve the situation. But, because the basis for
recruitment of new trusting agents in such a situation is quite limited, the consumer may feel that s/he has few other options than to approach the food system individually in a cautious, conscious, critical and, in Beck’s term, reflexive way. Distrust in public authorities may then contribute to a growing interest and demand for alternative forms of production and distribution. Arrangements like farmers’ markets and organic foods, where the distance from
the producer to the consumer is conceived as shorter, more transparent and
surveyable, will gain popularity.
This latter interpretation could signify that distrust mainly directed towards
the market, will be met by consumers looking for new trusting agents
(Luhmann), agents that can reduce complexity by introducing an independent
sceptical eye, like stronger intervention by public authorities or specific and
trustworthy actors in the market. But when, or if, distrust continues to grow
and distrust directed towards public authorities reaches a certain level, then
reflexive, but also disillusioned, consumers may emerge in greater numbers
(Beck).
According to our material, the majority of the Nordic consumers were still
rather trusting when the data were collected in 1997. As shown in Figure 8-2,
in all the Nordic countries far less consumers distrust the public authorities
than the food market. One important implication of these results is that distrust in the food system should not be treated as one coherent quantity. Distrust in the market and distrust in the state seem to have quite different implications.
Berg & Kjærnes: Consumers and Food Safety Policies in the Nordic Countries
251
We also asked whether differing wishes for regulations and information
among the Nordic countries could be explained by different levels of distrust
directed towards the food market and state interference in the market. If the
level of distrust was the only explanation and our variables gave an accurate
measure of distrust, all effects from country to the dependent variables would
be indirect (through distrust in the market and/or distrust in the state), which
they are not. Nevertheless, while we cannot expect such accuracy from our
variables, the results do indicate traces of such expected indirect effects: Norwegians tend to distrust the market more than Danes (-.06**), Swedes (.10**) and Finns (-.20**), which should trigger a higher wish for both information and regulation among Norwegians. Denmark is also distinguished by a
higher degree of distrust in public authorities (.11**), which also reduces the
wish for information and regulations. In other words: one explanation for a
higher wish for public information and public regulations in Norway than in
the other three Scandinavian countries seems to be that Norwegians distrust
the food market more, and/or distrust public interference less, than citizens of
the other Nordic countries.
In this chapter, we have not emphasised the implications of gender, education
and age, even though we expected such individual characteristics to affect the
level of trust and the wish for public regulations and information. Somewhat
unexpectedly, our analysis shows that among the individual variables only age
has significant effects71 on the level of distrust. Older people are more sceptical than younger people, both towards the market and towards public authorities. Older generations may prefer other, more personal trust relations rather
than the abstract, system based trust indicated by these questions. When it
comes to preferences for protective measures against food safety problems,
we find no effects from gender or education on the wish for information, and
only small (and opposite) indirect effects from age through the intervening
variables measuring distrust.
The individual variables do show differences in impact on the wish for public
regulations: Women, lower educated people and older people (indirectly)
want the state to regulate the food market through prices and directives more
often than others. This finding was repeated in a survey among Norwegian,
Belgian and British consumers in 1999 (Berg 2000a, Berg 2000b).
71
As reported we only present estimates equal or greater than .06**. Actually there are small,
but statistically still significant, effects from gender (.05**) and education (.04**) on distrust in
state, and a small effect from education (-.04**) on distrust in market.
Eating Patterns
252
8.8
Concluding remarks
Consumer distrust in the food system has received a lot of attention in the media in recent years. This situation of change and turbulence has emerged in a
period with an immense increase in the knowledge about food, knowledge
that is highly specialised and largely abstract for the consumers. The food system is also increasingly complex, involving a lot of actors, the range of foods
offered in the market has provided a variety that was difficult to imagine just a
few years ago, and transparency becomes more and more problematic. While
opening for individual creativity, this situation can also become a problem and
evoke a feeling of uncertainty (Giddens 1991, Luhmann 1979).
There are two classical political solutions that are expected to strengthen consumer safety and security. One is consumer protection, which implies public
regulation of the market. This solution is meant to prevent inferior and harmful products from being marketed. The other strategy has been operationalised
as the commonly expressed opposite to this solution, namely consumer
information. By this means, the consumer becomes more capable of actively
choosing between different products and properties, in other words increasing
personal control and individual responsibility. Our analysis indicates that
many people in the Nordic countries think that information and public regulations should be supplementary rather than alternative food policy strategies.
One interpretation is that consumers, on the one hand, want public authorities
in order to ensure a basic level of safety. But, on the other hand, for their own
feeling of security they still want information, education and openness in order to make possible some sort of individual control. They want a considerable degree of freedom to choose - when this does not threaten their basic
safety. One could call this a typical attitude among citizens in and proponents
of the Scandinavian, social democratic, welfare state.
The food system is complex and trust is a problematic concept. In our analysis, we found it fruitful to distinguish between distrust in market conditions,
on the one hand, and distrust directed towards public interference in the market, on the other. According to our analysis, trust in these institutions appear
to have opposite effects: The more the consumer distrusts the food market,
and the more the consumer trusts the way public authorities interferes in the
market, the more common it is to wish for public regulations and information.
This means, for example, that if you distrust the way regulations of the market
operate today, you are less likely to wish for public solutions like strict regulations and public information. In other words, trust in public authorities seems
Berg & Kjærnes: Consumers and Food Safety Policies in the Nordic Countries
253
to be a significant condition for beliefs in public information and regulatory
measures of the food system. While information strategies leave more freedom to choose for the consumer, the use of information has to be founded on
trust in the sender (which in this case is public authorities). On the other hand,
distrust in market institutions, which is quite common according to our data, is
associated with a wish for public interference in the food system. National
public authorities are important as trust building institutions in the Nordic
countries, while scepticism towards the market as well as international institutions seems to be more widespread.
According to our analysis, Norwegian consumers seem to be somewhat more
in favour of public regulations and information than consumers in the other
Nordic countries. We asked whether differences in preferences for protective
measures could be explained by varying levels of trust in public authorities
and in the market, respectively. The analysis supports such a hypothesis, but
we cannot claim that the level of trust explains all differences between the
Nordic countries. Higher distrust in market mechanisms and higher trust in
public interference in the food market explain partly why Norwegian consumers more often than Swedish, Danish and Finnish consumers are in favour of
public protective measures against food insecurity.
We do not have one solid explanation for why the measured levels of trust in
the Nordic countries vary. We have described how the Nordic countries, while
having many similarities concerning norms and institutions, also differ in several important respects, both concerning the particular policies in question and
more general conditions of trust.
One may distinguish between three different levels of interpretation of national variation. Although lacking good, comparative studies, it may be anticipated that the experiences differ in terms of concrete food scandals. While
hardly any scandals have appeared in Norway, there have been quite a few in
Denmark in connection with contaminated meat, perhaps more than in
Finland and Sweden as well. This may, at least partly, explain the relatively
higher level of trust and belief in public solutions in Norway (and Sweden and
Finland), compared to Denmark.
The second level is related to diverging food policies. Food safety policies are
not that divergent in the Nordic countries, at least in principle. However,
again, Norwegians have generally been most eager to regulate the food market
for various purposes, Danes, with their export-oriented agriculture, the least.
This is, for example, reflected in their respective nutrition policies. While having long established nutrition policies in Sweden and Finland, their eagerness
254
Eating Patterns
for market regulation has been lower than in Norway. Agricultural policies
also seem to have been more rapidly liberalised in these countries compared
to Norway. In Denmark, a liberal food policy seems to have been marked by a
rather high level of conflict, consumer and environmental groups being
among the strongest critics (Holm et al 1993). Yet, scepticism towards market
solutions seems to be more widespread in Norway, expressing a general antiliberal attitude towards the food market, while this topic is more controversial
in Denmark. We have seen in our study that distrust in market actors is associated with a higher wish for information and regulations.
The third level of explanation is related to political trust in general. In studies
that are not related to particular issues, Norwegians usually show a high level
of trust, while Finns and (perhaps also Swedes) represent the other extreme,
Denmark taking an intermediate position (Listhaug and Wiberg 1995, Miller
and Listhaug 1998).
Only one single cause or mechanism can very rarely explain a social phenomenon. In comparative studies, the picture is even more complicated: Different causes and mechanisms often have different explanatory powers in different countries, simply because the national conditions vary. In this particular
empirical analysis, we have investigated whether differences in preferences
for protective measures against food insecurity could be partly explained by
national differences in distrust directed towards the food systems. We do find
support for this assumption.
The results also indicate that in order to maintain a well-functioning and integrated food system, distrust in how the public authorities intervene in the food
system seems to be far more serious than distrust in the food market.
9
Discussion and Conclusions
Unni Kjærnes
9.1
The aim of the study
There is an ongoing debate about the role of meals and eating in modern everyday life. The present project has analysed differences and similarities in
food habits in four quite similar Nordic countries. By using a comparative
approach, we hoped to reveal cultural variations and shared features of contemporary eating patterns.
We wanted to discuss contradictory assumptions about the structure of modern eating, as characterised either by continuity and clear patterns, or by tendencies of dissolution and anomie. Our main focus has been on the existence
of structures for the individual, i.e., the existence of ‘meals’ (as opposed to
‘grazing’) – leniency vs. strictness - on the one hand, and culturally and socially shared patterns and formats of eating – variety vs. uniformity - on the
other. There are also diverging assumptions about eating patterns as solely a
matter of individual choice of lifestyle, or as strongly influenced by social
structure and societal institutions. One hypothesis has been that differing ways
of shaping social institutions such as work and family will be reflected in differing eating patterns, but not (necessarily) in terms of strictness or uniformity. We wanted to describe everyday reality in a field where many generalised
assumptions lack an empirical foundation.
We have described several dimensions of everyday eating habits, including
the composition of meals and selection of foods, the daily patterns and the
social aspect of eating. We were also interested in the social division of labour
Eating Patterns
256
with regard to food preparation and cooking - between family/home and public arenas, and the impact of gender roles for the division of labour within the
family.
This final chapter will combine the findings of the various chapters and discuss them in relation to the overall questions posed in chapter 1. We will examine the structure of eating more literally – in terms of eating patterns and
formats as well as the social context of eating. We will also discuss and bring
together the various dimensions of ‘proper meals’. This notion, which is used
both in everyday language and in the literature, has formed an important reference in public fears about disappearing meals. As methodology has been a
central concern in this project, we will also discuss some methodological aspects, particularly with regard to generalisation and the validity of the findings.
9.2
Grazing vs. structured eating in the Nordic countries
‘Grazing’ is a popular notion, probably originating from market research
(Caplan 1997). We have interpreted grazing as indicating an anomic situation
where eating is no longer socially shared or co-ordinated and not individually
patterned (Fischler 1988). According to Fischler, cultural norms for what
should be eaten when and together with whom, disappear. We have used the
term ‘de-structured’ to describe this situation with regard to concrete eating
practices. Opposed to ‘de-structured’ ways of eating, there must be ways of
eating where certain arrangements or ‘structures’ can be identified. ‘Structure’
is thus used simply in contrast to non-systematic variation, both at the individual and an aggregate level (see also our analytical model described in
chapter 1). However, observed structures need not necessarily be in accordance with explicit expectations and conventions. The question of conventions is discussed in the next section, which deals with ‘proper meals’.
We have investigated the degree and types of structures of eating along three
main dimensions: eating patterns, the food, and the social context of eating.
9.2.1 Eating rhythms
The regularity of eating patterns has been operationalised as the timing of
various types of eating events during the day and the number of each type of
event. This is relevant both in terms of the degree of uniformity of eating
Kjærnes: Discussion and Conclusions
257
rhythms in the populations and population groups (i.e., whether manifest patterns can be identified in the samples), and with regard to the degree of strictness at the individual level (i.e., whether eating is concentrated in relatively
few, distinct events).
In chapter 3, quite evident national patterns in the daily rhythm of eating were
identified. There were commonly shared eating times and patterns in each
country, and eating was concentrated at shared peak times: breakfast, lunch
and dinner. Eating hours were most uniform in Denmark, least so in Finland
and Sweden. In all four countries, there were additional, but smaller peaks
during mid-morning, mid-afternoon and in the evening, but these varied considerably between the countries. Eating hours seemed to vary between the
countries, particularly with regard to the time of dinner. Eating hours were
quite different and showed much more heterogeneity on Sundays.
Gronow and Jääskeläinen suggest in their chapter that the rhythms of eating
represent socially shared routines that have emerged from the need for social
co-ordination, particularly in relation to work and family life. In our study,
this was most clear for breakfast and lunch hours, where working hours have a
direct impact. Family meals at home in the evening (dinner) require coordination within the household, but not at a community level. But dinnertime is also influenced by change within the work sphere, as illustrated by the
shift in time for dinner in Norway after a change in working hours in the late
1970s (Statistics Norway 1992). Eating rhythms are by no means compulsory
(except for those living in some kind of institution); they represent very flexible conventions with a lot of freedom for the individual (Gronow and Warde
2001). People may establish their own routines and they may make exceptions. This was emphasised by the finding that throughout the week, a considerable proportion of the respondents had had something to eat at any hour during the day, from 4 o’clock in the morning till 12 at night.
The types of eating events that dominate at different hours were quite uniform
with regard to the main eating events, thus reflecting strong cultural elements
in these patterns. This was clearly demonstrated with regard to the cold vs. hot
lunch traditions in Denmark and Norway, Finland and Sweden, respectively.
Diverging opinions about the midday meal have been most visibly expressed
in debates about school meals. Since the inter-war period, the Norwegian debate has focused on the “packed lunch”, with sandwiches regarded as being
practical as well as nutritious, while the Swedish debate has been very concerned about school children needing a proper hot meal at that time of the
day.
258
Eating Patterns
The alternatives to observing shared conventions are either to establish individual routines or to skip predictable meal patterns altogether. ‘Grazing’ implies that food is eaten more or less randomly with regard to time, place and
contents, according to immediate preferences and concerns. In this study, we
found that for the majority in all countries, eating was concentrated in relatively few events72. Between eighty and ninety per cent had had 3-5 eating
events on the day before (and somewhat less on Sundays). The total number
of eating events was slightly lower in Norway than in the other countries, the
main difference being fewer snack events. Ten per cent or less had no hot eating events on the day before, and a majority of the cold events consisted of
sandwiches, breakfast cereals etc. All this indicates that minor snacks eaten
randomly represent additions to rather than a replacement of a general meal
pattern consisting of 1-2 hot meals plus sandwiches. The extremes – with very
few or more than 5 events per day –seem to represent only what might be expected from the variation and flexibility of ordinary everyday life (illness,
extra treats etc.).
On the basis of this study alone no conclusions can be drawn about the frequency and general occurrence of very small 'in between' snacks. What can be
concluded, however, is that no respondent ate only such minor snacks. It
should also be kept in mind that in these kinds of surveys small groups with
their peculiar eating habits (e.g. some religious or ethnic minorities or people
living in closed institutions) are not discernible, nor are people with strongly
'deviant' eating habits, such as those suffering from eating disorders. Nationwide surveys give information only about the majorities.
Prättälä (2000) has collected studies of Nordic meals conducted during the
1980s and 1990s. These studies include mainly questions about habits, asking
what people normally do. Concerning breakfast, our findings confirm those of
earlier studies, in that more than 90 per cent have had a morning meal. Regarding lunch and dinner, these studies report somewhat higher frequencies
than those from our study. This could be interpreted as indicating a trend towards a more relaxed and heterogeneous eating pattern. However, as already
discussed in chapter 3, this is also a methodological issue, as questions about
habits (what you “usually” do) generally produce higher figures than questions like ours, about yesterday’s eating. The fact that all figures are lower,
including the total number of meals and snacks, supports this interpretation.
Concerning the timing of lunch and dinner, on the other hand, the tendencies
are less uniform. While the time of the midday meal is similar to earlier stud72
The problem of recording very small events is discussed in chapter 1 and in section 9.4 in
this chapter.
Kjærnes: Discussion and Conclusions
259
ies, dinner seems to start somewhat later in our study (not in Finland), perhaps
indicating a changing pattern. A third aspect of the meal pattern is whether
lunch is hot or cold. All studies confirm the pattern of hot lunches in Finland
and Sweden, cold lunches in Denmark and Norway. However, according to
Prättälä (2000), there seems to be a higher proportion of hot lunches in Finland than in Sweden, while they are quite similar in our study, rather like
earlier Swedish studies. This may indicate a shift in the Finnish eating pattern,
but not in the Swedish.
Contrary to common belief, people in the youngest age groups had as many
(Finland, Norway) or fewer (Denmark, Sweden) eating events, compared to
the older groups. There was a slight tendency for the middle age groups to
have higher averages, while the eldest again had eaten less often. These shifting patterns between age groups may point towards a life phase phenomenon.
The age effect was evident when looking at cold events in particular (not significant in Norway). For hot meals the tendency was opposite in Sweden:
younger people had more hot meals. When hot meals at lunch-time were analysed separately, however, there were few age differences when other factors
were controlled for. Concerning dinner-time hot meals, young and middleaged people more often had a hot meal in Finland. In Denmark, more elderly
respondents had their one hot meal at midday, rather than in the evening. According to our data, the smaller eating events occurred mainly in the middle
and older age groups.
The impact of life phase was also evident when looking at household type,
where couples with children had higher averages in Finland and Norway. This
group had more hot meals, especially when compared to people living alone
(not significant in Denmark). Employment status, on the other hand, had far
less impact. The exception was in Sweden, where the occupationally active
had more hot meals. This is probably related to the institutionalised hot lunch
system in Sweden. These findings are in accordance with the assumption that
variations in eating practices may be connected to differences in everyday
living (Rotenberg 1981). However, contrary to Rotenberg, we found that in
the privatised Nordic eating systems family status had considerably stronger
influence on the number and types of eating events than work had. The primary impact of work was on time schedules.
Gender may, as already indicated, signify different everyday lives as well as
different conventions with regard to food. Our findings show that the effect of
gender was highly significant in Finland, Norway and Sweden: women had
more cold eating events, while men had more hot events. This might be related to dissimilar practices at lunchtime, due to differences in occupational
260
Eating Patterns
activity. The hypothesis fits for Sweden, where the gender difference in hot
lunch eating disappeared in the regression analysis, while occupational status
was significant. In Norway, women had more cold eating events, while the
number of hot events was equal between the genders (no hot midday meal).
There were no gender differences in Denmark. A particular analysis of hot
dinner eating confirmed that this was a social event within the family, with no
difference between the genders when family status was controlled for.
Socio-economic factors such as education, income and social status were,
with very few exceptions, insignificant as indicators of variations in meal patterns. Variation according to social background is limited and to a minor degree affected by such “indirect” factors, in contrast to the direct influence of
work and family. The persistent contrasts between social groups that have
been identified for food habits (see for example Warde 1997, Roos and Prättälä 1999) seem not to be reflected in eating patterns and rhythms in our data.
Social status may, however, indicate work conditions where eating is organised differently among industry workers, service workers, executives, etc.
With few exceptions, we do not find indications of such correlations.
It has been suggested that urbanisation processes have led to a reduction in
the number of meals, and less hot meals in particular (Prättälä and Helminen
1990). In this study, we used the size of the municipality as an indicator of the
degree of urbanisation. We found no association between the number of hot
eating events and the size of the municipality. On the other hand, in Finland
there were regional differences: respondents from Northern Finland had more
hot meals than respondents from more southern parts of the country.
To sum up, our findings reveal eating rhythms that reflect both some common
patterns at the national level with regard to the number, timing and types of
meals and other eating events, and considerable variation individually and
between social groups. This variation is, moreover, quite dissimilar between
the countries and therefore difficult to condense. The variation is associated
most clearly with the factors emerging as important for most food questions:
age and gender. Age accounts for most of the variation in eating rhythms in
Denmark. In Finland and Norway, on the other hand, gender seems to be more
significant. Both gender and age are significant in Sweden. Factors that determine our everyday time schedules have some relevance, in particular the
family situation (not in Denmark), while work related indicators seem to be of
less importance (influencing hot meals in Sweden).
Kjærnes: Discussion and Conclusions
261
9.2.2 Homogeneity in the contents of eating events
Variation in eating habits has been discussed in relation to the choice of foods
and dishes (Mennell 1985, Warde 1997). Food habits that are complex lend
more room for variation (whether systematic or random). In chapter 2, we
found that the Nordic everyday eating patterns included a relatively limited
and quite homogeneous selection of foods and beverages, with several distinguishing characteristics between the countries.
The potential range of foods and beverages that we can buy, cook and eat or
drink is extremely wide, and we must expect all kinds of combinations in a
representative study of eating. This was also confirmed in our study, where
there was no single standard set of foods or combinations of foods and beverages. However, according to the findings presented in chapter 2, the range of
foods and beverages taken at breakfast and cold eating events later in the day
was not very large on an everyday basis. The limited range holds for both the
number of items at each individual event and the variation within each national sample. Variety may still be large at the level of labels and flavours.
But this was not reflected when looking at types of foods and dishes - as we
do in this study. The questionnaire included long lists of foods and beverages,
of which several options were hardly selected at all73. For instance, coffee was
the most popular drink at cold eating events, but varieties like espresso and
cappuccino, so frequently spoken of, were rarely reported by the respondents.
The sales figures for these varieties have increased considerably, but our data
indicate that their relative position compared to “ordinary” coffee is small in
the population at large.
The number of different dishes and components included in hot meals was
considerable, with no possibility of identifying a single pattern shared by a
substantial proportion of the respondents. Yet, it has been possible to classify
most of the meals and meal components according to a common classification
scheme. The basic format designated in the questionnaire seemed to fit in
most cases. A main ingredient of meat or fish accompanied by a staple and
often also a vegetable side dish seemed to represent the typical Nordic hot
meal. About one in ten hot meals included only one element (pizza, soup, a
casserole including several ingredients etc.). A minor proportion of the hot
main courses included more than three elements, the highest frequency appearing in Sweden. Homogeneity can also be seen in the use of vegetables,
73
This finding is to some degree a reflection of our methodology, where we identify the most
common features of everyday eating rather than distinctive patterns of specific groups or culturally important foods and dishes that appear on the menu infrequently and in more ritualised
settings.
262
Eating Patterns
where the variety in each country was quite limited. In Norway, which represents an extreme, about one third of all hot meals included carrots. Few hot
meals included more than one course. Starters of any kind appeared very
rarely, while one fourth of the weekday hot meals included a dessert, mainly
fruit or ice-cream.
Although still allowing for numerous combinations, this limited range of
types of foods and beverages appearing on the Nordic everyday menu must be
expected to curb the heterogeneity of eating. Experimenting and elaborate
practices may very well occur, but they are not a typical feature. Even though
variation may be great in a historical perspective, these findings do not seem
to support Mennell’s thesis of increasing variety74 (Mennell 1985) in terms of
types of foods and dishes (again in the everyday setting), perhaps rather the
mass consumption of industrial societies suggested by Ritzer (1993). However, the Nordic countries were not quite identical in this respect. Swedish
respondents were repeatedly found to have more complex eating habits (more
varied cold eating events, more elements of hot meals, a larger variety of
beverages, etc.). For the most part Finland and Norway had less variation
along these dimensions.
9.2.3 The social context of eating
Different social contexts are dominated by different sorts of people. Warde
observes: “If eating is the sum effect of many situated events, the sociologically appropriate question is whether there is a social logic to the situations
in which people find themselves.” (Warde 1997) As a consequence, neither
items of consumption nor situations should be observed separately. Rather,
the focus should be on the sequence of situations and bundles of items. By
doing so, Warde points to the role of the organisation and institutionalisation
of eating as part of everyday life. The organisation of eating depends on factors like the job and family situation, as well as by social forces associated
with events, like the frequency of attendance at celebrations or the habit of
taking Sunday lunch. We can add the role of how the provision of food and
meals is organised: through shops for individual purchase, market-based restaurants and fast-food outlets, or public catering (Fine and Leopold 1993,
Warde 1992).
The discussion about the structuration of eating has included several elements
related to the social context, with regard to both the social company at meals
74
His parallel thesis of diminishing contrasts will be discussed below.
Kjærnes: Discussion and Conclusions
263
and where eating takes place. There are two main underlying dimensions, first
social vs. individual meals, second public vs. private meals. Several authors
have included social company in the classification of the meal system. Eating
taking place alone is, for some authors, by definition a snack (Prättälä and
Helminen 1990, Rotenberg 1981). Simmel (1957) saw the meal as a socially
shared event. Shared meals are “more stylised, aesthetic and supraindividually regulated”. An increasing number of events of eating alone might signify
a process of individualisation – and assumptions about less patterned ways of
eating. Worries about de-structuration also focus on eating as a separate event
vs. something you do as part of or in between other activities. The question of
what goes on while eating may point to the degree of focus on the meal as a
social event in itself.
The second aspect was the location of eating. Rotenberg (1981) has suggested
that the social context of eating changes according to transformations of society, work and family life, leading to shifts between home, work and other
places, like cafés. The location influences the social company. While the
home is associated mainly with meeting the family at the dining table, and
eating at work assembles colleagues, eating out in cafés and restaurants may
be socially more open. Eating outside these established spheres, like at a fast
food outlet, walking about or in bed, may be regarded as “disorderly” or “less
structured”. However, as pointed out by Fine (2000), the location is indicative
of distribution systems and the social division of labour with food (see also
section 1.2.2).
Holm (chapter 5) found that the eating system included a mixture of events
taking place alone, with other household members, with colleagues and with
friends. A considerable proportion of breakfast eating took place alone, even
in non-single person households. Lunch, however, most often took place with
family or colleagues. For those who lived with a family, dinner was usually
shared with them. This was the case also for the cold eating event in the evening. Eating with friends and others constituted a minor proportion of everyday eating events at all hours. While individual eating took place throughout
the day, the peak meal hours were, to a high degree, social events. Moreover,
on Sundays, when meals are usually not structured by working hours, family
eating dominated, even for a considerable proportion of those who lived
alone. While about one third had had no eating events with family members,
this proportion was reduced to about ten per cent in households including at
least two adults.
Considering the association between meals and family life, the division between private and public eating is important in debates about contemporary
264
Eating Patterns
eating habits. In this study, the majority of eating events took place at home,
eating at the workplace coming second. Eating in other people’s homes or at
cafés/restaurants constituted only minor proportions. Chapter 5 states that
very few studies can give comparative information about other countries. The
home-centred Nordic eating habits seem, however, to be very different from
the American style, where far less eating takes place in the home.
‘Eating in passing’ has been suggested as a term to describe modern eating
habits in Denmark, indicating a tendency where people grab something to eat
in between or while carrying out other activities (Andersen 1997). The questionnaire allowed detailed records of such types of eating, operationalised as
eating at a fast-food outlet, in the street etc. and, in the home, eating in other
places than the kitchen or at a dinner table. We found that such less structured
(or less morally acceptable) eating places outside the home constituted a negligible proportion. There was, however, considerable variation concerning
places of eating at home. The most frequent alternative to the dinner table was
a coffee table/sofa (including also some events other than coffee or snacks).
Television sets are usually placed within view from the sofa in the Nordic
countries, and evening meals and snacks seem to be brought to the coffee table. This was confirmed when inquiring what went on while eating, where
about one fifth of the events took place while watching TV. Very little eating
took place at other sites (at a work desk, in bed etc.) or without sitting down at
all.
‘Eating in passing’ also indicates that there is no clear beginning or end to the
eating event, and it may be very short. But a very long duration may also indicate a less structured situation. As was the case with the social company, the
duration of the eating events varied considerably, probably indicating that
events have very variable status. Most people had had several events of medium duration, but also one or two events on the day before that were very
brief or went on for quite some time.
The two main dimensions of the social context, place and company, have been
combined, showing that while social events at home dominated in Denmark,
there were slightly more individual eating events at home in Norway and
Sweden. Third in all countries were social events outside the home, while individual events outside the home constituted 3-5 per cent (eating alone at
work, in a café or a fast-food outlet). Social eating events lasted considerably
longer than the individual ones, particularly compared to individual events
outside the home. Social events at home also took place more often without
other things going on simultaneously.
Kjærnes: Discussion and Conclusions
265
There was a linear relationship with age with regard to eating at home in
Denmark, Norway and Sweden (Finland excluded). Younger people ate at
home less often, elderly people most often. Older people also ate alone more
often. Eating with family members, something that required both being at
home and having a family, was most frequent in the middle age groups. Taken
together, these relationships indicate – again – that age first of all implies life
phase.
Yet, other factors influencing everyday life were also important. The occupationally active had fewer eating events at home and less alone. In Denmark
this group also ate more often with family members. Women, being less
occupationally active, had more eating events at home, but there was no gender difference with regard to eating alone.
Probably because they are more occupationally active, people with higher
education had less eating events at home. These effects were significant only
in Denmark, with more eating events at home and with family members. The
degree of urbanisation did not have any clear impact.
Taken together, we found that most people’s eating practices included organised social events as well as less organised, individual events. The social context of eating was influenced first of all by the organisation of everyday life,
by age, family and employment, while more indirect factors played a minor
role. The major factor indicating a process of individualisation, with more
eating alone, was whether the respondent had a family or not. Eating outside
the home was mainly associated with work and thus the respondent’s employment status. Public, individual eating events seem to have a marginal
status in Nordic daily eating patterns.
9.2.4 Contemporary Nordic eating systems – uniformity and/or individual strictness?
The presentation shows that the Nordic eating systems were varied and multifacetted with regard to meal types, eating rhythms, the place and social context (see section 1.4.1). Yet, some quite clear patterns of eating could be deducted, concerning the when and what of eating and its social context.
When? The time schedules of eating events show that a significant proportion
of the respondents in each country had eaten something at all times of the day,
from very early morning till late at night. Yet, eating as a social activity is
reflected in clear peaks at conventional eating hours, thus showing a certain
266
Eating Patterns
degree of uniformity. The number of eating events also varied, but the majority had between three and five events, pointing towards some individual
strictness in eating patterns. Grazing would imply many (probably small, as
the amount of food that can be consumed is limited) events spread arbitrarily
throughout the day. It has been difficult to identify individuals characterised
by that kind of eating pattern.
What? It has been possible to identify certain formats that predominated at
the various hours of the day. A large majority had eaten something in the
morning, with a limited selection of foods and beverages. The next main event
was lunch-time eating. We found mainly two types of meals: solely cold food
(mostly sandwiches) in Denmark and Norway, and a considerable proportion
of hot food events in Finland and Sweden. Most respondents had a hot meal in
the afternoon/evening. However, although usually quite simple in terms of the
number of components, the format of hot eating events varied considerably.
Minor events, or snacking, seemed to constitute additions to the pattern of
larger meals, rather than substitutions. The patterns were evident at the country level, and very few deviated considerably. This picture of structured elements combined with more leniency could be identified also at the individual
level. Most respondents had had at least a couple of conventional meals as
well as less structured snacks.
Where and with whom? The overall finding about social context is that the
home is the dominant place of eating in all four countries and for all social
groups. The majority of conventional meals (breakfast, lunch, dinner) did not
seem to have been very lax with regard to the social context – taking place at
home or in a canteen, at a dining table, with medium duration, and when feasible with social company. People living alone had kept a considerable proportion of these structuring features. Yet, most respondents had one or several
individual eating events, and these tended to be less strict in terms of duration
and what went on while eating.
Referring to our analytical model (paragraph 1.4.2), we wanted to identify
social factors that influenced eating patterns. These factors were grouped into
two categories: organisational and cultural background, where the country of
residence could influence both. The organisational background was meant to
include factors with direct impact on everyday life, in particular family and
work, while the cultural background included factors that might influence eating patterns more indirectly, through resources, norms and expectations (gender, generation, social status, education).
Kjærnes: Discussion and Conclusions
267
Eating patterns seem to represent a very flexible system. At all hours and in
all situations there were quite a few cases that were not in accordance with the
general code of practice. Yet, it has been difficult to identify social groups
with particularly ‘destructured’ food habits in term of contents, rhythm and/or
social context. There were differences between social groups, where age was
the dominant factor, and in many instances also gender. The organisational
background had some influence, in particular the type of household. Individualised eating seemed first of all to express the living situation, i.e., living
alone. Employment had direct (and perhaps rather trivial) impact on eating
hours and the place of eating. Status variables had in general very little explanatory power for variation in eating patterns.
We are thus left with the two factors that dominate in almost all studies of
eating: age and gender, each of which may contain both organisational and
cultural aspects. Age seemed in many instances to indicate life phase (thus
practical, everyday considerations), where increasing age indicated more uniformity in terms of hours and meal types, but with old age again representing
a new stage where this adherence to general conventions was supplanted by
new, more relaxed habits. It was, on the other hand, difficult to find indications of a general tendency of less individual strictness among the younger
generations. More time spent away from home was not reflected in more eating in public places. However, not all age variations could be interpreted in
this way. Meal formats, in particular, showed differences that might indicate
generational changes.
Gender differences were evident in several, but not all analyses. This was
linked to differences in everyday life, for example more women eating at
home because they spend somewhat more time at home. As a consequence,
women to a lesser degree took part in the institutionalised hot lunches in
Finland and Sweden. However, many women’s preference for cold food and
men’s for hot food could not only be explained by such direct factors. Traditional social events where women meet over a cup of coffee and a piece of
cake must be considered. We also know from other studies that women, when
not eating together with men, often prefer something lighter than meat dishes,
for example a salad. Men, on the other hand, more often want a “substantial”
cooked meal. This will be discussed further in the next section.
Eating Patterns
268
9.3
Different dimensions of ‘proper meals’
In discussions about the status of meals, ‘proper meals’ have been an important concept. The notion seems to be an integrated part of the everyday conceptualisation and classification of eating, representing a more or less explicit
expression of what is acceptable as a meal. In Norwegian, expressions like
‘ordentlige måltider’ and ‘skikkelig mat’ – meaning ‘square meals’ and
‘proper food’ – appear frequently in qualitative interviews (Bugge 1995,
Guzman et al. 2000), similarly in Denmark (Iversen and Holm 1999), in Sweden (Ekström 1990) and in Finland – “kunnon ateria” (Mäkelä 2000).
This everyday notion was introduced as an analytical concept by Murcott,
based on the British understanding of ‘proper food’ (see chapter 4). The British descriptions of a proper meal (and even of meal formats in general) concentrate on hot meals. However, this study has shown that conventions may
be very strong even in cases where the food eaten is cold, as in the case of the
Danish and Norwegian “mat/d-pakke”, i.e., open-faced sandwiches brought
from home for lunch (see also Døving, in press). This lunch meal is highly
institutionalised as part of everyday life, as demonstrated in chapter 3. In the
Swedish and Finnish discourse, however, it is hot meals that count (of course
hot meals of a certain format), while taking a sandwich is regarded as substituting a potentially proper meal for a snack, expressed as a fear of ‘breakfastisation’ (‘frukostisering’).
Yet, a more general multi-facetted discussion is allowed with regard to
cooked, hot meals, and our analyses of proper meals have concentrated on hot
meals. Conceptions of a ‘proper meal’ seem to contain several dimensions:
• As indicated by Murcott, a ‘proper meal’ should contain certain foods and
dishes. It is a question, nevertheless, whether this food should indicate ordinary, weekday practices, like the Norwegian fish, or weekend events,
i.e. something extra, like the British Sunday roast. We can call that the
food dimension.
• The social context of the meal has also been emphasised in Nordic studies. A ‘proper meal’ has to be eaten together with the family, even as a
kind of material proof of the existence of a (proper) family – thus there is
a family dimension. It has been suggested that the couple is the essential
element (Charles and Kerr 1988, Murcott 1983), while others emphasise
the presence of children (Wandel et al. 1995).
• A ‘proper meal’ is not only a matter of food and eating, but also how the
food has been prepared and who prepared it. A cooking dimension points
to the role of preparing the food ‘properly’ – i.e., at home from raw and
Kjærnes: Discussion and Conclusions
•
269
fresh ingredients – often expected to be done by the mother of the family
(Ekström 1990, Murcott 1983).
Then there is a health dimension. The foods included and the preparation
of a ‘proper meal’ are also supposed to be healthy. And when there is no
family, there is no ‘proper meal’ and hence no healthy meal either
(Guzman et al. 2000). But there are tensions within the health issue. Is
health (understood as nutrition) decisive, or can a ‘proper meal’ be somewhat extravagant, for example including a rich sauce that your mother
used to make?
From an analytical point of view, one can expect that the balance between
these dimensions may vary in the national food discourses and in everyday
life. There may also be tensions and tradeoffs between these dimensions or
concerns. Analyses of ‘proper meals’ within the Nordic setting have been carried out in our study in order to get a better grasp of meal conventions – in
addition to the analyses of schedules and social context that have no explicit
normative reference. We have no information about norms and expectations,
but aspects of shared practices that have been discussed in the ‘proper meal’
debate can be of relevance.
These analyses do not represent an evaluation or ranking of the eating patterns
that we find, merely the degree of adherence to established conventions. It is
difficult to operationalise such conventions (which may vary between countries and social groups and also change with time) in a quantitative study. This
is particularly relevant when describing the contents and formats of meals.
The solution (which is not optimal) has been to apply the formats described in
the (British) literature, modified according to Nordic studies. The findings
from various chapters of this report will be used to explore three of these dimensions, their relative importance in the various Nordic countries and in different social groups. The health dimension requires other types of data than
those included in this study and is omitted from the following discussion.
9.3.1 The food dimension
The existence of proper hot meals has been discussed in this report on the basis of the meal format as described within the British tradition, a tradition that
seems to hold even for the Nordic food cultures (Ekström 1990, Mäkelä
1996). The analyses in chapter 4 confirm that this ‘proper meal’ format (as it
has been defined here) does exist as a significant phenomenon in all four Nordic countries, but it does not constitute any clear directive for everyday eating.
Using this rather broad definition of the ‘proper’ hot meal format (centre +
270
Eating Patterns
staple + vegetable as a minimum), about 40 per cent of the hot meals fall into
that category. There was, however, considerable variation between the countries. Looking at the different time intervals (‘lunchtime’ and ‘dinnertime’, –
respectively), the proportions of hot meals that were ‘proper’ were lower in
Finland and Denmark, compared to Sweden and Norway. Regarding the day
as a whole, however, the proportion of respondents who had at least one
proper hot meal was similar in Finland to that in Norway, with Swedish figures being higher and leaving the Danes at the lower end.
This analysis is built on a structural conceptualisation of the food and the
meal format, and it does not imply any normative or health-related evaluation
of the meals. Among the hot meals that were not categorised as ‘proper
meals’, there may very well be meals that in the Nordic everyday languages
would be regarded as ‘proper’, such as a mutton casserole with cabbage
(‘fårikål’) served with potatoes (a centre – including a vegetable – plus a staple), and also dishes that are perhaps not traditionally ‘proper’ but that are
nutritious all the same, such as a pasta dish with a vegetable sauce.
Among the meals with only one or two components, a considerable proportion
had a centre in the ‘other’ category. As mentioned before, the list of ‘other’
centres includes traditional as well as modern dishes (i.e., pizza and soup).
Moreover, a meat centre with only additional bread, which could indicate a
hamburger or hotdog with bread, constitutes only a minor proportion in all
countries except Finland, where bread is a traditional part of hot meals. Thus,
it is not possible to discern whether the identified simplicity of Nordic hot
meals is old or new, as we have little basis for comparison. Neither can the
Swedish tendency towards somewhat higher complexity and more ‘proper
meals’ be related towards either traditionalism or more recently introduced
habits by analysing the food content alone.
The three-partite structure – with three courses - indicated by Douglas is
rarely found in Nordic meal practices. Desserts seem to have been more
common a couple of decades ago (Prättälä 2000), perhaps also entrées in the
form of soups. However, Murcott’s idea of “a plateful” seems to be quite
relevant. Contemporary Nordic meal norms and the discourse around them are
concerned merely with the main course, not with any multi-course ideal.
9.3.2 The family dimension
The family meal has been explored in depth in chapter 6. To bring the family
together at a meal is regarded as an indication, or even an accomplishment, of
Kjærnes: Discussion and Conclusions
271
the family as such. In our study, a family meal was operationalised as a hot
meal eaten (by one who does not live alone) at home, with all household
members present. It was found that about six out of ten respondents not living
alone had had a family meal on the day before, the highest figure appearing in
Norway, the lowest in Finland. There were differences between the days of
the week, with slightly more family meals on Sundays, except in Denmark
where figures were equally high on weekdays. The findings indicate that the
family meal still exists as a dominant phenomenon in all the Nordic countries,
even on weekdays, but it is not a very strict convention or practice.
Let us now introduce an additional analysis of the data, combining the findings about ‘proper meal’ formats and family meals. The question is whether
there is any association between these two dimensions of ‘proper meals’, as
they have been defined here.
60
49
50
Per cent
40
43
42
40
35
35
55
53
50
47
50
38
30
20
10
0
De nmark
Living alone
Finland
Family, not all
Norway
Swe de n
All family
Figure 9-1 The proportions of ‘proper meals’ in different social contexts. Per cent.
Figure 9-1 shows the proportion of ‘proper meals’ among those living alone,
those where the hot meal did not include the whole family, and those where
all family members have eaten together, respectively. The overall tendency
seems to be that ‘all-family meals’ are more often ‘proper’, whereas single
person households have slightly less ‘proper meals’. However, the differences
are small and not quite consistent. Those who had a family had somewhat
more proper dinners in Denmark and Finland. But the differences were not
great, and in Sweden there was nearly no difference. Somewhat surprisingly,
the tendency was opposite in Norway.
272
Eating Patterns
Thus our findings do not indicate any clear-cut association between these two
dimensions. One reason could be the variable nature of everyday eating,
where we have recorded several types of events. Another explanation may be
that our operationalisation of ‘proper meals’ does not capture the phenomenon
very well. However, a third explanation is of course that there is no association, that ‘proper’ family meals do not imply particular meal formats, as opposed to other social contexts. This is particularly relevant for Denmark,
where family meals were more prominent than in the other countries, while
they had the lowest proportion of ‘proper meals’.
9.3.3 The cooking dimension
As indicated above, the cooking dimension includes the act of cooking per se
as well as the person(s) or institution that carry out the cooking. It is difficult
to assess the extent of (home) cooking from our data. We included a question
about whether the main dish had been wholly or partially home cooked or if it
was readymade. The results show that two thirds of the hot meals eaten at
home were home cooked (data not shown). However, what did the respondents understand by a homemade dish? Was heating sausages or a pizza regarded as cooking? The interpretation is so difficult that we have not analysed
this question further. The presentation of main ingredients of hot meals can,
however, provide some information. A considerable proportion of the meat
was served as meatballs, sausages etc. Such dishes are very often readymade,
meaning that you only heat them at home. In Norway, this was the case also
for fish dishes, where readymade fish cakes and fish pudding are very popular
dishes. The major preparation method was pan-frying. Pan-frying is simple
and quick for small portions, but less so than the use of microwave oven.
Moreover, complex vegetable stews or sauces as side dishes were not used
very often.
Taken together, our data indicate that cooking typically implied:
• Pan-frying the main ingredient: pork chops, meat balls, sausages, fish
cakes or a fish filet,
• Boiling potatoes (or rice/pasta),
• Perhaps adding boiled carrots or some tomatoes.
The outcome is thus a meal that would count as ‘proper’, but where the cooking was quite simple. In some cases cooking was limited to the heating of a
Kjærnes: Discussion and Conclusions
273
frozen pizza in the oven75. The answers were varied, however, and in many
cases considerably more time and effort was probably involved.
Chapter 7 has analysed the question of who did the cooking. Our study confirms that cooking is still clearly gender divided. If a ‘proper meal’ is perceived as a meal cooked by the mother in the family, this was very often the
case in all the Nordic countries. Men had done some of the cooking, while
major contributions from other persons (children, a maid etc.) were rare. We
must add that this question was of course only relevant in non-single households. Most men as well as women living alone did have dinner at home – and
thus did some cooking.
However, the findings emphasise the uncertainty about what cooking is, as
there was no direct agreement in the Nordic data sets between the genders
concerning who did the cooking. There was considerable overlap. In chapter 7
this is discussed as a question of different perceptions between men and
women of what cooking is and what the main operations are. Food preparation is a many-dimensional task where men and women may have different
experiences and expectations. This is clearly demonstrated when the question
about yesterday’s cooking was combined with one about interest in cooking.
Among the male respondents that had cooked, there were more who were interested in cooking, compared to those who had not cooked, while female respondents had cooked whether they were interested or not.
Some methodological considerations should also be mentioned. First, in questions where respondents are asked to evaluate their own (positive) contribution there is a general tendency towards overestimation. We must expect this
effect to be present in our study as well. Moreover, as a considerable proportion of the contacted men were not willing to participate in the present study
(more than among women), those men who did comply may be particularly
interested in food and cooking and in that respect may not be quite representative for the male population in general (see also chapter 1).
9.3.4 ‘Proper meals’ and social differences
We found that the various ‘proper meal’ dimensions varied considerably in
different population groups, the most significant being age and gender:
75
Takeaway food (pizza, Chinese food etc.) is an even more simple alternative with regard to
cooking. It is difficult to assess the extent of takeaway in our study, but other questions (who
did the cooking) as well as the reported frequency of relevant dishes indicate that this was a
marginal phenomenon.
274
Eating Patterns
1) factors that influence the organisation and structure of everyday life
(household type, work situation), have some impact on the various dimensions of ‘proper meal’ eating; likewise
2) demographic variables, age and gender, were significant, while
3) the influence of factors such as educational level, social status and
urbanisation (‘cultural background’) was much more limited.
Age was generally the most important variable in all these analyses:
• The youngest age groups had fewer meals with a ‘proper meal’ format
and fewer family meals, indicating more meals eaten outside the home
and more meals eaten alone.
• The relationship between family meals and age was strong and linear in
all four countries (we must remember that people living alone were excluded from this particular analysis).
• The same relationship pertained with regard to the number of meals eaten
at home and meals eaten alone.
• The number of meals eaten with family members related to age in a curvilinear way, with maximum values in the age group 35-44 years (this
analysis included the total sample).
• The ‘proper meal’ format was linked to age in an even more complex
way, with high values in this same age group (35-44 years) as well as
among the highest age group. The exception was here Norway, where
there was a more clear curvilinear relationship, with maximum values
among those aged 45-54 years.
Age variations may to some degree imply generational effects. The younger
generations have views and practices that differ from their elders, and they
will maintain these as they grow older. The findings in the four countries consistently confirm that age is an important variable with regard to ‘proper
meals’, but they indicate that this is no simple phenomenon. It is a complex
dynamic where both life phase and generation are involved: the whole ‘proper
meal’ concept implies that proper meals are more relevant in some life situations than in others. But younger generations may still handle this in a different way compared to the elder ones. Perhaps the different dimensions of a
‘proper meal’ may be valued and weighed differently? The definition of a
‘proper meal’, though important and persistent, may also change; there is no
reason to expect it to be static.
Variables indicating the family situation confirmed the importance of life
phase. When ‘proper meals’ are perceived as family meals, it is almost tautological that people living alone more often eat alone. (It should be emphasised
Kjærnes: Discussion and Conclusions
275
that this does not mean that those living alone always eat alone, even at dinner
time.) In households with two or more people, there was some variation between those that consisted of adults only and those with children (except in
Denmark), with the somewhat surprising finding that adults without children
tended to have slightly more meals with a ‘proper’ format than those with
children. The youngest generation (the children) may, after all, have some
influence on the menu, with slightly more pizzas, pancakes and pasta dishes,
dishes that often come without the elements required for a ‘proper’ hot meal.
But the additional analyses did not confirm that those household groups which
have more family meals also consistently have more ‘proper meals’ in terms
of food content, even though there was a slight tendency in that direction.
Lunch-time ‘proper meals’ (which could only be analysed in Finland and
Sweden) were not related to the family situation.
Daily life is influenced not only by age and family, but also by employment
status: whether you work outside the home, go to school or not. We did not
find this to be a significant factor with regard to variations in the eating of
dinners with a ‘proper’ format. There was a tendency for those who worked to
have more proper lunches in Finland, but this was not confirmed in Sweden.
With regard to family eating, however, the employment factor was significant,
as the non-active ate at home more often and family meals occurred more often in all four countries. However, this group also had relatively more meals
alone in Denmark and Norway.
Household type and employment affect daily life directly. Age may, as discussed above, have both direct and indirect impacts. Gender is similarly complex, affecting roles in the family and the labour force as well as socialisation,
knowledge and conventions about food and cooking. In all social scientific
studies on food, cooking and eating, gender emerges as a major factor (chapter 7, O'Doherty, Jensen and Holm 1997). The convention is that women
should cook the ‘proper meals’, but what about eating them? More women
than men had ‘proper’ dinners in Denmark and Finland, but not in Norway
and Sweden. Invariably, women had more meals at home, but they did not eat
alone more or less often, and family meals showed no gender difference.
There are probably two factors behind this. First, the cultural aspect of women
taking more responsibility for proper and healthy eating in the family can explain that women also tend to eat more ‘proper meals’ – and more often at
home. On the other hand, ‘proper meals’ are supposed to be eaten together
with the family, thus of course indicating that both genders share the same
type of meal. The somewhat confusing outcome may be explained by such
counteracting forces.
276
Eating Patterns
We have included education and social status in the cultural dimension of
eating. There might very well be class and status distinctions in the role of
‘proper meals’, ‘proper meals’ being mainly a characteristic of ‘proper’ middle class living or – the other way around – a phenomenon in a working class
sticking more to collective conventions compared to a middle class concerned
with individual freedom. A general finding in this study was the lack of any
clear correlation. For a number of variables social status has a strong impact
in Finland, but the concrete variation for this variable is difficult to interpret.
The relation to education in Sweden may similarly be part of a larger picture
showing the role of education. Social status made a difference for ‘proper’
dinners only in Finland (middle class respondents having fewer proper dinners), education only in Sweden (respondents with university level education
having more proper meals). When it came to the social context, however,
there were clear distinctions. Those with higher education had fewer eating
events at home, but less events alone. Family meals appeared more often
among those with lower education in Denmark, with even stronger distinctions in Sweden, not in Finland or Norway.
The degree of urbanisation may imply somewhat differing daily lives, but
with regard to ‘proper meals’, it is more reasonable to treat urbanisation as a
cultural factor, representing degrees of modernisation. Urbanisation had no
impact on the meal format, while there were variations with regard to the social context in Denmark and Finland, but not in Norway and Sweden. The
Danish respondents living in big cities had somewhat fewer eating events at
home and clearly less events together with family members – but not significantly less family meals. In Finland, the difference in family meals was considerable between urban and rural areas.
9.3.5 ‘Proper meals’: ideals or practices?
From qualitative studies we know that ‘proper meals’ are important as a convention and an ideal. We have found that ‘proper meals’ clearly exist in the
Nordic food cultures, but not as a predominant everyday practice. While meal
formats are often simple (this may not be a recent phenomenon), families eat
together at home and mothers do the cooking. However, this is not always
possible or relevant. The occurrence of ‘proper meals’ seems to be influenced
by living situation and life phase, as indicated by age and household type.
The association with age is first of all related to the format, rather than the
other dimensions. While the ‘proper meal’ convention links the various di-
Kjærnes: Discussion and Conclusions
277
mensions of ‘proper meals’ closely together (food, cooking and context), this
is not as evident in everyday practices. Even having pizza while sitting on the
sofa watching TV may take place together with family members and at dinner
time. And those living alone have meals with a ‘proper’ format almost as often as those living in a family situation.
Comparing the countries, the differences were not very great, thus confirming
that ‘proper meals’ represent a rather uniform convention in these countries.
Yet, the various dimensions seemed to be emphasised somewhat differently.
The ‘proper meal’ format appeared most frequently in Sweden and Norway,
least so in Denmark. The somewhat ‘disciplined’ or moralistic definition of
the ‘proper meal’ (for example including vegetables as a necessary component, but not a sauce) seems to have more success in Norway and Sweden, in
particular compared to Denmark. Family meals, on the other hand, were most
frequent in Denmark, most rare in Finland. Concerning cooking, the gender
division of labour was least pronounced, but also most disputed, in Norway.
As food habits do change, albeit slowly, it must be expected that norms and
ideals about ‘proper meals’ also change, particularly with regard to the contents and format of the meal. The vegetable component may, for instance, be a
rather recent addition. There seem to be few Nordic studies that can give insights into how ‘proper meal’ ideals and practices change over time. It has
been suggested that British ideals as well as practices concerning the contents
of hot meals are shifting (Kemmer et al. 1998, Marshall 2000, see also chapter
4). A recent Norwegian study of single young people did not show similar
trends concerning changing ideals. In this group, however, proper meal ideals
were closely associated with the couple situation, and perceived as not relevant for their practices – yet (Guzman et al. 2000). Our own findings related
to age may indicate that changes are taking place in terms of norms and/or
practices, in particular with regard to the meal format.
9.4
What is a meal?
Studies of eating systems have concentrated on meals and hot meals in particular (Wood 1995). This is also evident in discussions of the degree of structuration of meals. Hence the operationalisations of a meal and a snack are important. Central in our choice of methodological approach was that we wanted
to record all types of eating and then categorise them afterwards. The findings
confirm that several types of hot and cold eating events could be identified.
Two matters of interest will be discussed here. First, the distinction between a
278
Eating Patterns
meal and a snack will be problematised on the basis of our findings. Second,
the recording of very small eating events will be discussed, a question that
touches upon our methodological approach.
Meal or snack? According to Douglas and Nicod (1974), four terms describe
the different forms of eating. A food event is an occasion when food is eaten, a
structured event is a social occasion organised by rules concerning time, place
and sequence of action. Food eaten as part of a structured event is a meal. A
meal observes the rules of combination and sequence. A snack is an unstructured food event without any rules of combination and sequence. This is simply a classificatory hierarchy, and the conventions and rules may vary. It reflects varying degrees of institutionalisation with regard to the contents – food
and cooking, the meal times and the focus on the eating activity.
The distinction between hot and cold eating events was made by the respondent. We have characterised all hot eating events as ‘hot meals’. It could be
that the hot food was only a hot dog eaten ‘in passing’. But our study has confirmed that such eating events constitute a very minor part of the hot eating
events (see chapters 2 and 4). The hot food (even if it was a hot dog or a hamburger) was usually consumed at home, with the family if there was one, most
commonly also at a dinner table. Thus, while the food did not in all cases satisfy rules of combination and sequence, the social context and timing did indicate a ‘meal’ according to Douglas’ analytical scheme.
What we have named ‘cold food eating events’ have the most ambiguous
status with regard to the distinction between snacks and meals. Most of these
events included an open sandwich and a beverage (hot or cold). Sandwiches
represented a larger proportion of all cold eating events in Denmark and Norway, compared to Finland and Sweden. As described above, there were clear
patterns in the number and periodicity of such events. Such patterns were particularly evident in Denmark and Norway, where certain types of combination
and sequence (although simple) also predominated. The eating of sandwiches
around noon was highly institutionalised in these countries in terms of the
contents of the meal, the hours, and the eating being the main activity. As
shown in chapter 5, this was largely also a social event. The sandwiches frequently eaten in the evening in Norway may be characterised in the same
manner, but the structure was less clear. The total number of such cold food
eating events varied between one and two a day (in addition to breakfast).
In Finland and Sweden, on the other hand, the patterns were far less obvious,
with regard to both the contents and the timing of ‘cold food eating events’.
Kjærnes: Discussion and Conclusions
279
We have therefore characterised the ‘cold food eating events’ as ‘meals’ in
Denmark and Norway, as ‘snacks’ in Finland and Sweden (see chapter 2).
Food or non-food? At the level of the contents and complexity of each event,
snacking can be represented by very small and simple events: events that require no preparation at all, food that can be eaten anywhere at any time, without a plate, knife, fork or spoon. It may be a biscuit, an apple or a chocolate
bar76. In our study, we have distinguished small snacks eaten alone (perhaps
with a beverage) from other types of cold eating events, and we found that
they constituted a minor part of the Nordic menus. Moreover, we found no
cases where such events dominated throughout the day. They constituted
more or less regular additions to the major meals in almost all cases.
There is, however, a question about the magnitude of this category. In a quantitative telephone survey it is impossible to record every single glass of water
or piece of chewing gum. It is likely that eating events recorded as ‘snacks’
are underreported. Even though written and oral instructions to interviewers
included a specified account of what should be recorded (chapter 1), the
distinction may still have been difficult in practice.
One observation may be of interest here. That is the particularly low frequency of ‘snack’ events in Norway compared to the other countries, which
also contributed to a lower total number of eating events. Our first interpretation was that this was a reflection of a culture of puritanism and simplicity
surrounding eating in Norway. Returning to table 1-5, chapter 1, however, we
found that according to gross statistics, the consumption of a number of foods
and beverages that are commonly used as snacks was systematically higher in
Norway (fruit, chocolate, salty snacks, soft drinks), not lower. One explanation might be that the Norwegian interviewers were particularly sloppy, but
this is not very likely. Hence we searched for other explanations.
In designing our study, we were concerned about avoiding meal concepts in
the questions in order to circumvent normative expectations about the various
types of meals. It was not possible, however, to omit words like ’eating’ and
’food’. But words like ‘food’ are also parts of a classification system (Murcott
1995, Douglas 1983), where food/non-food constitutes a major distinction. It
is likely that ‘eating’ has similar connotations.77 Døving has found that in
Norway, chocolate is not included in commonly used food or meal categories,
neither within the private sphere nor commercially (Døving, in press). Choco76
The term ‘snacking’ may, as discussed above, also include more substantial foods.
The respondents were asked “When did you eat…?”
77
Eating Patterns
280
late is not ‘food’, but ‘a temptation’, ‘a pleasure’ or ‘a kiosk item’. Salty
snacks and soft drinks are classified in a similar manner. It might be that to
the Norwegian respondents (and/or interviewers), this type of consumption
was not recalled or regarded as ’eating’ at all.
Similar discussions of the classification of foods and meals in the other Nordic countries are lacking, and we cannot rule out the existence of this phenomenon there as well. Our findings may indicate, however, that it is particularly evident in the Norwegian food culture, perhaps even as a tacit response
to or a reflection of a puritan meal culture that does not allow for the ‘pleasure’ of chocolates.
9.5
Method and generalisation
Several methodological choices followed from our overall research questions:
- representative surveys rather than qualitative in-depth interviews (in order
to be able to generalise to the populations at large)78,
- a record of yesterday’s eating rather than the more common food frequency method (in order to be able to study the order, sequence and context of eating events),
- a deconstruction of ‘eating events’ rather than questions about particular
meals and snacks (in order to avoid confusion between temporal shifts
and social variations in conceptions, on the one hand, and practices of eating, on the other)
- a comparative rather than a nation-based design (in order to reveal features of eating that may be self-evident within a national frame of reference).
The validity of yesterday’s eating
The choice of method was made in order to grasp the entirety of eating events,
and to record various aspects that appear as part of a concrete situation. A focus on one particular day makes that possible, while at the same time –
allowing for the inclusion of various types of events throughout the day and
throughout the week.
The validity and reliability of various ways of recording eating patterns have
been discussed at length in dietary research (Nelson 1997). Eating patterns are
complex and flexible, and they vary considerably from one day to the next.
78
Representativity has been discussed in section 1.5.6.
Kjærnes: Discussion and Conclusions
281
Therefore, a record of one day of eating (‘24 h recall’ in dietary research)
does not necessarily give a valid picture of more general patterns at the individual level. However, the method is relatively valid at a group level when it
comes to food consumption. This is particularly relevant for the present study,
which had no quantification of food intakes. Yet, when we compare the ranking of various foods consumed in the various countries in our study with gross
statistics from these countries (see chapter 1), there are clear deviations, as
already indicated above. The most evident example is the consumption of
sweets, salty snacks, soft drinks etc., where our study indicates that Norwegians consume less than people in the other Nordic countries, while the opposite is the case according to gross statistics.
However, it is relatively easy to remember what you had to eat yesterday,
compared to records of longer periods of time. Moreover, frequency questions
of the respondents’ usual eating patterns will entail even greater problems because the respondents have to make estimates that may involve not only recollection errors, but also ideas about what their diet should be like.
Social scientific studies of eating must address other methodological concerns
as well (see Hellevik 1991). In particular, conceptual validity has been a major concern in our study. The concept ‘eating event’ was introduced because
we wanted to avoid preconceived ideas about what a particular meal or meals
in general should involve. We wanted to be as specific as possible in recording the concrete acts of eating rather than anticipations about them. As
already mentioned, this task became particularly relevant within a comparative context, where the process of developing questions for the questionnaire
revealed that very little could be taken for granted.
The subsequent analytical process of reconstructing the various types of eating events as meals and snacks has produced a classification based on our
own theoretical framework. Meals are precisely defined in terms of foods and
periodicity. But we cannot say, for example, to what degree our figures about
the first eating event of the day correspond with Nordic people’s experience
of having had ‘breakfast’ on that particular day. Therefore, particular caution
should be observed in the use of findings on the occurrence of various types
of meals.
The comparative design
The comparative analyses including the Nordic countries have revealed that
although they are quite similar in terms of political history, culture etc., their
food habits are basically very different, even under the present influence of
282
Eating Patterns
modern welfare institutions and globalised markets, technologies and communication. Our idea was to study supposedly traditional national differences
as opposed to more general characteristics of modern societies. The findings
may, however, be interpreted as indicating that the Nordic eating systems
clearly have both traditional and modern elements, which are modified not
only by nation-specific cultures but also by institutions. The clearest example
is the organisation of lunch eaten outside the home, which is strongly institutionalised and includes hot meals in Finland and Sweden, but not in Denmark
and Norway. The emergence of these two systems is closely associated with
the organisation of work in industrial society. On the other hand, most women
are employed outside the home and thus cannot have dinner ready when male
family members come home from work. Still, dinner is usually eaten quite
soon after the end of a normal working day and the time and social organisation vary considerably between the countries.
9.6 Conclusions: contemporary eating patterns in a Nordic
comparative perspective
Our main aim has been to analyse the structure of eating as an integrated and
dynamic part of modern everyday life. Regarding contradictory assumptions
about the structuration of eating, the findings in this study indicate that most
people’s daily menu consists of several types of eating events in terms of
structure and strictness. The eating events included a considerable proportion
of meals that were highly structured in terms of meal format, which took place
at a dinner table and were not either very brief or very long. Those who lived
with a family also had one or two meals together with the rest of the family.
But most respondents also had events that were more casual, including perhaps a snack eaten alone in front of the television set. The proportion of the
respondents whose menu consisted only of such events was, however, negligible. Individual events tended to be more casual than eating events with others. But those who lived alone did not exhibit a clearly less structured eating
pattern.
The answer to the question of socially shared patterns of eating – variety vs.
uniformity – is not very clear. The existence of a shared pattern in all the Nordic countries was most clear for the first eating event of the day (breakfast).
National differences were mainly associated with cold vs. hot meals at lunch
hour. A third finding was that most people had a hot meal in the afternoon/evening. Apart from that, we found considerable variation concerning
eating hours, meal formats and social company. Least variation was found
with regard to the location of eating, where the home was definitely the most
Kjærnes: Discussion and Conclusions
283
common place of eating in all four countries and for all social groups, eating
at work coming second.
We have also studied the various dimensions of eating with regard to variation
between social groups. From our findings it is very difficult to state the degree
of influence of social structure and societal institutions, as the consistency of
the findings is low. The most general finding is that age has considerable impact. The eldest and the youngest age groups were in many cases quite different. This was most evident for the foods eaten. In questions that were related
to family meals the middle age groups were, perhaps not surprisingly, the
ones with highest frequencies. However, we could not find any tendency
where the eating pattern of the youngest age groups could be characterised in
terms of ‘grazing’ or the like.
Gender is a major factor in most studies of food and eating. We could identify
differences associated with differing living situations as well as the division of
work and responsibility for food. But, again, it has been difficult to find different degrees of structuration of eating based on gender.
As integrated aspects of everyday life, it is not surprising that employment
and household type have impact. Such factors seem, however, to have mainly
direct effects, deciding where and with whom you eat. With some exceptions,
they seem to indicate special life styles only to a minor degree.
The effect of social stratification variables was usually small and quite unsystematic. This was the case for educational level as well as class and income.
Country of residence is of relevance both for what we have called the ‘institutionalisation of eating’, i.e., various forms of food provision and food policy,
and through more general features of social structure and cultural norms. For
the factors we have studied, the general finding has been that country of residence has larger impact than the respondent’s social position and living situation within each country. The findings seem to indicate a complex interrelation between cultural conventions and institutionalised forms of food provision – through retailing as well as public catering. This was most evident for
the solutions to lunch away from home during work/school hours. Eating out
in restaurants, cafés and fast food outlets plays a marginal part in everyday
eating patterns in all four countries. Thus, the division between private and
public eating, being relevant primarily for eating at work, was influenced
mainly by nation-specific institutional solutions.
284
Eating Patterns
Finally, there are indications of quite consistent cultural differences, not only
in terms of the particular selections of foods and dishes:
• Danish eating patterns seem in many cases to be more family oriented.
• Differences between regions and between social groups are relatively larger in Finland.
• Inconsistencies and particular features of the Norwegian responses (i.e.,
more ‘proper meals’, underreporting of snacks, the gender division of
cooking) may indicate that the relation between norms and practices of
eating is more controversial.
• Swedish records show a general tendency towards somewhat more heterogeneity and complexity.
We cannot draw any definite conclusions on these particular features and how
they may be explained. They should rather be subject to further studies.
Literature
Ahrne G. and Roman C. 1997, Hemmet, barnen och makten. Förhandlingar om arbete och pengar i familjen. [The Home, Children and Power. Negotiations about Work
and Money in the Family]. SOU 1997:139, Fritzes, Stockholm.
Andersen J. 1997, Hverdagens centrifuge - Det daglige liv og den moderne livsform
[The centrifuge of everyday life - Modern daily living], Hovedland, Copenhagen.
Andersen N. L., Fagt S., Groth M. V., Hartokopp H. B., Møller A., Ovesen L., and
Warming D. L. 1996, Danskernes kostvaner 1995. Hovedresultater [Danish dietary
habits 1995. Main results], Levnedsmiddelstyrelsen, Søborg.
Askegaard S. and Madsen T.K. 1995, European food cultures: An exploratory analysis of food related preferences and behaviour in European regions, MAPP Working
Paper no 26, Århus.
Aymard M., Grignon C., and Sabban F. 1996, Food allocation of time and social
rhythms. Introduction. Food and Foodways 6 [3-4], 161-185.
Barthes R. 1979, Towards a psychosociology of contemporary food consumption, in
Food and drink in history, eds. R. Forster and O. Ranum, Johns Hopkins University
Press, Baltimore, pp. 166-173.
Beardsworth A. and Keil T. 1997, Sociology of the menu. An invitation to the study of
food and society, Routhledge, London, New York.
Beck-Gernsheim E. 1998, On the way to a post-familial family. From a community of
need to elective affinities, Theory, Culture and Society 15 [3-4]: 53-70.
Beck U. 1992, From industrial society to the risk society: questions of survival, social
structure and ecological enlightenment, Theory, Culture and Society 9: 97-123.
286
Eating Patterns
Beck U. 1994, The reinvention of politics: Towards a theory of reflexive modernization, in Reflexive modernization. Politics, tradition and aesthetics in the modern social order, eds. U. Beck, A. Giddens, and S. Lash, Polity Press, Cambridge, pp. 1-56.
Becker W., Enghardt H., and Robertson A.K. 1994, Kostundersökningar i Sverige
1950-1990 [Dietary surveys in Sweden 1950-1990], Livsmedelsverket, Uppsala.
Bell D. and Valentine G. 1997, Consuming geographies. We are where we eat,
Routledge, London, New York.
Berg L. 1997, Studieløpet - om tidsbrukvalg, faglige valg og kunnskapsteoretiske valg
[The study course – choice of time use, discipline and knowledge theory]. Rapport
3/97, NIFU, Oslo.
Berg L. 2000a, Tillit til mat i kugalskapens tid. En komparativ kartlegging, med fokus
på forbrukertillit og matsikkerhet i Norge, England og Belgia. [Trust in food in the
age of mad cow's disease], SIFO Report no.6, The National Institute for Consumer
Research, Lysaker.
Berg L. 2000b, Trust in food in Europe. Focus on consumer trust in Norway, England
and Belgium. Working Paper No 15, The National Institute for Consumer Research,
Lysaker.
Bjørkum E., Ekström M. P., Gronow J., Holm L., Kjærnes U., and Mäkelä J. 1998,
Utespising i Norden [Eating out in the Nordic countries]. Seminar on Nordic consumer Research, Lillehammer, Norway, 11.11.1998.
Bolin O. 1993, Agricultural policies in confrontation with public health interests, in
Regulating markets - regulating people. On food and nutrition policy, eds. U.
Kjærnes, L. Holm, M. Ekström, E.L. Fürst, and R. Prättälä, Novus Press, Oslo, pp.
65-76.
Bonke J. 1997, Dilemmaet arbejdsliv - familieliv i Norden. [The dilemma of working
life - family life in the Nordic countries], Institute for Social Research, Copenhagen.
Bourdieu P. 1986, Distinction. A social critique of the judgement of taste, Routledge
& Kegan Paul, London.
Braidotti R. 1990, The politics of ontological difference, in Between feminism and
psychoanalysis, eds. R. Braidotti and T. Brennan, Routledge, London, pp. 89-105.
Bugge A. 1995, Mat til begjær og besvær. Forbrukernes vurderinger og kunnskaper
om helse, miljø og etiske aspekter ved mat. [Food for desire and trouble. Consumers’
evaluation and knowledge about health, environment and ethical aspects of food],
Working Paper no 6, The National Institute for Consumer Research, Lysaker.
Literature
287
Bugge A. and Døving R. 2000, Det norske måltidsmønsteret - ideal og praksis [The
Nordic meal pattern - ideal and practice]. SIFO Report no.2, The National Institute
for Consumer Research, Lysaker.
Burnett J. 1989, Plenty and want. A social history of food in England from 1815 to the
present day, Routledge, London.
Campbell, C. 1996, Detraditionalization, character and the limits of agency. In Detraditionalization. Critical reflections on authority and identity, eds. P. Heelas, S. Lash,
and P. Morris, Blackwell, Cambridge, Mass, pp. 149-169.
Caplan P. 1997, Food, health and identity, Routledge, London.
Caplan P., Keane A., Willetts A., and Williams J. 1998, Studying food choice in its
social and cultural contexts: approaches from a social anthropological perspective, in
The Nation's Diet. The social science of food choice, ed. A. Murcott, Longman, London, New York, pp. 168-182.
Carlsson C., Esseveldt J., Goodman S., and Widerberg K. 1983, Om patriarkat: en
kritisk granskning [On patriarchy: a critical review], Kvinnovetenskaplig tidskrift 1
[4]: 55-69.
Charles N. and Kerr M. 1988, Women, food and families, Manchester University
Press, Manchester.
Counihan C. M. 1988, Female identity, food and power in contemporary Florence,
Anthropological Quarterly 61: 51-62.
DeVault M. L. 1991, Feeding the family. The social organization of caring as gendered work, The University of Chicago Press, Chicago, London.
Dinkins J. M. 1992, Expenditures on food away from home, Family Economics Review 5 [3]: 9-17.
Douglas M. 1975, The sociology of bread, in Bread. Social, nutritional and agricultural aspects of wheaten bread, ed. T. Spicer, Applied Science Publishers Ltd., London, pp. 7-24.
Douglas M. 1975, Implicit meanings. Essays in anthropology, Routledge & Kegan
Paul, London, Boston.
Douglas M. 1983, Culture and food, in The pleasures of anthropology, ed. M.
Freilich, Merton, New American Library, New York, pp. 101.
288
Eating Patterns
Douglas M. and Gross J. 1981, Food and culture: Measuring the intricacy of rule systems, Social Science Information 20 [1]: 1-35.
Douglas M. and Nicod M. 1974, Taking the biscuit. The structure of British meals,
New Society 30: 744-747.
Dulsrud A. 1996, Konsentrasjon i nordisk dagligvarehandel [Concentration in Nordic
retailing]. Nord 1996:19, Nordic Council of Ministers, Copenhagen.
Dumangan J. C. and Hackett J. W. 1995, Almost half of the food budget is spent on
eating out, Food Review pp. 37-39.
Dunne, A. G. 1998, A passion for 'sameness'?: sexuality and gender accountability, in
The new family?, eds. E. Silva and C. Sment, Sage, London.
Døving, R. (in press) Sjokolade – et søtt og syndig tegn på en matrifokal husholdsstruktur [Chocolate – a sweet and sinful sign of a matrifocal houshold structure]. In
Mat som totalt sosialt fenoment [Food as a total social phenomenon]. Dissertation.
University of Oslo, Oslo.
Ekström M. 1990, Kost, klass och kön [Food, Class and Gender], University of
Umeå, Department of Sociology, Umeå.
Ekström M. 1991, Class and gender in the kitchen, in Palatable worlds. Sociocultural
food studies, eds. E.L. Fürst, R. Prättälä, M. Ekström, L. Holm, and U. Kjærnes,
Solum forlag, Oslo, pp. 145-160.
Ekström M., Jonsson L., Renström Å., and Shanahan H. 1996, Matvanor och el i små
lägenhetshushåll [Food habits and electricity in small households living in apartments]. Research Report 20, Göteborg University, Department of Crafts and Home
Economics, Gothenburg.
Elster J. 1989, Nuts and bolts for the social sciences, Cambridge University Press,
Cambridge.
Esping-Andersen G. 1990, The three worlds of welfare capitalism, Polity Press, Cambridge.
Esping-Andersen G. 1999, Social foundations of postindustrial economies, Oxford
University Press, Oxford.
Fagt S., Groth M. V., and Andersen N. L. 1999, Danskernes kostvaner 1995. Mad og
måltider. [Dietary habits in Denmark 1995. Food and meals], FødevareRapport
1999:06, Fødevaredirektoratet, Copenhagen.
Literature
289
Falk P. 1994, The consuming body, Sage Publications, London, Thousand Oaks, New
Dehli.
Feldbæk I. 1998, Nordic statistical yearbook 1998. Nord 1998:1, Nordic Council of
Ministers, Copenhagen.
Fine B. 2000, Whither the welfare state: Public versus private consumption? Working
Paper Series No:92, School of Oriental and African Studies, University of London,
London.
Fine B., Heasman M., and Wright J. 1996, Consumption in the age of affluence. The
world of food, Routledge, London.
Fine B. and Leopold E. 1993, The world of consumption, Routledge, London, New
York
Fischler C. 1988, Food, self and identity, Social Science Information 27: 275-292.
Fischler C. 1990, L'homnivore, Odile Jacob, Paris.
Flood L. and Gråsjö U. 1997, Tid för barn, tid för arbete [Time for children, time for
work], in SOU [Government Report] 1997:138 Familj, makt och jämställdhet [Family, power and equality], eds. G. Ahrne and I. Persson, Fritzes, Stockholm, pp. 159187.
Food Facts – Ruotatieto 1999, Ed. Tapionlinna, U., Elintarviketieto Oy, Helsinki.
Fürst E. 1981, Samfunn, rasjonalitet og sanselighet. Om sosialisering alment og kvinnespesifikt. [Society, rationality and sensuality. On socialisation in general and for
women in particular.] ISO Working Paper No.159, University of Oslo, Department of
Sociology, Oslo.
Fürst E. 1985, Vår matkultur: Konflikt mellom det tradisjonelle og det moderne.
[Food culture: A conflict between tradition and modernity], The National Institute for
Consumer Research, Lysaker.
Fürst E.L. 1995, Mat - et annet språk. Rasjonalitet, kropp og kvinnelighet [Food – a
different language. Rationality, body and femininity], Pax forlag, Oslo.
Fürst E.L. 1997, Cooking and femininity, Women's Studies International Forum 20
[3]: 441-449.
Gallup Finland 1974, A household survey 1973, Unpublished report.
290
Eating Patterns
Gibney M. J. and Wolever M.S. 1997, Periodicity of eating and human health: present
perspective and future directions, British Journal of Nutrition 77 [supp.1]: S3-S5.
Giddens A. 1991, Modernity and self-identity. Self and society in the late modern age,
Polity Press, Cambridge.
Giddens A. 1994, Replies and critiques. Risk, trust, reflexivity, in Reflexive modernization. Politics, tradition and aestethics in the modern social order, eds. U. Beck, A.
Giddens, and S. Lash, Cambridge, pp. 184-197.
Giddens A. 1994, Living in a post-traditional society, in Reflexive modernization.
Politics, tradition and aestetics in the modern social order, eds. , Polity Press, Cambridge, pp. 56-109.
Gilligan C. 1982, In a different voice: psychological theory and women's development
Harvard University Press, Cambridge Mass.
Gisselberg M. 1985, Att stå vid spisen och föda barn. Om hushållsarbete som kvinnoarbete [Women's place in the home. About housework as women's work], University of Umeå, Department of Sociology, Umeå.
Glefjell S. 1985, Arbeidsdeling og matlaging. [Division of labour and cooking], Report 3 from the project 'Mat: arbeid og kultur' [Food: work and culture], The National
Institute for Consumer Research, Lysaker.
Glefjell S. 1986, Mat. Veien til barnas trygghet og mannens hjerte? [Food. The way
to the safety for your children or the way to your husband's heart?], The National
Institute for Consumer Research, Lysaker.
Government Report 1979, Kvinnors arbete. Om hemarbetande kvinnors situation,
kvinnors försörjning, barns omsorg och mäns delaktighet i det oavlönade arbetet.
[Women's work. On the situation of women, working in the home, the care of children
and on men's participation in unpaid work.], SOU 1979:89, Liber, Stockholm.
Government Report 1998, Ty makten är din... Myten om det rationella arbetslivet och
det jämställda Sverige. Betänkande från Kvinnomaktutredningen. [The power is
yours. The myth of the rational working life and the equal Swedes], SOU 1998:6,
Fritzes, Stockholm.
Gronow J. 1997, The sociology of taste, Routledge, London, New York.
Gronow J., Mäkelä J., Kjærnes U., Ekström M. P., Holm L., and Bjørkum E. 1998, A
comparative study of Nordic meal patterns, in Culinary arts and sciences II. Global
and national perspectives, eds. J. S. A. Edwards and D. Lee-Ross, The Worshipful
Company of Cooks Centre for Culinary Research, Bournemouth University, Bournemouth.
Literature
291
Gronow J. and Warde A. 2001, Ordinary consumption, Routledge, London and New
York.
Groth M. V. 1998, Danskernes måltider [The meals of Danes], Fødevarenyt pp. 1215.
Grünert K.G., Baadsgaard A., Larsen H.H., and Madsen T.K. 1996, Market orientation in food and agriculture, Kluwer Academic Publishers, Boston, Dordrect, London.
Grønseth E. 1986, Kjønnspolitiske modeller for likestilling og frigjøring [Models of
gender policy for equality and women's liberation], Cappelen forlag, Oslo.
Guzman M., Bjørkum E., and Kjærnes U. 2000, Ungdommers måltider. En studie av
livssituasjon, mat og kjønn [Young people’s meals. A study of living situation, food
and gender]. SIFO Report No.6, The National Institute for Consumer Research, Lysaker.
Haavind H. 1982, Makt og kjærlighet i ekteskapet. [Power and love in marriage], in
Kvinneforskning. Bidrag til samfunnsteori. [Research on women. A contribution to a
theory of society.], eds. R. Haukaa, M. Hoel, and H. Haavind, Universitetsforlaget,
Oslo.
Haavind H. 1987, Liten og stor. Mødres omsorg og barns utviklingsmuligheter.
[Children and grownups. The care for children and the possiblity for growth], Universitetsforlaget, Oslo.
Haastrup L. 1992, Food cultures, household types, and life-modes, Ethnologia Scandinavica 22: 52-66.
Haraldsdottir J., Holm L., Jensen J. H., and Møller A. 1987, Danskernes kostvaner
1985. 2. Hvem spiser hvad? [Danish dietary habits 1985], Levnedsmiddelstyrelsen,
Søborg.
Haraldsen G. and Kitterød H. 1992, Døgnet rundt. Tidsbruk og tidsorganisering
1970-1990. Tidsnyttingsundersøkelsene [Round the clock. Time use and the organisation of time, 1970-90]. Sosiale og økonomiske studier Nr.76, Statistics Norway, OsloKongsvinger.
Harmaja L. 1928, Husmoderns ekonomiska gärning [The housewife's economical
deeds], Kooperativa förbundets bokförlag, Stockholm.
Hartsock N. 1983, The feminist standpoint: Developing the ground for a specifically
feminist historical materialism, in Discovering reality. Feminist perspectives on epistemology, metaphysics, methodology and philosophy of science, eds. S. Harding and
H. B. Hintikka, D.Reidel Publishing Company, Dordrecht, Holland, pp. 283-310.
292
Eating Patterns
Hellevik O. (1991), Forskningsmetode i sosiologi og statsvitenskap [Research methodology in sociology and political science], Universitetsforlaget, Oslo.
Holm L. 1990, Sociale og kulturelle aspekter af en sund kost [Social and cultural aspects of a healthy diet], Tidsskrift for Landøkonomi 177: 114-121.
Holm L. 1996, Food and identity among families in Copenhagen. A review of an interview study, in Essen und kulturelle Identität. Europäische Perspektiven, eds. H. J.
Teuteberg, G. Neumann, and A. Wierlacher, Akademie Verlag, Berlin, pp. 356-371.
Holm L., Bredsdorff N., and Rasmussen F.D. 1993, The weaknesses of nutrition policies, in Regulating markets - regulating people. On food and nutrition policy, eds. U.
Kjærnes, L. Holm, M. Ekström, E.L. Fürst and R. Prättälä, Novus Forlag, Oslo, pp.
47-64.
Holm, L. and Møhl, M. 1994, Kød i hverdagens madkultur [Meat in everyday food
culture]. Research Institute for Human Nutrition, Copenhagen.
Holm L. and Kildevang H. 1996a, Consumer's views on food quality. A qualitative
interview study, Appetite 27: 1-14.
Ilmonen K. and Stø E. 1997, The 'consumer' in political discourse: Consumer policy
in the Nordic welfare states, in Constructing the new consumer society, eds. P.
Sulkunen, J. Holmwood, H. Radner , and G. Schulze, Macmillan Press, Houndmills
and London, pp. 197-217.
Iversen T. and Holm L. 1999, Måltider som familieskabelse og frisættelse [Meals for
the making of family and individualisation], Tidsskriftet Antropologi 39: 53-64.
Jacobsson L., Johansson B., and Larsson F. 1996, Handeln i Norden in på 2000-talet.
Handelns långtidsutredning [Retailing in the Nordic countries in the 21st century.
Retailers’ long-term report], Handelns utredningsinstitut, Stockholm.
Jakobsen L. and Karlsson J. C. 1993, Arbete och kärlek. En utveckling av livsformsanalys. [Work and love. A contribution to the analysis of life forms], Arkiv, Lund.
Jansson S. 1988, Maten och myterna [The food and the myths], Vår Föda 40: 1-203.
Jansson S. 1993, Maten och det sociala samspelet. Etnologiska perspektiv på matvanor [Food and social interaction. Ethnological perspectives on food habits], Sveriges
Lantbruksuniversitet, Uppsala.
Johansson V. 1996, Vem gör vad, när och var? Omsorgens organiseringsformer i
kommuner och kvinnors och mäns omsorgsarbetsdelning [Who does what, when and
Literature
293
where? The organisation of care in municipalities and women's and men's division of
caring work], Publica, Gothenburg.
Jul M. 1988, Mad, ernæring og politik i Norden [Food, nutrition and politics in the
Nordic countries]. Rapport til det Permanente Nordiske Udvalg for Næringsmiddelspørsmål, Statens Livsmedelsverk, Uppsala.
Kaluzynska E. 1980, Wiping the floor with theory - a survey of writings on housework, Feminist Review 2: 46-57.
Kearney M., Gibney M.J., Martinez J.A., de Almeida M.D.V., Friebe D., Zinft H.J.F.,
Widhalm K., and Kearney J.M. 1997, Perceived need to alter eating habits among
representative samples of adults from all member states of the European Union,
European Journal of Clinical Nutrition 51: 30-35.
Kemmer D., Anderson A. S., and Marshall D. W. 1998, Living Together and Eating
Together: Changes in Food Choice and Eating Habits During the Transition from
Single to Married/Cohabitating, Sociological Review 46 [1]: 48-72.
Kjærnes U. 1985, Spedbarns og småbarns ernæring. En utredning om behov og muligheter for nordisk samarbeid [Nutrition for infants and young children. A report on
needs and possibilities for Nordic co-operation]. Rapport 1985:2, Nordic Council of
Ministers, Oslo.
Kjærnes U. 1997, Framveksten av ernæringspolitikk i Norden [The emergence of
nutrition policies in the Nordic countries], in Utfordringer i ernæringspolitikken.
Nord-Rapport 619, ed. U. Kjærnes, Nordisk Ministerråd, Copenhagen, pp. 73-88.
Kjærnes U. 1999, Food risks and trust relations, Sosiologisk tidsskrift 7: 265-284.
Kjærnes U., Ekström M., Gronow J., Holm L., and Mäkelä J. 2000, Utvalg og representativitet i 'En matdag i Norden'. Oversikt over bakgrunnsspørsmål, omkoding og
sammenligning med relevant statistikk [Sample and representativity in ’A day in the
lives of Nordic peoples. A review of background questions, recoding and comparisons
with relevant statistics]. Working Paper no.1, The National Institute for Consumer
Research, Lysaker.
Kocktürk-Runefors T. 1991. A model for adaptation to a new food pattern: the case of
immigrants, in Palatable worlds. Sociocultural food studies, eds. E.L. Fürst et al,
Solum forlag, Oslo, pp. 185-192.
Labovitz S. 1967, Some observations on measurements and statistics, Social Forces
46: 151-160.
Latouche K., Rainelli P., and Vermersch D. 1998, Food safety issues and the BSE
scare: some lessons from the French case. Food Policy 23[5]: 347-356.
294
Eating Patterns
Leira A. 1976, Kvinner i lønnsarbeid og ulønt arbeid. [Women, wage work and unpaid work], in Kvinnekunnskap [Women's knowledge], eds., T. Støren and T. Schou,
Gyldendal norsk forlag, Oslo, pp. 119-137.
Levenstein H. A. 1988, Revolution at the table. The transformation of the American
diet, Oxford University Press, New York/Oxford.
Lévi-Strauss C. 1964, The raw and the cooked, Jonathan Cape, London.
Lévi-Strauss C. 1966, The culinary triangle, New Society 22: 937-940.
Lien M., Bjørkum E., and Bye E. K. 1998, Kjøtt, holdninger og endring [Meat: attitudes and change]. SIFO Report No 4, The National Institute for Consumer Research,
Lysaker.
Listhaug O. and Wiberg M. 1995, Confidence in political and private institutions, in
Citizens and the state, eds. H.-D. Klingemann and D. Fuchs, Oxford University Press,
Oxford, pp. 298-322.
Longnecker M. P., Harper J. M., and Kim S. 1997, Eating frequency in the nationwide food consumption survey (U.S.A.), 1987-1988. Appetite 29: 55-59.
Luhmann N. 1979, Trust and power: two works , John Wiley & Sons, Chichester.
Lupton, D. 1996. Food, the body and the self. Sage, London.
Marshall D. 2000, British meals and food choice, in Dimensions of the meal. The science, culture, business and art of eating, ed. H. L. Meiselman, Aspen Publishers,
Gaitherburg, pp. 202-220.
Maula J. 1995, Elintarvikkeiden kulutus ja kulutusmuutokset Suomessa 1950-1993
[Changes in Finnish food consumption], Julkaisuja 10/1995, Kuluttajatutkimuskeskus, Helsinki.
Mennell S. 1985, All manners of food: eating and taste in England and France from
the Middle Ages to the present, Blackwell, Oxford.
Mennell S., Murcott A., and van Otterloo A. 1992, The sociology of food. Eating, diet
and culture, Sage, London, Newbury Park, New Dehli.
Miller A.H. and Listhaug O. 1998, Policy preferences and political distrust: A comparison of Norway, Sweden and the United States, Scandinavian Political Studies 21:
161-187.
Literature
295
Miller D. and Reilly J. 1995, Making an issue of food safety: The media, pressure
groups, and the public sphere, in Eating agendas. Food and nutrition as social problems, eds. D. Maurer and J. Sobal, Aldine de Gruyter, New York, pp. 305-336.
Ministry of Food, Agriculture and Fisheries 1999, Måltidets hus - fra råvarer til madkultur med måltidet i centrum [House of Meals - From raw products to food culture,
Focusing on the meal], Copenhagen.
Mintz S.W. 1996, Tasting food, tasting freedom. Excursions into eating, culture, and
the past, Beacon Press, Boston.
Mitchell P. 1997, Struggle worldwide to regain consumers' faith in food, The Lancet
350: 1234-1235.
Mogelonsky M. 1998, Food in demand, American Demographics, pp. 57-60.
Murcott A. 1982, On the social significance of the 'cooked dinner' in South Wales,
Social Science Information 21: 677-696.
Murcott A. 1983, ’Its a pleasure to cook for him’: Food, mealtimes and gender in
some South Wales households, in The Public and the Private, ed. E. Gamarnikow,
Heinemann Educational Books, London, pp. 78-90.
Murcott A. 1986, Opening the ‘Black Box’: Food, eating and household relationships,
Sosiaalilääketieteellinen Aikakauslehti 23 [2]: 85-92.
Murcott A. 1995, Raw, cooked and proper meals at home, in Food choice and the
consumer, ed. D. Marshall, Blackie Academic & Professional, London, pp. 219-235.
Murcott A. 1997, Family meals - a thing of the past?, in Food, health and identity, ed.
P. Caplan, Routledge, London and New York, pp. 32-49.
Murcott A. 1998, The nation's diet. The social science of food choice, Longman, London/New York.
Murcott A. and Prättälä R. 1993, Comparing nutrition policies. A very small start, in
Regulating markets, regulating people. On food and nutrition policy, eds. U. Kjærnes,
L. Holm, M. Ekström, E.L. Fürst, and R. Prättälä, Novus Press, Oslo, pp. 37-46.
Mäkelä J. 1991, Defining a meal, in Palatable worlds. Sociocultural food studies, eds.
E.L. Fürst, R. Prättälä, M. Ekström, L. Holm, and U. Kjærnes, Solum forlag, Oslo, pp.
87-96.
Mäkelä J. 1995, The structure of Finnish meals, in Current research into eating practices. contributions of social sciences. Proceedings of the European Interdisciplinary
296
Eating Patterns
Meeting, 1-16 Oct 1993, Potsdam, eds. E. Feichtinger and B.M. Köhler, Umschau
Zeitschriftenverlag, Frankfurt am Main, pp. 112-114.
Mäkelä J. 1996, Kunnon ateria. Pääkaupunkiseudun perheellisten naisten käsityksiä
[A proper meal: Exploring the views of women with families], Sosiologia 33 [1]: 1222.
Mäkelä, J. 1996A. „Kyllähän sitä aina yrittää syödä enemmän salaattia."
Ravitsemussuositukset ja pääkaupunkilaisnaisten näkemykset terveellisestä
syömisestä ["You should eat more salad." Nutrition recommendations and women's
views on healthy eating in Helsinki metropolitan area], Sosiaalilääketieteellinen Aikakauslehti 33 [1]: 17-23.
Mäkelä J. 2000, Cultural definitions of the meal, in Dimensions of the Meal. The Science, Culture, Business, and Art of Eating, ed. H. L. Meiselman, Aspen Publication,
Gaithersburg, pp. 7-18.
Mäkelä J., Kjærnes U., Ekström M. P., Fürst E. L., Gronow J., and Holm L.1999,
Nordic meals. Methodological notes on a comparative survey. Appetite 32: 73-79.
Möller C. Slå vakt om den goda måltiden [Guard the good meal]. Dagens Nyheter 2710-2000.
Mørch M. 1996, Yearbook of Nordic statistics. Nord 1996:1, Nordic Council of Ministers, Copenhagen.
National Encyclopedia 2000, Ordbok [Dictionary] Språkdata/Bra Böcker, Gothenburg/Höganäs.
Niemi I. and Pääkkönen H. 1989, Ajankäytön muutokset, 1980-luvulla. [Changes in
time use in the 1980s], Central Statistical Office of Finland, Helsinki.
Nelson M. 1997, The validation of dietary assessment, in Design concepts in nutritional epidemiology, eds. B.M. Margetts and M. Nelson, Oxford University Press,
Oxford, New York, Tokyo, pp. 241-272.
Nordisk Ministerråd 2001, Food consumption in the Nordic countries 1965-1998.
National, annual food balance sheets. TemaNord 2001:527, Nordic Council of Ministers, Copenhagen.
Nyberg A. 1997, Makt, kön och BNP. [Power, gender and GNP], in Familj, makt och
jämställdhet [Family, power and equality], SOU 1997:138, eds. G. Ahrne and I. Persson, Fritzes, Stockholm.
Literature
297
O'Doherty Jensen K. and Holm L. 1997, Mad og køn i socialt og kulturelt perspektiv.
[Food and gender in a social and cultural perspective], Research Institute for Human
Nutrition, Copenhagen.
Otnes P. 1991, What do meals do, in Palatable worlds. Sociocultural food studies,
eds. E. L. Fürst et al., Solum Forlag, Oslo, pp. 97-110.
Prokop U. 1978, Kvindelig livssammenhæng. Om strategiernes inskrænkethed og de
umådelige ønsker. [Weiblicher Lebenszusammenhang], GMT Forlag, Århus.
Przeworski A and Teune H. 1970, The logic of comparative social inquiry, John
Wiley, New York.
Prättälä R. 2000, North European meals: Observations from Denmark, Finland, Norway, and Sweden, in Dimensions of the meal. The science, culture, business, and art
of eating, ed. H.L. Meiselman Aspen Publishers, Gaithersburg, Maryland, pp. 191201.
Prättälä R. and Helminen P. 1990, Finnish meal patterns, in Nutritional adaptation to
new life-styles. Bibl. Nutritio et Dieata No 45, eds. J.C.Somogyi and E.H. Koskinen,
Karger, Basel, pp. 80-92.
Prättälä R., Pelto G., Pelto P., Ahola M., and Räsänen L. 1993, Continuity and change
in meal patterns: The case of urban Finland, Ecology of Food and Nutrition 31: 87199.
Ritzer G. 1993, The McDonalidization of society. An investigation into the changing
character of contemporary social life, Pine Forge Press, Thousand Oaks, London,
New Dehli.
Roman C. 1999, Familjelivets organisering - ekonomiska resurser, kön och manlig
dominans. [The organisation of family life. Economical resources, gender and male
dominance], Kvinnovetenskaplig tidskrift 20 [1]: 3-20.
Roos E. and Prättälä R. 1997, Meal pattern and nutritient intake among adult Finns.
Appetite 29: 11-24.
Roos E. 1998, Social patterning of food behaviour among Finnish men and women.
Publications of the National Public Health Institute A 6/1998, Helsinki.
Roos G. and Prättälä R. 1999, Disparities in food habits. Review of research in 15
European countries. Report B24, The National Public Health Institute, Helsinki.
Rotenberg R. 1981, The impact of industrialization on meal patterns in Vienna, Austria, Ecology of Food and Nutrition 11: 25-35.
298
Eating Patterns
Senauer B., Asp E., and Kinsey J. 1991, Food trends and the changing consumer,
Egan Press, Minnesota.
Simmel G. 1957, Soziologie der Mahlzeit, in Brücke und Tür. Essays des Philosophen
zur Geschichte, Religion, Kunst und Gesellschaft, ed. M. Landmann, K.F.Koehlers
Verlag, Stuttgart, pp. 243-250.
Simmel G. 1994, The sociology of the meal, Food and Foodways 5 [4]: 345-351.
Sjögren de Beauchaine A. 1988, The bourgeoisie in the dining-room. Meal rituals and
cultural process in the Parisian families of today, Institutet för folklivsforskning,
Nordiska museet and Stockholms universitet, Stockholm.
Statistics Norway 1992, Tidsbruk og tidsorganisering 1970-90 [The use and organisation of time 1970-90]. NOS C 10, Oslo.
Statistics Sweden 1992, At all times. How women and men use their time. 1990/91.
Report No.79, Stockholm.
Sulkunen P. 1997, Introduction: the new consumer society - Rethinking the social
bond, in Constructing the new consumer society, eds. P. Sulkunen, J. Holmwood, H.
Radner, and G. Schulze, Macmillan Press, Houndmills, Basingstoke, pp. 1-20.
Sørensen Ø. and Stråth B. 1997, The cultural construction of Norden, Scandinavian
University Press, Oslo.
The European Commission 2000, Eurostat Annual Report. Ten year overview 19881998, The Office for Publications from The European Community, Luxembourg.
Theophano J. and Curtis K. 1991, Sisters, mothers and daughters: Food exchange and
reciprocity in an Italian-American community, in Diet and domestic life in society,
eds. A. Sharman et al., Temple University Press, Philadephia, pp. 147-171.
Toivonen T. 1994, Does consumption determine social class? On the changing pattern
of consumption determination, Journal of Consumer Studies and Home Economics
18: 45-63.
Tranøy B.S. 1993, Komparativ metode - mellom ideografiske og nomotetiske idealer
[Comparative methodology – between ideographic and nomothetical ideals], Sosiologi i dag 23 [4]: 17-40.
USDA 1998, Eating out in America. Impact on food choices and nutrient profiles.
Press release. United States Department of Agriculture, Washington DC.
Literature
299
Varjonen J. 2000, Changing habits of meal management in Finnish households. Paper
for IFHE World Congress, Accra, Ghana, July 2000.
Wandel M. 1994, Understanding consumer concern about food-related health risks,
British Food Journal 96: 35-40.
Wandel M., Bugge A., and Ramm J.S. 1995, Matvaner i endring og stabilitet
[Change and Stability of Food Habits]. SIFO Report No 4, The National Institute for
Consumer Research, Lysaker.
Warde A. (1992), Notes on the relationship between production and consumption, in
Consumption and class. Divisions and change , eds. R. Burrows and C. Marsh, St.
Martin's Press, New York p 15-31
Warde A. 1997, Consumption, food and taste, Sage, London.
Warde A. and Martens L. 1998, A sociological approach to food choice: the case of
eating out, in The Nation's Diet. The social science of food choice, ed. A. Murcott,
Longman, London, New York, pp. 129-144.
Warde A. and Martens L. 2000, Eating out. Social differentiation, consumption and
pleasure, Cambridge University Press, Cambridge.
Weber M. 1968, Economy and society. An outline of interpretative sociology, Bedminster Press, New York.
Whit W. C. 1995, Food and society: A sociological approach, General Hall Inc., New
York.
Wold B. 1985, Mat: husholdsforbruk og husholdsarbeid [Food: work and consumption in the household]. Report no.1 from the project Mat: arbeid og kultur [Food:
work and culture]. Melding nr.90, The National Institute for Consumer Research,
Lysaker.
Wood R.C. 1995, The sociology of the meal, Edinburgh University Press, Edinburgh.
Wærness K. 1978, Kvinnekultur og husmorrolle. [Women's culture and the role as
housewife], in Hvis husmoren ikke fantes.[If there was no housewife], ed. R. Grenness, Pax forlag, Oslo.
Ziehe T. 1989, Kulturanalyser. Ungdom, utbildning, modernitet [Cultural analyses.
Youth, education, modernity], Symposion Bokförlag, Stockholm/Stehag.
Åkerman B. 1983, Den okända vardagen. Om arbetet i hemmen. [The unknown working day. On housework], Akademilitteratur, Stockholm.
300
Eating Patterns
List of figures and tables
Figures
Figure 1-1 The eating system ..................................................................................... 40
Figure 1-2 The analytical model................................................................................. 42
Figure 1-3 The record of eating events (x 10) ............................................................ 46
Figure 1-4 Employment status among men and women in the Nordic countries (1996)
............................................................................................................................ 56
Figure 2-1A visualisation of the main components of a hot meal .............................. 68
Figure 2-2 Types of foods at breakfast on weekdays. Per cent. N: Denmark 737,
Finland 975, Norway 958, Sweden 904. More than one choice possible .......... 70
Figure 2-3 Types of foods at cold eating events on weekdays. Per cent. N: Denmark
1462, Finland 1540, Norway 1666, Sweden 1412. More than one choice was
possible............................................................................................................... 72
Figure 2-4 Types of foods at cold eating events on Sundays. Per cent. N: Denmark
417, Finland 313, Norway 275, Sweden 239. More than one choice was possible
............................................................................................................................ 73
Figure 2-5 Beverages at cold eating events on weekdays and Sundays. Per cent. N:
Denmark 1462/417, Finland 1540/313, Norway 1666/275, Sweden 1412/239.
More than one choice was possible .................................................................... 74
Figure 2-6 Beverages at hot meals on weekdays and Sundays. Per cent. N: Denmark
686/237, Finland 1376/260, Norway 976/222, Sweden 1222/202. More than one
choice was possible ............................................................................................ 83
Figure 3-1 The distribution of eating events during the day on weekdays in the
Nordic countries, 1 hour intervals (per cent) ...................................................... 94
Figure 3-2 The distribution of eating events during the day on Sundays in the
Nordic countries, 1 hour intervals (per cent) ...................................................... 95
Figure -3-3 Eating activities during the day in Norway on Mondays to Thursdays,
1971-72, 1980-81, and 1990-91 (per cent) ......................................................... 96
Figure 3-4 The distribution of various cold eating events during the day on weekdays.
1 hour intervals (per cent)................................................................................. 104
Figure 5-1 Time and place of eating. The distribution during the day on weekdays.
Percentages of populations eating at home, at work/in school or at other places.
.......................................................................................................................... 167
302
Eating Patterns
Figure 5-2 Time and place of eating. The distribution during the day on Sundays.
Percentage of populations eating at home, at work/in school or at other places.
..........................................................................................................................169
Figure 5-3 Time and social context of eating. The distribution during the day on
weekdays. Percentage of populations eating alone, with family, with
colleagues/business contacts or with others. ....................................................171
Figure 5-4 Time and social context of eating. The distribution during the day on
Sundays. Percentage of populations eating alone, with family, with
colleagues/business contacts, friends and others..............................................173
Figure 5-5 Social context and location of eating. Per cent of total number of eating
events. (N= Denmark: 3266, Norway: 3363, Sweden: 3920)............................175
Figure 5-6 Eating out. Per cent of populations with different frequencies of eating out.
N= Denmark: 1187, Finland: 1200, Norway: 1178, Sweden: 1244..................176
Figure 5-7 Eating arrangements – where to sit. Percentage of eating events at home in
different kinds of sitting arrangements. N= Denmark: 3160, Norway: 2945,
Sweden: 3170). .................................................................................................186
Figure 5-8 Eating at the workplace. Percentage of eating events in different places.
N= Denmark 437, Norway: 553, Sweden: 607)................................................186
Figure 5-9 Eating arrangements – how long did the eating event take? Percentage of
eating events at home with different durations. N= Denmark: 4792, Finland:
5159, Norway: 4221, Sweden: 5043)................................................................187
Figure 5-10Types of eating arrangements and their duration. Percentage of eating
events in different contexts and locations by duration. N: Denmark 3257,
Norway 3334, Sweden 3917. ............................................................................189
Figure 5-11 Eating arrangements – what went on while eating? Percentage of eating
events at home with different kinds or no simultaneous activity. N= Denmark:
3268, Norway: 3028, Sweden: 3323. ................................................................190
Figure 5-12 Types of eating arrangements at home: what went on while eating?
Percentage of eating events in different contexts and location with simultaneous
activities. N= Denmark: 2454, Norway: 2378, Sweden: 2694..........................191
Figure 6-1Family meals and household types (N: Denmark 810, Finland 912,
Noreway 811, Sweden 744). .............................................................................205
Figure 6-2 Family meals and age groups (N: Denmark 890, Finland 993, Norway 888,
Sweden 820)......................................................................................................206
Figure 6-3 Family meals and day of the week (N: Denmark 890, Finland 996, Norway
888, Sweden 820)..............................................................................................207
Tables
Table 1-1 Types of eating events ................................................................................35
Table 1-2 Factors influencing eating patterns .............................................................41
Table 1-3 Samples and response rates in the survey ...................................................51
Table 1-4 Proportion of household expenditure spent on food (per cent)..................58
Table 1-5 Annual consumption of foods and beverages (wholesale level) in the
Nordic countries (1997) (kg per capita per year). Main source: FAO Food
Balance Sheets ....................................................................................................59
List of figures and tables
303
Table 2-1 The recoding of foods at breakfast ............................................................ 66
Table 2-2 The recoding of cold eating events ............................................................ 66
Table 2-3The recoding of hot meal components ........................................................ 67
Table 2-4 The recoding of beverages at hot and cold eating events ........................... 67
Table 2-5 The number of components at breakfast on weekdays and Sundays. Per cent
............................................................................................................................ 69
Table 2-6 Beverages at breakfast on weekdays and Sundays. Per cent. More than one
choice was possible ............................................................................................ 71
Table 2-7 The main dish of hot meals on weekdays and Sundays. Per cent .............. 75
Table 2-8 Meat and fish preparation methods. Per cent ............................................. 77
Table 2-9 Types of staples at hot meals on weekdays and Sundays. Per cent. More
than one choice was possible.............................................................................. 78
Table 2-10 Vegetables as side dishes at hot meals on weekdays and Sundays. Per
cent. More than one choice was possible............................................................ 79
Table 2-11 Preparation of vegetables as side dishes on weekdays and Sundays. Per
cent ..................................................................................................................... 80
Table 2-12 The top five and bottom of the vegetable choice list. Per cent of total
number of hot meals ........................................................................................... 81
Table 2-13 Trimmings at hot meals on weekdays and Sundays. Per cent. More than
one choice was possible...................................................................................... 81
Table 2-14 The percentage of starters and dessert on weekdays and Sundays. Per cent
of all hot meals ................................................................................................... 83
Table 2-15 Desserts at hot meals on weekdays and Sundays. Per cent ...................... 84
Table 2-16 The typical features of different eating events in four Nordic countries.
Items used in 15 per cent or more of the respective types of eating events ........ 86
Table 3-1 The number of eating events in the Nordic countries (per cent).
Weekdays*) and Sundays*)................................................................................ 98
Table 3-2 The number of cold eating events in the Nordic countries (per cent).
Breakfast not included. Weekdays*) and Sundays*).......................................... 98
Table 3-3 The number of hot eating events in the Nordic countries (per cent).
Weekdays*) and Sundays*)................................................................................ 99
Table 3-4 The mean number of different cold eating events in the Nordic countries on
weekdays .......................................................................................................... 102
Table 3-5 The mean number of different kinds of cold eating events eaten on
weekdays and Sundays by men and women in all four countries..................... 103
Table 3-6 The mean number of different types of eating events in Finnish regions. All
days................................................................................................................... 107
Table 3-7 The percentage of those persons who had eaten hot food at dinner time (4-6
p.m.) in Finnish regions.................................................................................... 107
Table 3-8 The number of eating events in the Nordic countries on weekdays. Linear
regression analysis. A cleaned model**........................................................... 108
Table 3-9 The relative proportion of the most common daily patterns of eating events
in the Nordic countries on weekdays and on Sundays (per cent) ..................... 110
Table 3-10 The number of persons who had eaten cold and hot food at lunch hour
(from 11 a.m. to 1 p.m.) on weekdays in the Nordic countries (per cent) ........ 111
304
Eating Patterns
Table 3-11 The percentage of people who had eaten hot and cold food in Finland and
Norway at 4-6 p.m., Denmark at 6-8 p.m. and Sweden at 5-7 p.m. on weekdays
..........................................................................................................................115
Table 3-12 Mean numbers of all eating events in the Nordic countries in different
population groups. Weekdays ...........................................................................118
Table 3-13 Mean numbers of cold eating events in the Nordic countries in different
population groups..............................................................................................119
Table 3-14 Mean numbers of hot eating events in the Nordic countries in different
population groups..............................................................................................120
Table 3-15 An explanatory analysis of cold lunch eating (11am-1pm) in the Nordic
countries on weekdays. Logistic regression of all countries together and each
separately ..........................................................................................................121
Table 3-16 A description of hot lunch eating (11 a.m. –1 p.m.) in Nordic countries on
weekdays according to different background variables. Logistic regression of all
countries together and each separately..............................................................122
Table 3-17 An explanatory analysis of hot dinner eating in the Nordic countries.
Relative hot dinner eating in the Nordic countries on weekdays (odds ratio.
Logistic regression of all countries together and each separately .....................123
Table 4-1 Meal components......................................................................................133
Table 4-2 Hot meal combinations on weekdays. Per cent of all hot meals ..............134
Table 4-3 Mean and number of components in hot meals on weekdays and Sundays.
Per cent of all hot meals....................................................................................137
Table 4-4 Number of staples in hot meals on weekdays and Sundays. Per cent of all
hot meals ...........................................................................................................138
Table 4-5 Eating hot meals on weekdays. Per cent of all informants .......................139
Table 4-6 Informants who had eaten hot meals between 11 a.m. and 3 p.m. and
between 3 p.m. and 9 p.m. on weekdays. Per cent of all informants ................140
Table 4-7 Informants who had eaten proper meals on weekdays. Per cent of
informants who had eaten hot meals .................................................................140
Table 4-8 Types of centres in hot meals on weekdays. Per cent of proper and other
types of meals ...................................................................................................141
Table 4-9 Description (%), relative eating (odds ratio) and full model (odds ratio,
logistic regression) of eating a proper meal between 11 a.m. and 3 p.m. and
between 3 p.m. and 9 p.m. of those who have eaten a hot meal. Weekdays .....144
Table 4-10 Description (%), relative eating (odds ratio) and full model (odds ratio,
logistic regression) of eating a proper meal between 11 a.m. and 3 p.m. of those
who have eaten a hot meal. Weekdays. Finland and Sweden ...........................146
Table 4-11 Description (%), relative eating (odds ratio) and full model (odds ratio,
logistic regression) of eating a proper meal between 3 p.m. and 9 p.m. of those
who have eaten a hot meal. Weekdays. Denmark, Finland, Norway, and Sweden
..........................................................................................................................147
Table 4-12 Description (%), relative eating (odds ratio) and full model (odds ratio,
logistic regression) of eating at least one proper meal during the day of those
who have eaten hot meal(s)...............................................................................148
List of figures and tables
305
Table 4-13 Description (%), relative eating (odds ratio) and full model (odds ratio,
logistic regression) of eating one or more proper meal during the day of those
who have eaten hot meal(s). Separate analyses for men and women ............... 150
Table 4-14 Eating one proper meal between 11 a.m. and 3 p.m. of those who have
eaten a hot meal. Weekdays. Description (%), relative eating (odds ratio) and full
model (odds ratio, logistic regression). Men .................................................... 151
Table 4-15 Eating one proper meal between 3 p.m. and 9 p.m. of those who have
eaten a hot meal. Weekdays. Description (%), relative eating (odds ratio) and full
model (odds ratio, logistic regression). Separate analyses for men and women152
Table 5-1 Eating at home. Average number of eating events at home in population
groups and proportion of population who had less than 2 eating events at home
on the previous day. ......................................................................................... 178
Table 5-2 Eating with family members. Average number of eating events in
population groups and proportion of population who had no eating events with
family members on the previous day................................................................ 181
Table 5-3 Eating alone. Average number of individual eating events in population
groups and proportion of population who had no individual eating events a day.
.......................................................................................................................... 183
Table 6-1 Family meals.Percent who ate a hot meal at home with the whole
household. Relative family meal eating, odds ratio. (Persons living alone
excluded). ......................................................................................................... 208
Table 7-1 Persons answering the question "Who did the cooking?" living together
with someone, and who have had one hot meal or more during the day. Per cent.
.......................................................................................................................... 219
Table 7-2 Who did the cooking? Persons living together with someone, and who have
had one hot meal in the afternoon (2-9 p.m.). Per cent..................................... 221
Table 7-3 "I did the cooking." Men's and women's answers about who did the
cooking in different social groups. Per cent. .................................................... 223
Table 7-4 Division of labour according to cooking. Divided into the categories
"Modern", "Normal" and "Traditional". Per cent. ............................................ 227
Table 7-5 Levels of over- or underreporting or differences in men's and women's
understanding of division of labour concerning cooking. Percentage points. .. 230
Table 7-6 Relations between cooking and interest. Per cent. ................................... 231