Beyond the activist ghetto: A deductive blockmodeling approach to
understanding the relationship between contact with environmental
organisations and public attitudes and behaviour
Paper presented at ECPR Glasgow to the panel on Social Movements and
Social Networks. Note that this paper has now been published in
Social Movement Studies 13(1) (2014)
Clare Saunders, Milena Büchs, Anastasios Papafragkou, Rebecca Wallbridge
and Graham Smith
Abstract
Current research on the behavioural impacts of social movements tends to focus on their
influence on those most intensely involved. Consequently it overlooks the impacts that social
movement organisations might have on those outside the activist ghetto. To begin to address
this gap in the literature, this article examines the relationship between contact with
environmental organisations and public attitudes and behaviour. Monitoring the electricity use
of 72 households has facilitated analysis of its association with their environmental attitudes
and contact with environmental organisations. Although standard statistical approaches fail to
uncover a relationship between contact with environmental organisations and attitudes and
behaviour, a deductive blockmodeling approach tells a different story. Low household
electricity use is associated with households sharing pro-environmental attitudes and contact
with environmental organisations. High energy use is associated with households not sharing
any of these; and moderate energy use is associated with a moderate degree of sharing. Our
findings reveal the need for systematic studies of environmental movement organisations’
impact on the public’s pro-environmental behaviours.
Key words: social movements, environmental movements, blockmodeling, cultural impacts,
energy use, climate change.
Introduction
Since the 1980s, there has been a significant cultural turn in the study of social movements
(Johnston and Klandermans 1995, Meyer et al 2002). This work has made great inroads into
supplementing a dominant interpretation of social movements as instrumental actors seeking
to influence policy (Gamson 1990, Giugni et al 1999). Although the emphasis given to
expressive and identity aspects of social movements has done much to enrich the field,
1
assessment of the relationship between the outreach work of social movements and the
attitudes and behaviour of the public has continued to be overlooked.
We suggest that this gap in the literature exists because scholars who look beyond
instrumentally oriented outcomes often restrict themselves to the study of the more visible
cultural aspects of movements and/or to the behavioural impacts upon those most intensely
involved. In much of the work on the cultural impacts of movements, for example,
movements are seen either as challengers of hegemonic culture that produce cultural artefacts
(such as fashion items and music), as nurturers of their own (sub)culture (Whittier 2002: 293)
or as exhibitions of expressive identities (Hetherington 1998). Hetherington’s (1998) work on
expressive identities emphasises the liminal, visible, yet system challenging, aspects of
movements. Importantly, expressive identities, incorporating fashion, music and sub-cultures
are visible to outsiders and are to some degree shared amongst a group. But this body of
research tells us little about the broader behavioural impacts of social movements upon those
who exist outside of a group of intensely interacting activists. It also fails to address broader
social movement impacts – like household energy use, which we focus upon in this article –
that are invisible to an outside observer.
In this article we seek to extend the literature on the cultural impacts of social movements.
We focus on the broader outreach work of movements directed at the wider public, using the
environmental movement as a case study. We are interested to understand whether
individuals who are not intensely involved in environmental movements but who have contact
with environmental organisations tend to save more energy in their homes or not. To date,
there has been a dearth of literature on the relationship between weak forms of engagement
with environmental organisations and energy use. This is somewhat surprising given
important policy implications. The Department of Environment, Food and Rural Affairs
(DEFRA 2008), for example, which seeks to promote pro-environmental behaviours, has
funded several projects involving third sector organisations (e.g. the Environmental Action
Fund, the Third Sector Fund and the Greener Living Fund); and the Department of Energy
and Climate Change (DECC 2012) is currently working on an Energy Demand Reduction
Project and the Low Carbon Communities Challenge. Policy in this area has drawn mostly on
academic studies which argue that changing peoples’ attitudes will lead to a corresponding
change in their behaviours (Büchs et al 2012). But little is known about how contact with
environmental organisations might mediate attitudes and behaviour, besides fairly tentative
2
suggestions that they have potential to do well at facilitating sustainable behaviours amongst
the public (Seyfang et al 2010).
Using data drawn from the Community Based Initiatives in Energy Saving Project funded by
the UK Research Council’s (RUCK) Energy Programme, we explore the relationship between
contact with environmental organisations and public attitudes and behaviour. We draw on
social survey and household electricity use data from 72 households in South East England.1
Our broad-ranging data allows us to measure individuals’ contact with and paid support of
environmental organisations; their attitudes to peak oil, climate change and the efficacy of
individual action; and their household electricity use.
We proceed by introducing the literature on the broader cultural impacts of social movements,
focusing particularly on environmentalism. In addition to highlighting the extant literature’s
over-emphasis on visible cultural artefacts and the behaviour change of the committed to the
detriment of the less visible and less engaged, we draw on aspects of the debate which suggest
that being part of a network helps to foster pro-environmental behaviours. Although we are
unable to prove cause and effect, we postulate that less intense networks of contact with
environmental organisations – outside of the activist ghetto – might have effects on the
energy-use behaviours of the public. After that, we introduce our methodology, which uses
innovative inductive blockmodeling (Saunders 2011). Our results suggest that there is a
complex relationship between shared contact with environmental organisations and energy
use, and between shared contact with environmental organisations and pro-environmental
attitudes.
Cultural impacts of social movements
Cultural approaches to the study of social movements, which took off in the 1980s, sought to
look beyond instrumental policy-based movement outcomes. There are two main strands of
this relatively new body research. The first is internally focused on the behaviours of those
most intensely involved in social movements and the second is externally focused on visible
outcomes and changes in values and discourses amongst broader publics. Internally,
movements have been said to have influenced the behaviours of those involved in a number
of ways. For example, it has been suggested that movements allow people to engage in new
lifestyles, facilitate new forms of expressivism (Hetherington 1998), develop senses of
belonging (Melucci 1989) and experiment with new organisational forms (Jasper 1997).
Externally, movements have been shown to have helped spread beliefs across publics
3
(d’Anjou 1996, Inglehart 1977), encourage new institutional practices (Epstein 1988), spark
fashions (McAdam 1986), and shape discourses (Gamson and Modigliani 1989).
Despite this impressive body of evidence, cultural approaches to social movement outcomes
have, in some regards, been limited. Studies of behaviour change, unlike those that focus on
value change, have largely been empirically confined to the internal realm – that is they are
focused on the minority of the public who are deeply involved in cultural reproduction within
social movements or other restricted communities (see, for example, Kanter 1972 and Barnes
and Starr Sered 2005). Put differently, the focus has been on the activist ghetto. This is
especially true of many studies that have considered the behavioural impacts of involvement
in environmental organisations or environmentalist sub-cultures.
Horton (2005), for example, conducted ethnographic research to uncover the everyday lives
of green activists in the city of Lancaster (1998-2002). He found that ‘where one shops, and
how one moves around both locally and farther are afield are things that really matter’ to
environmental activists (Horton 2005: 134, emphasis in the original). Horton suggests that
people learn to live green lifestyles through becoming immersed in cultural codes via
networking within formal and informal activist groups and meetings. Arenas for such
networking include vegetarian cafes, radical bookshops, green gatherings and festivals, and
through friendships with other activists. Similarly, Haluza-DeLay (2008) finds that the
internal culture of environmental organisations provides the social space through which a
logic of practice consistent with movement goals emerges. In Ingalsbee’s (1996:266)
research, lifestyle changes and the behavioural practices that accompany them form part of
the ‘identity praxis’ of Earth First! activists. Their biocentric belief system and commitment
to radical activism is given expression through their everyday lives in which they strive to
‘live more ecologically conscious lifestyles of voluntary simplicity’ (ibid.:268, see also Ergas
2010). Similarly, Saunders (2008a) found that the encompassing collective identity of an
environmental direct action group allowed movement culture to pervade the activists’
everyday lives. In these examples, what might be called ‘submerged cultural networks’
(Whittier 2002) have influenced the behaviour of those deeply involved in environmental
activism.
The above examples focus on lifestyle changes amongst activists. But individuals need not be
activists in order to ‘consciously and actively promote a lifestyle or way of life’ (Haenfler et
al 2012: 2). Indeed, lifestyles can be viewed as a ‘primary means to foster social change’ (ibid;
4
see also Evans and Jackson 2007). Consequently, Haenfler et al (2012), who bemoan the
state-level focus of social movement research, have drawn a distinction between lifestyle
movements and social movements. Unlike social movements that seek policy change,
lifestyle movements like veganism and New Age Travellers measure success in terms of
changes to everyday behaviours and lifestyles (Cherry 2006:156). Returning to the
importance of networking, Cherry shows how behavioural changes needed to maintain a
vegan lifestyle are not ‘dependent on individual willpower, epiphanies, or simple norm
following’ but are ‘dependent on having social networks that are supportive of veganism’
(ibid.:157), highlighting once more importance of movement based social networks in
supporting behaviour change.
Physical networking is therefore considered to be important in motivating behaviour change
in the literature on environmentalist subcultures and lifestyle movements. But what about
external impacts and external networking that is less physical – such as having contact with
environmental organisations through emails, websites, brochures/leaflets or making a
donation? How might that be related to the broader behaviours of the non-activist public?
These are important but as yet unanswered questions. So whilst it is promising to see a turn
away from emphasis on instrumental outcomes of social movements, research has not gone
quite far enough in exploring the relationship between contact with environmental/social
movements and the everyday behaviour of the non-activist public. This gap persists despite
the emergence of a body of literature emphasising environmental organisations’ potential to
play a significant role in promoting environmentally sustainable behaviours (Hale 2010;
Seyfang and Smith 2007, Hargreaves et al 2008, Middlemiss 2008). Several reasons have
been given as to how and why environmental organisations may fulfil this role: grassroots
organisations can introduce and support ‘social innovations’ – novel or changed practices –
because they are operating outside of the social mainstream (e.g. Seyfang and Smith 2007);
their collective nature and engagement in small-group activities is conducive to establishing
and maintaining new social norms and practices (Heiskanen et al 2010); and they may have
better chances of influencing citizens because they are more trusted by and closer to citizens
than government or private corporations (Steward 2009; House of Lords 2011). Thus, it is not
surprising that direct involvement in environmental organisations seems to positively
influence participants’ pro-environmental attitudes and behaviours (Georg 1999, Hobson
2003, Hargreaves et al 2008 on the Global Action Plan programme; Howell 2009, on Carbon
Rationing Action Groups; and DEFRA 2009 on the Environmental Action Plan).
5
As we have shown, the existing literature on environmental organisations and behaviour
change focusses largely on direct and/or intense forms of engagement. But it has also stressed
the need (and related potential difficulties) that environmental organisations may encounter
when reaching out to the wider public to ‘scale up’ social innovation (Georg 1999, Seyfang
and Smith 2007). Furthermore, it has not systematically assessed whether less intensive forms
of engagement such as being a recipient of environmental organisations’ electronic
communication or leaflets, or being a paid supporter, may influence attitudes and behaviours
of the broader public.
By focusing on these overlooked dimensions, our research also contributes to debates on the
concept of chequebook membership. ‘Chequebook members’ are thought to barely participate
in the real action within social movements (Jordan and Maloney 1997). Their consciences are
thought to be satisfied by knowledge that they are paying someone else to take action on their
behalf. Whilst it might be true that the majority of paid members of environmental
organisations do little in the way of activism, no attention has been given to the question of
whether chequebook members take action to solve environmental problems in their everyday
lives. We suggest that it is short-sighted to overlook the fact that many environmental
organisations aim to influence the attitudes and/or everyday behaviours of their wider
audience of supporters, not just the cultures of the most committed. The 10:10 campaign2, for
example, suggests simple changes that people can make to their lifestyles such as insulating
their homes, turning their appliances of standby and driving less. In addition to moving
beyond the activist ghetto, our study also develops the literature by looking at household level
behaviours that are not visible to the outside world, and which are familial rather than shared
amongst a cultural group. The example we focus on is domestic energy use – specifically,
electricity base load.
Methods
Measuring electricity use
Electricity use data was collected from 72 households located in the south of England. These
72 households are located in areas classified by the Office for National Statistics (ONS) as
4a2: ‘Prospering young families’. Areas with this classification are close to the national
average for the proportion of people working at home, the number of people per room and
population density. Households in these areas are more likely to own two or more cars and to
live in detached properties than the national average; but homes are less likely to be occupied
6
by retired people. Similar areas exist across the UK. Our sample consists of households
located within a village and town willing to sign up to our broader project, which offers free
insulation and an energy use monitor. Thus, our emphasis is on the electricity use behaviour
of average families, rather than, as in previous studies, on the behaviour of activists deeply
involved in environmental activism. We measure electricity use objectively with monitoring
equipment, allowing us to avoid reliance on often inaccurate self-reported behaviour.
To measure household electricity use we installed split core clamp meter readers on the
electricity meter of each dwelling to record real time electricity consumption data at one
second resolution. Each meter reader is linked to a hub to which data is transmitted wirelessly.
The hub is connected to the internet via each household’s broadband.3 The hub within the
dwelling stores data from the meter reader and every two hours connects to a remote server
and saves the dataset, before clearing its cache. An Application Programming Interface (API)
was developed in order to connect to a remote server and obtain the data. Due to the large
volume of data, a Structured Query Language (SQL) database was designed which ensured a
fast and efficient data management and data post-processing. The monthly electricity
consumption and the base-load over a month were calculated for each household. In order to
minimise the impact of irregular electricity use, such as electric heating, and untypical
patterns of occupancy, such as school holidays, data was analysed from the month of May.
We use base-load rather than overall electricity consumption data to fit the examined
households into one of three electricity use phenotypes – low energy use, moderate energy
use and high energy use. The base-load is the minimum amount of power used by a household
through appliances that are always left switched on (such as appliances left on standby),
excluding refrigerators and freezers. In households with day-time working patterns, the baseload is revealed by night-time energy consumption, less the cyclical traces from fridges and
freezers. Figure 1 shows the electricity consumption profile of a household, where the baseload has been identified.4
Figure 1. An example of a dwelling’s daily electricity consumption profile (1 second
resolution data)
7
We focus our analysis of electricity consumption on base-load because overall electricity
demand is not necessarily a reflection of pro-environmental electricity use behaviour. Instead,
it is strongly influenced by a number of factors, including the number of occupants and their
social background / relative wealth, appliances installed, length of holidays and so on.
Research has shown that overall electricity use depends more on socio-demographic
characteristics than attitudes (Abrahamse/Steg 2009, 2011) and is influenced by habits,
routines, more general social norms and infrastructures (Shove 2003, Gram-Hanssen 2011).
Some of these factors may lead it to appear as if households have high electricity demand
even when they are making conscious efforts to save energy. Defining the level and type of
dependence of each factor on overall electricity consumption is a hugely complex undertaking,
requiring a highly detailed profile of each household. The main advantage of performing the
analysis on the basis of base-load is that it better reflects occupants’ behaviour and underlying
energy use as it is less dependent on the factors we discuss above. Exploratory data analysis
on our data-set confirms this. It reveals that there are either no or very weak associations
between the following three variables and base-load: floor space, number of occupants and
employment status.5 Instead, base-load could be said to reflect a conscious effort to save
energy. Whilst a single occupancy household is expected to consume less electricity than a
household with two occupants, their base-load usage could still be similar as stand-by items
such as broadband routers, telephones and TV sets are more likely to be shared amongst
several household members. Our premise is that a household with high base-load has not
adopted pro-environmental energy use if they fail to switch appliances off stand-by. A
number of organisations, including Energy Saving Trust (EST), Department of Energy and
8
Climate Change (DECC) and International Energy Agency (IEA), have stressed the
importance of switching appliances off standby to save energy and reduce carbon emissions.
IEA has raised the importance of standby power since the early 1990s through publications,
conferences and policy advice to governments. In 1999, the IEA proposed that all countries
harmonise energy policies to reduce standby power use to no more than one watt per device
by 2010 and no more than half a watt per device by 2013 (Meier and LeBot, 1999). EST
includes switching appliances off standby as one of their top ten tips to save energy (EST,
2012). Recently, the “Powering the Nation” study jointly commissioned by DEFRA, DECC
and the EST demonstrated the importance of standby power consumption by revealing that
total standby consumption can amount from 9 to 16% of domestic power demand (EST,
2012b). Thus, there is a clear contradiction if a household believes they are ‘green’ in their
behaviour and yet has a high base-load.
Although low base-load could be economically rather than environmentally motivated (i.e. to
save money), our survey data finds that those who claim to be most motivated to prevent
climate change (measured by an agreement scale of 1-5) are more likely to claim to switch
appliances off standby (frequency measured on a scale from 1-5) than those who are
motivated to save money (agreement scale from 1-5).6 Recall, also, that our sample of
households is largely ‘prospering young families’, meaning that the lowest income houses,
whose electricity use is influenced more by pricing than higher income houses (see, for
example Jasamb and Meier 2010), are excluded.
Figure 2 shows the base-load of all 72 households we analyse. Classification of households
into one of the three electricity use phenotypes is performed based on the quartiles of the
base-load distribution. The first and the third quartile of the base-load distribution are
indicated with the two horizontal dashed lines. Eighteen households with a base-load lower
than the 1st quartile were fitted into the first electricity use group (low energy use) and 18
households with a base-load higher than the 3rd quartile were fitted into the third electricity
group (high energy use). The remaining 36 households, from the middle two quartiles, were
fitted into the moderate energy use group.
9
Figure 2. Electrical base-load of all 72 households used in this study
We elect to focus on electricity (baseload) rather than gas use because household electricity
consumption is more dependent on occupants’ behaviour than natural gas consumption. Gas
consumption is strongly related to the level of insulation in a dwelling, which is more difficult
for occupants to control. Furthermore, minimum thresholds of thermal comfort may differ
amongst occupants of different households because they depend on personal factors such as
age, health and level of activity within dwellings. Such differences may have a significant
impact on overall gas consumption leading to false conclusions with respect to proenvironmental energy use. Measuring gas use in inadequately heated dwellings can also lead
to false conclusions, since low natural gas consumption could be the result of fuel poverty,
rather than a conscious attempt to save energy.
Encounters, paid support and involvement
We operationalize two forms of contact with environmental organisations: ‘encounters’ and
‘paid support’. Encounters are forms of contact with environmental organisations that can
occur without being directly involved or being a paid supporter. Thus, encounters include
receiving an email, seeing a magazine, brochure or leaflet and viewing a web-page.
We distinguish this from ‘paid support’. We also collected data on involvement in
environmental organisations, measured by volunteering or attending meetings. The density of
the involvement network was very low (only six respondents were involved in any of the
organisations listed and the density of the actor-by-actor affiliation network was only 1.1%7)
so we did not conduct further analysis. In any case, excluding involvement in environmental
organisations and concentrating instead on encounters allows us to move beyond the focus on
10
sub-cultural networks that has limited previous research. Including ‘paid support’ allows us to
shed new light on the behaviour of chequebook members.
Network analysis
We analyse social survey and electricity use data using two approaches: standard statistical
measures of association and a positional approach to network analysis (Saunders 2007).
Standard statistical approaches and positional network analysis are likely to yield different
results because of their different assumptions. Simple individual-level statistical approaches
view each individual in any sample as a separate agent. In the absence of a multi-level
framework, it can be argued that this overlooks that individuals are embedded in relational
contexts. In contrast, positional network analysis recognises that individuals ‘are inseparable
from the transactional contexts in which they are embedded’ (Emirbayer 1997: 287). As
Mitchell (1989:46) puts it, the ‘essential idea’ of this approach is that:
…the variations in behaviour of people in any one role relationship may be traced to
the effects of behaviour of other people to whom they are linked in one, two or more
steps in some quite different relationship.
Our particular focus in this article is not on links between people, but instead on shared
contact with a network of environmental organisations. In this context, the positional
approach presupposes that the energy use behaviour of people in their homes is related to the
shared structural positions that they have in a network of encounters or of paid support for
environmental organisations (see also Wellman 1988). We use two social network methods to
explore relationships between our variables of interest: network correlation and relatively
novel deductive optimized user-defined block modelling (de Nooy et al 2005; Saunders 2011).
Often in social movement research, blockmodels are applied inductively, error scores are not
reported, and statistical tests of association between error free blockmodels and hypothetical
blockmodels are entirely absent (Saunders 2011). We illustrate a different approach to
blockmodeling which does not suffer from these same pitfalls.
Networks of attitudes, encounters and paid support
To measure attitudes, we analyse responses to three five-point likert-scale ‘agree-disagree’
questions which were dichotomised as follows: (1) ‘Human induced climate change is
happening’ (agree or strongly agree = 1; otherwise = 0); (2) ‘Current rates of global oil
production can continue in the long-term’ (disagree or strongly disagree = 1; otherwise = 0)
and (3) ‘Individual action can make a significant contribution to solving climate change’
11
(agree or strongly agree = 1; otherwise =0). Each respondent is subsequently given a score
out of three, where a three represents a respondent who believes that anthropogenic warming
is occurring, global oil production rates cannot continue in the long-term and that individual
action can make a difference. We include this final variable as measure of individual efficacy.
It is important to include this because of the alleged importance of a sense of efficacy in
bridging the attitude-behaviour gap (Ajzen and Fishbein 1980). The mean score for attitudes
was 2.041. The 2-mode matrix (72 households against 3 attitudes) was converted to a valued
affiliation matrix. This matrix shows the number of attitudes that each pair of respondents
share (for example, if respondent 23 shared 2 attitudes with respondent 4, the cell in the 23rd
row and 4th column of the matrix would show a 2). The network was dichotomised at greater
than or equal to two (GE2).8 In other words, if respondents 4 and 23 share two or more
attitudes, the two in the 23rd row and 4th column of the valued matrix would be converted to a
one; if only one attitude, then a zero.
For encounters and paid support, we draw on the survey question which asked: ‘In the past 12
months, what contact have you had with the following national organisations?’. We list a
number of significant environmental organisations: Friends of the Earth, Greenpeace,
National Trust, RSPB, Wildlife Trust, Campaign Against Climate Change, WWF, Oxfam,
World Development Movement9, Stop Climate Chaos, Transition Towns10 and 10:10. At the
time at which the survey was administered, all of these organisations had active campaigns on
climate change and/or provided advice on behaviour change. Each organisation identifies
climate change as a serious and pressing threat which individuals should be acting to counter.
In contrast, peak oil is only identified as an issue by the Transitions Network. Furthermore, of
the 12 organisations, eight offer advice on how individuals can reduce their carbon footprint
through saving energy in their homes. 11
A similar approach is taken to create affiliation matrices for encounters. Having an encounter
equates with having at least one of these three forms of contact: receiving an email, viewing a
website, seeing a brochure/magazine/leaflet. The affiliation network for encounters is
dichotomised at GE1 (density 23.7%).12 For paid support, the affiliation network is
dichotomised at GE2 (density 24.4%).13
The encounters and paid support matrices are each correlated against the attitudes matrix
using the quadratic assignment procedure (QAP). This allows us to check for association
between positions in the networks of each of encounters and paid support, and positions in the
12
attitude network. Secondly, we test these networks – of attitudes, encounters and paid support
– against a hypothetical block model, using deductive blockmodeling (Saunders 2011, see
Batagelj et al 2004 for more detail on the procedure). It is necessary to deploy a deductive
approach because otherwise blockmodels can be interpreted as little more than ‘highly
idealized patterns of interaction from the complex interweaving of thousands of paired
relationships’ (Breiger 1976:134). Indeed, social network analyst experts Wasserman and
Faust suggest that the inductive approach using CONCOR, frequently deployed by social
movement scholars (see for example Diani 1995, Diani and Bison 2004), should only be used
‘with a great deal of caution’ (Wasserman and Faust 1995: 381). This is partly because of the
‘fuzzy’ way it converts actors into approximately structurally equivalent actors (Scott 1992:
154).
Consequently, we prefer to use an entirely different algorithm: the optimized approach using
the software package Pajek (Batagelj and Mrvar 1998). This involves creating a hypothetical
blockmodel based on theoretical assumptions, loading it as an existing partition and asking
the software to optimize that partition. The software provides an error score which tell us how
many of the cells in our measured network matrix do not fit the anticipated pattern. The
optimize routine then moves as few actors around as possible to find a matrix with the
minimal number of errors. During the optimisation routine, the number of actors in each block
changes, but the hypothetical structure remains intact. Once actors have been allocated to
blocks, standard measures of association (we used Kendall’s tau) can be used to calculate the
strength of association between the hypothesized position of actors (in the loaded partition)
and their actual network position (in the final position).14
More explicitly, we start with a hypothetical block model derived from the proposition that
sharing contact with environmental organisations, and/or sharing attitudes are likely to relate
to household electricity use. We expect those with low electricity use to be in a ‘complete’
block. This means that we anticipate that all of them share at least two attitudes, one
encounter and or/paid support in two or more environmental organisations, and also share
attitudes regarding climate change, peak oil and individual efficacy. Those with highest
electricity use are expected to be in a block that is ‘null’. This means that, for each of our
three networks (attitudes, encounters and paid support), we expect none of the actors with
high electricity use to share any attitudes, encounters and/or paid support with other actors in
the network. Those with moderate energy use are expected to be in a ‘non-null’ block – that is,
we expect some of them, but not all of them, to share attitudes, encounters and paid support
13
with low energy users and other moderate electricity users. See Table 1 for the full
hypothetical block model.
Table 1: Hypothetical block model
Low electricity use
Moderate electricity use
High electricity use
Low electricity
use
Complete
Non-null
Null
Moderate electricity
use
Non-null
Non-null
Null
High electricity use
Null
Null
Null
Our hypothetical blockmodel has the advantage of allowing us to bypass a central weakness
of many approaches to blockmodeling: that blockmodels assume ‘pure’ structural equivalence
when it does not really exist. Frequently, blockmodelers assume that their blocks are either
complete (with ones in all cells of a block) or null (zeros in all cells of a block), when in
reality they are not. Rarely does pure structural equivalence occur in real social networks.
Instead, we hypothesise non-null cells (consisting of a mixture of 1s and 0s) in our
blockmodel. Doing this makes our results more realistic and increases the possibility of fitting
error-free matrices to the data. However, although we consider the approach preferable to
inductive use of the CONCOR algorithm, it should be noted that each actor ends up allocated
to one of several acceptable possible positions. In other words, the final error-free
blockmodels derived from the algorithm do not constitute a unique solution: others are
possible (Doreian et al 2005: 231). This is particularly the case when non-null blocks are prespecified.
Unfortunately, as with standard statistical analysis, our network analysis cannot confirm cause
and effect. We cannot be certain that encounters and/or paid support directly effect proenvironmental attitudes or behaviour. It may be that those with low electricity use join or
actively seek encounters with environmental groups because they are already environmentally
interested and have low electricity use. All we are able to identify is whether there is an
association between our variables of interest. But if an association is present this becomes
good grounds for future research on the broader impacts of social movements on the public,
and particularly on the role of environmental organisations in fostering pro-environmental
behaviours through less intense forms of engagement.
Results
Frequencies
14
Table 2 shows frequencies we derive from our survey data for attitudes and contact with
environmental organisations. Over two-thirds of our respondents agree that human-induced
climate change exists; half believe that individual action can make a difference; and a large
majority believe in peak oil. All but 11 respondents claim to have some form of contact with
the environmental organisations we listed in our survey. Over half have at least one
encounter and nearly three quarters pay to support more than one environmental organisation,
but very few were involved more intensely through meetings or volunteering. A similar
pattern emerges if we count the number of organisations respondents have contact with: on
average, respondents have encounters with 1.8 organisations, have financially supported 1.1
organisations and are involved in 0.1 organisations. We also correlate the three attitudes
against one another, and find a statistically significant relationship between those who agree
that climate change exists and that individual action can make a difference (Kendall’s Tau-b
0.34, p<0.01). In other words, most of the respondents who agree that climate change is
happening also believe that individual action can make a difference. Conversely, those that
disagree with one of those two statements tend to disagree with the other.
Table 2. Frequencies for attitudes and contact with environmental organisations
Survey item
n
%
Agree or strongly agree that human induced climate change exists
49
68.1
Agree or strongly agree that individual action can make a difference
36
50.0
Disagree or strongly disagree that current rates of oil production can continue in the long-run
62
86.1
At least one encounter with an environmental organisation
42
58.3
At least one paid support with an environmental organisation
52
72.2
Involved in at least one environmental organisation
6
8.3
Attitudes
Contact with environmental organisations
The average base-load of households in the sample is 121 watts (sd 73.13). T-tests do not
reveal any significant differences of mean base-loads between respondents who agreed or
disagreed with the three attitude statements or for respondents who were and were not in
contact with any of the national environmental organisations. In some cases, mean base-loads
are higher for respondents with pro-environmental attitudes or involvement (Table 3).
Table 3. Mean base-loads (in watts) and predictor variables
15
No
Yes
Anthropogenic
climate change
is happening
Peak oil is
happening
Individual
action can make
a difference
At least
one
encounter
Paid support in
at least one
organisation
Involvement
in at least one
organisation
128
118
119
123
133
121
109
119
109
129
121
121
To determine whether there is a relationship between contact with environmental
organisations and environmental attitudes, we conduct Wilcoxon rank sum tests using the
ordinal attitude variables and dummy ‘contact’ variables. Those who have contact with
environmental organisations rank relatively higher on the human-induced climate change
statement and all of those who made donations to environmental organisations rank higher on
the environmental attitude statements than their counterparts. However, none of these
differences are significant at the 5% level.
Network analysis
The QAP correlation scores between the attitude matrix (dichotomised at GE2 and GE3) and
each of the encounters (dichotomised at GE1, GE2 and GE3) and paid support (dichotomised
at GE2 and GE3) matrices are very low (<1%) and mostly insignificant. There is one
significant (p<0.05) correlation between the GE2 attitudes and GE2 paid support networks,
yet it is very weak (r=0.098). It tails off when the paid support network is dichotomised at
GE3. Thus, it is not the case that encounters or paid support in a higher number of
environmental organisations corresponds neatly with the sharing of any number of our three
attitudes. We therefore separately test the two networks (of encounters and paid support)
against the hypothetical block model instead of building a single multiplex network.
Attitudes and energy use
Table 4 shows the number of errors that the real data has in comparison to the hypothesized
blockmodel for the attitudes network. Of the 5184 cells in the matrix, 4104 are correctly
placed, with 1044 errors. In other words, 20% of the cells have errors. Recall that an error is
recorded if the cells in the matrix have 1s or 0s in places not anticipated by the hypothetical
blockmodel (Table 4).
Table 4: Initial ‘error matrix’ for the attitudes blockmodel
1
2
16
1
190
0
2
0
0
3
164
221
3
164
221
Initial error = 1044
84
The final blockmodel, which the software Pajek has optimized, has zero errors. This means
that all actors in the network neatly fit into one of the three structural blocks. It places three
low energy users in block 1, five high energy users in block 3, and the rest in block 2.
Twenty-nine out of 72 households have a structural position that differs from what we would
expect in our hypothetical block model. Put differently, 60% of actors in the network are
placed as we would expect them to be. We found a statistically significant association
between the position in the hypothetical block model (the initial matrix) and the final (errorfree) block model (Kendall’s tau-b = 0.451; p<0.01)
Encounters and energy use
The encounters network has 830 errors in comparison to the hypothesized blockmodel. Put
differently, 16% of the cells have errors. Of the 5184 cells in the encounters matrix, 4354 are
placed as we expected (Table 5). The final optimized blockmodel with zero errors places a
low energy user in block one, six high energy users in block 3, and all the other actors in
block 2. Thirty of our 72 households have a structural position in the encounters network that
differs from the hypothetical block model. 59.9% of the actors in the network are placed as
we hypothesized. There is a statistically significant association between the initial and final
block models (Kendall’s tau-b = 0.439, p<0.01).
Table 5: Initial ‘error matrix’ for the encounters block model
1
2
3
1
2
340
0
0
0
85
135
Initial error = 830
3
85
135
50
Paid support and energy use
Of the 5184 cells in the paid support matrix, 4670 have the structural position we
hypothesized, with 541 errors (Table 6). This means that fewer than 10% of the cells have
errors. The final error-free blockmodel misplaces a single high energy user in block one, but
correctly locates 12 high energy users in block 3. All of the other actors are placed in block 2.
Twenty-three of our 72 households have a structural position in the encounters network that
17
differs from the hypothetical block model. 68.1% of the actors in the network are placed as
we hypothesized. The high match between the initial and final matrices is confirmed by a
statistically significant association (Kendall’s tau-b = 0.546, p<0.01).
Table 6: Initial ‘error matrix’ for the paid support block model
1
2
3
1
2
460
0
0
0
7
17
Initial error = 514
3
7
17
6
Discussion
In our sample, a very high proportion of our respondents have encounters with at least one
environmental organisation that we asked about in our survey, despite very few being
involved in meetings or as a volunteer. Thus, the environmental organisations we examined
appear to effectively distribute their messages to a broad audience. This finding alone
highlights the importance of moving beyond the sub-cultural focus that we argue has
restricted research on the broader cultural impacts of social movements.
However, the absence of significant correlations between networks of attitudes and
encounters; and between attitudes and paid support suggests that, although environmental
organisations appear to have broad reach, their audience’s shared positions in the encounters
and paid support networks do not straightforwardly translate into shared positions in the
attitudes matrix. Nonetheless, our statistical analysis suggests that in cases where there is a
clear mismatch between attitudes and behaviour, encounters with environmental organisations
are conspicuously absent; and where there is more coherence, environmental encounters are
more prolific, though by no means ubiquitous. This relationship highlights a need for further
research. Although we are in no position to determine the causal mechanism, for example,
whether pro-environmental attitudes or behaviour lead to contact with environmental
organisations or vice versa, we have at least illustrated that our research agenda is worth
pursuing further.
Our blockmodeling results show that positions in the attitudes, encounters and paid support
networks each individually map well onto our hypothesized blockmodel. Whilst the three
attitudes we examine have little bearing on base-load electricity use when considered
18
separately (see Table 3), when we put actors into structurally equivalent blocks sharing at
least two attitudes, there is a close and statistically significant fit with our hypothetical block
model. Those with the lowest base-loads have proclivity to share two or more of our three
attitudes with all of the other low energy users. Thus, low energy users have a close fit with
the ‘complete’ cell of our hypothesized blockmodel (Table 1). Those with moderate energy
use tend to have at least two shared attitudes with some low energy users and with other
moderate energy users, but they infrequently share two attitudes with relatively high energy
users. Those with high energy use have a pattern of not sharing two or more attitudes with
anyone else in the network. The significant association between believing in
anthropogenically induced climate change and agreeing that individuals can make a
contribution towards solving climate change is crucial to understanding the structural
positions of actors in the attitudes matrix. Most of the respondents who had at least two
shared attitudes and therefore shared a position in the attitudes network (which, recall, was
dichotomised at GE2) shared these particular two attitudes. Thus, whilst belief in climate
change alone seems to make only a modest difference to base-loads (Table 3), belief in
climate change combined with belief in individual efficacy maps neatly on to our
hypothesised blockmodel. Thus, our analysis offers further empirical evidence to support
Ajzen and Fishbein’s (1980) work which suggests that gaps between attitudes and behaviours
are more likely to be bridged when individuals think that their behaviour will make a
difference.
The encounters network has even fewer errors than the attitudes network when compared to
the hypothetical blockmodel, and the paid support network is the best fitting of them all. This
is despite our exploratory statistical analysis (see Table 3) which suggests two unexpected
statistical relationships. Firstly, it indicates that the mean base-load of households with at least
one encounter with a specified environmental organisation is higher than households with no
encounters. Secondly, it suggests that households with and without paid support of at least
one of the environmental organisations we asked about have identical mean base-loads.
Central to explaining this apparent contradiction is the notion of shared positions in the
encounters network; and shared positions in the paid support network. This stresses the
importance of using a relational approach to the data: measures of association look only at
individuals, not at their position vis-à-vis others in the sample and can therefore miss
interesting patterns (Wellman 1988). Whereas straightforward linear analysis misses the
relationship between electricity use and contact with environmental organisations, we find
that our variables of interest are associated in more complex ways. Recall that households
19
with pro-environmental attitudes, but also those with high base-loads, were generally more
likely to have not had any contact with environmental organisations. Clearly there is a
relationship between contact with environmental organisations, attitudes and behaviour, even
if is not a straightforward linear one.
Supportive social networks that materialise through physical encounters are often deemed
necessary for shaping environmental behaviours (Horton 2005, Ingalsbee 1996). They are
also considered crucial for shaping norms that help people transfer their attitudes into actions
(Blake 1999). However, the strong association we have found between positions in our
encounters matrix and our hypothetical blockmodel suggests that physical encounters may not
be necessary to generate certain types of environmentally friendly behaviours. Non-physical
encounters such as sharing contact with or sharing paid support of environmental
organisations are strongly associated with low base-load electricity use. Of course, this
suggestion comes with caveats regarding our inability to determine causation.
Finally, the strong association between shared positions in the paid support and energy use
networks suggests that some accounts of chequebook membership may have overlooked the
action that chequebook members take outside of actual activism. Thus, whilst it might be true
that chequebook members avoid active political participation, preferring to pay for someone
else to campaign on their behalf, it seems less true that chequebook members are passive in
their environmental behaviour. Those who are structurally equivalent in a complete block in
the paid support network tend to have the lowest electricity base-loads, and those who are
structurally equivalent in the null block in the paid support network have higher electricity
base-loads. Thus it seems that chequebook members do actually make changes in their
everyday lives. This should not be belittled, for such lifestyle changes – no matter how small
– are a force for broader social change (Stern 1999). And, as our review of the literature
revealed some have even labelled lifestyle changes as social movements in their own right
(Cherry 2006).
Methodologically, this article has made a contribution to the literature by illustrating the
practice of deductive blockmodeling. In this research, hypothesising a non-null block
(consisting of a mixture of 1s and 0s) not only made theoretical sense, but also allowed us to
produce a series of error-free matrices, recognising that pure structural equivalence rarely
occurs in real data. Social movement scholars might learn from and extend the application of
deductive blockmodeling. The software Pajek allows not only identification of null, non-null
20
and complete blocks, but also of row equivalence and column equivalence, which could work
well for testing other theories related to social movement networks. For example, it could be
used to uncover relationships between network embeddedness and any number of cultural
outcomes. It could even be deployed in policy network studies to help understand the more
instrumental outcomes of movements. In particular, it would be well-suited to addressing the
question of whether organisations that interact most intensely with policy-makers are the most
likely to have their demands realised.
Conclusion
Our methodological innovation has produced fascinating results, which confirm our
hypothetical blockmodel. Although we are unable to prove cause and effect, it seems certain
that there is a relationship between the sharing of environmental attitudes and of contact with
environmental organisations in a sample of households in ONS4a2. Whilst we are unable to
generalise beyond this population category or beyond environmentalism, we have at the very
least established the need to address the broader impacts that social movements might have on
the behaviour of the public beyond the activist ghetto. The need to expand the scope of
studies of social movement impacts applies to a host of social movements, not just
environmental ones.
Our future research aims to expand the exploratory research we have offered in this article.
Our analysis here is restricted to 72 households due to the complexity of preparing the data at
the time of writing. But, as part of a larger interdisciplinary research team, we are collecting
data on the consumption practices, social networks and attitudes of over 180 households in
the southeast of England over a three-and-a-half year period. This means that we will have a
larger sample size, longitudinal data and information on a broad range of behaviours that go
beyond electricity base-load. We hope to set an agenda that takes seriously systematic
research on the impact of environmental organisations on the pro-environmental behaviours
of the public.
References
Abrahamse, W. and Steg, L. ( 2009) ‘How do socio-demographic and psychological factors relate to
households’ direct and indirect energy use and savings?’ Journal of Economic Psychology, 30 (5), 711720.
Abrahamse, W. and Steg, L. (2011) ‘Factors related to household energy use and intention to reduce it:
The role of psychological and socio-demographic variables’, Human Ecology Review, 18 (1), 30-40.
Ajzen, I. and Fishbein, M. (1980) Understanding Attitudes and Predicting Social Behaviour, Prentice-
21
Hall, Englewood Cliffs, NJ.
Barnes L.L. and Starr Sered, S. (2005) Religion and Healing in America, Oxford: Oxford University
Press.
Batagelj V. and Mrvar A (1998) ‘Pajek – Program for Large Network Analysis’, Connections, 21(2),
47-57.
Batagelj, V., Mrvar, A., Ferligoj, A. and Doreian, P. (2004) ‘Generalized Block Modelling with
Pajek’, Metodološki zvezki 1(2): 455-67.
Blake, J. (1999) ‘Overcoming the ‘value-action gap’ in environmental policy: Tensions between
national policy and local experience’, Local Environment, 4, 3: 257-278.
Breiger, R. (1976) ‘Career Attributes and Network Structure: A Blockmodel Study of a Biomedical
Research Speciality’, American Sociological Review 41(1): 117-135.
Büchs, M.; Edwards, R; Smith, G (2012) ‘Third sector organisations’ role in pro-environmental
behaviour change – a review of the literature and evidence’, Third Sector Research Centre, Working
Paper 81.
Cherry, E. (2006) ‘Veganism as a Cultural Movement: A Relational Approach’, Social Movement
Studies, 5(2), 155-170.
D’Anjou, L. (1996) Social Movements and Cultural Change: The First Abolition Campaign Revisited,
New York: Walter de Gruyter.
De Nooy, W., Mrvar A. and Batagelj, V.(2005) Exploratory Social Network Analysis with Pajek ,
Cambridge, Cambridge University Press.
Diani, M. (1995) Green Networks: A Structural Analysis of the Italian Environmental Movement.
Edinburgh: Edinburgh University Press.
Diani, M., and I. Bison. 2004. ‘Organizations, Coalitions and Movements’, Theory and Society 33(34): 281-309.
Doreian, P., V. Batagelj and A. Ferligoj (2005) Generalized Blockmodeling, Cambridge: Cambridge
University Press.
Emirbayer, M. (1997) ‘Manifesto for relational sociology’, American Journal of Sociology, 103(2):
281-317.
Energy Saving Trust (2012) [online]. Accessed 6th September 2012. Available from World Wide Web:
http://www.energysavingtrust.org.uk/Take-action/Save-money2
Energy Saving Trust (2012b) [online]. ‘Powering the nation – household electricity-using habits
revealed’, EST, London. Available from World Wide Web:
http://www.energysavingtrust.org.uk/Publications2/Corporate/Research-and-insights/Powering-thenation-household-electricity-using-habits-revealed
Epstein, S. (1998) Impure Science: AIDS, Activism and the Politics of Knowledge, Berkeley and Los
Angeles: University of California Press.
Ergas, C. (2010) ‘A Model of Sustainable Living: Collective Identity in an Urban Ecovillage’,
Organization and Environment, 23(1), 32-54.
Evans, D. and Jackson, T. (2007) ‘Towards a Sociology of Sustainable Lifestyles’, RESOLVE Working
Paper, 03-07.
Gamson W. and Modigliani, A. (1989) ‘Media discourses and public opinion on nuclear power: A
constructionist approach’, American Journal of Sociology, 95(1): 1-37.
Gamson, W. (1990) The Strategy of Social Protest, Belmont, CA: Wadsworth Publishing. Company
Georg, S. (1999), The social shaping of household consumption. Ecological Economics, 28, 3: 455466.
Gram-Hanssen, K. (2011). ‘Understanding change and continuity in residential energy consumption’,
22
Journal of Consumer Culture ,11(1): 61-78.
Haenfler, R., Johnson, B., and Jones, E. (2012) ‘Lifestyle Movements: Exploring the Intersection of
Lifestyle and Social Movements’, Social Movement Studies, 11(1): 1-20.
Hale, S. (2010) ‘The new politics of climate change: why we are failing and how we will succeed’,
Environmental Politics, 19:2, 255-274.
Hargreaves, T., Nye, M. and Burgess, J. (2008) ‘Social experiments in sustainable consumption: An
evidence-based approach with potential for engaging low-income communities’, Local Environment:
The International Journal of Justice and Sustainability, 13 (8), 743 – 758.
Heiskanen, E., Johnson, M., Robinson, S. Vadovics, E. and Saastamoinen, M. (2010). ‘Low-carbon
communities as a context for individual behaviour change’, Energy Policy, 38: 7586-7595.
Hetherington, K. (1998) Expressions of Identity: Place Power and Politics, London: Sage.
Hobson, K. (2003) ‘Thinking Habits into Action: the role of knowledge and process in questioning
household consumption practices’, Local Environment: The International Journal of Justice and
Sustainability, 8(1): 95 – 112.
Horton, D.( 2005) ‘Demonstrating Environmental Citizenship? A Study of Everyday Life Among
Green Activists’, in Andrew Dobson and Derek Bell (eds), Environmental Citizenship, MIT Press
pp.127-150.
House of Lords (2011) ‘Behaviour change, Science and Technology Select Committee, 2nd Report of
Session 2010–12, July 2011’, London: The Stationery Office.
Howell, R. (2009) ‘The Experience of Carbon Rationing Action Groups: Implications for a Personal
Carbon Allowances Policy’, Final Report. Environmental Change Institute, University of Oxford; UK
Energy Research Centre.
Ingalsbee, T. (1996). ‘Earth First! Activism: Ecological Postmodern Praxis in Radical
Environmentalist Identities’, Sociological Perspectives, 39(2), 263-276.
Inglehart, R. (1997) Modernization and Postmodernization: Cultural, Economic and Political Change
in 43 Societies, Princeton: Princeton University Press.
Jasamb, T. and Meier, H. (2010) ‘Household energy expenditure and income groups: Evidence from
Great Britain’, Electricity Policy Research Group Working Paper 1003, Cambridge Working Paper in
Economics 1011, University of Cambridge.
Jasper. J.M. (1997) The Art of Moral Protest: Culture, Biography and Creativity in Social Movements,
Chicago: University of Chicago Press.
Johnston, H. and Klandermans, B. (1995) Social Movements and Culture, University of Minnesota
Press.
Jordan, G. and Maloney, W. (1997) The Protest Business? Mobilizing Campaign Groups in Britain,
Manchester: Manchester University Press.
Kanter, R.M. (1972) Commitment and Community: Communes and Utopias in Sociological
Perspective, Harvard University Press.
McAdam, D. (1986) ‘Recruitment to high-risk activism: The Case of Freedom Summer’, American
Journal of Sociology, 92(1): 64-90.
Meier, A. and LeBot, B. (1999) ‘One Watt Initiative: a Global Effort to Reduce Leaking Electricity’,
Lawrence Berkeley National Laboratory: Lawrence Berkeley National Laboratory. Retrieved from:
http://escholarship.org/uc/item/9m94709c
Melucci, A. (1989) Nomads of the Present, Philadelphia: Temple University Press.
Meyer, D., Whittier N. and Robnett B. (2002) Social Movements, Identity, Culture and the State,
Oxford: Oxford University Press.
Middlemiss, L. K. (2008) ‘Influencing Individual Sustainability: a review of the evidence on the role
of community-based organisations’, International Journal of Environment and Sustainable
23
Development 7(1): 78-93.
Mitchell, J.C (1989) The Concept and Use of Social networks, in Mitchell J.C (ed) Social Networks in
Urban Situations, Manchester, Manchester University Press.
Saunders, C. (2007) ‘Using social network analysis to explore social movements: a relational
approach’, Social Movement Studies, 6, (3), 227-243.
Saunders, C. (2008) ‘Double edged swords? Collective identity and solidarity in the environment
movement’, British Journal of Sociology, 59, (2), 227-253.
Saunders , C. (2011) ‘Unblocking the pathway to effective use of blockmodels in social movement
research’, Mobilization, 16(3), 283-302
Scott, J. (2000) Social Network Analysis: A Handbook, London: Sage.
Seyfang, G. and Smith, A. (2007) ‘Grassroots Innovations for Sustainable Development: Towards a
New Research and Policy Agenda’, Environmental Politics, 16(4), 584-603.
Shove, E. (2003) Comfort, Cleanliness and Convenience. The social organization of normality,
Oxford: Berg.
Stern, P. (1999) ‘Toward a Coherent Theory of Environmentally Significant Behavior’, Journal of
Social Issues, 56(3), 407-424.
Steward, F., S. Liff, et al. (2009) ‘Mapping the Big Green Challenge. An analysis of 355 community
proposals for low carbon innovation,’ Appleby and Brook Lyndhurst for NESTA.
Wasserman S., and Faust, C. (1994) Social Network Analysis: Methods and Applications. Cambridge:
Cambridge University Press.
Wellman, B. (1988) Structural Analysis : From Method and Metaphor to Theory and Substance, in
Wellman B and Berkowitz S.D (eds), Social Structures : A Network Approach, Cambridge, Cambridge
University Press, 19-61.
Whittier, N. (2002) Meaning and structure in social movements, in Meyer, D., Whittier N. and
Robnett B. (eds) Social Movements, Identity, Culture and the State, Oxford: Oxford University Press:
289-309.
Acknowledgements
This research forms part of the Community-based Initiatives in Energy Saving project that is
funded by the UK Research Council’s (RCUK) Energy Programme. The Energy Programme
is a RCUK cross council initiative supported by EPSRC, ESRC, NERC, BBSRC and STFC.
The authors are grateful to the support and insights of their colleagues on the research project,
in particular Nicholas Bardsley, Patrick James, Tom Rushby, Nick Woodman and all of the
part-time researchers who have helped collect the data.
Notes
1
Overall the project will collect data from over 180 households, but at the time of writing we only have data
available from 72 dwellings.
2
See http://www.1010global.org/uk/about/faq, last accessed 08/02/12.
3
Household interviews confirmed that the broadband routers of households in our sample would have been left
switched on regardless of our monitoring equipment. The electricity consumed by routers would therefore have
contributed to the base-load level independently from the research.
24
Note that the term ‘base-load’ was consciously not mentioned in the social survey, to avoid influencing
respondents’ behaviour by driving attention to the base-load. Although the authors acknowledge the fact that the
survey itself could potentially influence participants, it could be argued that the process of installing the
monitoring equipment is a form of intervention with much greater impact, which is unavoidable.
5
Spearman’s rho for a correlation of household floor space against base-load is 0.3 (p<0.05). Spearman’s rho for
occupants and baseload is 0.1 (p=0.4) whilst it is significant for electricity use (rho=0.5, p<0.01). A t-test of
baseload and employment status dichotomised to focus on those at home (i.e. the unemployed, retired or house
wives/husbands) is also not significant at the 5% level.
6
Kendall’s tau(b) = 0.32 for switching appliances off standby and the aim to prevent climate change (p<0.01)
and 0.19 for switching appliances off standby and the aim to save money (not significant at 5% level).
7
This means that only 1.1% of possible links between our 72 households were realised.
8
We dichotomised the network at greater than or equal to 2 (GE2) because it represents the number of shared
attitudes above the mean (which was 2.0). This choice also made sense from a density point of view. Density
was 46% with GE2; with GE1, it was 81%. In other words, 81% of all possible ties were linked at GE1, making
a very dense. Such a high density would make it very difficult to distinguish between the positions of actors in
the network.
9
We are aware that Oxfam and World Development movement are not, strictly speaking, environmental
organisations. However, they have both played significant roles in campaigns against climate change.
10
Although Transition Towns has a local focus, there is a national hub for the network.
11
These are 10:10, Transition Towns, Oxfam, WWF, People and Planet, RSPB, Greenpeace and Friends of the
Earth.
12
Although the mean number of encounters was 1.8, we dichotomised this network at GE1 rather than GE2
because our respondents had, amongst them, fairly even frequencies of encounters with our 12 organisations,
except for World Development Movement, Stop Climate Chaos, Transition Towns and 10:10, which only one or
two of our respondents had encounters with. This made for a very sparse network at GE2, providing little detail
for analysis.
13
The GE1 network for paid support was even more dense (34.3%) because over 1/3 of the sample were paid
supporters of the National Trust and almost half the sample were paid supporters of Oxfam. We decided to
dichotomise this network at GE2 because of the distorting effect of membership of Oxfam and the National Trust.
Of the National Trust supporters, five also supported the RSPB and two the Wildlife Trust. Of the Oxfam
supporters, six supported RSPB and one supported the Wildlife Trust and Worldwide Fund for Nature.
14
In personal correspondence (01/03/12) Wouter de Nooy, co-author of the main text book on Pajek (de Nooy et
al 2005), suggested to the principal author to ‘use any statistic measure of association for two ordinal variables
(e.g., gamma, Somers'd, Kendall's tau) to calculate the strength of the association’ in order to yield a reasoned
indication of the fit between the blockmodel and energy use.
4
25