At what age is it OK?

advertisement
“AT WHAT AGE DO YOU THINK IT’S OK?”: THE SOCIAL
CLOCK FOR DRINKING AND DRUG USE AMONG ONTARIO
TEENAGERS1
Robin Room* and Angela Paglia**
*Centre for Social Research on Alcohol and Drugs, Stockholm University, Stockholm, Sweden
**Centre for Addiction and Mental Health, Toronto, Canada
BACKGROUND
Part of growing up is to try out and to take on new behaviours. While the process is often
fraught with anxiety for the person growing up, it is often even more anxiety-producing for
parents and other adults in the vicinity. The anxiety or disapproval may be about trying out the
behaviour at all. But often it is also about the age at which the behaviour is taken on. Behaviour
which is seen as too “grown up” for one age may be accepted without too much fuss if it occurs
at a later age.
In the context of discussions of social problems and youth, the focus tends to be on
behaviours that are taken on “too young”. But in a wider frame, there is also growing unease if a
young person does not try out and take on a behaviour at what is felt to be an appropriate age. It
may be seen as equally inappropriate to fail to have a full-time job by the age of 25 as it would
be to hold a full-time job at age 12. Sociologists talk of these normative standards for when a
behaviour or status should be taken on as the “social clock” (Neugarten et al., 1965). The
normative standards for the social clock for any given behaviour or status are likely to vary in
time and by cultural group.
In an earlier report (Paglia and Room, 1998), we considered the opinions of Ontario
residents aged 25 or older about contested behaviours. Altogether, 15 behaviours were asked
about, including not only cigarette smoking, beer drinking, and trying marijuana, but also a
variety of other behaviours -- e.g., driving a car alone, getting a fulltime job, having sex with a
girl/boyfriend, and moving in with a girl/boyfriend. When asked how old someone should be
before it’s OK to engage in each behaviour, some respondents answered “never” for most
behaviours, with the proportion ranging up to 52% for trying marijuana. Where ages were given,
there was a fairly close clustering in the mean ages for different behaviours, ranging from 17.4
for buying a lottery ticket to 20.1 for moving in with a girl/boyfriend. A random half of the
sample was asked the questions concerning males, and the other half concerning females. For
most behaviours, respondents were slightly more likely to say “never” concerning behaviours for
females, but the mean acceptable age tended to be slightly lower for females. We noted that for
cigarette smoking, alcohol consumption and marijuana use, the mean age of initiation for those
who did begin to use each substance was about 6 years below the mean age found acceptable by
adults.
Our findings raised the question of the views of young people themselves about the
“social clock” for contested behaviours. Did the discrepancy between when behaviours were
first tried on and when adults felt they were OK indicate a systematic normative difference
Prepared for presentation at an international research conference, “Youth cultures and subcultures: functions and
patterns of drinking and drug use”, Skarpö, Sweden, 23-26 April, 2001.
1
1
across the generational gap? Ed Adlaf and his colleagues afforded us the opportunity to
investigate this question with data collected from Ontario teenagers in the course of the 1997
Ontario Student Drug Use Survey (Adlaf et al., 1997b).
METHODS
Participants and Procedure
A total of 3,390 students from 168 high schools across the province of Ontario was
surveyed during the spring of 1997 by means of an anonymous self-administered questionnaire.
The cross-sectional probability survey is conducted biannually by the Centre for Addiction and
Mental Health (into which the Addiction Research Foundation was merged in 1998) and is used
to track consumption of alcohol, tobacco and other drugs by those in Grades 7, 9, 11, and 13.
The survey design was a stratified (grade by region) single-stage cluster sample of home-room
classes. The survey participation rate was 77%.
Students in grade 7 are usually aged 12 or 13, and students in each succeeding grade
studied are about two years older. Grade 13, now being phased out in Ontario, was viewed as a
preparation for university study, and is only attended by about 40% of the number of students
who are enrolled in Grade 11 (Adlaf et al., 1997b, p. 149). Students enrolled in Grade 13 are
thus a subgroup of their cohort, selected to a considerable extent on the basis of academic
achievement, whereas most members of their age cohort are among the students in Grades 7, 9
and 11.
Only a random half of the total sample received the questions examined in this report,
thereby reducing the sample size to 2,031. Further, the data were weighted by an average design
effect of 1.40 to adjust for the complex sampling procedures (i.e., non-simple random sampling).
As a result the effective sample size was 1,439. The mean age for the sample was 15.0 years
(SD=2.1); 53% were female.
Results from this student survey are compared in this paper with results from a generalpopulation survey of 1189 Ontario adults, interviewed by telephone by the Institute for Social
Research of York University in June, July, September and November, 1996 (Adlaf et al., 1997a).
The estimated response rate for the survey, conducted by random digit dialling, was 64%. Those
over 65 and with lesser education were somewhat underrepresented in the final sample.
Measures
Following other standard survey items about demographics and drug consumption,
students were presented with a series of questions in the form, “Regardless of what the law says,
at what age do you think it is OK for a male/female...” to engage in each behaviour. Students
had the choice of writing an age, choosing “it is never OK”, or the “don’t know” option. The
behaviours about which respondents were asked were:
•to smoke a cigarette
•to have a drink of beer
•to buy a six-pack of beer (6 cans packed together)
•to try some marijuana
These four behaviours were asked for a male protagonist and then for a female. The responses
concerning a female protagonist may thus have been influenced by the respondent’s prior
answers concerning a male protagonist.
Included in the analyses as independent variables were demographic measures, including
gender, grade, rural/urban residence, family socioeconomic status, school marks, and whether or
not a part-time job was held. Measures of the respondent’s history of substance use (i.e., ever
2
using tobacco, alcohol, or marijuana) were also used in the multivariate analyses.
In the adult survey, the same four questions about normative age were asked, in the
course of a series of 15 behaviours altogether. Unlike the student sample, adults were not
explicitly offered a “don’t know” choice, though it was coded if volunteered. A random half of
the adult sample was asked the questions for a male protagonist, while the others answered for a
female protagonist. Thus in this sample, the responses concerning males and concerning females
are independent.
RESULTS
Whether the behaviour is acceptable at all: students and adults
The left-hand portion of Table 1 shows the proportion of students in each grade who
responded it was “never OK” to engage in each of the four behaviours. On a lifetime basis,
drinking alcohol was clearly the most acceptable behaviour, and trying marijuana the least
acceptable. Overall, the students were somewhat less accepting of each behaviour than the
adults, except that adults were slightly less accepting of female cigarette smoking.
In the adult sample, acceptance of ever smoking a cigarette, trying marijuana, or buying a
six-pack of beer decreased with increasing age; for having a drink of beer, there was no
significant trend. In the student sample, 7th graders showed a relatively high intolerance of each
behaviour -- in 7 of 8 comparisons, a higher level than among adults aged 55 and older. But the
trend from grade 7 to grade 11 was quite strongly towards greater acceptance. For the alcohol
variables, this trend continued in grade 13, but not for the cigarette and marijuana variables. But
the selectivity involved in being in the grade 13 sample should be kept in mind in interpreting the
results.
Those in the student sample made little differentiation, on average, in the acceptability of
the behaviours for a male and for a female protagonist. This contrasted with the tendency
particularly for older adults to be less accepting of a female than of a male smoking cigarettes
and buying a 6 pack, and of younger adults to be less accepting of a female trying marijuana.
However, as noted above, the greater gender equality in the student sample may be artifactually
influenced by the fact that the respondent answered for a female directly after answering for a
male.
Mean acceptable age: students and adults
Turning to the mean acceptable age, among those giving an age (Table 2), among the
students the average ages when it is acceptable to smoke a cigarette, try marijuana, and have a
drink of beer are about the same, while the acceptable age for buying a six-pack of beer is over a
year older. The students give an acceptable age for each behaviour that is on average between
about one year and about 2½ years lower than that given by adults. The disparity in average ages
is highest for trying marijuana, and for males to have a drink of beer, while it is least for females
to buy a six-pack of beer.
Across the table as a whole, the mean acceptable age shows a curvilinear form for all
behaviours, with the nadir for all behaviours at the 9th or 11th grade. For smoking a cigarette,
trying marijuana and having a drink of beer, the mean acceptable ages given by 9th and 11th
graders is an age at which the protagonist would be in 9th or 10th grade. For buying a six-pack of
beer, a protagonist with the mean acceptable age would be in the 12th grade.
Standard deviations from the mean acceptable age tend to be slightly larger for the
student sample than for the adult sample, particularly for alcohol behaviours.
3
Predicting acceptability of behaviour among students
In Tables 3-5 , we turn to the question of what characteristics of students predict whether
or not they see cigarette smoking, trying marijuana, and having a drink of beer as ever
acceptable. These analyses are confined to grades 7, 9 and 11, given the selectivity inherent in
the grade 13 sample.
The results for cigarette smoking are shown in Table 3. The first column shows the
univariate odds ratio. Thus, for example, urban students are only two-thirds as likely as rural
students to find cigarette smoking ever acceptable. Students with lower school marks, and with a
lower than average family socioeconomic status, are significantly more accepting of cigarette
smoking than other students. As might be expected, approval of smoking among adults, and
whether the respondent has ever smoked, are both highly predictive of acceptance of smoking.
The first multivariate logistics regression (Model A) confines the predictors to variables
which are usually prior to the respondent’s own choices. It remains true, for the multivariate
analysis, that older students, rural students, and those from lower socio-economic status families
are more accepting of smoking. Males are somewhat less accepting than females, though the
relationship is not significant. At the bottom of the column, Nagelkerke’s R2 attempts to
quantify the proportion of explained “variance” in a logistic regression model. (Unlike multiple
regression, the variance here is restricted. Thus, this statistic must be interpreted with caution
and should only be taken as a general estimate of a model’s goodness-of-fit.) These logically
prior variables do not explain much of the variance in acceptance of smoking.
Lower school marks and the student working part-time, when added to the model (Model
B), significantly predict acceptance of smoking, while family SES loses its significance. This
might be interpreted in terms of family SES exerting an effect through lower school marks and
part-time working status. When variables closer to the dependent variable are added in Models
C, D, and E, the student’s grade-level, school marks, and part-time work status lose predictive
power, suggesting that the smoking attitude and behaviour variables intermediate between these
variables and acceptance of smoking. It is only with the addition of the smoking attitude and
behaviour variables that the variance explained becomes substantial.
Table 4 shows results with the same series of analyses for acceptance of drinking beer.
Both grade level and rural residence are stronger predictors of acceptance of drinking beer than
they are for acceptance of cigarette smoking. On the other hand, gender and family socioeconomic status have little predictive power. The relationship of school marks with acceptance
of drinking is reversed, although the differences are not significant. As with acceptance of
smoking, approval of daily drinking among adults, and consumption of alcohol oneself, are both
highly predictive of acceptance of drinking. But the total explained variance, after they are
included, is less than for acceptance of cigarette smoking.
Table 5 shows the results with acceptance of trying marijuana as the dependent variable.
In the univariate odds ratio, females are significantly more likely than males to accept trying
marijuana. Again, grade level and rural residence significantly predict higher acceptance. Low
school marks fairly strongly predict acceptance of marijuana, while family socioeconomic status
shows no significant effect. The attitude and behaviour variables are particularly strongly
predictive of accepting trying marijuana, intermediating the effects of the other variables
sufficiently that only the odds ratios for school marks remain significant. The variance
explained, once the attitude and behaviour items have been entered, is considerably higher than
for cigarette smoking and beer drinking.
4
Predicting OK to use at an age under 19
In Tables 6-8, we examine the predictors of acceptance of use at younger ages, among
those who accept use at any age. For cigarette smoking (Table 6), school grade and rural
residence predict acceptance of a younger age, while family socio-economic status and sex do
not show significant differences in odds ratios. Low school marks have some association with
acceptance of a lower age for cigarette smoking, but the effect is not as strong as for smoking at
all. Once the attitude and behaviour items are added, only the odds ration between school grade
levels remain significant. The predictive power of the variables for acceptance of a younger age
of smoking are roughly the same as those for acceptance of smoking at all.
For beer drinking (Table 7), school grade predicts acceptance of a younger age, as do
lower school marks, and, somewhat more weakly, rural residence. Gender and family socioeconomic status show little relation to acceptance of a younger age of drinking. The predictive
power of the variables for acceptance of a younger age is slightly higher than for acceptance of
drinking at all.
For trying marijuana (Table 8), school grade makes a particularly strong prediction of
acceptance of a younger age, and having low school marks is also predictive. The relationship
with family socio-economic status is curvilinear, with those from average SES families being
least accepting of trying marijuana at a younger age. While females were significantly more
accepting of trying marijuana at all, it is males who tend to be more accepting of trying it at a
younger age. The predictive power for acceptance of trying marijuana at a younger age is less
than for acceptance of trying it at all, but higher than the predictive power of the models for
cigarette smoking and for drinking beer.
DISCUSSION
As others have found, we found that students in 7th Grade were among the segments of
society least accepting of tobacco, alcohol or marijuana use, but that acceptance had increased by
11th Grade, for alcohol to levels equivalent to those among adults. In terms of the normative age
at which use was acceptable, we found that teenagers tended to set the age a little lower than
adults, and that the age was set lowest (15.8-17.6) by teenagers who were themselves at about
those ages.
Teenage users actually initiate use at ages which are considerably lower than the ages
they name as normative -- at a mean age of 12 for cigarettes and beer, and 14 for marijuana. In
the other direction, the legal age for cigarette smoking in Ontario is 18 and for drinking 19.
Teenagers are thus initiating these contested behaviours not only well below the legal age, and
below the age at which adults would consider it acceptable, but also at below the age thought
acceptable by teenagers themselves. The “social clock” expectations, to a considerable extent
shared between teenagers and adults, along with the legal age limits, may influence when and
how “precocious” behaviours occur. But the patterns we have described mean that use even of
legal psychoactive drugs is initiated outside the bounds of the public norms to which the
teenagers themselves subscribe. In terms of the relative timing in the individual life-course, it
seems as if new behaviour leads to new norms, rather than vice versa.
In terms of the predictors of acceptance at all, and of acceptance of early use, the patterns
differ quite considerably between smoking cigarettes, drinking beer, and trying marijuana. The
two constants are a greater acceptance at Grade 9 than at Grade 7, and an even greater
acceptance at Grade 11; and a greater acceptance in rural than in urban Ontario. Acceptance of
beer drinking, and of drinking at a relatively early age, does not differ much by gender or by
social class; lower school grades do predict acceptance of beer drinking at an earlier age.
5
Acceptance of cigarette smoking, on the other hand, is predicted by lower family socio-economic
status as well as by lower school marks, although these relations are weaker for acceptance of
smoking at a relatively early age. Acceptance of marijuana smoking shows yet another pattern:
females are more accepting of it at all, though males seem somewhat more likely to accept it at a
younger age. Though again there is a relation with low school marks, it is those with average
family socio-economic status who are least accepting of it.
These differences among teenagers in the acceptance of the different substances at all,
and in acceptance of using them at an early age, are presumably indications of styles and
preferences in different youth subcultures. These need to be studied directly with other methods
than a student drug survey. However, studies of youth cultures might well pay attention, too, to
the issue of the “social clock” of expected and acceptable behaviours at different ages, and to the
nuances of when and how people come to behave in ways that even they themselves in principle
do not support.
REFERENCES
Adlaf, E.M., Ivis, F.J., Ialomiteanu, A., Walsh, G. and Bondy, S. (1997a) Alcohol, Tobacco and
Illicit Drugs Use among Ontario Adults: 1977-1996: The Ontario Drug Monitor, 1996.
Toronto: Addiction Research Foundation, ARF Research Document 135.
Adlaf, E.M., Ivis, F.J. and Smart, R.G. (1997b) Ontario Student Drug Use Survey 1977-1997.
Toronto: Addiction Research Foundation, ARF Research Document 136.
Neugarten, B.L., Moore, J.W. and Lowe, J.C. (1965) Age norms, age constraints, and adult
socialization, American Journal of Sociology 40:710-717.
Paglia, A. and Room, R. (1998) How unthinkable and at what age?: adult opinions about the
“social clock” for contested behaviour by teenagers, Journal of Youth Studies 1:295-314.
6
Table 1. Percentage responding it is “Never OK” to engage in various behaviours, by grade (1997 Ontario student
survey), and by adult age group (1996 Ontario survey).
Grade
Adult Age Group
Behaviour
Total
7
9
11
13
Total
18-24
25-39
40-54
55+
Smoke a cigarette
30
37
30
24
32**
20
5
11
32
30***
Try marijuana
56
77
56
43
50***
46
22
35
53
70***
Have drink of beer
12
23
12
7
5**
3
2
1
3
6
Buy 6-pack of beer
14
27
14
8
5***
4
0
1
4
12***
Smoke a cigarette
32
37
33
26
34**
36
17
28
42
49***
Try marijuana
57
78
59
43
50***
53
42
46
56
68***
Have drink of beer
14
24
14
9
5***
4
5
2
7
6
Buy 6-pack of beer
16
30
16
10
6***
9
4
2
13
19***
1439
382
434
440
182
1214
159
416
355
253
By A Male:
By A Female:
Overall N:
Note: Ns vary by item due to “don’t know” or missing responses. For the adult survey, a random half of
respondents received questions about a male protagonist only, while the other half received questions about a
female. Comparisons between the students’ and adults’ opinions (Total columns) revealed significant differences
for all items (2 tests, p<.001), with two exceptions: no difference was found for female smoking a cigarette, and a
marginal difference (p<.09) was found for female trying marijuana.
**
Significant difference among subgroups in that row, based on 2 test, p<.01.
***
Significant difference among subgroups in that row, based on 2 test, p<.001.
7
Table 2. Mean (sd) acceptable age given, by grade (1997 Ontario student survey), and by adult age group (1996
Ontario survey).
Grade
Adult Age Group
Behaviour
Total
7
9
11
13
Total
18-24
25-39
40-54
55+
16.4
(2.5)
16.3
(2.6)
16.7
(2.6)
18.0
(2.3)
17.1a
15.9b
16.1b
17.1a
17.4
18.1
18.2
18.3
17.7a
15.9b
15.8b
17.0a
18.5a
18.8a
18.5a
20.4b
17.7a
16.4b
16.2b
17.3a
18.2a
19.0b
18.9b
19.0b
19.0a
17.9bd
17.6bc
18.3d
18.1
(2.4)
18.9
(2.3)
18.9
(1.6)
19.6
(1.7)
18.7a
19.8b
19.7b
19.6b
16.3
(2.9)
16.3
(2.6)
16.8
(2.8)
18.1
(2.3)
448948
17.1a
15.9b
15.8b
17.1a
17.9
17.6
18.2
17.8
17.2a
15.9b
16.0bc
17.0ac
18.1
18.5
18.6
19.2
17.7a
16.4b
16.2b
17.4a
18.3
18.6
18.8
18.5
19.0a
17.9bc
17.6b
18.3ac
19.0
19.2
19.6
19.3
57-186
130288
187337
71-149
4983
102210
By A Male:
Smoke a cigarette
Try marijuana
Have drink of beer
Buy 6-pack of beer
By A Female:
Smoke a cigarette
Try marijuana
Have drink of beer
Buy 6-pack of beer
N Range:
17.9
(2.5)
18.6
(2.2)
18.6
(1.8)
19.3
(1.7)
280577
70-172 34-113
Note: Means with the same subscript are not significantly different at p <. 05, based on the Scheffe comparison test.
Comparisons between the students’ overall means and adults’ overall means revealed significant differences (t-tests,
p<.001) for all items. N ranges in size due to the “never OK,” “don’t know” options or missing responses.
8
Table 3. Sequential/Hierarchical Logistic Regressions Predicting the Opinion that it is “OK” to Smoke (versus
Never OK), Ontario Student Drug Use Survey, 1997, N=1004 (entries are Odds Ratios).
Univariate
Odds Ratio
Model A
Model B
Model C
Model D
Model E
Male (vs Female)
0.91
0.92
0.80
0.80
0.76
0.84
Grade 9 (vs Grade 7)
Grade 11 (vs Grade 7)
1.38D
1.93A
1.43B
1.96A
1.41D
1.67B
1.48C
1.71B
1.28
1.42
1.03
1.00
Family SES:
Average (vs Above average)
Below Average (vs Above)
1.19
1.90C
1.15
1.81C
1.07
1.71
1.01
1.39
1.02
1.20
1.08
1.23
Urban (vs Rural):
0.66C
0.64B
0.63B
0.54B
0.61C
0.68D
School Marks:
B Average (vs A average)
C, D, or F Average (vs A)
1.61B
1.95A
1.62B
2.14A
1.83A
2.30A
1.64B
1.90B
1.49C
1.38
Part-time Work (vs no work):
1.50B
1.33D
1.37D
1.39D
1.26
1.26
1.60
1.38
1.00
1.05
1.12
1.00
0.90
1.21
4.85A
3.42A
Predictor
Perceived Risk from
Smoking:
No Risk (vs great risk)
Slight risk (vs great risk)
Medium risk (vs great risk)
1.06
1.56
1.29
Approve of Smoking Among
Adults (vs Disapprove):
4.77A
Ever Smoked
(vs Never Smoked)
5.04A
Model Improvement
Goodness of Fit Test *
Pseudo R2 (Nagelkerke’s)
3.60A
--
--
--
.04
2(3)=22.7,
p<.001
.06
2(3)=4.3,
p=0.23
.08
2(1)=54.3,
p<.001
.16
2(1)=47.6,
p<.001
.23
Note: Reference categories in brackets. All predictors were entered simultaneously within each model.
Significance was based on the Wald Test.
* Compares the fit of the model with the fit of the previous model.
A
p<.001.
B
p<.01.
C
p<.05.
D
p<.08.
9
Table 4. Sequential/Hierarchical Logistic Regressions Predicting the Opinion that it is “OK” to Drink Beer (versus
Never OK), Ontario Student Drug Use Survey, 1997, N=976 (entries are Odds Ratios).
Univariate
Odds Ratio
Model A
Model B
Model C
Model D
Model E
Male (vs Female)
1.08
1.13
1.13
0.92
0.83
0.94
Grade 9 (vs Grade 7)
Grade 11 (vs Grade 7)
2.23A
3.73A
2.54A
3.93A
2.36A
3.51A
2.48A
4.16A
2.30A
3.98A
1.90C
2.54B
Family SES:
Average (vs Above average)
Below Average (vs Above)
1.05
1.21
1.03
1.15
1.05
1.28
1.05
1.26
0.97
1.33
0.90
1.13
Urban (vs Rural):
0.55C
0.50B
0.45B
0.44C
0.47C
0.47C
School Marks:
B Average (vs A average)
C, D, or F Average (vs A)
0.90
0.94
0.81
0.91
0.86
0.97
0.84
0.84
0.74
0.75
Part-time Work (vs no work):
1.50C
1.18
1.28
1.28
1.17
1.43
2.36C
1.06
0.61
0.48
0.85
0.68
1.44
0.81
4.17A
3.44A
 (1)=18.5,
p<.001
.14
3.27A
 (1)=20.7,
p<.001
.19
Predictor
Perceived Risk from
Daily Drinking:
No Risk (vs great risk)
Slight risk (vs great risk)
Medium risk (vs great risk)
Approve of Daily Drinking
Among Adults (vs
Disapprove):
Ever Had a Drink of Alcohol
(vs Never):
Model Improvement
Goodness of Fit Test *
Pseudo R2 (Nagelkerke’s)
1.05
2.06C
0.96
3.81A
5.04A
--
--
--
.08
 (3)=1.1,
p=0.78
.07
2
 (3)=6.2,
p=0.10
.09
2
2
Note: Reference categories in brackets. All predictors were entered simultaneously within each model.
Significance was based on the Wald Test.
* Compares the fit of the model with the fit of the previous model.
A
p<.001.
B
p<.01.
C
p<.05.
10
2
Table 5. Sequential/Hierarchical Logistic Regressions Predicting the Opinion that it is “OK” to Try Marijuana
(versus Never OK), Ontario Student Drug Use Survey, 1997, N=954 (entries are Odds Ratios).
Univariate
Odds Ratio
Model A
Model B
Model C
Model D
Model E
Male (vs Female)
0.79D
0.76C
0.66B
0.65B
0.83
0.85
Grade 9 (vs Grade 7)
Grade 11 (vs Grade 7)
2.44A
4.36A
2.48A
4.54A
2.59A
4.52A
1.68C
2.22A
1.18
1.30
0.98
0.86
Family SES:
Average (vs Above)
Below Average (vs Above)
0.92
1.17
0.85
1.02
0.78
0.91
0.79
0.81
0.77
0.80
0.87
0.88
Urban (vs Rural):
0.74D
0.63B
0.62B
0.67C
0.76
0.71
School Marks:
B Average (vs A average)
C, D, or F Average (vs A)
1.44B
2.11A
1.48B
2.51A
1.57B
2.45A
1.60C
2.60A
1.48D
1.80C
Part-time Work
(vs no work):
1.21
0.95
1.02
1.04
0.87
15.4A
4.49A
2.08B
3.30A
2.14B
1.54
1.96C
1.81C
1.44
15.03A
10.65A
 (1)=187.4
p<.001
.53
6.01A
 (1)=45.3,
p<.001
.58
Predictor
Perceived Risk from Trying
Cannabis:
No Risk (vs great risk)
Slight risk (vs great risk)
Medium risk (vs great risk)
18.72A
5.36A
1.96B
Approve of Trying Cannabis
Among Adults
(vs Disapprove):
22.27A
Ever Used Cannabis
(vs Never Used)
Model Improvement
Goodness of Fit Test *
Pseudo R2 (Nagelkerke’s)
18.60A
--
--
--
.11
 (3)=23.2,
p<.001
.14
2
 (3)=131.1
p<.001
.32
2
2
Note: Reference categories in brackets. All predictors were entered simultaneously within each model.
Significance was based on the Wald Test.
* Compares the fit of the model with the fit of the previous model.
A
p<.001.
B
p<.01.
C
p<.05.
D
p<.08.
11
2
Table 6. Sequential/Hierarchical Logistic Regressions Predicting the Opinion that it is “OK” to Smoke Under 19
Years of Age (versus OK at Age 19 or Older), Ontario Student Drug Use Survey, 1997, N=698 (entries are Odds
Ratios).
Univariate
Odds Ratio
Model A
Model B
Model C
Model D
Model E
Male (vs Female)
1.04
1.06
0.99
1.07
1.04
1.24
Grade 9 (vs Grade 7)
Grade 11 (vs Grade 7)
2.84A
4.14A
2.80A
4.09A
2.80A
3.92A
2.68A
4.10A
2.46A
3.50A
2.03B
2.43B
Family SES:
Average (vs Above)
Below Average (vs Above)
0.99
1.20
0.97
1.20
0.87
1.03
0.85
0.97
0.82
0.88
0.90
0.92
Urban (vs Rural):
0.74
0.61A
0.64D
0.60D
0.67
0.77
School Marks:
B Average (vs A average)
C, D, or F Average (vs A)
1.12
1.83C
1.10
1.74D
1.04
1.48
0.90
1.20
0.77
0.89
Part-time Work
(vs no work):
1.38
1.07
1.00
1.01
0.92
Perceived Risk from
Smoking:
No Risk (vs great risk)
Slight risk (vs great risk)
Medium risk (vs great risk)
1.53
1.64
1.28
1.60
2.06
1.50
1.34
1.39
1.21
1.50
1.20
1.27
4.50A
3.63A
 (1)=33.3,
p<.001
.17
3.27A
 (1)=22.7,
p<.001
.22
Predictor
Approve of Smoking
Among Adults
(vs Disapprove):
5.42A
Ever Smoked (vs Never)
Model Improvement
Goodness of Fit Test *
Pseudo R2 (Nagelkerke’s)
5.21A
--
--
--
.09
 (3)=2.8,
p=.42
.09
2
 (3)=4.4,
p=.22
.09
2
2
Note: Reference categories in brackets. All predictors were entered simultaneously within each model.
Significance was based on the Wald Test.
* Compares the fit of the model with the fit of the previous model.
A
p<.001.
B
p<.01.
C
p<.05.
D
p<.08.
12
2
Table 7. Sequential/Hierarchical Logistic Regressions Predicting the Opinion that it is “OK” to Drink Beer Under
19 Years of Age (versus OK at Age 19 or Older), Ontario Student Drug Use Survey, 1997, N=822 (entries are Odds
Ratios).
Univariate
Odds Ratio
Model A
Model B
Model C
Model D
Model E
Male (vs Female)
1.07
1.12
0.99
0.95
0.80
0.95
Grade 9 (vs Grade 7)
Grade 11 (vs Grade 7)
2.75A
4.41A
2.83A
4.43A
2.94A
4.12A
2.99A
4.07A
2.71A
3.76A
2.14B
2.26B
Family SES:
Average (vs Above)
Below Average (vs Above)
0.98
1.23
0.98
1.25
0.90
1.17
0.90
1.13
0.89
1.23
0.87
1.11
Urban (vs Rural):
0.81
0.69
0.68D
0.74
0.80
0.82
School Marks:
B Average (vs A average)
C, D, or F Average (vs A)
1.55B
2.19A
1.52C
2.18B
1.35
1.94B
1.31
1.75C
1.26
1.68D
Part-time Work
(vs no work):
1.45A
1.11
1.22
1.26
1.13
Perceived Risk from Daily
Drinking:
No Risk (vs great risk)
Slight risk (vs great risk)
Medium risk (vs great risk)
2.06
1.54
0.87
3.02C
1.82C
0.95
1.92
1.18
0.78
1.95
1.10
0.72
3.22A
2.84A
 (1)= 24.3,
p<.001
.17
4.23A
 (1)=37.0,
p<.001
.23
Predictor
Approve of Daily Drinking
Among Adults
(vs Disapprove):
3.14A
Ever Used Alcohol
(vs Never Used)
Model Improvement
Goodness of Fit Test *
Pseudo R2 (Nagelkerke’s)
5.74A
--
--
--
.10
 (3)=9.0,
p=.03
.11
2
 (3)=12.4,
p=.01
.12
2
2
Note: Reference categories in brackets. All predictors were entered simultaneously within each model.
Significance was based on the Wald Test.
* Compares the fit of the model with the fit of the previous model.
A
p<.001.
B
p<.01.
C
p<.05.
D
p<.08.
13
2
Table 8. Sequential/Hierarchical Logistic Regressions Predicting the Opinion that it is “OK” to Try Marijuana
Under 19 Years of Age (versus OK at Age 19 or Older), Ontario Student Drug Use Survey, 1997, N=399 (entries
are Odds Ratios).
Univariate
Odds Ratio
Model A
Model B
Model C
Model D
Model E
Male (vs Female)
1.49
1.44
1.49
1.56
1.52
1.55
Grade 9 (vs Grade 7)
Grade 11 (vs Grade 7)
4.79A
7.89A
5.15A
7.71A
5.31A
7.99A
4.05A
4.75A
2.81B
3.11B
2.37C
2.29D
Family SES:
Average (vs Above)
Below Average (vs Above)
0.51C
1.25
0.55C
1.29
0.48C
1.62
0.38B
1.55
0.39B
1.46
0.45C
1.50
Urban (vs Rural):
0.88
0.71
0.68
0.56
0.66
0.62
School Marks:
B Average (vs A average)
C, D, or F Average (vs A)
0.91
2.20D
0.87
2.85A
0.90
2.28
0.95
2.20
0.78
1.55
Part-time Work
(vs no work):
1.23
0.95
0.86
1.00
0.85
9.91A
2.71C
1.39
6.36A
2.06
1.28
4.25C
1.88
1.28
2.97B
2.31C
 (1)=8.1,
p<.001
.33
4.91A
 (1)=14.3,
p<.001
.38
Predictor
Perceived Risk from Trying
Cannabis:
No Risk (vs great risk)
Slight risk (vs great risk)
Medium risk (vs great risk)
15.81A
3.79A
1.49
Approve of Trying Cannabis
Among Adults
(vs Disapprove):
Ever Used Cannabis
(vs Never Used)
Model Improvement
Goodness of Fit Test *
Pseudo R2 (Nagelkerke’s)
6.79A
12.21A
--
--
--
.18
 (3)=6.1,
p=.11
.21
2
 (3)=23.6,
p<.001
.30
2
2
Note: Reference categories in brackets. All predictors were entered simultaneously within each model.
Significance was based on the Wald Test.
* Compares the fit of the model with the fit of the previous model.
A
p<.001.
B
p<.01.
C
p<.05.
D
p<.08.
14
2
Download