Scandinavian Journal of Psychology, 1992, 33, 315-325
Adolescents’ alcohol use related t o perceived norms HENRIK AAS and KNUT-MGE KLEPP Research Center for Health Promotion, Department of Psychosocial Sciences, University of Bergen, Bergen, Norway
H. Aas & K.-I. Klepp, (1992). Adolescents’ alcohol use related to perceived norms. Scandinavian Journal of Psychology, 33, 31 5-325.
A questionnaire survey was camed out among 898 Norwegian adolescents age twelve to twenty. The study focused on the relation between adolescents’ alcohol use on one side and estimated drinking norms (peers’ drinking) together with attributed opinion norms (parents’ and friends’ acceptance of adolescents’ alcohol use) on the other. On average, 5”h of students in 71k,V h and 91hgrade reported weekly alcohol use. They overestimated the number of fellow students using alcohol every week seven-fold. Students’ estimates were significantly related to self-reported alcohol use. Even in high-school (loth, 1 I t h and 12Ih grade) where 29% reported weekly alcohol use, students overestimated number of weekly drinkers among friends and fellow students by two-fold. Contrary to the results from a previous study conducted in Norway, in this study opinions attributed to both friends and parents about adolescents’ alcohol use were significantly related to the students’ own alcohol use. Estimated behavior norms and attributed opinion norms explained 46% of the observed variance in students’ self-reported frequency of drinking. In a health promotion context, these results point to the importance of correcting student misconceptions about how often friends and feilow students drink alcohol. Key wordr: Alcohol, norms, adolescents, peers, parents. Henrik Aas. HEMIL-Center, Department of Psychosocial Sciences, University of Bergen, Oisteinsgate 3, 5007 Bergen, Norway
Alcohol consumption among Norwegian adolescents has remained at epidemic proportions throughout the 1980s (Irgens-Jensen, 1990; Klepp et al., 1991). Attempting to improve primary prevention programs, social psychological theory has been applied in the search for important predictors of adolescent alcohol use. In a review of primary prevention programs, Moskowitz (1989) stresses the importance of influencing social norms regrading drinking. He points out that our knowledge is too limited about how to influence these norms, and that most primary prevention programs have employed educational (“awareness”) approaches to influence attitudes toward drinking as the means of changing behaviors. These programs have had little short-term direct effect on drinking behavior (Moskowitz, 1989). In a metanalysis of school-based alcohol prevention programs, Rundall & Bruvold ( 1988) reached basically the same conclusions. They found that only 19 out of 31 alcohol studies reported some success in changing students’ attitudes, and that interventions relying upon social reinforcement, social norms, and developmental behavioral models were more effective than traditional “awareness” programs designed primarily to inform about health risks. Adolescent alcohol use is formed mainly through the social influences of culture, family and friends (Abrams & Niaura, 1987). Hollander (1976, p. 6) defines social norms as “standards of behavior expected from members of a given social group or an entire society”. In this paper, social norms encompass the pressure that arises from misinterpretations of significant others’ opinions and behaviors. Overestimation of peers’ (meaning both friends and other adolescents at comparable age) alcohol use creates pressure to conform to standards for alcohol consumption that is unrealistically high.
3 16
H. Aus and K . 4 Klepp
Scand I Psycho1 33 (1992)
In this paper alcohol use is conceived as learned behavior, which is influenced and changed through the process of socialization. Adolescents are actively and deliberately seeking information on alcohol use from their culture, family and peers (Bandura, 1969, 1986). In psychological literature the peer group’s influence is considered to be one of the most important determinants of adolescent development (Fisher & Bauman, 1988). Parents and peers can influence the onset of adolescent’s drinking behavior both directly and indirectly; indirectly through their attitudes, expectations and beliefs, and directly through their modeling of alcohol consumption and social reinforcement of drinking (Abrams Br Niaura, 1987). Adolescents develop their drinking behavior on the basis of others’ behavior, but also based on how they believe others drink. Through their search for social reinforcement, adolescents respond to what their significant others express, or to what the adolescents believe the significant others conceive as appropriate drinking behavior. Most cross-cultural and other studies on alcohol have found that parents and peers represent the two most important sources of influence for adolescents (Bank et al., 1985; Kandel, 1980; Kandel & Andrews. 1987). While the relationship between parental attitudes and behaviors toward alcohol and adolescent alcohol use appear to be a strong one (Barnes et af., 1986; Plant et ol., 1989). it is not necessarily linear (Johnson, 1986). However, recent studies have suggested that the commonly reported association between parental alcohol use and adolescent alcohol use may be strong for adolescents’ estimates of parents’ drinking only, and weak or non-existent for parents’ self-reported alcohol use (Hansen et af., 1987; Ruela & Fromme, 1991). Age tends to moderate the relationship between parental attitudes toward alcohol and adolescent alcohol use. Peer influences generally assume gradually greater importance as an individual moves from childhood through adolescence (Abrams & Niaura, 1987). Adolescents’ estimation of signiJicant others’ drinking
Several studies have concluded that adolescents tend to overestimate friends drinking (Baer et al., 1991a; Hauge, 1967; Marklund, 1984). Marklund (1984) suggests the following reasons for why adolescents often exaggerate when trying to estimate alcohol use among their friends and other adolescents: Young people do not talk seriously about alcohol and why they do or do not use it. Many adolescents brag about how much they drink, and how much they can drink without getting sick. Also, mass media coverage concerning alcohol and drug use among young people tends to focus upon problems associated with alcohol and delinquent teenagers in such a way that this minority of young people may be understood as being representative of most adolescents. Adolescents’ attribution of norms
~
10 parents
ond friendr
Among adolescents and in peer groups, some of the norms and rules for behavior are in harmony with, and some are opposed to the governing adult norms and regulations. Marklund (1984) stresses the lack of clarity that adolescents meet through their socialization toward adulthood by their significant others, concerning norms and expectations. As parents give more room to the adolescents’ developing personality, opportunities open up for other influencing authorities. Studies in the U.S. concerned with the normative inffuence of parents and friends on adolescent alcohol use have indicated that parents have greater influence (Biddle et al., 1980). There is, however, reason to be cautious about generalizing from American studies to the Norwegian population. In a carefully designed crosscultural interview study, Bank and colleagues (1985) report that norms attributed to parents had a significantly positive effect on adolescent drinking in samples in the U.S. and Australia, but that no such significant relation was found in either Norway or France. It is clear, however, that parents’ norms were
S a n d J Psycho1 33 (1992)
Adolescent alcohol use and norms
positively, although not significantly related to adolescents’ self-reported drinking in Norway. Norms attributed to friends had a significantly positive effect in all four countries. Relative importance of parents’ and friends’ norms was reversed from the U.S. (parents most important) to France (friends most important). In Norway, friends’ norms were more strongly than parents’ norms associated to self-reported frequency of drinking, but the difference between friends’ and parents’ influence was not significant. Research questions
The first part of this study focuses on norms for alcohol use developed from perception of-and beliefs about-others’ drinking, and addresses the following questions: a) Do many adolescents overestimate friends’ and other adolescents’ use of alcohol? b) Is the estimated alcohol consumption of friends and fellow students related to the adolescents’ own drinking? c) Are the answers to these questions different for junior high-school and high-school, and are they gender-specific? The second part deals with adolescents’ attribution of norms concerning adolescent alcohol use to parents and friends, and examines the following questions: a) Is acceptance of adolescents’ alcohol use, which is attributed to friends and parents, associated with drinking among the adolescents? b) If so, are norms attributed to parents or to friends the better predictor? c) Are the associations between the adolescents’ alcohol use and these attributed norms changing from 7‘h grade through 81h and gthand up to high-school, and are they gender-specific? We will also examine to what extent estimates of other’s alcohol use and attributed opinion norms separately and in combination can predict adolescents’ self-reported alcohol use.
METHOD Subjects and procedures A total of 898 students (49.9% male), grades 7 through 12 were included in the analyses in this study. Mean age of students was 13.0 years (SD 0.2) in 7Ih grade, 13.9 (SD 0.3) in 8Ih grade. 14.9 (SD 0.4) in 9Ih grade and 17.6 years (range I5 to 20 years, SD 1.2) in high-school (which includes grades 10, 1 I and 12). There were relatively few high-school students participating, and we therefore combined all students from high-school classes in this study. Subjects were recruited from 23 schools located in both rural and urban districts of Hordaland county in western Norway. Participation in the study was voluntary and anonymous. Teachers, who had signed up for a drug related course, administered a questionnaire to all students present in the classrooms the day of the survey. We d o not have an exact figure for the number of absent students, but data from similar studies [Wold, 1989) show that approximately 5% of students are absent on any given day. Students older than age 20 ( n =34) and those who gave inconsistent answers on key questions (n = 1) or had missing data on gender or on alcohol use ( n = 10) were excluded from all analyses in this study. After exclusions, our sample consisted of 95‘X of the 943 students who returned the questionnaire. Insrrument
The questionnaire included 5 demographic questions, 26 alcohol related questions, 3 questions assessed smoking, and 4 assessed other drug use. Less than 40 minutes was needed to complete the questionnaire. Alcohol use was measured by the following question: “How often have you been drinking alcohol the last three months?’. The 8 answer categories ranged from “every day” to ”not at all“. Information about the estimated drinking by friends and other reference groups was gained by the way of 2 questions: “How many of your friends [. . . of the students at your school.. .] do you think drink alcohol every week?’ Each question had 7 possible response categories ranging from ”all” (coded as 100%) to “none” (coded as O’%).’ The question about “students at your school“ was chosen because ‘Estimates of friends’ and of other students‘ weekly alcohol use were coded as percent-scores based on seven point scales to obtain comparable data to frequency of self-reported alcohol use. As a check for this coding, regression on frequency of self-reported alcohol use resulted in almost identical adjusted R’ using both percent-scores scales and 1-7 scales.
317
318
H . Aas and K.-I.Klepp
S a n d J Psycho1 33 (1992)
fellow students are believed to have an important role for setting standards for appropriate drinking behavior at different grades among adolescents. To obtain a measure of what norms the subjects attributed to parents and friends, the subjects were asked-2 questions: “How do you think your friends [, . . your parents. . .] feel about people at your age drinking alcohol?” Both questions had the same 4 response options; “approve”, “accept”, “disapprove” and “strongly disapprove”.
RESULTS Table 1 presents self-reported frequency of drinking over the last three months for boys and girls at each grade level. The table shows that frequency of drinking increased with grade ( x 2 = 274.88, df = 12, p < 0.001). Boys reported drinking slightly more frequently than did girls ( x 2 = 9.70, df = 4, p < 0.05), due to more frequent drinking by boys than by girls in high-school. The average proportion of students who reported consuming alcohol at least once a week was 5% in junior high-schools and 29% in high-schools. Frequency of drinking alcohol last 3 months was chosen as dependent measure in the following analyses as it was highly correlated with other measures of alcohol use including; typical quantity of alcohol consumed last 3 months ( r =0.81), and frequency of drinking occasions last 3 months on which more than 5 drinks were consumed (r = 0.86). Table 2 presents ranges, means and standard deviations for the independent variables for both boys and girls. Pearson correlations and partial correlations between each independent variable and self-reported frequency of drinking are also displayed. As Table 2 shows, all four independent variables were positively correlated with both boys’ and girls’ self-reported alcohol use. Estimated number of “students at school drinking weekly” was the only independent variable with a small, but significant gender difference on oneway analysis of variance (ANOVA) ( F = 9.42, df = 1, 790, p < 0.01). ANOVAs for 71h, 81h, gthgrade and high-school separately revealed that only girls in 8 I h grade had significantly higher estimates than boys ( F = 5.43, df = 1, 212, p c 0.05). As expected, the partial correlations presented in Table 2 were somewhat lower than the ordinary zero-order Pearson correlations. They did, however, demonstrate that all relationships between the independent variables and alcohol use stayed significant ( p < 0.001) when grade was controlled for. The first column in Figs. 1 and 2 shows “percentage of self-reported alcohol use”, which is percentage (based on the actual number) of boys and girls in each age group that reported drinking alcohol at least once a week (see Table I). The next two columns Figs. 1 and 2 present means of “estimated percentage of friends that drink weekly” and “estimated percentage of students that drink weekly” based on the answers from each gender separately.
Table 1. Percent self-reported frequency of alcohol use during the last three months, for boys and girls at each grade level (N = 898) Boys Category
Girls
71h
8th
n=64
n=124
9‘h n=154
High S. Number 71h n = 1 0 6 ofboys n = 6 2
63 26 5 4 2
43 35 14 7 1
14 26 23 22 15
Not in 3 months 84 Max 1 per month 12 2-3 times a month 0 Once a week 2 More often 2
213 122 52
40 21
79 15 3 3 0
Sth 91h High S. Number n = 114 n = I53 n = 121 ofgirls 76 17 4 3 0
39 42 14 4 1
13 43 22 17 5
21 1 145 54 33 7
Adolescent alcohol use and norms
S a n d J Psycho1 33 (1992)
Table 2. Mean and standard deuiation of the independenr variables by gender, and correlation and partial correlation
10
self-reported frequency of drinking
Boys
Girls
Variable
Range
Mean
SD
r*
rp**
Mean
SD
r*
rp**
Friends drinking weekly Students drinking weekly” Friends’ approval Parents’ approval
(1-7t) (1-7t) (I-4+) (1-4+)
3.3 3.4 2.9 1.8
5 1.9 kl.2 k0.8 +0.7
0.62 0.29 0.48 0.38
0.50 0.17 0.36 0.23
3.3 3.6 2.8 1.7
k1.9
0.64
0.57 0.18 0.33 0.23
k1.1 0.29 k0.7
k0.7
0.46 0.40
*Pearson correlation coefficient between independent variables and alcohol frequency. **Partial correlations in which the effects of grade are removed. All correlations were significant at p
Adolescents’ alcohol use related t o perceived norms HENRIK AAS and KNUT-MGE KLEPP Research Center for Health Promotion, Department of Psychosocial Sciences, University of Bergen, Bergen, Norway
H. Aas & K.-I. Klepp, (1992). Adolescents’ alcohol use related to perceived norms. Scandinavian Journal of Psychology, 33, 31 5-325.
A questionnaire survey was camed out among 898 Norwegian adolescents age twelve to twenty. The study focused on the relation between adolescents’ alcohol use on one side and estimated drinking norms (peers’ drinking) together with attributed opinion norms (parents’ and friends’ acceptance of adolescents’ alcohol use) on the other. On average, 5”h of students in 71k,V h and 91hgrade reported weekly alcohol use. They overestimated the number of fellow students using alcohol every week seven-fold. Students’ estimates were significantly related to self-reported alcohol use. Even in high-school (loth, 1 I t h and 12Ih grade) where 29% reported weekly alcohol use, students overestimated number of weekly drinkers among friends and fellow students by two-fold. Contrary to the results from a previous study conducted in Norway, in this study opinions attributed to both friends and parents about adolescents’ alcohol use were significantly related to the students’ own alcohol use. Estimated behavior norms and attributed opinion norms explained 46% of the observed variance in students’ self-reported frequency of drinking. In a health promotion context, these results point to the importance of correcting student misconceptions about how often friends and feilow students drink alcohol. Key wordr: Alcohol, norms, adolescents, peers, parents. Henrik Aas. HEMIL-Center, Department of Psychosocial Sciences, University of Bergen, Oisteinsgate 3, 5007 Bergen, Norway
Alcohol consumption among Norwegian adolescents has remained at epidemic proportions throughout the 1980s (Irgens-Jensen, 1990; Klepp et al., 1991). Attempting to improve primary prevention programs, social psychological theory has been applied in the search for important predictors of adolescent alcohol use. In a review of primary prevention programs, Moskowitz (1989) stresses the importance of influencing social norms regrading drinking. He points out that our knowledge is too limited about how to influence these norms, and that most primary prevention programs have employed educational (“awareness”) approaches to influence attitudes toward drinking as the means of changing behaviors. These programs have had little short-term direct effect on drinking behavior (Moskowitz, 1989). In a metanalysis of school-based alcohol prevention programs, Rundall & Bruvold ( 1988) reached basically the same conclusions. They found that only 19 out of 31 alcohol studies reported some success in changing students’ attitudes, and that interventions relying upon social reinforcement, social norms, and developmental behavioral models were more effective than traditional “awareness” programs designed primarily to inform about health risks. Adolescent alcohol use is formed mainly through the social influences of culture, family and friends (Abrams & Niaura, 1987). Hollander (1976, p. 6) defines social norms as “standards of behavior expected from members of a given social group or an entire society”. In this paper, social norms encompass the pressure that arises from misinterpretations of significant others’ opinions and behaviors. Overestimation of peers’ (meaning both friends and other adolescents at comparable age) alcohol use creates pressure to conform to standards for alcohol consumption that is unrealistically high.
3 16
H. Aus and K . 4 Klepp
Scand I Psycho1 33 (1992)
In this paper alcohol use is conceived as learned behavior, which is influenced and changed through the process of socialization. Adolescents are actively and deliberately seeking information on alcohol use from their culture, family and peers (Bandura, 1969, 1986). In psychological literature the peer group’s influence is considered to be one of the most important determinants of adolescent development (Fisher & Bauman, 1988). Parents and peers can influence the onset of adolescent’s drinking behavior both directly and indirectly; indirectly through their attitudes, expectations and beliefs, and directly through their modeling of alcohol consumption and social reinforcement of drinking (Abrams Br Niaura, 1987). Adolescents develop their drinking behavior on the basis of others’ behavior, but also based on how they believe others drink. Through their search for social reinforcement, adolescents respond to what their significant others express, or to what the adolescents believe the significant others conceive as appropriate drinking behavior. Most cross-cultural and other studies on alcohol have found that parents and peers represent the two most important sources of influence for adolescents (Bank et al., 1985; Kandel, 1980; Kandel & Andrews. 1987). While the relationship between parental attitudes and behaviors toward alcohol and adolescent alcohol use appear to be a strong one (Barnes et af., 1986; Plant et ol., 1989). it is not necessarily linear (Johnson, 1986). However, recent studies have suggested that the commonly reported association between parental alcohol use and adolescent alcohol use may be strong for adolescents’ estimates of parents’ drinking only, and weak or non-existent for parents’ self-reported alcohol use (Hansen et af., 1987; Ruela & Fromme, 1991). Age tends to moderate the relationship between parental attitudes toward alcohol and adolescent alcohol use. Peer influences generally assume gradually greater importance as an individual moves from childhood through adolescence (Abrams & Niaura, 1987). Adolescents’ estimation of signiJicant others’ drinking
Several studies have concluded that adolescents tend to overestimate friends drinking (Baer et al., 1991a; Hauge, 1967; Marklund, 1984). Marklund (1984) suggests the following reasons for why adolescents often exaggerate when trying to estimate alcohol use among their friends and other adolescents: Young people do not talk seriously about alcohol and why they do or do not use it. Many adolescents brag about how much they drink, and how much they can drink without getting sick. Also, mass media coverage concerning alcohol and drug use among young people tends to focus upon problems associated with alcohol and delinquent teenagers in such a way that this minority of young people may be understood as being representative of most adolescents. Adolescents’ attribution of norms
~
10 parents
ond friendr
Among adolescents and in peer groups, some of the norms and rules for behavior are in harmony with, and some are opposed to the governing adult norms and regulations. Marklund (1984) stresses the lack of clarity that adolescents meet through their socialization toward adulthood by their significant others, concerning norms and expectations. As parents give more room to the adolescents’ developing personality, opportunities open up for other influencing authorities. Studies in the U.S. concerned with the normative inffuence of parents and friends on adolescent alcohol use have indicated that parents have greater influence (Biddle et al., 1980). There is, however, reason to be cautious about generalizing from American studies to the Norwegian population. In a carefully designed crosscultural interview study, Bank and colleagues (1985) report that norms attributed to parents had a significantly positive effect on adolescent drinking in samples in the U.S. and Australia, but that no such significant relation was found in either Norway or France. It is clear, however, that parents’ norms were
S a n d J Psycho1 33 (1992)
Adolescent alcohol use and norms
positively, although not significantly related to adolescents’ self-reported drinking in Norway. Norms attributed to friends had a significantly positive effect in all four countries. Relative importance of parents’ and friends’ norms was reversed from the U.S. (parents most important) to France (friends most important). In Norway, friends’ norms were more strongly than parents’ norms associated to self-reported frequency of drinking, but the difference between friends’ and parents’ influence was not significant. Research questions
The first part of this study focuses on norms for alcohol use developed from perception of-and beliefs about-others’ drinking, and addresses the following questions: a) Do many adolescents overestimate friends’ and other adolescents’ use of alcohol? b) Is the estimated alcohol consumption of friends and fellow students related to the adolescents’ own drinking? c) Are the answers to these questions different for junior high-school and high-school, and are they gender-specific? The second part deals with adolescents’ attribution of norms concerning adolescent alcohol use to parents and friends, and examines the following questions: a) Is acceptance of adolescents’ alcohol use, which is attributed to friends and parents, associated with drinking among the adolescents? b) If so, are norms attributed to parents or to friends the better predictor? c) Are the associations between the adolescents’ alcohol use and these attributed norms changing from 7‘h grade through 81h and gthand up to high-school, and are they gender-specific? We will also examine to what extent estimates of other’s alcohol use and attributed opinion norms separately and in combination can predict adolescents’ self-reported alcohol use.
METHOD Subjects and procedures A total of 898 students (49.9% male), grades 7 through 12 were included in the analyses in this study. Mean age of students was 13.0 years (SD 0.2) in 7Ih grade, 13.9 (SD 0.3) in 8Ih grade. 14.9 (SD 0.4) in 9Ih grade and 17.6 years (range I5 to 20 years, SD 1.2) in high-school (which includes grades 10, 1 I and 12). There were relatively few high-school students participating, and we therefore combined all students from high-school classes in this study. Subjects were recruited from 23 schools located in both rural and urban districts of Hordaland county in western Norway. Participation in the study was voluntary and anonymous. Teachers, who had signed up for a drug related course, administered a questionnaire to all students present in the classrooms the day of the survey. We d o not have an exact figure for the number of absent students, but data from similar studies [Wold, 1989) show that approximately 5% of students are absent on any given day. Students older than age 20 ( n =34) and those who gave inconsistent answers on key questions (n = 1) or had missing data on gender or on alcohol use ( n = 10) were excluded from all analyses in this study. After exclusions, our sample consisted of 95‘X of the 943 students who returned the questionnaire. Insrrument
The questionnaire included 5 demographic questions, 26 alcohol related questions, 3 questions assessed smoking, and 4 assessed other drug use. Less than 40 minutes was needed to complete the questionnaire. Alcohol use was measured by the following question: “How often have you been drinking alcohol the last three months?’. The 8 answer categories ranged from “every day” to ”not at all“. Information about the estimated drinking by friends and other reference groups was gained by the way of 2 questions: “How many of your friends [. . . of the students at your school.. .] do you think drink alcohol every week?’ Each question had 7 possible response categories ranging from ”all” (coded as 100%) to “none” (coded as O’%).’ The question about “students at your school“ was chosen because ‘Estimates of friends’ and of other students‘ weekly alcohol use were coded as percent-scores based on seven point scales to obtain comparable data to frequency of self-reported alcohol use. As a check for this coding, regression on frequency of self-reported alcohol use resulted in almost identical adjusted R’ using both percent-scores scales and 1-7 scales.
317
318
H . Aas and K.-I.Klepp
S a n d J Psycho1 33 (1992)
fellow students are believed to have an important role for setting standards for appropriate drinking behavior at different grades among adolescents. To obtain a measure of what norms the subjects attributed to parents and friends, the subjects were asked-2 questions: “How do you think your friends [, . . your parents. . .] feel about people at your age drinking alcohol?” Both questions had the same 4 response options; “approve”, “accept”, “disapprove” and “strongly disapprove”.
RESULTS Table 1 presents self-reported frequency of drinking over the last three months for boys and girls at each grade level. The table shows that frequency of drinking increased with grade ( x 2 = 274.88, df = 12, p < 0.001). Boys reported drinking slightly more frequently than did girls ( x 2 = 9.70, df = 4, p < 0.05), due to more frequent drinking by boys than by girls in high-school. The average proportion of students who reported consuming alcohol at least once a week was 5% in junior high-schools and 29% in high-schools. Frequency of drinking alcohol last 3 months was chosen as dependent measure in the following analyses as it was highly correlated with other measures of alcohol use including; typical quantity of alcohol consumed last 3 months ( r =0.81), and frequency of drinking occasions last 3 months on which more than 5 drinks were consumed (r = 0.86). Table 2 presents ranges, means and standard deviations for the independent variables for both boys and girls. Pearson correlations and partial correlations between each independent variable and self-reported frequency of drinking are also displayed. As Table 2 shows, all four independent variables were positively correlated with both boys’ and girls’ self-reported alcohol use. Estimated number of “students at school drinking weekly” was the only independent variable with a small, but significant gender difference on oneway analysis of variance (ANOVA) ( F = 9.42, df = 1, 790, p < 0.01). ANOVAs for 71h, 81h, gthgrade and high-school separately revealed that only girls in 8 I h grade had significantly higher estimates than boys ( F = 5.43, df = 1, 212, p c 0.05). As expected, the partial correlations presented in Table 2 were somewhat lower than the ordinary zero-order Pearson correlations. They did, however, demonstrate that all relationships between the independent variables and alcohol use stayed significant ( p < 0.001) when grade was controlled for. The first column in Figs. 1 and 2 shows “percentage of self-reported alcohol use”, which is percentage (based on the actual number) of boys and girls in each age group that reported drinking alcohol at least once a week (see Table I). The next two columns Figs. 1 and 2 present means of “estimated percentage of friends that drink weekly” and “estimated percentage of students that drink weekly” based on the answers from each gender separately.
Table 1. Percent self-reported frequency of alcohol use during the last three months, for boys and girls at each grade level (N = 898) Boys Category
Girls
71h
8th
n=64
n=124
9‘h n=154
High S. Number 71h n = 1 0 6 ofboys n = 6 2
63 26 5 4 2
43 35 14 7 1
14 26 23 22 15
Not in 3 months 84 Max 1 per month 12 2-3 times a month 0 Once a week 2 More often 2
213 122 52
40 21
79 15 3 3 0
Sth 91h High S. Number n = 114 n = I53 n = 121 ofgirls 76 17 4 3 0
39 42 14 4 1
13 43 22 17 5
21 1 145 54 33 7
Adolescent alcohol use and norms
S a n d J Psycho1 33 (1992)
Table 2. Mean and standard deuiation of the independenr variables by gender, and correlation and partial correlation
10
self-reported frequency of drinking
Boys
Girls
Variable
Range
Mean
SD
r*
rp**
Mean
SD
r*
rp**
Friends drinking weekly Students drinking weekly” Friends’ approval Parents’ approval
(1-7t) (1-7t) (I-4+) (1-4+)
3.3 3.4 2.9 1.8
5 1.9 kl.2 k0.8 +0.7
0.62 0.29 0.48 0.38
0.50 0.17 0.36 0.23
3.3 3.6 2.8 1.7
k1.9
0.64
0.57 0.18 0.33 0.23
k1.1 0.29 k0.7
k0.7
0.46 0.40
*Pearson correlation coefficient between independent variables and alcohol frequency. **Partial correlations in which the effects of grade are removed. All correlations were significant at p
