Reference group influence on adolescent alcohol use.

Abstract

This study assessed the relative influence of the perceived level of support for an adolescent's alcohol use by best friends, most students in school, live-in parents/guardians and most adults in neighorbhood/community on adolescent current drinking, heavy drinking and frequent drinking prevalence rates. Anonymous self-reported questionnaires were completed by 83,669 students, most of the 6th, 8th, and 11th grades enrolled in Iowa's public schools. Best friends perceived level of support proved to be most highly correlated with the decision by adolescents to use alcohol, but there was evidence that the influence of best friend's unequivocal approval of drinking could be significantly moderated by the perceived unequivocal disapproval of drinking by the other reference groups. There were some differences in the findings by sex and grade in school, as well as the different kinds of drinking behavior (any use and extent of use) investigates, and some, but not most, may have practical prevention planning consequences. Primary and secondary alcohol prevention programs that are designed to encourage unequivocal disapproval of drinking by any of the four reference groups have the potential to significantly reduce adolescent drinking-related prevalence rates. Prevention programs that include all four reference groups would be expected to have the most impact on adolescent drinking behaviors.

**********

Reference Group Influence on Adolescent Alcohol Use

Several studies have consistently confirmed that various attributes in four social environments (peer, school, home and neighborhood/community) can have significant direct and mediating effects on adolescents' drinking related behaviors (Mason, & Windle, 2001; Ouellete, Gerrard, Gibbons, & Reis-Bergan, 1999; Bahr,, Marcos, & Maughan, 1995). Adolescent alcohol prevention programs have targeted each of these environments with varying levels of success and that has suggested the possibility that the effect of an attribute in one environment may be offset by attributes in one or more of the other environments. Recent studies designed to assess the relative significance of attributes in each of these environments have most often found peer attributes to have the most influence, followed closely by home attributes (Beal, Ausiello, & Perrin, 2001; Olds, & Thombs, 2001; Urberg, Degirmenciouglu, & Pilgrim, 1997), but there is certainly a good deal of inconsistency of results between these two environments in particular and all four environments in general (Eccles, & Gootman, 2002; Mason, & Windle, 2001).

There are several plausible explanations for these conflicting results. First, the existing studies include a wide range of attributes from the various environments and none, as far as the authors know, include the same attribute from each environment. The former makes it impossible to make cross study comparisons and the latter makes it impossible to make a common attribute comparison within any one study. The existing studies have also included a wide variety of adolescents, and the relative importance of these four environments might vary depending on the kind of adolescents included in the analyses (e.g., young adolescents may give more credence to their home environment while other adolescents might pay more heed to their peers). Also, the several studies have included varying outcome indicators and it may well be that the relative influence of the attributes in these environments varies to some extent depending on the alcohol-related outcome (any drinking, frequent drinking or heavy drinking) analyzed. Finally, several of these studies have used multiple regression analyses. Multiple regression analyses produce prediction models that account for any intercorrelation between the independent (predictor) variables included in the model. Unfortunately the results of such regression analyses can be misleading. That is, predictors A and B might have an equal influence on the outcome variable, but if A and B are highly correlated the prediction model might well drop either A or B from the prediction model, leaving the impression that the dropped predictor has little or no influence on the outcome variable. Further, when both predictors A and B have a nearly equally significant influence on the outcome variable, rather minor idiosyncrasies in the samples analyzed (including demographics, measurement instruments and/or random error) can produce conflicting results; a drop of predictor A in one sample and predictor B in another or at least a change in the relative significance of predictor A compared to predictor B.

The present study, we believe, is the first to assess the relative influence of each of the four environments (peer, home and neighborhood/community) based on a common attribute. The common attribute is the perceived appropriateness of an adolescent's drinking by reference groups in each environment, respectively; best friends, most students in my school, live-in parents/guardians, and most adults in my neighborhood/community. As well, the sample in this study is sufficiently large and diverse to reliably assess how differences in the kinds of adolescents included and the drinking behavior outcomes included might influence the results. Finally, bivariate correlation and drinking outcome prevalence rate comparison analyses will avoid any misleading conclusions that might be drawn from multiple regression analyses. Statistical significance is of little consequence because the sample in this study is so large, but the magnitude of the correlations should provide an unbiased estimate of the relative influence of each environment.

Analysis is designed to address these specific questions: 1) What is the relationship between the level of perceived support for alcohol use from each reference group in each environment and an adolescent's drinking behavior? 2) What is the relationship between the combined levels of perceived support for alcohol use from the four reference groups and an adolescent's drinking behavior? 3) Are the relationships the same whether the drinking behavior to be predicted is any drinking at all or extent of drinking? 4) Are the relationships affected by the adolescent's sex or grade in school? 5) Is there any evidence that the relationship between best friends' attitudes (perceived appropriateness of adolescent drinking) and the respondent's drinking behavior can be modified by the perceived "drinking appropriateness" attitudes attributed to the other three reference groups?

Method

Study participants. The findings in this study are based on data obtained from students who participated in the 1999 Iowa Youth Survey (IYS), Data were collected anonymously with a self-reported questionnaire in a classroom setting in the Fall of 1999. The survey's objective was to obtain youth-development-related data from all of Iowa's public schools, and all their students in grades 6, 8, and 11; as well as 14-18 year old students enrolled in alternative school programs. Participation in the survey was voluntary, both for the schools and students, and 329 of Iowa's 375 school districts elected to participate, as did one private school. A total of 85,426 students, with the passive consent of their parents, consented to participate in the 1999 IYS. The study sample included the majority of students enrolled in each grade: 76% of 6th graders, 80% of 8th graders and 69% of 11th graders. Alternative school enrollment figures were not available in Iowa so there was no way to know what proportion of these students participated in the 1999 IYS. The participating students were a sample of convenience, but there was no indication of any systematic biases.

The results of this study relied on each student's ability to read and honestly respond to the survey questionnaire. Every effort was made to assure students that their responses would remain anonymous and that their responses would be used to help design and implement programs that would benefit the students in Iowa's schools. Under these conditions, it is generally acknowledged that most students provide valid and reliable responses (Harrison, & Hughes, 1997). Still, self-reported behaviors and attitudes/beliefs are always subject to falsification, either intentionally (denial, boasting or just mischievousness) or unintentionally (mistakenly filling in the wrong circle, misreading, etc.). In order to minimize such problems each questionnaire was reviewed for evidence that would support a claim that the respondent had little or no intention of making an honest effort to complete the survey. This evidence consisted of such things as inconsistent responses (e.g., indicating use of substances on one question but denying ever using them on another), improbable responses (e.g., using every illegal drug every day), and patterned responses (answering a series of questions exactly the same way). A total of 27 such validity checks were made and a total of 125 questionnaires (0.2%) failed five or more of these checks. These questionnaires were deleted from the data file used in this analysis. To further ensure that only valid questionnaires were included in this analysis, all the respondents that admitted drinking on one question but denied drinking on another (1,632) were also eliminated (1.9%). This left a total of 83,669 respondents in this study.

The participating students, like Iowa's population, were not very ethnically diverse. Nearly 88% of participants described themselves as white. Slightly more students identified themselves as males (49.8% vs. 49.5%). A large majority (78%) reported a rural background (27% live in the country and 51% live in a city with a population of 30, 000 or less), and it could be argued that no large metropolitan areas exist in the State.

Constructs and measures. The study had three dependent variables (drinking outcomes): current (30 days prior to interview) drinking, heavy drinking and frequent drinking. Current drinkers and current frequent drinkers were identified by responses to the question, "In the past 30 days, on how many days have you had at least one drink of alcohol (glass, bottle or can of beer, glass of wine, liquor or mixed drink)?" Students who responded they had consumed alcohol on one or more days are identified as current drinkers and assigned a score of "1," students who responded 0 days assigned a score of "0," and all others are assigned a "missing data" score. Respondents who reported that they drank alcohol on 10 or more days in the 30 days prior to interview were identified as current frequent drinkers and assigned a score of "1," students who responded less than 10 days were assigned a score of "0," and all others were assigned a missing data score. Current heavy drinkers were identified by responses to the questions, "During the last 30 days, on how many days did you have 5 or more drinks of alcohol (glasses, bottles or cans of beer, glasses of wine, liquor, mixed drinks) in a row, that is within a couple of hours?" Students who responded they drank that much on one or more days are assigned a score of "1," students who responded 0 days are assigned a score of "0," and all others were assigned a missing data score.

The independent variables used in this study were based on the responses to four 1999 IYS questions: "Thinking of your best friends, how wrong would most of them feel it would be for you to drink beer, wine, or hard liquor (for example vodka, whiskey, gin);" "How wrong would most of the students in your school (not just your best friends) feel it would be for you to drink beer, wine, or hard liquor (for example vodka, whiskey, gin);" "How wrong would your parents/guardians feel it would be for you to drink beer, wine, or hard liquor (for example vodka, whiskey, gin) without their permission;" and "How wrong would most adults in your neighborhood and/or community feel it would be for you to drink beer, wine or hard liquor (for example vodka, whiskey, gin)." Each of these questions was included in a separate section of the questionnaire and the response options were identical for each: "Very wrong," "Wrong," "A little wrong," "Not wrong at all," and "Don't know."

Five scales were created: Best Friends, Most Students in School, Live-In Parents/Guardians, Most Adults in Neighborhood/Community and Total Perceived Level of Appropriateness of Drinking Scales. The first four scales were based on the students' responses with respect to each reference group based on the following assigned scores' a score of "1" to the "very wrong" responses, "2" to the "wrong" responses, "3" to "A little wrong," "4" to "Not wrong at all" and missing data to "Don't know" responses and no response. The fifth scale compiled the scores from each of the four reference groups. If any one of the four reference groups was assigned a missing data score, this combined scale was assigned a missing data score. The Total Perceived Level of Appropriateness of Drinking Scale scores range in value from 4 to 16 (Cronbach Alpha for raw scores=.81). A score of 4 indicated the respondent believed that persons in each of the four reference groups would think it was very wrong for the student to drink alcohol and a score of 16 indicated the respondent believed that persons in each reference group would see nothing wrong at all with the student drinking alcohol. Two dichotomous variables were also identified: unequivocal approval and unequivocal disapproval. The "unequivocal approval" dichotomy consisted of "Not wrong at all" response assigned a score of "1" no response or "Don't Know" scored missing data, and the remaining responses scored "0." The "unequivocal disapproval" dichotomy consisted of the "Very Wrong" response assigned a score of "1," no response or "Don't Know" scored missing data, and the remaining responses scored "0."

The proportion of respondents (followed in parentheses by the number of students) who qualified as current drinkers, frequent drinkers and heavy drinkers respectively was 25% (20,541), 4% (3,156) and 18% (15,272). Male prevalence rates were slightly higher than female rates and all three prevalence rates increased substantially from 6th to 11th grade: the respective current drinker, frequent drinker and heavy drinker prevalence rate for 6th graders was 5%, 0.5% and 3%; for 8th graders 20%, 2% and 13% and for 11th graders 49%, 8% and 39%. The Best Friends, Most Students in School, Live-In Parents/Guardians, Most Adults in Neighborhood/Community and Total Perceived Level of Appropriateness of Drinking Scale score means and standard deviations respectively were: 2.59 [+ or -] 1.17, 2.81 [+ or -] 1.17, 3.66 [+ or -] 0.71,3.51 [+ or -] 0.75 and 12.96 [+ or -] 3.11. All the scales were skewed in the directions of disapproval, most especially the Live-In Parents/Guardians and Most Adults in Neighborhood/Community Scales.

Statistics. Independent bivariate correlations and drinking outcome prevalence rate comparisons were used to identify the potential role that each reference group's attitudes might have on the respondent's drinking behavior. Both nonparametric correlations (Kendall tau b) and parametric (Pearson) correlations were computed. No significant differences were observed and only the result for the Pearson Correlation Coefficients are reported.

Results

Table 1-3 indicate a strong correlation between each alcohol use indicator (current drinking, frequent drinking and heavy drinking) and each Perceived Level of the Appropriateness of Drinking Scale (best friends, most students in school, live-in parents/guardians, most adults in neighborhood/community and total). This is true for the total sample, each sex and each grade.

Clearly each of the reference groups from their respective environments (peer, school, home and community) has the potential to influence adolescent alcohol use. The finding that the Total Level of the Appropriateness of Drinking Scale is consistently most highly correlated with each drinking behavior, indicates that the consistency of the message across reference groups, as well as the level of wrongness in the perceived message from each reference group, influences adolescent drinking behaviors.

Tables 1-3 also reveal some interesting differences both with respect to the drinking behaviors involved and the relative significance of the reference groups on those behaviors. With respect to current (past 30 days) drinking and heavy drinking, the correlations are consistently highest for the Total and Best Friends Perceived Level of Appropriateness of Drinking Scale. This is true for the total sample, as well as for males and females and students at each grade level. The Most Adults in Neighborhood/Community Perceived Level of Appropriateness of Drinking Scale is also most consistently the lowest correlated scale with these two drinking behaviors. However, for the 11th grades there is not much difference between the correlations for the Most Adults in Neighborhood/Community and Most Students in School Scales.

The findings for current (past 30 days) frequent drinking are similar in many respects to the findings for current drinking and heavy drinking, including: the Total Perceived Level of Appropriateness of Drinking Scale continues to be one of the highest correlated scales and the pattern of findings is consistent for the total sample, males, females, and each grade level. Still, the findings for frequent drinking are somewhat different in other respects. First, the correlations, in general, tend to be a little lower. Much of this difference can be attributed to the fact that there is less variance (a smaller proportion of the respondents indicate they are engaged in this kind of drinking behavior) in this drinking outcome than the others. Still, there is also evidence from crosstabs of the frequent drinking outcome with the perceived level of appropriateness of drinking scales that at least some of the reduction in correlations for this drinking outcome can be attributed to something other than the reduced variance in this outcome. Second, the range from their highest to lowest correlation is much smaller. Third, though the differences are small, the correlations with live-in parents/guardians generally surpass the best friend's correlations for this drinking outcome. Finally, the differences in correlations between the Most Students in School and the Most Adults in Neighborhood/Community Scales are generally relatively small, and in many instances, especially for 11th graders, it is the Most Students in School Scale that is least correlated with current frequent drinking.

The differences between the current drinking/heavy drinking and current frequent drinking correlations are not large. However, they do tend to suggest that: 1) predictors other than the perceived appropriateness of drinking scales may play a more significant role for the current frequent drinking outcome than the current drinking and heavy drinking outcomes, 2) the reference groups in each environment have a more equal opportunity to influence the current frequent drinking outcome, and 3) live-in parental/guardian attitudes are more significant with respect to the current frequent drinking outcome.

Tables 4-6 focus on the students who indicated that their best friends would feel is was "not wrong at all" for them to drink alcohol (best friends unequivocal approval). The tables show the current (past 30 days) drinking, heavy drinking and frequent drinking prevalence rates for these students based on the number of the other three reference groups (most students in school, live-in parents/guardians and most adults in neighborhood/community) that the students indicated would find it "very wrong" for them to drink alcohol (unequivocal disapproval). A "0" indicates none of the three reference groups would unequivocally disapprove of the respondent drinking, a "1" indicates on of the three reference groups would unequivocally disapprove of the respondent drinking, "2" indicated two of the reference groups would unequivocally disapprove and "3" indicates all the reference groups, except for best friends, would unequivocally disapprove.

Tables 4-6 clearly indicate that the drinking outcome (current drinking, heavy drinking and frequent drinking) prevalence rates associated with best friend's unequivocal approval of respondent drinking can be moderated by the perceived unequivocal disapproval of the respondent's drinking by reference groups from the school, home and neighborhood/community environments. The presence of even one unequivocal disapproving reference group significantly reduces the prevalence rates for all three drinking outcomes, and generally speaking, the more reference groups who are perceived to be unequivocally disapproving, the lower the drinking outcome prevalence rates. When all the reference groups, except best friends, are perceived to unequivocally disapprove of respondent drinking, the prevalence rates for all three drinking outcomes are reduced to nearly half or less of the prevalence rate for students who acknowledge best friends unequivocal approval, and no other reference group unequivocal disapproval, of respondent drinking. This holds true for the total sample, as well as males and females and students at each grade level (6th, 8th, and 11th).

Tables 4-6 also identify some significant differences in results depending on the sample and drinking outcome used. First, for the current drinking and heavy drinking outcomes, by far the largest reductions in prevalence rates occur when all three of the school, home and neighborhood/community reference groups are perceived to unequivocally disapprove of the respondent drinking. However, for the current frequent drinking prevalence rate outcome in general, and especially for the 6th graders, 11th graders and females, the prevalence rate reduction is similar regardless of the number of reference groups that unequivocally disapprove of the respondent drinking. Also, it would appear that all three drinking outcome prevalence rates for the females who perceive that their best friends unequivocally approve of their drinking are somewhat less impacted than the males by the perceived unequivocal disapproval of the other reference groups. On the other hand, it would also appear that, regardless of the drinking outcome, the drinking prevalence rates for 6th graders who perceive that their best friends unequivocally approve of their drinking are somewhat impacted than 8th or 11th graders by the perceived unequivocal disapproval of the other reference groups.

Finally, an analysis of those students who indicated that their best friends would feel it was "not wrong at all" for them to drink alcohol (best friend unequivocal approval) and the unequivocal disapproval of each of the school, home and neighborhood/community reference groups individually reveals some interesting patterns regarding the relative moderating influence of each reference group. For the current drinking and heavy drinking outcomes in each sample, except 6th graders, it is generally the case that unequivocal disapproval by most students in school has the most impact, followed by live-in parents/guardians and most adults in neighborhood/community. The current frequent drinking prevalence rates, like the other drinking outcome prevalence rates, are significantly reduced by unequivocal disapproval of one or more of the other reference groups (most students in school, live-in parents/guardians or most adults in neighborhood/community), but there is virtually no difference in the extent of the influence of any one of these reference groups. The same is true for 6th graders for all the drinking outcomes.

Discussion

This study of Iowa students is not without its limitations. The most notable are: First, the results are based on a sample of students from only one state that is mostly rural and not very ethnically diverse. Our results may not hold for adolescents outside of America or even Americans in a more urban and ethnically diverse population. Second, there is the issue of attributing causal connections between the level of perceived support for alcohol use from each of four reference groups (best friends, most students in school, live-in parents/guardians and most adults in neighborhood/community) and the respondent's drinking outcome (reporting they are a current drinker, heavy drinker and frequent drinker). Our analysis is based on correlations and it might be that the decision to drink alcohol precedes affiliation with best friends who approve of the respondent drinking. The decision to use alcohol might also influence how the respondent perceives the opinions of the other reference groups. There is some evidence, however, that adolescents' reports of their peer and parent drinking norms correspond quite closely to the actual peer and parent reported norms (Wilks, Callan, & Austin 1989). Still, longitudinal studies must be completed in order to ensure that the reference group perceptions precede the drinking behaviors. Finally, it is not know what impact other reference groups from these environments (e.g., perceived level of support for respondent alcohol use from administrators and faculty in the school environment, siblings or extended family in the home environment or any number of others in the neighborhood/community environment) might have on students' decisions to use alcohol. In the absence of information from these other reference groups, it could be that the combined impact of all the reference groups in any one of these environments would be larger than the impact of best friends. Still, despite its shortcomings, this study does offer some rather convincing evidence that supports two basic conclusions: 1) each reference group's (best friends, most students in school, live-in parents/guardians and most adults in neighborhood/community) perceived level of support for an adolescent's alcohol use, as well as their combined total, does have a potential to influence an adolescent's decision to use alcohol and; 2) the influence of best friend unequivocal approval of drinking can be reduced significantly by school, home and neighborhood/community reference group unequivoal disapproval.

The results of this study have several primary and secondary prevention-related implications. The most significant include; 1) prevention programs that encourage disapproval, especially unequivocal disapproval, or respondents drinking by reference groups in any one of the four environments (best friends, school, home and neighborhood/community) have the potential to reduce adolescent drinking prevalence rates; 2) comprehensive prevention programs that are designs to encourage disapproval, especially unequivocal disapproval, by all four reference groups are likely to reduce the risk of adolescent alcohol use the most; 3) prevention programs that are designed to encourage unequivocal disapproval of drinking by most students in school, live-in parents/guardians and/or most adults in neighborhood/community have the potential to reduce the risks of drinking precipitated by the unequivocal approval of drinking by best friends; 4) disapproval of drinking by the four reference groups is likely to have the most impact on current (past 30 days) drinking and heavy drinking prevalence rates and the least impact on current frequent drinking prevalence rates; 5) compared to the other reference groups, live-in parental/guardian disapproval of drinking would appear to have a limited impact on adolescent current drinking and heavy drinking prevalence rates, but it is one of the most influential for current frequent drinking prevalence rates; 6) the drinking prevalence rates of adolescents who perceive that their best friends unequivocally approve of their drinking can generally be reduced the most by prevention programs that encourage unequivocal disapproval of drinking by all three remaining reference groups (most students in school, live-in parents/guardians and most adults in neighborhood/community), but for 6th graders in general and the frequent drinking prevalence rate in particular, reductions are similar regardless of the number of the remaining reference groups that are perceived to unequivocally disapprove of the adolescent drinking; 7) drinking prevalence rates by males and 6th graders who perceive that their best friends unequivocally approve of their drinking are more likely to be reduced by unequivocal disapproval of the other reference groups than are their female and 8th and 11th grade counterparts; and 8) the drinking prevalence rates by adolescents who perceived that their best friends unequivocally approve of their drinking are likely to be reduced the most by unequivocal disapproval by most students in school, followed by live-in parents/guardians and most adults in neighborhood/community, but for 6th graders for all the drinking outcomes and the current frequent drinking outcome in particular there is virtually no difference in the extent of the influence of these three reference groups.

Table 1.
Correlations of Best Friends, Most Students in School, Live-In
Parents/Guardians, Most Adults in Neighborhood/Community and Total
Perceived Level of Appropriateness of Drinking Scales with Current
(Past 30 Days) Drinking.

Pearson Correlation Coefficients

Appropriateness of   Total
drinking scale       sample   Males   Females

Best of friends      .61      .60     .62

Most Students
in school            .50      .52     .48

Live-In
Parents/guardians    .45      .47     .43

Most adults in
neighborhood/
community            .37      .40     .35

Total                .62      .62     .61

Pearson Correlation Coefficients

Appropriateness of   6th       8th       11th
drinking scale       graders   graders   graders

Best of friends      .35       .53       .51

Most Students
in school            .30       .41       .25

Live-In
Parents/guardians    .31       .38       .36

Most adults in
neighborhood/
community            .25       .32       .23

Total                .40       .54       .49

Note. All correlations significant p<.0001.

Table 2.
Correlations of Best Friends, Most Students in School, Live-In
Parents/Guardians, Most Adults in Neighborhood/Community and Total
Perceived Level of Appropriateness of Drinking Scales with Current
(Past 30 Days) Frequent Drinking.

Pearson Correlation Coefficients

Appropriateness of   total
drinking scale       sample   Males   Females

Best friends         .27      .28     .24

Most students in
school               .21      .25     .17

Live-in
parents/guardians    .29      .31     .23

Most adults in
Neighborhood/
community            .23      .26     .17

Total                .30      .34     .25

Pearson Correlation Coefficients

Appropriateness of   6th       8th       11th
drinking scale       graders   graders   graders

Best friends         .17       .21       .22

Most students in
school               .13       .18       .12

Live-in
parents/guardians    .19       .23       .25

Most adults in
Neighborhood/
community            .12       .18       .19

Total                .20       .25       .27

Note. All correlations significant p<.0001.

Table 3.
Correlations of Best Friends, Most Students in School, Live-In
Parents/Guardians, Most Adults in Neighborhood/Community and Total
Perceived Level of Appropriateness of Drinking Scale with Current
(Past 30 Day) Heavy Drinking.

Pearson Correlation Coefficients

Appropriateness of   Total
drinking scale       sample   Males   Females

Best Friends         .56      .56     .56

Most students in
school               .46      .49     .42

Live-in
parents/guardians    .44      .46     .41

Most adults in
Neighborhood.
Community            .35      .38     .31

Total                .57      .59     .55

Pearson Correlation Coefficients

Appropriateness of   6th       8th       11th
drinking scale       graders   graders   graders

Best Friends         .31       .47      .48

Most students in
school               .26       .37       .23

Live-in
parents/guardians    .29       .37       .35

Most adults in
Neighborhood.
Community            .23       .29       .23

Total                .35       .49       .46

Note. All correlations significant p<.0001.

Table 4.
Current (Past 30 Days) Drinking Prevalence Rates (In Percent) for
Students Who Responded Their Best Friends Would Feel it was "Not
Wrong At All" for Them to Drink Alcohol by Number of Other Reference
Groups (Most Students in School, Live-In Parents/Guardians and Most
Adults in Neighborhood/Community) Who Respondents Felt Would Say it
Was "Very Wrong" for Respondent to Drink Alcohol.

Current (past 30 days) drinking prevalence rates (in percent)

Number
reference

groups say very     Total
wrong to drink      sample       Males       Females

       0          82 (5,834)   81 (3,433)   83 (2,367)
       1          69 (3,170)   67 (l,696)   71 (1,461)
       2          60 (2,469)   56 (1,319)   65 (1,139)
       3          38 (183)     35 (116)     44 (66)

Current (past 30 days) drinking prevalence rates (in percent)

Number
reference

groups say very     6th        8th         11th
wrong to drink    graders    graders     graders

       0          74 (108)   80 (940)   83 (4,520)
       1          47 (59)    66 (644)   70 (2,357)
       2          37 (93)    56 (623)   62 (1,674)
       3          26 (50)    45 (87)    44 (41)

Note. Number in parentheses is number of students prevalence rate
is based on.

Table 5.
Current (Past 30 Days) Frequent Drinking Prevalence Rates (In Percent)
for Students Who Responded Their Best Friends Would Feel it was "Not
Wrong at All" for Them to Drink Alcohol by Number of Other Reference
Groups (Most Students in School), Live-In Parents/Guardians and Most
Adults in Neighborhood/Community) Who Respondents Felt Would Say it
Was "Very Wrong" for Respondent to Drink Alcohol.

Current (past 30 days) drinking prevalence rates (in percent)

Number
reference

groups
say very
wrong to      Total
drink         sample       Males       Females

    0       20 (5,834)   23 (3,433)   16 (2,367)
    1       12 (3,170)   14 (1,696)   11 (1,461)
    2       09 (2,469)   10 (1,319)   08 (1,139)
    3       05 (183)     03 (116)     09 (66)

Current (past 30 days) drinking prevalence rates (in percent)

Number
reference

groups
say very
wrong to      6th        8th         11th
drink       graders    graders     graders

    0       24 (108)   18 (940)   20 (4,520)
    1       05 (59)    11 (644)   13 (2,357)
    2       06 (93)    08 (623)   09 (1,674)
    3       04 (50)    05 (87)    10 (41)

Note. Number in parentheses is number of students prevalence rate
is based on.

Table 6.
Current (Past 30 Days) Heavy Drinking Prevalence Rates (In Percent)
for Students Who Responded Their Best Friends Would Feel it was "Not
Wrong At All" for Them to Drink Alcohol by Number of Other Reference
Groups (Most Students in School, Live-In Parents/Guardians and Most
Adults in Neighborhood/Community) Who Respondents Felt Would Say it
Was "Very Wrong" for Respondent to Drink Alcohol.

Current (past 30 days) drinking prevalence rates (in percent)

Number
reference

groups say very   Total
wrong to drink    sample       Males        Females

       0          71 (5.888)   72 (3,465)   70 (2,389)
       1          57 (3,230)   57 (1,714)   56 (1,476)
       2          48 (2,477)   45 (1,328)   51 (1,138)
       3          27 (186)     25 (117)     32 (68)

Current (past 30 days) drinking prevalence rates (in percent)

Number
reference

groups say very   6th        8th        11th
wrong to drink    graders    graders    graders

       0          58 (110)   64 (951)   73 (4,555)
       1          29 (59)    51 (655)   58 (2,377)
       2          26 (93)    41 (626)   51 (1,679)
       3          22 (51)    27 (89)    39 (41)

Note. Number in parentheses is number of students prevalence rate
is based on.

References

Bahr, S.J., Marcos, A.C., & Maughan, S.L. (1995). Family, educational and peer influences on the alcohol use of female and male adolescents. Journal of Studies on Alcohol, 56, 457-469.

Beal, A.C., Ausiello, J., & Perrin, J.M. (2001). Social influence on health-risk behaviors among minority middle school students. Journal of Adolescent Health, 28, 474-480.

Eccles, J., & Gootman, J.A. (Eds,). (2002) Community Programs to Promote Youth Development. Washington, DC: National Academy Press.

Harrison, L., & Hughes, A. (Eds.). (1997). The validity of self-reported drug use: Improving the accuracy of survey estimates. (NIDA Publication No. 97-4147). Rockville MD: National Institute on Drug Abuse.

Mason, W.A., & Windle, M. (2001). Family, religious, school and peer influences on adolescent alcohol use: A longitudinal study. Journal of Studies on Alcohol, 62, 44-53.

Olds, R.S., & Thombs, D.L. (2001). The relationship of adolescent perceptions of peer norms and parent involvement to cigarette and alcohol use. Journal of School Health, 71, 223-228.

Ouellette, J.A., Gerrard, M. Gibbons; F.X., & Reis-Bergan, M. (1999). Parents, peers, and prototypes: Antecedents of adolescent alcohol expectancies, alcohol consumption, and alcohol-related life problems in rural youth. Psychology of Addictive Behaviors, 13, 183-197.

Urberg, K.A., Degirmencioglu, S.M., & Pilgrim, C. (1997). Close friend and group influence on adolescent cigarette smoking and alcohol use. Developmental Psychology, 33, 834-844.

Wilks, J., Callan, J.J., & Austin, D.A. (1989). Parent, peer and personal determinants of adolescent drinking. British Journal of Addiction, 84, 619-630.

Author Note

J. L. Fitzgerald, Iowa Consortium for Substance Abuse Research and Evaluation, University of Iowa, Stephan Amdt, Iowa Consortium for Substance Abuse Research and Evaluation, Department of Psychiatry and Department of Biostatistics, University of Iowa.

Major funding for data collection was provided by a grant from the United States Department of Education (S184G980007). Additional funds were provided by the Iowa Department of Public Health, Division of Substance Abuse and Health Promotion; Iowa Department of Education,; Iowa Office of Drug Control Policy, and Iowa Department of Human Rights, Criminal and Juvenile Justice Planning and Statistical analysis Center.

Correspondence concerning this article should be addressed to J. L. Fitzgerald, M110 Oakdale Hall, University of Iowa, Iowa City, Iowa 52242.

STEPHAN ARNDT 

University of Iowa

J. L. Fitzgerald, Send all Correspondence concerning this article should be addressed to:

J. L. Fitzgerald

M110 Oakdale Hall

University of Iowa

Iowa City, Iowa 52242