Drug Abuse

Correlates of Conformity in the Consumption of Illicit Drugs and Alcohol

William O. Bearden UNIVERSITY OF SOUTH CAROLINA
Randall L. Rose UNIVERSITY OF SOUTH CAROLINA
Jesse E. Teel UNIVERSITY OF SOUTH CAROLINA

Journal of business research 30, 25-31 (1994)

The results of our studies regarding conformity in illicit drug and alcohol consumption among high school and college students are reported. Correlational evidence regarding the relationships between nine hypothesized antecedents of compliance and a measure of conformity intentions is provided through a series of median split comparison tests and estimates of bivariate and multivariate correlations. The results reveal that perceptions of the group (e.g., perceived group attractiveness), attributions about reasons for the group's behavior, and external influences (e.g., self-reports of family and friend use) are related to compliance as predicted. Overall, the results provide insight into factors associated with conformity in the illicit consumption of alcohol and drugs and the conditions under which normative pressures are likely to be operative.

 

 

 

Introduction

One problem facing researchers interested in predicting and understanding consumer behavior is the explanation of the determinants of normative influences (Bearden and Rose, 1990; Miniard and Cohen, 1983). This issue is particularly relevant regarding decisions of young people to engage in the illicit consumption of alcohol and drugs, because conformity to peers in reaction to normative pressures is considered one of the hallmarks of adolescent behavior. Peers endorse illicit drug and alcohol use, facilitate the behaviors by making substances available, and provide appropriate social settings and even instructions for use (Kaplan, Martin, and Robbins, 1984).

Although the study of drug and alcohol use among adolescents and young adults has been extensive and frequently includes peers as a determinant of behavior, most prior research has not examined the nature of the referent(s) and many of the cognitive process variables that may operate when normative pressures on the individual are present. That is, interpersonal influence from peers often arises from perceptual and attributional thoughts of the individual being influenced (Rose, Bearden, and Teel, 1992). These thoughts concern the characteristics and behaviors of the influencing referent or peer group, and research regarding the nature of the peer pressures adolescents perceive is needed (Brown, Clasen, and Eicher, 1986). These limitations then restrict our understand ing of how peer influences operate to affect conformity.

The overall objective of the present research was to examine the extent to which a series of social psychological beliefs, perceptions of the group, and behaviors are related to individual intentions to comply with the expectations of referent others regarding the illicit consumption of alcohol and marijuana. Specifically, our aims were two-fold. Our first objective was to examine the usefulness of the hypothesized predictors for explaining compliance with group expectations. Our second objective was to embed evidence regarding the role played by attributional reasoning in conformity decisions (Rose et al., 1992) into a traditional framework of conformity determinants that includes group processes and external influences. As such, the paper attempts to incorporate research in the consumer behavior literature with related drug abuse research to offer a test of multiple influences that other individuals may have on decisions to use illicit drugs and alcohol. In the research, three sets of explanatory variables are examined: perceptions of the group (e.g., perceived group attractiveness); attributions about reasons for the group's behavior (e.g., perceived intemal/external locus of causality); and external influences (e.g., self-reports of family and friend alcohol and drug abuse). A total of nine correlates allocated across these three categories is examined. While the tripartite categorization is our own configuration (see Figure 1) and the assignment of variables to the categories is somewhat arbitrary, all of the hypothesized antecedents have theoretical justification for being considered as antecedents of compliance to normative pressures involving the illicit consumption of alcohol and drugs. However, to our knowledge, this paper is the first to include measures of group perceptions or characteristics and measures of attributional thoughts about the reasons for group behavior with the more frequently studied variables of self-reported peer and family usage.

Hypothesized Antecedents to Compliance Decisions

The overall model that guided our research is presented in Figure 1. The assumption underlying our thinking was that decisions to consume drugs or alcohol are often made in response to group and peer influences (e.g., behaviors, pressures, requests for compliance). This model suggests three categories of influences on individual compliance judgments but is intended as an organizational framework rather than a predictive model of conformity. The three types of factors represent interpersonal, cognitive, and behavioral categories (cf. Stein, Newcomb, and Bender, 1987), and are not intended to portray latent constructs. First, it is expected that perceived similarity of the group to the individual, perceived group attractiveness, and the ability of the group to control rewards may each affect decisions to comply with group expectations. Second, attributions made by the individual regarding the reasons for the group's behavior are expected to affect the likelihood of compliance as well. Third, and as often studied in the drug-use literature, external influences from peer approval

in addition to family and friend usage are posited also as correlates of compliance decisions. Although it is not clear a priori which of these antecedents or sources of influence will be most important, each correlate has precedent for inclusion in the present research (cf. Kandel, Kessler, and Margulies, 1978). The rationale and justification for the consideration of each of these predictors follows.

Perceptions of the Group

As shown in Figure 1, three group perception variables are posited to influence compliance-group attractiveness, perceived group similarity, and perceptions of the group's reward structure. These effects are assumed to operate through the normative processes of identification and conformity (Deutsch and Gerard, 1955; Kelman, 1961). In considering person group relationships, one of the most important factors is the degree of attraction the group holds for the person (Allen, 1965; Witt, 1969). Group attractiveness is thought to operate through the process of identification. The individual is motivated to attain a gratifying self-concept or enhanced self-esteem through identification with an attractive referent (Bumkrant and Cousineau, 1975). McGuire (1969) argues that attractiveness effects, and hence identification, are further enhanced to the extent that the group is perceived as similar. Normative social influences are also enhanced to the extent that individuals are motivated to realize rewards or avoid punishments (Bumkrant and Cousineau, 1975). In the contexts studied here, these rewards and punishments involve approval, threats of ridicule, and thoughts about group acceptance.

Attributions About Reasons for the Group's Behavior

Decisions to conform to the expectations and behavior of others (i.e., peers, referent groups) may be preceded by complex judgmental processes. These processes often involve reasoning about the group's behavior. The outcomes from these processes include attributions about the referent other's behavior. That is, the behavior of others (i.e., the consumption of marijuana or alcohol by friends) serves as input for the observers (i.e., the person to be influenced) attribution processes. The nature of the attributions resulting from these cognitive processes is hypothesized here as affecting the conformity judgments of adolescents and young adults exposed to pressures to conform. The reasons for these predictions are as follows.

Expectations for individual behaviors are often categorized into normative and personal dimensions (cf. Ajzen and Fishbein, 1980). It is predicted that attributing a referent's behavior (in this case a group of friends' consumption of alcohol or drugs) to normative causes (i.e., the desire to fit in or "be cool") will be associated with lower compliance. In contrast, attributions involving personal consequences (i.e., the desire to get high) will be associated with greater compliance. These predictions are based upon the premise that normative attributions can be construed as being primarily external in nature and, therefore, providing more powerful bases for dissent. That is, the referent's illicit behavior can more easily be attributed to a different set of consequences (than those faced by the potential dissenter) when the explanations for the behavior are normative as opposed to personal (Rose, Bearden, and Teel, 1992; Ross, Bierbrauer, and Hoffman, 1976). Thus, the availability of attributions accounting for referent others: behavior reduces pressures toward conformity and lessens the risks of dissent. As a corollary and for similar reasons, we expect that, on an internal/external locus of causality dimension (cf. Miller, Smith, and Uleman, 1981; Smith and Miller, 1982), perceptions of the group's behavior as being influenced primarily by external or situational factors will be inversely correlated with compliance.

External influences

Our model and subsequent studies also enabled tests of the relationships among three additional variables that are often hypothesized in the drug-related literature to be significant influences on drug and alcohol use among youthful consumers. As shown in Figure 1, the external influences of perceived peer approval and peer behavior along with self-reports of family member behavior (e.g., siblings, parents) are depicted as antecedents of compliance. As such, inclusion of these frequently studied correlates allows some comparison with the more process-oriented antecedents discussed above (i.e., group characteristics and attributions).

A substantial body of literature dealing with adolescent drug and alcohol use demonstrates the importance of peer and family influences on the illicit consumption of drugs and alcohol (e.g., Brook, Whiteman, and Gordon, 1983; Downs, 1987; Kaplan, Martin, and Robbins, 1984; Marcos, Bahr, and Johnson, 1986; Needle et al., 1986, Stein, Newcombe, and Bentler, 1987; Swadi, 1988). Researchers have consistently reported from both cross-sectional and longitudinal studies that: (1) adolescents and young adults whose peers use drugs and alcohol are more likely to use drugs and alcohol and consume them more frequently than youth whose friends are nonusers; (2) individuals from families in which one or more members drink or take drugs are more likely to use substances than those whose families do not; and (3) peers, relative to parents, have a greater influence on consumption and usage behavior (Needle et al., 1986, p. 740). In two examples, Kandel et al. (1978) found association with marijuana-using peers to be the strongest predictor of initiation to marijuana rise, while Marcos et al. (1986) report that the best single predictor of drug use is association with drug-using friends.

These peer and family influences are thought to operate through a number of processes. First, it is well-accepted that young people often model or imitate the behavior of their siblings and parents (cf. Halebsky, 1987; Gfroerer, 1987). This generalized imitation of adult behavior is a frequently cited source of influence on the illicit consumption of drugs and alcohol. Second, as individuals associate with peers who share social values favorable to the performance of certain activities (including those that are counter to societal norms), the individual is more likely to engage in those activities and behaviors (Kaplan et al., 1984). Obviously, such favorable attitudes often persist among ongoing peer groups and subcultures. Such influences serve as reference standards, and young people who are motivated to maintain or gain approval from those who share their environment will often adapt the standards and the behaviors of these referent others.

Methods Overview

Data were collected to test the hypothesized correlates in a series of four studies involving both university and suburban high school student subjects. The subjects were given one of two written descriptions, one involving the use of marijuana at a friend's apartment and the other involving the consumption of alcohol in spite of being underage. In both circumstances, the subject encountered close friends already engaged in the illicit behavior and were offered either beer or marijuana by their friends at a party. Subjects were asked to consider the situation carefully, to place themselves in the role of the student described in the scenario, and to imagine that they and their own friends were involved. Subjects first read the scenario and then completed a brief questionnaire. The sensitive nature of the topic and the illegality of the behaviors for high school and college students necessitated the use of scenarios and subjective self-reports. The obvious limitations of these procedures are acknowledged in the Discussion section.

Subjects

For each study, subjects participated in a classroom setting. Study 1 involved 266 students from a large state university and the consumption of marijuana. The second study assessed the reactions of 108 high school juniors and seniors (who were all under the legal drinking age) to a beer consumption scenario. In Study 3, 88 different students from the same high school participated in a Marijuana consumption context. Last, 100 students from a small urban college participated in a study involving the use of marijuana.

Measures

The key measure was an 11-point compliance scale anchored, for example, by "certainly would drink the beer" (11) and "there is no chance I would drink the beer" (1). Mean scores (and standard deviations) for this dependent measure across the four studies were: (1) 2.91 (2.98)-266 university students, marijuana; (2) 6.04 (3.65)-108 high school students, beer; (3) 3.19 (3.08)-88 high school students, marijuana; and (4) 1.95 (2.17)-100 college students, marijuana.

Varying combinations of the hypothesized nine antecedents or correlates of compliance were investigated across the four studies. As shown in Figure 1, three variables each were designed to reflect: (1) perceptions of the group (e.g., perceived group similarity); (2) attributions about the behavior of the group (e.g., reasons tied to perceived peer pressure); and (3) potential external influences (e.g., family substance abuse). In Studies 3 and 4, all nine antecedents were examined. Five and six of the nine predictors were examined in Studies 1 and 2, respectively. (See Tables 1 and 2.) In most instances; the hypothesized antecedents were operationalized as nine-place bi scales reflecting reactions to statements designed to elicit subject opinions of the various predictors of conformity.

PERCEPTIONS OF THE GROUP.

Three variables regarding subject perceptions of the group were examined across each of the four studies. First, perceived group similarity was assessed as the sum of the responses to three statements (e.g., "How similar do you feel you are to' these other three' individuals?" Very Similar-Not Similar At, All). Coefficient alpha estimates of internal consistency reliability for the three-item measure across the four studies were: 0.73, 0:73, 0 ,82, and 0.75. Group attractiveness was operationalized as a single nine place agree-disagree scale following the statement: "In general I like the people described in the scenario." Last, perceptions of social rewards were measured via responses to a single agree-disagree scale following the statement: "These people would be rewarded differently for smoking the marijuana (drinking the beer) by their friends than 1 would be by mine."

ATTRIBUTIONS ABOUT THE BEHAVIOR OF THE GROUP.

Perceived normative pressure as an explanation for the behavior of the group was operationalized (in all four studies) as the sum of the responses of the subjects to three nine-place scales bounded by extremely unlikely (1)extremely likely (9) and following statements such as: "How likely is it that these people want to 'fit-in' with their friends?" The other two items reflected the likely effects of perceived peer pressure and the desire "to look cool" to impress others. The estimates of in ternal consistency reliability across the four studies were: 0.93, 0.93, 0.77, and 0.89. Three items used in Studies 3 and 4 only (and also operationalized using "extremely unlikely-extremely likely" bipolar scales) were included to assess perceptions regarding the personal consequences involved: "How likely is it that your friends in the group are smoking marijuana because they do not mind damaging their health,... want to get high, and ... do not fear getting in trouble with the law?" Reliability estimates for this three-item measure were 0.50 and 0.55 for Studies 3 and 4. These situation-specific normative and personal attribution measures were accompanied by a more abstract, single-item, internal/extemal locus of causality measure in Studies 3 and 4 (cf. Miller, Smith, and Uleman, 1981; Smith and Miller, 1982). After instructions regarding the nature of internal and external reasons for behavior, responses to the following statement were assessed by a nine-place bipolar scale anchored by completely internal (1)-completely external (9): "The reason my friends in the group are smoking marijuana is ......

EXTERNAL INFLUENCES.

Perceived approval from others was operationalized in each of the four separate studies via re sponses to a single statement followed by a nine-place scale bounded by disapprove-approve: "Do you think that your other friends (in other words, those not in the group in the scenario) would approve or disapprove of your smoking the marijuana (drinking the beer)?" For the last two studies, friend and family involvement with marijuana was assessed using a single dichotomous "yes-no" question (e.g., Do any of your friends smoke marijuana ... Yes-No-).

Results

CORRELATIONAL ANALYSES.

Simple correlations between each antecedent and the compliance measure are depicted in Table 1. The average absolute values of the intercorrelations among the nine antecedent or predictor variables for Studies, 1, 2; 3; and 4 were: 0.42, 0.27, 0.32, and 0.24, respectively. These intercorrelations ranged from a low of 0,01 for the relationship between perceived group similarity and the tendency to attribute the group's behavior to personal consequences in Study 4 to 0.71 for the relationship between perceived group similarity and group attractiveness in Study 3.

As shown in Table 1, 28 of the 36 simple correlations among the nine hypothesized predictors of compliance and the conformity scale were significant and consistent directionally with expectations. These results provide some insight into the magnitude of the effects involved. For example, the average correlations (across studies) between conformity and the measures of perceived group similarity, group attractiveness, and reward perceptions were 0.42, 0.45, and -0.37, respectively. Significant correlations were also obtained between the conformity measure and the attribution measures. Again, conformity was inversely related to the tendency to attribute the group's behavior to external reasons. All eight of the estimated correlations were significant (p < .05) and ranged in magnitude from 0.26 to 0.40. Consistent positive correlations were also observed between compliance and the three measures of external influence. For example, the average correlation between the perceived approval of others and conformity was 0.45. Self-reports of family and friends' behavior were, in general, significantly associated with compliance as well. Overall, these correlations are similar to the results of other crosssectional studies involving correlational predictors of Self reported drug and alcohol use (cf. Dembo et al., 1985; Marcos et al., 1986; McLaughlin et al., 1985).

GROUP DIFFERENCE TESTS.

The relationships between the different antecedents and the measure of compliance were also examined in a series of difference tests based upon median split group partitions. That is, each sample was split into high and low groups based upon the distribution of scores for each antecedent. Comparisons in means for each subgroup on the 11-place compliance measure were then examined. Although these analyses are somewhat redundant with the correlational analyses, it was felt that the means tests offered a view of data relationships that might be appreciated by some readers. These results are summarized in Table 2. In total, 29 comparisons were examined across the four samples. Of these tests, 24 were significant (p < .05, one-tailed tests) and consistent with the hypothesized effects.

The most consistent findings were observed for the group perception antecedents: perceived similarity, group attractiveness, and rewards. As hypothesized, those subjects who perceived themselves to be more similar to the individuals in the group depicted in the scenarios and those subjects reporting higher perceived group attractiveness scores also scored higher on the compliance measure. Likewise, those subjects who felt that the group members would be rewarded differently than they normally would tended to score lower as anticipated on the conformity measure.

Analysis of the three antecedents we have categorized as attributional (e.g., perceptions involving reasons for the group's behavior) provided, perhaps, the most intriguing results. First, those subjects making an internal attribution tended to report greater compliance. These findings are sup ported by the normative attribution difference tests across all four studies. That is, those students who reported external attributions to be more likely explanations for the group's behavior tended to comply less. In other words, subjects indicating that the reasons for the group's behavior may have been attributable to peer pressures or the desire to fit-in reported lower average scores.

Significant differences were also obtained for the third set, of median split comparisons involving "external influences." For each study and not unexpectedly, those students more likely to indicate that their current friends would approve of their agreeing to use marijuana (or drink the beer for the high school students) reported significantly higher conformity scores. Consistent with this finding and particularly for the high school marijuana study, self-reports regarding usage among both family and friends resulted in higher personal compliance mean scores.

Discussion

While there exists considerable evidence that family members, friends, and other peers both encourage and discourage illicit consumption of drugs and alcohol among young people, less effort has been devoted by researchers to understanding the cognitive processes underlying such influences. Our data provide additional support for the role of peers and family as models for drug and alcohol abuse but also highlight the importance of cognitive variables such as perceptions of peer groups and attributions for peer groups' illicit behavior. In particular, the finding that certain types of attributions made to account for a peer group's behavior carry different implications for conformity suggests the need to adopt a cognitive process orientation when studying deviant behaviors such as drug and alcohol abuse. (See Rose et al., 1992, for a more detailed treatment of this issue.)

Why might attributions carry different implications for conformity7 one explanation offered here is that normative attributions imply external or situational causes for the group's I behavior that may not affect the group and the potential dissenter equally. if potential dissenters can see that the group faces normative consequences that they do not face', then the risks of dissent are reduced, especially if the group can also see these differences in their situation and that facing the dissenters On the other hand, personal attributions do not imply differences in consequences between dissenter and group, Rather the attributor is forced to speculate about differences' in values, attitudes, and personality. Such attributions are less likely to be made with confidence because relevant peer groups for young people are likely to comprise youths similar to themselves (Ross et al., 1976).

Another explanation for the different effects on conformity is that attributions may carry different implications for group attractiveness. For example, drug use that is attributed to social pressures may be viewed less favorably than such behavior attributed to personal choice, In other words, young people may consider peers who use drugs or alcohol because they like the "high" better than those who succumb to external pressures. Clearly, research is needed that explores the link between attributions and conformity with the goal of identifying the chain of cognitive events that Read from peer pres sures to conformity or dissent. Without a firm understanding of the role of attributions, it would be premature to suggest intervention strategies based on the encouragement or discouragement of certain types of attributions for peer drug and alcohol use.

Previous research has suggested an inverted-U shaped function for the relationship between peer pressures and antisocial behavior (Brendt, 1979; Huba and Bentler, 1980). That is, conformity to peers in antisocial behavior may be more likely to occur during early adolescence than during preadolescence or late adolescence. Further, the relative contribution of family and peer influences to conformity may be moderated by characteristics of the sample such as urban versus rural location, age, sex, and nationality, and the nature of the deviant behavior (Brown et al., 1986). For example, our data revealed stronger family influences for high school students when the behavior involved drinking beer than when smoking marijuana was the target behavior.

In conclusion, our studies have extended previous correlational investigations of pressures toward conformity in drug and alcohol consumption by considering the influences of both perceptual and social modeling variables. Our conclusions are limited by the correlational nature of the data and by the use of role playing scenarios and self-report measures. A causative role for these variables should not be assumed. Nevertheless, across different samples (two high school and two college) and two types of substance abuse (beer and marijuana), our data demonstrate the importance of investigating the perceptual and attributional responses of young people exposed to peer pressures regarding the illicit consumption of drugs and alcohol in addition to identifying sources of influence in the social environment.

References

Ajzen, Rock, and Fishbein, Martin, Understanding Attitudes and Predicting Social Behavior, Prentice-Hall, Englewood Cliffs, NJ. 1980. Allen, Vernon L., Situational Factors in Conformity, in Advances in Experimental Social Psychology, Vol. 2, L. Berkowitz, ed., Academic Press, New York. 1965, pp. 133-•175.

Bearden, William O., and Rose, Randall L:, Attention to Social Comparison Information: An Individual Difference Factor Affecting Consumer Conformity. J. Consumer Res. 16 (March 1990): 461 471.

Brendt, T. J., Developmental Changes in Conformity to Peers and Parents. Dev. Psychol. 15 (1979): 606-616.

Brook, Judith S., Whiteman, Martin, and Gordon, Ann Scovell, Stages of Drug Use in Adolescence: Personality, Peer, and Family Correlates. Dev. Psychol. 19 (1983): 269-277.

Brown, B. Bradford, Clasen, Donna Rae, and Eicher, Sue Ann, Perceptions of Peer Pressure, Peer Conformity Dispositions, and Self-Reported Behavior Among Adolescents. Dev. Psychol. 22 (1986): 521-530.

Bumkrant, Robert F., and Cousineau, Alain, Informational and Normative Social Influence in Buyer Behavior, J. Consumer Res. 2 (December 1975):206-215.

Dembo, Richard, Allen, Nola, Farrow, Dana, Schmeidler, James, and Burgos, William, A Causal Analysis of Early Drug Involvement in Three Inner-City Neighborhood (1985): 1213-1237.

Deutsch, Morton, and Gerard, Harold B., A Study of Normative and Informational Influence Upon Individual (November 1955): 629-636.

Downs, William R., A Panel Study of Normative Structure, Adolescent Alcohol Use and Peer Alcohol Use, J. Stud. Alcohol, 48 (1987): 167-175.

Gfroerer, Joseph, Correlation between Drug Use by Teenagers and Drug Use by Older Family Members. Am. J. Drug Alcohol Abuse 13 (1987):95-108.

Halebsky, Mark A., Adolescent Alcohol and Substance Abuse: Parent and Peer Effects. Adolescence 22 (Winter 1987): 961-967.

Huba, G. J., and Bentler, P. M., The Role of Peer and Adult Models for Drug Taking at Different Stages in Adolescence. J. Youth Ado lescence, 9 (1980): 449-465.

Kandel, Denise B., Kessler, Ronald C., and Margulies, Rebecca Z., Antecedents of Adolescent Initiation Into Stages of Drug Use: A Developmental Analysis. J. Youth Adolescence 7 (1978): 13-40. Kaplan, Howard B., Martin, Steven S., and Robbins, Cynthia, Pathways to Adolescent Drug Use: Self-Derogation, Peer Influence, Weakening of Social Controls. J. Health Soc. Behav. 25 (September 1984): 270-289.

Kelman Herbert C., Processes of Opinion Change. Public Opinion Q. 25 (Spring 1961): 57-78.

Marcos, Anastasios C., Bahr, Stephen J., and Johnson, Richard B., Test of a Bonding Association Theory of Adolescent Drug Use. Social Forces 65 (September 1986): 135-161.

McGuire, William J., the Nature of Attitudes and Attitude Change, in The Handbook of Social Psychology, Second Edition, Vol. 3. Gardner Lindzey and Elliot Aronson, eds., Addison-Wesley Publishing Company, Reading, MA. 1969.

McLaughlin, RobertJ., Baer, Paul E., Burnside, Mary X, and Pokorny, Alex D., Psychosocial Correlates of Alcohol Use at Two State Levels During Adolescence. J. Stud. Alcohol 46 (1985): 212-218. Miller, Frederick D., Smith, Eliot R., and Uleman, James, Measurement and Interpretation of Situational and Dispositional(1981): 80-95.

Miniard, Paul W., and Cohen, Joel B., Modeling Personal and Normative Influences on Behavior. J. Consumer Res. 10 (September 1983): 169-180.

Needle, Richard, McCubbin, Hamilton, Wilson, Marc, Reineck, Robert, Lazar, Amnon, and Mederer, Helen, Interpersonal Influences in Adolescent Drug Use-The Role of Older Siblings, Parents and(1986): 739-766.

Rose, Randall L., Bearden, William 0., and Teel, Jesse E., An Attributional Analysis of Resistance to Group Pressure Regarding Illicit Drug and Alcohol Consumption. J. Consumer Res. 19 (June 1992): 1-13.

Ross, Lee, Bierbrauer, Gunter, and Hofman, Susan, The Role of Attribution Processes in Conformity and Dissent. Am. Psychol. 31 (February 1976): 148-157.

Smith, Eliot, and Miller, Frederick D., Latent-Variable Models of Attributional(1982): 221225.

Stein, Judith A., Newcomb, Michael D., and Bentler, P. M., An 8-Year Study of Multiple Influences on Drug Use and Drug(1987): 1094-1105.

Swath, Harith S., Adolescent Drug Taking: Role of Family and Peers. Drug Alcohol Depend. 21 (1988): 157-160.

Witt, Robert E., Informal Social Group influence on Consumer Brand Choice. J. Marketing Res. 6 (November 1969): 473-476.