Heavy episodic drinking among adolescents: a test of hypotheses derived from control theory.

Abstract

Despite a growing body of literature on the causes of heavy episodic drinking, little attention has been paid to this phenomenon in the sociological and criminological literature. This research assesses the extent to which a popular theory of crime and deviance, control theory, can explain heavy episodic drinking. Analysis of data collected from a convenience sample of 938 9th, 10th, and 11th grade students shows that attachment to parents, belief in the law, commitment to school, and self-control are all strong predictors of adolescent heavy episodic drinking. Thus, control theory contributes to a more comprehensive understanding of adolescent heady episodic drinking. The implications of the control theory notion that social integration is a deterrent to heady episodic drinking are discussed.

INTRODUCTION

Heavy episodic drinking among young people is increasingly perceived as an important social problem, and it has received considerable scholarly attention in recent years. The NIAAA created a Task Force on College Drinking in 1998, and in the past several years the journals Psychology of Addictive Behaviors and the Journal of Studies on Alcohol have devoted special issues to heavy episodic drinking. Research has revealed a number of important correlates of heavy episodic drinking, including developmental factors such as the transition to college, individual factors such as impulsivity and fraternity/sorority membership, and environmental factors such as perceived drinking norms and availability of alcohol (NIAAA, 2002). Despite the increasing focus on heavy episodic drinking among both scholars and the popular media, there has been little attention given to this issue in the sociological and criminological literature. Likewise, much research that has been conducted on heavy episodic drinking has not incorporated sociological or criminological perspectives or research findings. The primary purpose of this paper is to assess the extent to which a popular theory of crime and deviance, control theory, can explain heavy episodic drinking in a sample of high school students. We also explore the implications of our findings for further research on heavy episodic drinking incorporating the control theory perspective.

Social control theory (Hirschi, 1969) focuses on social forces restraining individuals from committing criminal or deviant acts, such as heavy drinking. The theory assumes that such behavior is a natural consequence of human nature, viewing deviance as a means to achieve desires with little effort. Thus, social control theory views deviance as nonproblematic, and instead aims to explain conformity. According to Hirschi (1969), most people conform most of the time due to their bond to society, which consists of four elements. An individual's emotional attachment to others, such as parents, deters deviation because individuals are concerned about the loss of respect that would likely result if their deviation were discovered. Commitment to long-term goals, such as educational or occupational pursuits, reduces the likelihood of deviation because it might jeopardize the individual's ability to achieve these goals. Involvement in conventional activities reduces the likelihood of deviance because time is limited--busy people have less time to get involved in deviant or illegal activity. Finally, belief in the moral validity of the law provides an additional restraint from deviance--to the extent that individuals believe that the law should be obeyed, they will be less likely to deviate from it. The four elements of the social bond are analytically separable, but Hirschi argued that they are empirically related--those with a strong bond in one area are likely to have strong bonds in other respects as well. In short, those who are more socially integrated are less likely to deviate from social norms.

A more recent version of control theory focuses less on social control and more on internalized control produced through early childhood socialization. Gottfredson and Hirschi's (1990) self-control theory starts with the same assumption as social control theory, or that people are naturally inclined to commit criminal acts and acts "analogous" to crime, such as heavy drinking, unprotected sex, and reckless driving. Gottfredson and Hirschi argue that these behaviors have important characteristics in common--they tend to provide immediate gratification and often have long-term negative consequences, require little skill or planning, and tend to be exciting or thrilling. Corresponding to these characteristics of criminal and analogous acts, Gottfredson and Hirschi argue that those with low self-control tend to pursue short-term gratification, to fail to consider the consequences of behavior, to be self-interested, and to prefer physical, risky, or thrilling activities to more intellectual or sedate ones. Thus, they argue that crime and analogous behaviors are the result of individuals pursuing immediate, self-interested goals when an easy opportunity to do so arises.

Gottfredson and Hirschi (1990) thus view deviant behavior as essentially unsocialized behavior, and argue that self-control is an acquired characteristic that is produced through early childhood socialization. Parents must consistently and appropriately supervise and punish children for disruptive or dangerous behavior. If they fail to do so, children will not develop self-control, and self-control is a characteristic that is unlikely to be learned later in life. Thus, Gottfredson and Hirschi highlight the stability of self-control over the life course, and emphasize the extent to which criminal and analogous behaviors also tend to remain stable over time.

Social and self-control theories make similar empirical predictions, albeit for slightly different reasons. For example, the two theories both point to the importance of family-related variables in reducing deviant behaviors. Social control theory explicitly predicts that current social bonds reduce delinquency due to the individual's desire to maintain those bonds (Hirschi, 1969). Self-control theory points to the importance of adequate early socialization in producing conformity, essentially arguing that social control in early childhood produces self-control later in life (Gottfredson & Hirschi, 1990). Because self-control theory sees stable adult relationships as one outcome of high levels of self-control, it predicts that those with stronger bonds to family and others should be less likely to engage in deviance, essentially viewing current social bonds as an indicator of one's level of self-control. Similarly, both theories predict that measures of commitment to school will be negatively associated with deviance. For social control theory, commitment provides a "stake in conformity" that individuals are unwilling to risk losing. For self-control theory, commitment can again be viewed as an indicator of one's underlying level of self-control--those who work hard at school do so largely because they have higher levels of self-control. Thus, while the causal mechanisms differ between the two theories, their empirical predictions are sufficiently similar to allow us to test the two theories simultaneously.

Control theory's major predictions have received substantial empirical support. Attachment to parents, commitment to educational and occupational pursuits, and belief in the validity of the law are robust correlates of delinquent behavior (Costello & Vowell, 1999; Hirschi, 1969; Wadsworth, 2000). Hirschi's (1969) predictions regarding involvement in conventional activities have been less well supported, likely because so many delinquent activities require very little time to complete (Hirschi, 1969). Self-control theory (Gottfredson & Hirschi, 1990) has also garnered substantial empirical support, with virtually all tests of the theory finding self-control to be negatively related to criminal or analogous behavior (Pratt & Cullen, 2000).

Studies examining control theory's ability to explain substance use have also been largely supportive of the theory. For example, Krohn and Massey (1980) found that social control theory explained 29% of the variation in alcohol/marijuana use among their sample of adolescents, and other studies of adolescent substance use find similarly supportive results (Lopez, Redondo, & Martin, 1989; Murray, 1986.) Keane, Maxim, and Teevan (1993) found that measures of low self-control were associated with drinking and driving for both males and females. However, most criminological research conceptualizes delinquency or deviance as a set of illegal or socially undesirable behaviors, and uses composite indexes of a variety of behaviors such as property offenses, violent crimes, and substance use. While this approach is consistent with control theory's focus on the commonalities between various types of deviant behavior, it limits our ability to draw conclusions about the theory's ability to explain drinking specifically.

This approach has also likely minimized the impact of criminological research on more specialized areas of research in psychology and other social sciences, including heavy episodic drinking research. The recent special issues of journals noted previously cite very few criminological or sociological studies, and there have been no studies to date on the ability of the most influential theories of crime to explain heavy episodic drinking. We believe that this lack of disciplinary integration has had negative consequences on the state of knowledge about heavy episodic drinking, in that criminological theories can provide a framework for better understanding the known correlates of heavy episodic drinking and can suggest new avenues for research that can inform both the heavy episodic drinking literature and the field of criminology.

For example, studies on the individual correlates of heavy episodic drinking sometimes include family-related variables such as parental norms related to alcohol or level of parental supervision (Beck & Treiman, 1996; Fitzgerald & Arndt, 2002; Lo, 1995; Perkins, 2002; Wechsler, Dowdall, Davenport, & Castillo, 1995). Such studies generally find that parents have only modest impact on their children's drinking (Perkins, 2002), and family-related factors are not a focus of most research on heavy episodic drinking among college students. However, social control theory suggests that the extent to which children, even adult children, are attached to their parents should influence the extent to which children feel the "psychological presence" of their parents and consider parents' likely reactions to their behavior (Hirschi, 1969). Thus, social control theory suggests that attachment to parents is an important additional consideration in studies of parental influence.

Self-control theory also provides an alternative framework for organizing some of the known correlates of heavy episodic drinking. For example, this perspective suggests an alternative to the view that heavy episodic drinking leads to negative consequences such as poor academic performance, problems with interpersonal relationships, or other forms of deviant behavior (D'Amico, McCarthy, Applebaum, Metrik, Frissell, & Brown, 2001; Jones, Oeltmann, Wilson, Brener, & Hill, 2001). Because self-control theory emphasizes the versatility of individuals' deviant behavior, it suggests that heavy episodic drinking and other correlated behaviors are both consequences of an underlying tendency toward reckless, self-interested, and shortsighted actions. Although self-control theory would not argue that heavy episodic drinking cannot be causally related to academic performance, it would argue that those students who drink heavily would be likely to show lower academic performance than their peers even if they were prevented from drinking. Similarly, self-control theory's generality helps us understand the correlation between heavy episodic drinking and smoking or other forms of substance use (Jones et al., 2001; Johnson, Boles, Vaughan, & Keller, 2000). This generality also helps account for findings showing that a number of different personality measures, such as impulsivity and sociability, are related to heavy episodic drinking (Baer, 2002).

For these and other reasons discussed below, we believe that the control theory perspective has much to offer studies of heavy episodic drinking. To empirically assess this claim, we hypothesize that measures of control theory constructs will be significantly related to heavy episodic drinking among our adolescent sample. From social control theory, we hypothesize that attachment to parents, belief in the law, and commitment to conventional goals will be positively related to each other, and negatively associated with heavy episodic drinking net of the effects of demographic characteristics, religiosity, and peers' delinquency. (1) From self-control theory, we hypothesize that self-control will be positively related to each of the social control measures noted above, and that self-control will be negatively related to heavy episodic drinking net of the effects of the other variables in the model.

METHODS

The data for this study were collected as part of an effort to replicate Hirschi's (1969) original test of social control theory. A self-administered questionnaire was completed by a convenience sample of 1,136 9th through 11th grade students in four schools in Fayetteville, Arkansas in 1997.

The control theory measures are summated-rating indexes (see Appendix for all questionnaire items.). Measures of background characteristics included gender, race, family's welfare status, and family structure. A measure of religiosity was included as a statistical control, given that prior research has shown it to be related to problem drinking and heavy episodic drinking (Donovan, Jessor, & Costa, 1999; Wechsler et al., 1995). Friends' delinquency was also included as a control variable because prior research on both heavy episodic drinking and delinquency shows a robust relationship between individuals' behavior and that of their friends (Bahr, Marcos, & Maughan, 1995; Costello & Vowell, 1999; Donovan et al., 1999; Perkins, 2002).

Heavy episodic drinking was measured with the question, "During the past 30 days, on how many days did you have 5 or more drinks of alcohol in a row, that is, within a couple of hours?" Gender specific measures of heavy episodic drinking were not available. The original response format included 5 ordered categories giving a range of days. We estimated the number of heavy episodic drinking days as the midpoint of the original category interval: "0 days" = 0, "1 to 3 days" = 2, "4 to 6" days = 5, "7 to 10 days" = 8.5, and "more than 10 days" = 20. To assess the degree to which our results might be sensitive to the somewhat arbitrary assignment of 20 to the 44 students who reported heavy episodic drinking on "more than 10 days," we replicated the analysis setting the last category equal to 10, 15, 25, and 30 drinking days. The results were substantively and statistically consistent with those we report.

Only 996 (87.4%) of the students answered the question on heavy episodic drinking, which was included near the end of the questionnaire. From previous research we know that students who fail to complete questionnaires tend to have more academic problems and are more likely to commit delinquent acts than completers. However, it is not our purpose to provide an accurate estimate of the prevalence of heavy episodic drinking among Arkansas youth, and the data include a sufficient number of cases to test our hypotheses after listwise deletion of missing cases. Our results are based on the 938 students for whom complete data are available.

As is typical of count variables, frequency of heavy episodic drinking was not approximately normally distributed; almost 2/3 (66.4%) reported zero heavy episodic drinking days in the past month. Therefore, we used negative-binomial regression to analyze the data. We standardized all continuous predictor variables prior to analysis. Following Long (1997), effects are expressed as the percentage change in the number of heavy episodic drinking days for a one-unit change in the predictor variable.

RESULTS

The sample was 46.5% male and 89.7% White, with 56.7% living with both biological parents. Approximately 3.9% were living in families currently receiving welfare benefits, and an additional 12.7% had received welfare benefits in the past. The average age of students in the sample was 15.7 years old, with about 38.5% in the 9th grade, 32.2% in the 10th grade, 29.3% in the 11th grade.

Those who answered the question on binge drinking (completers) did not differ significantly from non-completers with respect to age (t = .93, p = .350) but females (92.7%) were more likely ([chi square] = 27.18, p = .000) to complete the question than males (82.4%) and whites (89.1%) were more likely ([chi square] = 14.72, p = .000) to complete the question than minorities (77.7%). Children living in families that had ever received welfare (82.5%) were slightly though significantly ([chi square] = 7.34, p = .007) less likely to complete the question than those in families that had never received welfare benefits (89.4%), and those living in family with both biological parents present (89.9%) were slightly and significantly ([chi square] = 7.73, p = .005) more likely to complete the question than those in families with other structures (84.4%).

Overall, 33.6% of the sample reported at least one episode of heavy drinking in the month prior to survey administration. This is similar to the prevalence of heavy episodic drinking reported in a similar sample, in that D'Amico et al. (2001) report that 35% of their high school sample were "persistent binge drinkers" (2001: 341). National data show a heavy episodic drinking prevalence in 2000 of 26.2% among 10th graders in a 2-week time frame (U.S. Department of Health and Human Services, 2000). The similarity between the prevalence of heavy episodic drinking in our sample and these other studies lends confidence to the validity of our measure.

Because we expect the control theory measures to be significantly related to each other, we first performed a correlation analysis on these measures (see Table 1). As predicted, the social control and self-control measures are all significantly related, with particularly strong relationships between self-control and commitment, and self-control and belief. Factor loadings from exploratory factor analysis of the 35 items underlying the self-control and social-control indexes (estimated by iterative principal factoring, promax rotation, and 4-factor solution) supports the control theory prediction that these measures are empirically distinct despite the strong relationships between them. All items common to a specific index loaded at .38 or higher on the same factor (all but four items had loadings on the expected factor of .45 or higher) and only two items had loadings on other factors that exceeded the observed loading on the expected factor. These results do suggest caution, however, in interpreting the results of multivariate analyses including all of these independent variables.

The unadjusted (bivariate) effects on heavy episodic drinking are presented in the first two columns of Table 2. The results show that students more likely to report heavy episodic drinking are older, male, from families that receive welfare benefits, are not living in intact families, are attending religious services less frequently, and are more likely to have delinquent friends than those who have not drunk heavily in the month prior to the survey. Race was not significantly related to heavy episodic drinking. As hypothesized, all of the control theory measures were related to heavy episodic drinking, so that those with higher levels of attachment to parents, commitment to goals, belief in the law, and self-control were less likely to report drinking.

The results of the multivariate analysis are presented in the third and fourth columns of Table 2. In terms of demographic correlates, age is still positively related to heavy episodic drinking, but controlling for the effects of other variables in the model reduced the effects of sex and welfare status to nonsignificance. The control theory variables are all significantly and negatively related to the number of heavy drinking days, providing support for our major hypothesis. For example, for each standard deviation increase in commitment to long-term goals, the expected number of heavy drinking days decreases by 19.5%, and for each standard deviation increase in belief in the validity of the law, the expected number of heavy drinking days decreases by about 41%. Friends' delinquency persists as a major predictor of heavy episodic drinking, with a one standard deviation increase in the number of delinquent friends associated with a 57% increase in the number of heavy episodic drinking days.

DISCUSSION

The results support our hypothesis that control theories can help explain heavy episodic drinking. Measures of control theory have statistically significant and substantively important effects on heavy episodic drinking among this sample of 9th through 11th grade students, suggesting that future studies of heavy episodic drinking could benefit by including such measures.

While the current study had relatively modest goals, specifically to determine whether control theory can explain some of the variation in heavy episodic drinking, we believe that the implications of our study are more far-reaching. For example, self-control theory predicts stability in crime and analogous acts over time, consistent with research showing that rates of heavy episodic drinking in high school are strongly correlated with heavy episodic drinking in college (Lo, 1995; Wechsler et al., 1995). In general, this suggests that at least some of the underlying causes of heavy episodic drinking come into play before students enter college. Our results point to some of these possible causes, including the quality of relationships with parents, and assuming that self-control is learned in early childhood, parental childrearing practices. Thus, control theory offers one set of potential causes of heavy episodic drinking that have been under investigated in the literature. We therefore agree with Beck and Treiman's (1996) recommendation that the alcohol literature consider a wider range of parent variables as possible causes of alcohol misuse among young people.

Control theory's general focus on social integration as a deterrent to a variety of deviant behaviors also points to some new avenues of research on heavy episodic drinking. Although the current research is cross-sectional and thus cannot determine causal direction, control theory suggests that a number of variables often viewed as consequences of heavy episodic drinking might be usefully viewed as causes or merely correlates of it. For example, a low level commitment to academic goals may be viewed as a cause of heavy episodic drinking from the perspective of social control theory, suggesting that any efforts to reduce drinking to improve students' performance would be targeting the effect rather than the cause. Similarly, social control theory would argue that problems with relationships, rather than being consequences of drinking, are an indicator of a low level of attachment and a cause of heavy drinking (Hirschi, 1969). Self-control theory predicts a spurious correlation between relationship problems and heavy drinking--both are seen as consequences of low self-control (Gottfredson & Hirschi, 1990). Future research on heavy episodic drinking designed to test some of these hypotheses could contribute to the drinking literature as well as to the more general body of crime and delinquency literature.

The idea that social integration deters deviant behavior also suggests a different view of the correlation between individuals' and friends' deviance. Criminological learning theories (Akers, 1973; Sutherland, 1947) hold that friends' deviance is causally related to the individual's deviance due to processes of value transmission, so that members of deviant groups see deviant behavior as acceptable and are thus more likely to engage in it. This perspective is similar to normative explanations of college student drinking that argue that perceived group norms, or the actual behavior of peers, encourage students to drink to excess (Beck & Treiman, 1996; Perkins, 2002). It is also consistent with claims that students who are more socially integrated are more likely to drink heavily (Perkins, 2002). However, the argument that higher social integration leads to higher levels of deviant behavior is contrary to control theory, which explicitly argues that social integration is associated with lower levels of deviance.

Control theory assumes that acts that have negative consequences for oneself and especially for others are universally disapproved. Although the extent to which one believes in the norms of society can vary, this varies largely as a result of individuals' level of integration in other respects. Hirschi (1969:16) quotes Durkheim, "The more weakened the groups to which [the individual] belongs, the less he depends on them, the more he consequently depends only on himself and recognizes no other rules of conduct than what are founded on his private interests" (as cited in Durkheim, 1951:209). Thus, control theory predicts that what we might refer to as "deviant groups" tend to be only loosely knit, with members relatively low in concern for each other's welfare. This prediction suggests that in addition to using measures of social integration like fraternity membership or number of friends (see Perkins 2002 for a review), a more complete measure would also take into consideration the esteem with which these friends and associates are regarded. The conflicting views on the effect of social integration held by control theory and normative explanations of heavy episodic drinking provide the opportunity for a critical test between the two perspectives, and thus could contribute to the longstanding debate about these issues in the criminological literature.

Additional hypotheses can be derived from control theory's explanation for the correlation between peer deviance and individual deviance. Control theory explains this correlation in two ways--self-selection or opportunity. As some studies of heavy episodic drinking have noted (Baer, 2002; Bahr et al., 1995; Beck & Treiman, 1996), it is possible that students who drink heavily tend to seek out friends who also drink heavily, implying that the correlation between friends' and individuals' behavior is spurious. An alternative is that friends do have a causal effect on heavy episodic drinking, but only because they increase opportunity (Longshore & Turner, 1998; Osgood, Wilson, O'Malley, Bachman, & Johnston, 1996). Having friends who drink increases the availability of alcohol or provides the individual with temptations to drink that may not exist for the individual with few or no friends who drink. Control theory thus points out that a causal effect of peers on the individual's deviant behavior may simply be the result of a greater number of invitations to engage in deviant behavior (Costello, 1997).

Finally, our findings suggest that it would be of value to assess the relative merits of control theories and existing theories of heavy episodic drinking or deviant behavior. For example, there are some important similarities and differences between control theory and Jessor and Jessor's (1977) problem behavior theory. Both incorporate consideration of attachment to others, but problem behavior theory draws on elements of strain theories of delinquency (Merton, 1957) that are explicitly rejected by control theories. Both theories also account for versatility in offending, but control theory emphasizes stability over time and problem behavior theory takes a more developmental approach. It may also prove useful to assess the extent to which well-accepted predictors of heavy episodic drinking, such as alcohol expectancies, are themselves associated with lower levels of self-control or social control. Such approaches would go a long way toward greater cross-disciplinary integration in the study of heavy episodic drinking as well as other deviant or criminal acts, and could lead to new approaches to the prevention of heavy episodic drinking.

Clearly, we cannot resolve all of these issues with a single cross-sectional study of heavy episodic drinking. However, our results do suggest that social integration, broadly conceptualized, can reduce the likelihood of heavy episodic drinking, and thus that control theory holds promise as a new direction of heavy episodic drinking research.

APPENDIX

Questionnaire items and reliability of indexes

Unless otherwise noted, all measures were scored with a 5-point Likert response framework, and appropriate items were reverse scored so that a higher scale score equals a higher level of the variable measured.

Attachment to Parents (alpha = .81)

I share my thoughts and feelings with my mother.

I would like to be the kind of person my mother is.

I would like to be the kind of person my father is.

I talk over future plans with my parents.

My mother seems to understand me.

My parents know where I am when I am away from home.

My parents always know who I am dating.

Commitment (alpha = .72)

On average, how many hours per week do you spend studying or doing homework?

How important is getting good grades to you? (4-point scale ranging from completely unimportant to very important.)

What kind of grades do you get?

How much education do you plan to get? (5-point scale ranging from "some high school" to "college graduation")

Do you care what teachers think of you? (3-point scale ranging from "I care a lot" to "I don't care much")

I usually finish my homework.

Belief (alpha = .81)

Only fools tell the truth all the time.

To get ahead, you have to do some things that are not right.

It's okay to get around the law if you can get away with it.

Rules were made to be broken.

It's okay to take something from big business because they won't miss it.

I have lots of respect for the police.

There is no good reason for one person to hit another.

Most things people call delinquency don't really hurt anyone.

Fools and suckers deserve what they get.

Serf-Control (alpha = .72)

I am usually pretty cautious.

Sometimes, I take a risk just for the fun of it.

Hitchhiking is too dangerous for me.

I don't devote much thought and effort to preparing for the future.

I have had dates with more than one person in the same week.

I live for today and let tomorrow take care of itself.

An easy life is a happy life.

I try to save as much money as I can.

I lose my temper easily.

I try to get things I want even when I know that it's causing problems for other people.

I try hard in school.

I see no need for hard work.

Whatever I do I try hard.

Control Variables

Age in years

Sex (1 if male)

Race (1 if white, 0 otherwise)

Family Structure, composite of two questions asking, "Who currently acts as a father [mother] for you? Coded 1 if biological mother and father both lived at home, 0 otherwise.

Friends' Delinquency, response to the question "Have any of your friends been picked up by the police?" Ranges from 0 to 4 or more friends.

Welfare Status (1 if family had ever received welfare, 0 otherwise)

Religious Attendance, response to the question, "On average, how often do you attend religious services?" Response categories ranged from "never or almost never" to "once a week or more," on a 5-point scale.

Author Note

Barbara J. Costello, Department of Sociology and Anthropology, University of Rhode Island; Bradley J. Anderson, Division of General Internal Medicine, Rhode Island Hospital; Michael D. Stein, Brown University School of Medicine, Division of General Internal Medicine.

The data for this study were collected under NSF Grant SBR-9412316, "A Measurement Study of Control Theory Explanations of Delinquent and Deviant Behavior." We thank Travis Hirschi and Michael Gottfredson for making these data available for this study. An earlier version of this manuscript was presented at the 2001 annual meeting of the Research Society on Alcoholism.

Correspondence concerning this article should be addressed to Barbara J. Costello, Department of Sociology and Anthropology, University of Rhode Island, Kingston, RI 02822, Phone [phone]; Fax [phone]. E-mail: [e-mail].

References

Akers, R. L. (1973). Deviant behavior: A social earning approach. Belmont, CA: Wadsworth.

Baer, J. S. (2002). Student factors: Understanding individual variation in college drinking. Journal of Studies on Alcohol, Supplement No. 14, 40-53.

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.

Beck, K. H., & Treiman, K. A. (1996). The relationship of social context of drinking, perceived social norms, and parental influence to various drinking patterns of adolescents. Addictive Behaviors, 21, 5, 633-644.

Costello, B. J. (1997). On the logical adequacy of cultural deviance theories. Theoretical Criminology, 1, 403-428.

Costello, B. J., & Vowell, P. R. (1999). Testing control theory and differential association: A reanalysis of the Richmond youth project data. Criminology, 37, 815-842.

D'Amico, E. J., McCarthy, D. M., Applebaum, M., Metrik, J., Frissell, K. C., & Brown, S. A. (2001). Progression into and out of binge drinking among high school students. Psychology of Addictive Behaviors, 15, 341-349.

Donovan, J. E., Jessor, R., & Costa, F. M. (1999). Adolescent problem drinking: Stability of psychosocial and behavioral correlates across a generation. Journal of Studies on Alcohol, 60, 352-361.

Durkheim, E. (1951). Suicide. New York: The Free Press.

Fitzgerald, J. L., & Arndt, S. (2002). Reference group influence on adolescent alcohol use. Journal of Alcohol and Drug Education, 47, 42-56.

Gottfredson, M. R., & Hirschi, T. (1990). A general theory of crime. Stanford, CA: Stanford University Press.

Hirschi, T. (1960). Causes of delinquency. Berkeley, CA: University of California Press.

Jessor, R., & Jessor, S. L. (1977). Problem behavior and psychosocial development: A longitudinal study of youth. New York: Academic Press.

Johnson, P. B., Boles, S. M., Vaughan, R., & Keller, H. D. (2000). The co-occurrence of smoking and binge drinking in adolescence. Addictive Behaviors, 25, 779-783.

Jones, S. E., Oeltmann, J., Wilson, T. W., Brener, N. D., & Hill, C.V. (2001). Binge drinking among undergraduate college students in the United States: Implications for other substance use. Journal of American College Health, 50, 33-47.

Keane, C., Maxim, P. S., & Teevan, J. J. (1993). Drinking and driving, self-control, and gender: Testing a general theory of crime. Journal of Research in Crime and Delinquency, 30, 30-46.

Krohn, M. D., & Massey, J. L. (1980). Social control and delinquent behavior: An examination of the elements of the social bond. Sociological Quarterly, 21, 529-543.

Lo, C. C. (1995). Gender differences in collegiate alcohol use. Journal of Drug Issues, 25, 817-837.

Long, J. S. (1997). Regression models for categorical and limited dependent variables. Thousand Oaks, CA: Sage.

Longshore, D. L., & Turner, S. (1998). Self-control and criminal opportunity: Cross-sectional test of the general theory of crime. Criminal Justice and Behavior, 25, 81-98.

Lopez, J. M. O., Redondo, L. M., & Martin, A. L. (1989). Influence of family and peer group on the use of drugs by adolescents. International Journal of the Addictions, 24, 1065-1082.

Merton, R. K. (1957). Social theory and social structure. New York: Free Press.

Murray, G. F. (1986). Marijuana use and social control: A sociological perspective on deviance. International Journal of the Addictions, 21, 657-669.

National Institute on Alcohol Abuse and Alcoholism, Task Force on College Drinking (2002). High-risk drinking in college: What we know and what we need to learn. Retrieved from www.collegedrinking prevention.gov.

Osgood, D. W., Wilson, J. K., O'Malley, P. M., Bachman, J. G., & Johnston, L. D. (1996). Routine activities and individual deviant behavior. American Sociological Review, 61, 635-655.

Perkins, H. W. (2002). Social norms and the prevention of alcohol misuse in collegiate contexts. Journal of Studies on Alcohol, 14(Suppl.), 164-172.

Pratt, T. C., & Cullen, F. T. (2000). The empirical status of Gottfredson and Hirschi's general theory of crime: A meta-analysis. Criminology, 38, 931-964.

Sutherland, E. H. (1947). Principles of Criminology, 4th ed. Philadelphia: Lippincott.

United States Department of Health and Human Services (2000). Monitoring the future survey released: Moderating trend among teen drug use continues. HHS News, Dec. 14, 2000.

Wadsworth, T. (2000). Labor markets, delinquency, and social control theory: An empirical assessment of the mediating process. Social Forces, 78, 1041-1066.

Wechsler, H., Dowdall, G. W., Davenport, A., & Castillo, S. (1995). Correlates of college student binge drinking. American Journal of Public Health, 85, 921-927.

Barbara J. Costello, University of Rhode Island

Bradley J. Anderson, Rhode Island Hospital

Michael D. Stein, Brown University School of Medicine

(1) We do not test the social control theory prediction that involvement in conventional activities is associated with delinquency; due to the lack of empirical support for this prediction in previous research.

Table 1. Zero-order association between
measures of control theory constructs
(Spearman rank-order correlations) (a)

               Attachment
               to Parents   Commit.   Belief

Commitment        .405
Belief            .419       .481
Self-Control      .402       .618      .694

(a) All relationships significant at p < .000.

Table 2. Unadjusted (Bivariate) and Adjusted (Multivariate)
Estimated Effects for Negative Binomial Regression Models

                               UNADJUSTED EFFECTS

                            B      100 * (exp(b) 1) (b)

Age (a)                   .206 **         22.9%
Sex (Male)                .317 *          37.3%
Race (White)             -.32             -27.4%
Welfare (Yes)             .497 *          39.2%
Religious Attend. (a)    -.186 *          -17.0%
Fam. Struct. (Intact)    -.608 **         -45.6%
Friends' Delinq. (a)      .628 **         87.4%
Attach. to Parents (a)   -.301 **         -26.0%
Commitment (a)           -.703 **         -50.5%
Belief (a)               -.744 **         -52.5%
Self-Control(a)          -.724 **         -51.5%

                                ADJUSTED EFFECTS

                            B      100 * (exp(b) 1) (b)

Age (a)                   .246 **         27.8%
Sex (Male)               -.16             -14.8%
Race (White)              .041             4.2%
Welfare (Yes)             .092             8.8%
Religious Attend. (a)    -.083 **         -7.9%
Fam. Struct. (Intact)    -.383 *          -31.8%
Friends' Delinq. (a)      .451 **         57.0%
Attach. to Parents (a)   -.250 *          -22.1%
Commitment (a)           -.217 *          -19.5%
Belief (a)               -.526 **         -40.9%
Self-Control(a)          -.279 *          -24.3%

* p <.05; ** p < .01

(a) Effects for continuous predictor variables are semi-standardized.
Continuous predictor variables were standardized prior to analysis.

(b) Effects expressed as the expected percentage change in number
of heavy episodic drinking days for a unit change in the predictor
variable.