Adolescence-limited versus persistent delinquency: extending Moffitt's hypothesis into adulthood

The authors examined how neuropsychological, personality, and environmental risk factors and their interactions were related to trajectories of delinquent behavior from adolescence to adulthood. Four waves of longitudinal data from 698 male participants, ages 12-18 at Time 1 and ages 25-31 at Time 4, were included in the analyses. Using a growth mixture model approach, 4 trajectories were identified: nondelinquents, adolescence-limited delinquents, adolescence-to-adulthood-persistent delinquents, and escalating delinquents. Five risk factors distinguished escalating from persistent delinquents and 5 also distinguished nondelinquents from the 3 delinquency trajectories. Persistent delinquents scored significantly higher than adolescence-limited delinquents on only one risk factor, disinhibition. Overall, few of the factors that are related to childhood-to-adolescence persistence were associated with persistence in delinquency beyond adolescence.

Characteristics of girls with early onset disruptive and antisocial behaviour

BACKGROUND

Crime, particularly among juvenile females, has increased in recent years. Little is known, however, about the development and precursors in childhood of female delinquent behaviour. This is primarily due to a lack of consensus on how to define and assess female antisocial behaviour, and a lack of studies using sufficiently large samples.

METHOD

A community sample of 2451 girls between the ages of five and eight years were recruited into a longitudinal study following the enumeration of 103,238 households in the city of Pittsburgh. Data on disruptive and antisocial behaviours were collected from parents, teachers and children during the first wave of the study.

RESULTS

Prevalence rates of disruptive disorders varied by choice of informants and measurement thresholds. The prevalence of most disruptive behaviours was similar across the four age cohorts. Where there were differences, parents of younger girls tended to report fewer problematic behaviours compared with parents of older girls. Teachers reported more disruptive behaviours than parents and, by their reports, older girls were more likely to show oppositional/defiant behaviour and relational aggression than younger girls. Girls scoring highly on several domains relative to their peers were over-represented in disadvantaged neighbourhoods.

CONCLUSIONS

A range of disruptive disorders are present among a subgroup of females at an early age, particularly among girls in the most disadvantaged neighbourhoods. Longitudinal follow-up is required to examine the developmental trajectories and predictive utility of these behaviours. The implications for clinical interventions are discussed.

Adolescence-limited versus persistent delinquency: extending Moffitt's hypothesis into adulthood

The authors examined how neuropsychological, personality, and environmental risk factors and their interactions were related to trajectories of delinquent behavior from adolescence to adulthood. Four waves of longitudinal data from 698 male participants, ages 12-18 at Time 1 and ages 25-31 at Time 4, were included in the analyses. Using a growth mixture model approach, 4 trajectories were identified: nondelinquents, adolescence-limited delinquents, adolescence-to-adulthood-persistent delinquents, and escalating delinquents. Five risk factors distinguished escalating from persistent delinquents and 5 also distinguished nondelinquents from the 3 delinquency trajectories. Persistent delinquents scored significantly higher than adolescence-limited delinquents on only one risk factor, disinhibition. Overall, few of the factors that are related to childhood-to-adolescence persistence were associated with persistence in delinquency beyond adolescence.

Parental modeling and parenting behavior effects on offspring alcohol and cigarette use. A growth curve analysis

PURPOSE

This paper examined the effects of parental drinking and smoking, parental warmth and hostility, and their interactions on offsprings' drinking and smoking over time.

METHODS

We used four waves of prospective longitudinal data collected from 218 males and 214 females who were age 15 at Time 1 and age 28 by Time 4. Growth mixture modeling was used to develop offspring trajectory groups of cigarette smokers and alcohol drinkers. Logistic regression analyses were conducted to determine whether parent behaviors could predict offspring heavy drinking and smoking trajectories.

RESULTS

Four drinking and three smoking trajectory groups were identified for females and males. Parent drinking rather than parenting behavior predicted heavy drinking by offsprings and mothers' drinking was a slightly better predictor than fathers' drinking for both daughters and sons. Fathers' warmth and hostility was the best predictor of heavy smoking by sons. Neither modeling nor parenting significantly predicted female heavy smoking. For the most part, parent modeling did not interact with parenting behavior to predict smoking or drinking in offspring.

IMPLICATIONS

Parents affected their offspring's use of alcohol and cigarettes both through modeling and parenting behavior. However, the importance of modeling relative to parenting behavior differed by the type of substance. Prevention programs that focus on both the modeling of parental behaviors, as well as enhancing parenting skills, should be effective in influencing trajectories of substance use throughout adolescence and young adulthood.

Psychopathology as a predictor of adolescent drug use trajectories

The authors examined early psychopathology as a predictor of trajectories of drug use from ages 13-18 years. Six years of annual data were analyzed for 506 boys using a mixed effects polynomial growth curve model. They tested whether distinct measures of psychopathology and behavioral problems (i.e., attention-deficit/hyperactivity disorder, oppositional defiant disorder, conduct disorder, depression, and violence) assessed in early adolescence could prospectively predict level and change in alcohol and marijuana use. Higher levels of all of the types of psychopathology predicted higher levels of alcohol use, and higher levels of attention-deficit/hyperactivity disorder, conduct disorder, and violence predicted higher levels of marijuana use. Only conduct disorder predicted linear growth in alcohol use, and none of the measures predicted growth in marijuana use. The results suggest that drug use prevention programs should target youths with early symptoms of psychopathology.

Predictors of smoking cessation from adolescence into young adulthood

Although smoking cigarettes is hazardous to health and cessation has positive health benefits, few smokers are able to successfully quit. The purpose of this study was to examine the predictors of smoking cessation in a nonclinical sample of 134 male and 190 female, young adult, regular (daily) smokers within a social learning and maturing-out framework. Four waves of prospective, longitudinal data from a community sample followed from adolescence into young adulthood were analyzed. Logistic regression analyses were used to test the effects of differential associations, definitions, differential reinforcement, and changes in adult role status on smoking cessation in young adulthood. Becoming married to a nonsmoker and decreases in the proportion of friends who smoked were significant predictors of cessation. Current smokers and stoppers did not differ significantly in terms of prior intensity of cigarette use or alcohol abuse/dependence. They also did not differ in terms of psychological characteristics, including depression and prior coping use of cigarettes. Social networks were more important than social roles for predicting cessation in young adulthood. Thus, smoking cessation programs should focus on social learning processes.

The impact of childhood abuse and neglect on adult mental health: a prospective study

This paper examines the impact of three types of victimization in childhood--sexual abuse, physical abuse, and neglect--on lifetime measures of mental health among adults. In contrast to research that relies on retrospective recall of childhood victimization, this work uses a prospective sample gathered from records of documented court cases of childhood abuse and neglect in a midwestern city around 1970. These subjects were interviewed about twenty years later. In addition, this research compares outcomes of the 641 members of the abuse and neglect group with a matched control group of 510 persons who did not have documented cases of abuse or neglect. The results indicate that men who were abused and neglected as children have more dysthymia and antisocial personality disorder as adults than matched controls, but they did not have more alcohol problems. Abused and neglected women report more symptoms of dysthymia, antisocial personality disorder, and alcohol problems than controls. After controlling for stressful life events, however, childhood victimization had little direct impact on any lifetime mental health outcome. This research indicates the importance of adopting an approach that places childhood victimization in the context of other life stressors and of prospective changes over the life course.

Early adult outcomes of adolescent binge drinking: person- and variable-centered analyses of binge drinking trajectories

BACKGROUND

Many studies of the consequences of binge drinking take a variable-centered approach that may mask developmentally different trajectories. Recent studies have reported qualitatively different binge drinking trajectories in young adulthood. However, analyses of developmental trajectories of binge drinking have not been examined for an important period of drinking development: adolescence. The purpose of this study was to examine young adult outcomes of adolescent binge drinking using an approach that combines person-centered and variable-centered methods.

METHODS

Data were from the Seattle Social Development Project, an ethnically diverse, gender balanced sample (n = 808) followed prospectively from age 10 to age 21. Semiparametric group-based modeling was used to determine groups of binge drinking trajectories in adolescence. Logistic regression was used to examine how well the trajectory groups predicted young adult outcomes after demographics, childhood measures, and adolescent drug use were considered.

RESULTS

Four distinct trajectories of binge drinking during adolescence were identified: Early Highs, Increasers, Late Onsetters, and Nonbingers. These trajectories significantly predicted positive and negative outcomes in adulthood after controlling for demographic characteristics, early proxy measures of the outcome, and adolescent drug use.

CONCLUSIONS

This integrated person- and variable-centered approach provides more information about the effects of specific patterns of binge drinking than studies that employ variable-centered methods alone.

Personality, Environment, and Problem Drug Use

We tested a theoretical model that posited direct and Interactive relationships between personality and environment as predictive of problem drug use. Hierarchical regression analyses were performed on data from a nonclinical sample of males and females first tested in adolescence and followed into young adulthood. In main effects analyses, cross-sectional models for marijuana and a composite of “harder” drugs strongly supported our theoretical perspective: several environmental and personality variables significantly predicted concurrent problem use. Longitudinal analyses produced mixed results, with either smaller numbers of significant personality/environment predictors or personality predictors dropping out completely. Peer drug use and depression best predicted problem “hard” drug use, while motivations to use with others and disinhibition best predicted problem marijuana use. Personality/peer group interactions were significant both cross-sectionally and longitudinally, indicating that membership in a problem-using social network enhanced the predictive ability of, or helped bring to fruition, several personality predisposers.

Developmental associations between substance use and violence

This study examined the developmental associations between substance use and violence. We examined the trends in each behavior throughout adolescence, how the behaviors covaried over time, and the symmetry of associations taking into account frequency and severity of each behavior. We also examined whether changes in one behavior affected changes in the other behavior over time. Six years of annual data were analyzed for 506 boys who were in the seventh grade at the first assessment. Concurrent associations between frequency of substance use and violence were relatively strong throughout adolescence and were somewhat stronger for marijuana than alcohol, especially in early adolescence. Type or severity of violence was not related to concurrent alcohol or marijuana frequency, but severity of drug use was related to concurrent violence frequency. Depending, to some degree, on the age of the subjects, the longitudinal relationships between substance use and violence were reciprocal during adolescence and slightly stronger for alcohol and violence than for marijuana and violence. Further, increases in alcohol use were related to increases in violence: however, when early alcohol use was controlled, increases in marijuana use were not related to increases in violence. Only in early adolescence was the longitudinal relationship between marijuana use and later violence especially strong. The strength of the longitudinal associations between violence and substance use did not change when common risk factors for violence and substance use were controlled. Overall, the data lend more support for a reciprocal than for a unidirectional association between substance use and violence. Prevention efforts should be directed at aggressive males who are multiple-substance users in early adolescence.

Developmental aspects of delinquency and internalizing problems and their association with persistent juvenile substance use between ages 7 and 18

Analyzed longitudinal data from 3 samples of the Pittsburgh Youth Study on boys ages 7 to 18 to examine the co-occurrence of persistent substance use with other problem behaviors, including attention deficit hyperactivity disorder (ADHD), persistent delinquency, and persistent internalizing problems (i.e., depressed mood, anxiety, shy or withdrawn behavior). In preadolescence, persistent substance users also tended to be persistent delinquents, and half of this group displayed persistent internalizing problems as well. In adolescence, a third of the persistent substance users did not manifest other persistent problems. Across the samples, the least common substance users were those who manifested persistent internalizing problems only. Logistic regression analyses showed that persistent substance use in preadolescence was predicted by persistent delinquency and internalizing problems and in adolescence by persistent delinquency only. The combination of persistent substance use and delinquency was predicted by oppositional defiant disorder in middle childhood and by persistent internalizing problems in middle to late childhood. ADHD was not a predictor of persistent substance use (and delinquency) in any of the analyses. Results are discussed in terms of developmental models of multiproblem youth with an eye on improving early interventions.

How the negative and positive aspects of partner relationships affect the mental health of young married people

The relationship between marriage and positive mental health is one of the most established findings in the stress literature. Few studies, however, examine the problematic as well as the supportive impacts of marriage on mental health. This paper uses a cohort of young adults who were sampled at 18, 21, or 24 years of age and resampled seven years later when they were married at 25, 28, or 31. It examines what factors are associated with the quality of marital relationships, the relative impact of and balance between negative and positive partner relationships on mental health, and sex differences in the determinants and outcomes of marital quality. The results indicate that the structural strains of parenthood and financial need and their interaction predict problematic and supportive spousal relationships and the difference in the levels of these two relationships. Problematic relationships with spouses have considerably stronger impacts than supportive relationships on depression. However, the difference between the amounts of supportive and problematic relationships with spouses has a greater impact on mental health than levels of either considered separately. Finally, relational quality has a greater impact on the mental health of wives than husbands. These findings indicate the importance of considering how marriage affects mental health in complex, rather than in straightforward, ways.

Predicting problem drinking: a test of an interactive social learning model

This study tested a social learning model and explored the direct and interactive relationships between personality and environment in predicting problem alcohol use. We used longitudinal data from a nonclinical sample of males and females first tested in adolescence and followed into young adulthood. Hierarchial regression analyses were used to test main effects and interaction models. The cross-sectional data supported an interactive social learning model. Both personality and environment variables significantly predicted problem drinking. Two interactions between heavy drinking peer groups and personality variables were significant. Contrary to our hypothesis, the direction of the interaction was negative. In contrast, the longitudinal analyses did not provide strong support for our interactive model. Personality variables were significant predictors longitudinally, but in only one analysis did an environment variable significantly predict problem drinking. Furthermore, none of the interactions was significant predictors over time. Overall, the findings suggest that social learning models based on the interaction of personality and environmental influences may be more appropriate for predicting concurrent, as opposed to future problems, and that future research should include person-environment interactions. In addition, cultural tolerance of heavy drinking may be an important determinant of the role of psychological vulnerability in the development of problem drinking.

Longitudinal perspective on alcohol use and aggression during adolescence

While there is general agreement that alcohol use and aggression are related, few studies have examined this relationship among youth. This chapter reviews the literature on rates of alcohol use, aggression, and alcohol-related aggression among adolescents, as well as the cross-sectional and longitudinal associations among these behaviors. In general, the literature does not provide strong support for a unique association between alcohol use and aggressive behavior during adolescence. The observed relationship between alcohol use and aggression appears to be spurious because both behaviors are predicted by a similar set of individual, family, and environmental factors. Prevention programs that reduce these common risk factors should decrease both behaviors. Interventions with aggressive individuals, especially aggressive individuals who drink heavily, may be most indicated.

The use of multiple outcomes in stress research: a case study of gender differences in responses to marital dissolution

This study tests the hypothesis that the use of a single outcome variable distorts the mental health consequences of a stressor among different social groups. It uses the example of the impact of marital dissolution on the mental health of men and women to see whether rates of depression and alcohol problems rise disproportionately among women and men, respectively, who experience the same type of stressor. The sample compares 465 married subjects with 127 separated or divorced subjects drawn from a longitudinal study of 25-, 28-, and 31-year-olds. With controls for earlier rates of depression and alcohol problems, as well as for secondary stressors connected with separation and divorce, women undergoing marital dissolution show significantly greater increases in rates of depression compared to men who experience this stressor. Although men report far more alcohol problems than women, rates of these problems do not increase disproportionately among men, compared to women, during marital dissolution. The results indicate that the use of gender-typical mental health outcomes reduce, but do not eliminate,gender differences in the response to marital dissolution. They also indicate the need to use outcomes that typify how each group under study responds to stressful social conditions.

Cessation from cocaine use

This study explores factors that are related to cessation of cocaine use versus continued use in a non-clinical sample of American adolescents and young adults interviewed at three points in time. At time 3, cocaine stoppers (n = 104) and current users (n = 267) are compared in terms of age and sex, patterns of contemporary and prior drug use, life-style characteristics and a selected group of social learning variables. The data indicate that cocaine stoppers and users have similar patterns of alcohol, marijuana, cigarette, cocaine and other drug use at time 1 and time 2, but that users have higher time 3 frequencies of alcohol, marijuana and other drug use. In addition, those youth who stop are more likely to be married and have children, although the groups do not differ in terms of career/school status. The data lend partial support to a social learning perspective and indicate that differential associations (friends' use) and punishments (negative consequences) are most strongly related to cessation. In addition, users report more dependency symptoms than do stoppers.

Factors associated with inpatient and outpatient treatment for children and adolescents with serious mental illness

OBJECTIVE

This study describes the distribution of children and adolescents in psychiatric inpatient and outpatient facilities and identifies factors associated with the selection of individuals into inpatient versus outpatient care. SAMPLE DATA: The data are from a 1986 nationally representative sample surveyed by the National Institute of Mental Health.

RESULTS

Results indicate that the vast majority of children and adolescents with psychiatric problems receive outpatient treatment rather than inpatient care. Factors that predict psychiatric hospitalization rather than outpatient care are (1) public or private insurance coverage versus no insurance; (2) previous hospitalization; (3) psychiatric diagnosis of affective or psychotic disorders versus conduct disorders, adjustment disorders, drug and alcohol abuse, and other disorders; and (4) age, with adolescents more likely to be hospitalized than children.

CONCLUSIONS

Further research is needed to explore the role of insurance in mental health sorting processes. Moreover, systematic, controlled research is needed to determine how different financing strategies affect mental health outcomes for children and adolescents.

A longitudinal investigation of alcohol use and aggression in adolescence

Data from a prospective, longitudinal study of males and females tested at age 12, 15 and 18 years are used to study the relationship between alcohol use and aggression. Prevalence rates for alcohol use are similar for males and females. However, prevalence rates for aggressive behavior and alcohol-related aggression among females are lower than those for males and too low to permit meaningful analysis. Two series of nested structural equation models examine the interrelationships between alcohol use and aggressive behavior over time for all males in the sample and for male alcohol users only. The findings indicate that early aggressive behavior leads to increases in alcohol use and alcohol-related aggression, but that levels of alcohol use are not significantly related to later aggressive behavior. Thus, the data suggest that alcohol-related aggression is engaged in by aggressive people who drink. These data lend support to other research that indicates that early aggressive and antisocial behavior is predictive of later alcohol-related problems.

Self-attributed consequences of cocaine use

This study examines self-attributed consequences of cocaine use in a nonclinical sample of 1,270 young adults. A social learning perspective was used to examine the ability of motivational dimensions of use, as well as use patterns, to predict negative consequences of use. Four scales of negative consequences emerged: Negative Effects, Dependency Symptoms, Physical/Social, and Legal/Interpersonal consequences. Reasons for using cocaine, patterns of use, and their interactions significantly predicted negative consequences of use. Those subjects who endorsed coping reasons for use were more likely, while those who endorsed uplift reasons were less likely, to experience negative consequences.

The correspondence between saliva and breath estimates of blood alcohol concentration: advantages and limitations of the saliva method

Increased awareness of the devastating effects of alcohol misuse on our highways, workplaces and families, as well as on the individual, has resulted in increased social pressure to enforce driving-while-intoxicated laws and to develop educational, prevention and treatment programs. One aspect of this movement is to develop improved sobriety testing to ensure that laws are properly and fairly enforced and that there is compliance with abstinence in treatment. Although sophisticated blood and breath testing devices are available, field tests suggest that saliva alcohol tests based on alcohol-oxidase methodology offer advantages in portability, ease of administration, and cost and time efficiency. We evaluated the validity and reliability of a simple saliva test, based on the enzymatic oxidation of alcohol by alcohol oxidase, for estimating blood alcohol concentration. Ten subjects consumed various doses of alcohol and multiple saliva samples were obtained using alcohol sensitive saliva strips that change color in proportion to the concentration of alcohol. The reflectance values of reacted saliva strips were read by meter and estimates of blood alcohol concentration in the range of 10-90 mg/dl were compared to simultaneous estimates obtained from breath analysis using a Breathalyzer Model 900A. We also examined how alcohol levels changed over time in alcohol reacted saliva strips. The results of regression analysis indicated that the saliva strips and the Breathalyzer gave reasonably close estimates (r = 0.89-.90) of blood alcohol concentration. Correlation coefficents for the values of saliva samples read by meter measured at 10 minutes and at 18 days after collection ranged from .90 to 1.00, showing high test-retest reliability despite storage.(ABSTRACT TRUNCATED AT 250 WORDS)

Becoming married, depression, and alcohol problems among young adults

This paper examines three questions regarding the relationship between marriage and mental health, specifically depression and alcohol problems. First, does marriage lead to improved mental health compared to never marrying? Second, do any mental health benefits of marriage primarily accrue to men? Third, what qualitative aspects of marriage are related to psychological disorder? We explore these questions in a longitudinal sample of young adults sampled at age 21 and again at age 24. We find no indication that marriage reduces depression. Married people do report fewer alcohol problems than the never-married but this could be due to the selection of less problematic drinkers into marriage. We also fail to find that men receive disproportionate mental health benefits from marriage. Finally, we find that marital conflict is associated with problem drinking for men and depression for women. The results indicate the importance of considering stage in the life cycle and gender-sensitive indicators of psychological disorder in studies of marriage and mental health.

Risk factors of adolescent drug use: an affect-based interpretation

We examined the continuity and stability of relational patterns among putative risk factors and measures of total drug use over the age period of 12 to 18. Using Marlatt's (1987) affect-based model of moderation versus dependence, risk factors were classified on the basis of their assumed relationships to (a) positive and negative affect and (b) low versus high constraint. It was hypothesized that two distinct developmental pathways characterize the emergence of drug use in a normal population sample of adolescents: one indicative of low constraint and a predominance of positive affect, the other indicative of low constraint and a predominance of negative affect. Results of principal components analyses generally support the hypothesis. Theoretical and methodological implications are discussed.

An investigation of factors related to intoxicated driving behaviors among youth

This study assessed the prevalence of driving under the influence of alcohol and marijuana among a sample of 18 and 21 year olds and examined the across-time relationships between intoxicated driving and consumption, risk-taking/impulsive orientation, negative intrapersonal state, stress and use of alcohol and other drugs to cope with problems. Self-report data were collected from 556 men and women, ages 18 and 21, at two points in time. The data indicated that at least a minimum level of drinking and driving, as well as smoking marijuana and driving, is engaged in at least once for the majority of youth. Correlations between eight driving behaviors and consumption variables indicated that frequency of substance use was strongly related to frequency of driving while intoxicated (DWI). Regression analyses revealed that coping use of substances was the strongest predictor of driving under the influence. A path model examining the effect of stress, negative states and risk-taking orientations (T1) on driving under the influence as mediated through coping use (T2) was tested. Results showed that risk-taking orientation was the strongest predictor of DWI, both directly and indirectly (as mediated through coping use). Findings suggest that impaired driving may be part of a global syndrome of risk-taking behavior and is an activity engaged in most often by those who frequently use alcohol and other drugs to cope with problems.

Towards the assessment of adolescent problem drinking

Longitudinal data were obtained from a nonclinical sample of 1,308 male and female adolescents covering the age range from 12 to 21. Factor analyses of 52 symptoms and/or consequences of alcohol use yielded three problem dimensions. In addition, a unidimensional, 23-item scale (the Rutgers Alcohol Problem Index, RAPI) was constructed with an internal consistency of .92. Correlations between RAPI and alcohol-use intensity were moderately strong for all age groups at each test occasion (ranging from .20 to .57), yet low enough to suggest that identification of problem drinkers requires both types of measures. The results suggest that the RAPI may be a useful tool for the standardized and efficient assessment of problem drinking during adolescence.

Longitudinal patterns of cocaine use among adolescents

Developmental patterns of cocaine use were examined in a longitudinal sample of 1,308 male and female adolescents. The data indicate that substantial increases in use occurred between 15 and 18 and between 18 and 21 years old. Although there were developmental changes, there were no significant increases over time for same aged individuals. Gender differences within age groups were not statistically significant, although some differences were apparent. Examinations of intraindividual changes in cocaine use among continuous users indicate significant individual increases in all measures of cocaine use over a 3-year period. While continuous users increased their use of cocaine, they experienced decreases in their frequency of use of alcohol, marijuana, and other drugs. Analyses also suggest that those individuals who initiated cocaine use between T1 and T2 were already different from their age peers in terms of their cigarette, alcohol, and marijuana use at T1. Overall, the findings suggest that patterns of cocaine use may be more dependent upon the number of years of use rather than on the age of the user.

Longitudinal stability and dimensional structure of problem drinking in adolescence

A multivariate analysis of dimensions of problem drinking and their stability across time is conducted through a series of confirmatory factor analyses (as opposed to exploratory factor analyses used in previous studies), using self-report data from a longitudinal sample of adolescents and youth. Analyses are performed separately by age and gender. Results indicate that traditional measures of problem drinking represent at least two distinct dimensions--intensity of use and use-related problems--rather than a unitary construct for adolescent males and females. The results also suggest that dimensions of problem drinking remain relatively stable from 15 to 21 years of age, except that alcohol-related problems are unstable for males from 15 to 18 years of age.

An application of three deviance theories to adolescent substance use

This study tests the applicability of aspects of three theories of deviance--differential association, control, and strain--to the use of alcohol and drugs among a representative cross-section of 12-, 15-, and 18-year-olds. Regression analyses are conducted separately on each theory as well as on an overall model combining aspects of all three perspectives. Results show that differential association theory is a far more powerful predictor of adolescent alcohol and drug use than either the control or strain theories. The predictive power of the overall model is dependent upon the type of substance used as well as the age of the subject; however, the model is invariant between males and females.