Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
COPTER 2 ADOLE SCENT SEXUAL BEHAVIOR AS IT RELATES TO OTHER TRANSI TION BEHAVIORS IN YOUTH Margaret E. Ensminger The way in which adolescent sexual activity is linked to other adolescent behaviors is the subject of this report. Two distinct but related questions will be reviewed. First, how similar or different are the theoretical perspectives and empirical results that have been associated with studies of adolescent sexuality compared to those that are associated with other adolescent behaviors? Second, to what extent does adolescent sexual activity covary with certain adolescent problem behaviors such as drug, alcohol, and cigarette use or delinquent be- havior or other adolescent activities such as school performance or employment? These questions have several important policy and theoretical implications. First, if several kinds of adolescent behavior tend to covary together (referred to as a syndrome by Jessor, 1983) and have similar correlates and antecedents, then designing separate inter- ventions for each specific behavior may be inefficient. Second, by examining each behavior in isolation from the others, our understanding of the nature and origins of each may be limited. Third, by con- trasting and comparing the antecedents and correlates of each adoles- cent behavior, we may be more precise about the specific conditions, circumstances and processes that lead to adolescent sexuality and childbearing. Adolescent sexuality and other adolescent behaviors share certain common features. Adolescence is a developmental period in which an individual changes (over a varying length of time) from childhood into adulthood. In our society, many specific behaviors occur for the first time--drug, alcohol and cigarette use, the initiation of sexual activity, leaving home, and beginning to work. These transition be- haviors {Jessor, 1983) may have important similarities that examining them separately would miss. The social and cultural norms that either prohibit or postpone substance and alcohol use, delinquency, and sexual experience may have the effect of attaching certain symbolic meanings to their occurrence. Autonomy, independence, peer acceptance, and having rejected social conventions may all be signified by these tran- sition behaviors (Jessor and Jessor, 1977; Miller and Simon, 1980~. 36
37 Adolescent sexuality, substance use, or antisocial behavior are often considered as "problem" behaviors. From the life-course perspec- tive timing and sequencing of life transitions will influence the later life course (see Hogan, 1978~. An early birth is associated with less educational achievement and a high probability of living in an unstable family unit with poor economic resources. From the life-course per- spective, then, the ~problem" of early sexual intercourse is that it is off-time (normatively speaking) and may thus hinder the psychological and social development of the teenager and may result in childbearing that has detrimental consequence for both the mother and the child. With use of alcohol and drugs the "problem" is their illegality, their dependence producing nature, their physical and psychological harm to the user, and their negative effects on physical performance (including driving) and performance in social roles. With delinquency the problem is the possible harm to the individual and others in the environment and society. One of the basic distinctions between teen- age sexual behavior and other transition behaviors such as delinquency and substance use is that sexual activity is part of normal, adult life. The issue, then, for adolescent sexual intercourse is the appropriate timing for the initiation of this activity. This is also the case with some of the other transition behaviors although it is less so. While alcohol use and status delinquent offenses are legally prohibited for minors, they are not considered as part and parcel of normal adult life. An examination of the similarities in these adolescent n problems may indicate that they occur together, and the combined study of inter- ventions may increase our understanding and effectiveness. By com- paring these with other "nonproblem" transition behaviors we can begin to understand how the transition from youth to adulthood affects the behavior of adolescents, in general. The social context of the different behaviors may be the same. They often occur sequentially or at the same time. For example, alcohol and/or drugs are seen by many adolescents (and adults) as social cues for sexual activity. Delinquent activity often takes place concurrently with alcohol or drug use. There may also be a developmental progression, so that engaging in one kind of activity makes it more probable that another type of be- havior will be initiated. This progression approach would locate adolescent sexuality, then, on a continuum with other problem behaviors such as cigarette, alcohol, drug use and delinquency. The implication of this approach is that adolescents do not just go from conforming to nonconforming behavior, but one behavior serves as a stepping stone to another. The historical association of the attitudes toward adolescent sexuality and substance use also suggests that they may be linked. The recent societal change in values regarding adolescent sexual
38 activity and drug and alcohol use occurred at the same time. The period during the late 1960s and early 1970s was a time of increases in substance use and sexual activity by youth. Symbolically, this was important for young people because these two activities were proscribed by the society at large. While premarital sexual activity has always occurred in American society, the period of the late 1960s saw a more explicit acknowledgement and at least limited acceptance of its exis- tence. Drug use and sexual activity were associated in the perception of society, in rock music, and in the view that the United States had become more ~liberal" in its personal standards. The examination, then, of adolescent sexuality with other adolescent behaviors seems appropriate. First, I will review some of the theoretical perspectives that have been used to consider these behaviors. Second, the research results that have examined the covariation of these behaviors will be reviewed, and third, the similarities and differences in the association of these behaviors with gender, race, and social class will be summarized. It is hoped that this review will help us integrate adolescent sexual relations within the context of other adolescent activities. THEORETICAL FRAME:WORKS Several theoretical frameworks have been developed that are con- cerned with explaining an array of adolescent behaviors. These frameworks are largely social psychological with an emphasis on individual and family attributes. Problem Behavior Theory Jessor and colleagues (Jessor, Jessor, and Finney, 1973; Jessor, and Jessor, 1975; Jessor, 1976; Jessor and Jessor, 1977, 1978; Jessor, Chase and Donovan, 1980; Donovan and Jessor, 1984) have suggested "problem behavior theory to explain the variation in these behaviors among adolescents. The fundamental rationale of the problem be- havior perspective is the interpretation of many of the important tran- sitions that occur during adolescence as behaviors that depart from the regulatory norms defining what is appropriate for that age or stage in life (Jessor and Jessor, 19751. Early sexual experience, problem drinking, delinquency, and illicit drug use represent in adolescence a claim on more adult status or a transition in develop- ment, and engaging in such behaviors at a time that is considered too early constitutes a departure from regulatory norms. Within each of three systems--the personality, the perceived en- vironment, and the behavior--the proneness for problem behavior may be defined. The important personality constructs are favorable attitudes, values, beliefs and expectations to problem behavior. High value on independence and low expectation for academic goals are both con
39 ceptualized as favorable to problem behavior. In the perceived en- vironment system, low support and control from significant others and approval for and models for engaging in problem behavior are the impor- tant constructs. Within the behavior system, the degree of involvement in other problem behaviors on one hand, and in conventional behaviors, such as church attendance and school performance on the other, are expected to predict problem behavior. Variation in the time of initial intercourse was related to these personality and perceived environment variables (Jessor, Costa, Jessor, and Donovan, 1983~. In a nationwide sample of American adolescents, these same variables were related to both marijuana use and problem drinking (Jessor, Chase, and Donovan, 1980~. The problem behavior perspective has been criticized because the definition of what constitutes a problem is so connected to normative definitions, and the linking of sexual behavior with deviant and socially harmful behavior. Unlike most other "problem behaviors" such as alcohol or drug use, cigarette smoking, and delinquency, sexual behavior is expected to play a desirable role in adult life (Simon, 1985, personal communication). Social Control Theory A theoretical framework that has some similarity to the Jessors' problem-behavior theory in its focus on conforming and conventional behavior and the psychosocial context is Hirschi's social control theory. Social control theory has focused on explaining why people conform rather than why they deviate. Hirschi (1969) identified four elements of a social bond that constrain deviant behavior--attachment to others, commitment (dedication to pursuit of conventional means and goals), involvement in structural activities, and conventional beliefs. This perspective would argue that those adolescents who are strongly attached to societal institutions such as family, school, or peer group would be inhibited from engaging in deviant behavior. The research using a social control theory perspective has been limited primarily to examining the relationship between measures of delinquency and social bond measures; however, several studies have included drug use or sexual activity as one of the outcome measures of deviant behavior. In Krohn and Massey's (1980) research, minor delinquency (which included an item on sexual activity as well as other behaviors) was related to low attachment to mother, low grade point average {part of the commitment scale), unconventional beliefs for males, and to low scores on scale of commitment and conventional beliefs for females. Miller and Simon (1974) found that males high on parental involve- ment and low on peer involvement are least likely to report coital experience. While parental involvement was related for the females, peer involvement was not. These findings are similar to those found
40 for cigarette, alcohol and marijuana use in an urban black population (Ensminger et al. 1982--school attachment inhibited substance use for both males and females; strong family bonds inhibited use by females but not males; and strong peer bonds was related to more use by males. Both the research in the problem-behavior framework and the social control framework show involvement in conventional activities, parental involvement, and conventional values to be negatively related to adolescent problem or deviant behaviors. Critics of the social control perspective have pointed out that it implicitly assumes that the strength of one's social bonds are not influenced by gender, social class, race or other status placements (Kornhauser, 1978~. Others argue that attachment to those who are non- conforming contributes to deviant activity. Socialization Or Social Learning Theory A third framework focuses on the individual's interaction with significant others within the social context. For adolescents these social contexts include the family, peer group, and school. The influence of significant others on adolescent behavior may operate through modeling so that adolescents whose parents or friends use drugs or alcohol may be more likely to use them then adolescents whose parents and friends do not use these substances. Peer group or family norms and values may also influence children's and adolescents' be- havior. In terms of sexual behavior, Gagnon and Simon {1973) suggest that sexual scripts help to organize and communicate the social expecta- tions about sex. Interpersonal scripts provide the structures by which the individual presents him/herself to others and responds to others in ways that facilitate engagement. A traditional sexual script of the sixties might be that "nice girls don't.. Both parents and peers seem to be influential in adolescent be- havior. Inazu and Fox (1980) report that daughters whose mothers have lived with a man to whom she was not legally married during the daughter's lifetime were more likely to be sexually active. Jensen and Brownfield (1983) found that attachment to drug-using parents does not inhibit drug use, a finding which supports socialization ex- planations. Social control theory suggests that attachment inhibits deviance even if the attachment is to someone who is deviant. Modeling or social learning theory seems to provide a better explanation. Kandel and her colleagues (1973; Davies and Kandel, 1981) have studied the relative influence of parents and peers on drug use and future life plans. Handel (1973) found peer drug use was a more important influence than parental drug use on the adolescent's drug use, but that the aspirations of the adolescent are more influenced by parents than by friends (Davies and Kandel, 1981~.
41 A criticism of the socialization perspective is that it implies an overdetermination of behavior. Individuals often do not conform to how they have been socialized or taught. The tensions between the social prescription and individual wants or between tradition and social change are not reflected by this perspective. Developmental Perspective Adolescent behaviors, especially sexual behavior, are often con- sidered within a developmental perspective. This perspective sees adolescent experience as resulting from the integration of early infant and childhood experiences with an increased expectation of more orderly, rationalized and socially responsible behavior. Miller and Simon (1980) argue that the developmental perspective is the most coherent for discussing adolescent psychosexual development. However, they criticize this perspective because of its lack of attention to the sociocultural context. Adolescent sexual patterns cannot be ex- plained without reference to the existing definitions and expectations of adolescence. Jessor and Jessor {1977) conceptualized various adolescent be- haviors such as sexual intercourse and alcohol and drug use as tran- sitional activities. The onset of these behaviors can be anticipated by the adolescent's psychosocial development and by the other be- haviors he/she engages in. Robins and Wish (1977) apply this perspective to the development of deviant behavior. They argue that the initiation of one behavior is in part a function of past deviant behaviors and also makes more probable the initiation of additional deviant behavior. According to this perspective, the cessation of deviance is also part of the sequence with children/ adolescents being pressured by society to develop valued skills and to give up behaviors that conflict with societal goals. Robins and Wish suggest that differences between subcultures regarding the ages that are considered appropriate for various activities may be a key to value differences between the subcultures as well. Obviously, there are similiarities and overlapping among these perspectives, but they each focus attention on a different aspect of adolescent behavior. Problem behavior theory of Jessor and colleagues emphasizes the interrelationships among the perceived environment, personality and behavior and adolescence as a period of transition. Social control theory focuses our attention on the traditional in- stitutions and relationships that affect adolescence. The sociali- zation perspective focuses on the influence of significant others on adolescent behavior. The developmental framework views both past experiences and stage of life as crucial. Unconventionality of the studied behavior seems to be a common dimension to all the perspec- tives reviewed here.
42 THE RELATIONSHIPS AMONG VARIOUS ADOLE SCENT BEHAVIORS We now review some of the empirical work that examines the rela- tionships among various adolescent behaviors. Interrelationships Among Teenage Sexual Behavior, Substance Use, and Delinquency Jessor and Jessor (1977) suggest that drinking, marijuana use, delinquent behavior, and sexual intercourse may constitute a ~syndrome" of problem behavior in adolescence; they have focused on a broad array of adolescent behaviors rather than on a single outcome. Problem be- havior is "behavior that is socially defined as a problem, a source of concern, or as undesirable by the norms of conventional society...and its occurrence usually elicits some kind of social control responsen- (p. 33~. Support for the syndrome notion emerged from the Jessors' parallel longitudinal studies of two different samples: one of junior high school students and one of college students. Fifty-three percent of the randomly selected junior high school students from a small city in the Rocky Mountain states agreed to participate and were assessed four times in high school and twice in their twenties. For the college study, randomly selected college freshmen from the same city were con- tacted and sixty percent agreed to participate. They were assessed six times, each year of college and again six years and eight years later when they had reached their thirties. The results showed that the behaviors were positively associated in both samples; an index of the behaviors correlated negatively with measures of conforming or conventional behaviors, i.e., attendance at religious services and school performance; and the various problem behaviors correlated similarly with personality and social environment variables that reflected unconventionality fJessor and Jessor, 1977~. Donovan and Jessor (1984) have reanalyzed the earlier data using factor analytic models to test the hypothesis that the various problem behaviors reflect a single common factor. The results show that for the high school populations, sexual activity was significantly correlated with marijuana use, drinking behavior and a measure of general deviant behavior for males and females in both the third and fourth year of high school and for college females except for marijuana use in the fourth year. Further, the maximum likelihood test for factor analytic procedures demonstrated that only a single common factor was needed to account for the correlations among the behaviors for both the third and fourth high school years for males and females. For the fourth year college males, sexual behavior was not signifi- cantly correlated with any of the other behaviors. However, deviant behavior (at p ~.01) and frequency of marijuana use (at p ~.10) were correlated with sexual activity for year three college males. (r = .32 and .20 respectively.) For the college females, sexual behavior
43 was correlated (at least at the P6.10 level) with all the other n problem" behaviors in both years. Church attendance, one of two measures of conforming behavior that was included, was negatively correlated with the frequency of sexual experiences in both years for the college men and women and in the fourth year for the high school females and in the third year for the high school males. School performance was less correlated with sexual activity (or with the other problem behaviors) than any of the other measures--grades were significantly negatively correlated with sexual activity only for the year four high school males and the year three college females. A third population analyzed by Donovan and Jessor (1984} is from the 1978 National Study of Adolescent Drinking and consisted of a multi-stage stratified random sample of students in grades ten through twelve in the continental United States (Rachel et al., 1980~. Since this survey did not address sexual activity, its findings cannot inform us as to whether sexual activity fits into a "problem behavior syn- drome," but does include measures of cigarette, alcohol, and marijuana use and delinquent activity. The results from that study show signi- ficant positive correlations among measures of cigarette use, drinking, marijuana use, use of other illicit drugs, and general deviant be- havior, and significant negative correlations between these measures and measures of conventional behavior for both males and females. The single-factor model was tested in eight subsamples and in all analyses could account for much of the correlations among the problem behaviors. Both the high school populations and the college populations from the earlier studies were followed up in young adulthood (Donovan and JessOr, 1984~. The frequency of sexual behavior and school perfor- mance were no longer approppriate measures (i.e., not studied) of problem behaviors for these age groups, but drinking, marijuana use, other illicit drug use and general deviant behavior were. The single- factor model accounted for the correlations among these behaviors in young adulthood. Miller and Simon (1974) also studied the relationship of sexual intercourse with other adolescent behaviors in a random stratified sample of 2,064 white adolescents aged 14-17 living in Illinois house- holds. The survey collected data on sexual behavior, measured as the family, peer group, school adjustment, political attitudes, career and marriage aspirations, adolescent lifestyles and self-reported delin- quent activities. One question addressed was the relationship of sexual intercourse to other behaviors. The authors made a distinction between drug use as a Counter cultural" activity with explicit chal- lenges to the existing order and more traditional norm violations as indicated by delinquency. Clearly, coitus was associated with drug use in these data, but only five percent of the adolescents in the sample reported more than incidental drug use. Sexual intercourse was related to delinquent activity, especially for the males. The per
44 centage of 16 and 17 year old males who had had sexual intercourse was very much higher for those who reported moderately high, high, or very high delinquent involvement--than for males with no delinquent involve- ment. This pattern was not so strong for the females. In results similar to those of Jessor and colleagues, Miller and Simon also found that adolescents who have had sexual intercourse are less likely to aspire to advanced education and less likely to report being very ~ · ~ rellg loos. Using measures adapted from a study of delinquency (Kubik, Stein and Sabin, 1968), Zucker and colleagues (Zucker and Barron, 1973; Zucker and Devoe, 1975) in the early 1970s examined how problem drink- ing correlated with other problem behaviors in a population of 179 juniors and seniors in high school in a Middle Atlantic community. They found a strong association between the measures of drinking and the other problem behaviors measured: sexual behavior, drug use, de- linquency, and serious physical aggression. Sexual behavior, defined as was less related to drinking for the females than for the males (r = .28 for the girls; r = .59 for the boys). How drinking, sexual activity and contraceptive use were related was examined in a population of 370 undergraduate females in a large midwestern university by Zucker, Battistich, and Langer (1981~. They found that the women who were heavy drinkers (ranging from drinking 2 or 3 times a month, with 5-6 drinks every time to drinking 3 or more times a day with 1-2 drinks nearly every time) had more frequent sexual involvement than their peers who drank less. Further, despite their more frequent sexual activity, they were significantly less likely to use birth control. Robins and Wish {1977) addressed the question of whether deviance can be viewed as a developmental process in which one type of deviant act leads to another. They examined the ages and sequence of 13 kinds of behaviors in a population of 223 St. Louis black males born in the 1930s. Even though this study is based on a rather special population, it is summarized in detail here because it was the only study that systematically examined the sequencing of these adolescent behaviors. Precocious sexual experience as indicated by interview reports of first intercourse before the age of IS was one of the behaviors. Others included elementary school failure, behavior problems in elementary school, dropping out of school before high school graduation, juvenile arrest record, having drunk alcohol before the age of 15, and be- haviors commencing before the age of 18 including marijuana, bar- bituate, amphetamine, or opiate use, leaving home, marriage, and developing alcohol problems. The first four came from official records and the others were based on retrospective reports. Robins and Wish found that the behaviors were intercorrelated--of the 78 contingency tables that examined their interrelationships, 42 (54 percent) were statistically significant and positive, 21 times the number of significant positive relationships expected by chance.
45 Sexual intercourse before age 15 was related to early marijuana use, early drinking, dropping out of school, early use of barbituates, and development of alcohol problems; it was not significantly associated in this population with opiate or amphetamine use, elementary school academic problems, excessive elementary school absences, leaving home early, early marriage or juvenile arrests. They also examined in an actuarial test whether certain acts are plausible causes of others. They tested the significance of a dif- ference in rates of an act between persons with and without a prior behavior while instituting controls for age at risk. Among the be- haviors found to predict other behaviors, marijuana use, school absence, and early drinking were the most potent. Dropping out of school and alcohol problems were best predicted by the others. Pre- cocious sexual behavior was logically possible (based on age of initiation) as a predictor of all 12 of the other behaviors. It actually did significantly predict dropping out, barbituate use, marijuana use, and early drinking. These relationships remained sig- nificant when spuriousness was tested by controlling on other signi- ficant predictors. There were four logically possible antecedents of precocious sexual activity--early drinking was the only one found to be significant. Finally, they assessed whether certain acts are Necessary by establishing whether the second act almost never occured unless preceded by the first or ~sufficient" by examining whether the first act was almost always followed by the second. Sexual experience qualified as a "necessary. cause for marijuana use--in 78 percent of the cases, marijuana users had previously had sexual experience before their first use of marijuana. It is important to note that many of the studies just summarized are based on data collected over a decade or more ago when marijuana use and sexual intercourse by unmarried teenagers were considered more deviant than they are now. Hence, we must be cautious in generalizing these results to the current cohort of young people. However, these studies do present data collected from a wide diversity of young people in terms of social class, cohort, region, gender, race and age. The finding then that teenage sexual intercourse, substance use, and delin- quency are interrelated has been found among a variety of study popu- lations in studies conducted over a span of several decades. Delinquency, Alcohol and Drug Use Studies Researchers with a focus on one of the transition behaviors, such as delinquency or substance use, have often included its association with one or more of the other behaviors. We review some of this work here. They differ from the studies just reviewed in that their focus is not on the interrelationships among these adolescent behaviors, but on the explanation of a specific behavior.
46 Among researchers who study juvenile delinquency, frequency of sexual intercourse has often been used as one of many items measuring delinquency. For example, Krohn and Massey (1980) included alcohol and drug use and frequency of sexual intercourse in a delinquency checklist administered to a sample of 3065 adolescents in grades 7 through 12 in six communities in three midwestern states. A factor analysis was computed on all these items and four factors were iden- tified. Sexual intercourse loaded on a factor labeled Minor delin- quent acts" along with running away from home, truancy, and school suspension and/or expulsion. The scale composed of these items strongly correlated (the correlation coefficient ranged from .42 to .59) with the other measures of delinquency--alcohol and marijuana use, use of harder drugs, and the serious delinquency scale; the correlations among these scales were also strong. In a study of delinquent activities, Rubik, Stein and Sarbin (1968) administered a checklist of 52 delinquent behaviors to 505 boys attend- ing three high schools in the California Bay Area and 301 boys from eight California delinquency institutions. They were between the ages of 14 and 18. Sexual intercourse was one of the delinquency items. Cluster analyses were performed and four dimensions emerged. Sexual intercourse had a high factor loading on a cluster labelled delinquent role that also included items indicating alcohol use, gang activity, and school problems. The other three factor dimensions were drug use, parental defiance and assaultiveness. Adolescent drug use was correlated with sixteen other adolescent behaviors for 231 Ontario ninth grade students {Hundleby et al. 1982~. Most of these students attended a Roman Catholic school. Sexual be- havior was more strongly correlated with alcohol, tobacco, pain killers and marijuana use than were any of the other 15 adolescent behaviors. The other adolescent behaviors that were positively related to drug use were general delinquency, school misbehavior, and adoles- cent social behavior (visiting a friend's home, going to a party, other social events, etc.~. Studying/reading and academic achievement were negatively related to drug use. Behaviors that were not related in- cluded participation in team sports, religious behavior, musical/cul- tural activities, part-time employment, family excursions, and club participation. Adolescent Health Behavior Several studies have also examined the relationship of adolescent sexual activity to other adolescent health behaviors. While many are the same behaviors as those described above, they are considered in a ~health. framework along with additional health-related activities. Whether health behaviors are uni- or multidimensional has been a focus in several of these studies.
47 In a longitudinal study of 293 sixteen-year old students in the Haifa area of Israel, Epstein and Tamir (1984) studied the development of several health related behaviors--cigarette smoking, sexual be- havior, use of drugs and alcohol, and dropping out of school over a two year period. Initiation of sexual intercourse over the two years was related to both beginning to smoke and dropping out of school. Among males, 64.3 percent of those who began smoking had sexual inter- course for the first time during the study period, as compared to 20 percent of those who did not begin to smoke. For females, 38.5 per- cent of the smokers had intercourse compared to 13 percent who did not begin to smoke. Students who either smoked, had intercourse or drank alcoholic beverages at the first wave of the study were much less likely to remain in school. None of the students who smoked daily at age 16 or had used hashish continued on with school, and a substan- tially smaller proportion of those who sometimes smoked continued their schooling as compared to nonsmokers. The findings for males, were similar. Of the males, 30.9 percent who dropped out of school had had sexual intercourse by age 16, compared to 13.2 percent of those who continued their studies. Likewise, 26.6 percent of the males who stopped school drank alcoholic beverages compared to 13.7 percent of those who continued in school. A study of 1,546 males born in 1956 and residing in the Jyvaskyla military district in Finland investigated the interrelationships among health habits (Kannas 1981~. In 1975, mailed questionnaires were sent to this population with a return rate of 75.6 percent. Sixteen health behaviors were assessed including number of sexual partners, food habits, use of drugs, cigarette and alcohol use, personal hygiene, use of dental care services, sport activity, regularity of bedtime, use of traffic reflectors, and consumption of health information. Smoking was the behavior that correlated with the most other habits, and its second highest correlation was with number of sexual partners. Number of sexual partners was also highly correlated with excessive use of alcohol. In a factor analysis of the behaviors, these three behaviors received high loadings on the second factor which was labeled as n la dolce vita. n The author concluded that while certain health behaviors occurred together, in general, the results showed that a person be- having unhealthfully in one area does not necessarily act in a way that might harm or endanger health in all areas. Zabin (1984) recently examined the relationship of smoking to sexual behavior among 1200 female adolescents attending 32 contra- ceptive clinics in the U.S. Adolescent females are now more frequent smokers than adolescent males--twenty six percent of 17-19 year old females were smoking in 1979 compared to 19 percent of the males (Green, 1979~. Zabin found that those who initiated sexual inter- course in early teenage years smoked more than others in their age groups--23 percent of those girls 16 or less years old were smoking at least a half pack of cigarettes per day. Further, the clinics re- ported that over 70 percent of the teenage clients adopted the pill as their prescribed method of birth control. Given the known risk of
48 circulatory disease among pill users who smoke (Royal College of General Practitioners 1981), the future health implications of this pattern of smoking and pill use deserve some attention. With the increased attention (and perhaps, incidence) directed toward anorexia nervosa and bulimia, some have noted the connection between eating behaviors and other "motivated. behaviors {as opposed to reflexive or automatic behavior) such as drinking, sexuality and drug use (Andersen, 1984~. Given that the typical anorectic individual is an adolescent female, the relationship between eating behavior and other adolescent behaviors may be of interest. Andersen (1984) cites less sexual activity and "model. behavior as characteristic of anorexia nervosa in contrast to bulimia, in which more sexual activity and be- havioral problems are characteristic. However, there is very little research that has compared the behavior of teenagers with eating problems to those without, or research that examines eating behavior in the context of other behaviors. Summary of Research Findings In summary, while adolescent behaviors have largely been studied in isolation from each other there have been investigators who have been interested in the relationships among these behaviors. Jessor and colleagues have included adolescent sexual behavior in a n syndrome" of problem behaviors and have shown that it is positively associated with cigarette, alcohol, and drug use and deviance, and negatively correlated with conforming behaviors. Miller and Simon found that adolescent sexuality is related to drug use, delinquency, and low school aspiration and low religiosity. Robins and Wish have presented data that supports the developmental sequencing of problem behaviors with sexual exper fence as one of the stepping stones. The delinquency literature has used measures of adolescent sexual activity as a delin- quency indicator. It seems to be associated with other, more central measures of delinquency. The work of Bucker and colleagues found that adolescent sexual behavior is related to drinking; their work on heavy drinking among college females suggests that not only is heavy drinking related to more sexual activity, it is also related to less contracep- tive use. While Andersen suggests that "motivated" behaviors should be examined together, there is little research that has focused on adoles- cent sexual activity, eating behavior, and drug, cigarette and alcohol use from this biologically-based perspective of drives, appetites and satiation. Increasingly, adolescent sexual behavior has been studied as a health behavior in relation to other health behaviors. This literature shows that adolescent sexual behavior is related to cigarette, alcohol and substance use, but not necessarily to other risky health behaviors.
49 SIMILARITY AND DIFFERENCES IN BEHAVIORS FOR DIFFERENT ADOLESCENT BEHAVIORS FOR DIFFERENT SUBGROUPS OF ADOLESCENTS A number of studies have explored the social, psychological and family characteristics that seem to differentiate those adolescents that engage in specific behaviors from those who do not. Again, how- ever, these studies tend to focus on one or two adolescent outcomes rather than on a cluster of behaviors. While it would be very useful to compare and contrast the antecedents of these behaviors this task is beyond the scope of this review. However, examining whether the same subgroups of adolescents who are more likely to engage in one of the behaviors are the same as those more likely to engage in the other behaviors would be helpful in determining the relationships among the problem behaviors. In this section, we will examine whether subgroups of adolescents who are likely to engage in one of the transition be- haviors are also more likely to engage in the other behaviors. It may be that the patterns of relationships among these behaviors vary for different subgroups. Sex, race, and social class are major structural divisions within society and often reflect differences among people in social roles, behavior, attitudes, and opportunities. It is important to examine, then, how teenage behaviors vary according to these structural divisions and whether the same categories of adolescents who rank high on one behavior are the same as rank high on another. Sex Adolescence is such an important time for the initiation and crystallization of adult sex roles that we would certainly expect adolescent males and females to exhibit different behaviors from one another. Sex differences in adolescent behaviors can take several different forms (Ensminger, Brown and Kellam 1982~. First, males and females could differ in how much they engaged in certain behaviors. Second, the antecedents and correlates for the behaviors could differ for males and females, e.g., low parental supervision might be related to more sexual behavior for females but not for males. Third, an antecedent may be similar for males and females but it might occur more frequently for one sex than the other, e.g., school failure may be an antecedent to sexual activity for both sexes, but males may be more likely to fail school than females. In reviewing the antecedents to teenage behaviors, we will note whether they are related for both males and females or only one. Un- fortunately, studies often include only males or only females. For example, many studies of delinquency have only male subjects (see Harris 1977) while many studies of adolescent sexual behavior only include females (see Chilman 1978~.
so Male teenagers report more sexual experiences than females al- though increased rates of premarital sexual intercourse for females during the 1970s has decreased this gap (Zelnik and Kantner 19831. An important issue is whether the convergence in male and female rates of intercourse indicates that the meaning of premarital sexual behavior is similar for the two sexes. Simon et al. (1972) argue that increased levels of female sexual activity are not accompanied by radical changes in the meaning of this activity. While much of the sexual activity of males is directed toward the confirmation of masculinity, for females sexual activity is more likely to indicate emotional involvement with their partner. The results from a national probability survey done in 1979 of young women, aged 15-19, and young men, aged 17-21, support this contention {Zelnik and Shah 1983~. The young women (64.5 percent) were more likely than the young men (37.1 percent) to have been en- gaged or going steady with their first sexual partner while the young men were more likely to report being friends or to have only recently met their first partner (11.1 percent for the females compared to 43.0 percent for the males). Antonovsky et al. (1980) report similar findings from a nationwide sample of 5410 Israeli adolescents. While females raised on the kibbutz had similar rates of premarital sexual intercourse as kibbutz and non-kibbutz males, their pattern of sexual behavior was more similar to the nonkibbutz females {who had much lower rates of sexual experience) than to the males. That is, their first sexual experience was very likely with a steady boyfriend (88 percent) and they were more likely to have had only one partner (75 percent). These rates for the males were 46 percent and 42 percent (nonkibbutz) and 52 percent and 53 percent (kibbutz). These two studies suggest, then, that even though the previously found sex difference in age of initiation and frequency of sexual intercourse may be declining, sexual intercourse may still have a different meaning for teenage females then for teenage males. Differences in substance use rates between adolescent males and females have also declined over time (Johnston, Bachman and O'Malley 19821. However, there are different patterns in how different vari- ables related to use for males and females. While females are as likely to smoke cigarettes as males, female adolescent smokers have better grades, less truancy and more religious commitment than male smokers. However, these same variables are partly responsible for the lower use of alcohol and marijuana by females (Bachman et al. 1981~. In a population of urban black adolescents, Ensminger et al. (1982) also found different patterns of relationships for males and females between substance use (cigarettes, marijuana, and alcohol) and other social bonds. Strong bonds to their families reduced the level of use for females, but not males, while strong peer bonds was related to higher substance use for males but not females. Both males and females with strong attachment to school were less likely to report heavy substance use than adolescents with lower school attachment.
51 Male adolescents are much more likely to be delinquent than fe- males, whether official (police or court statistics) or self-report measures are used (Harris 1977), and much of the delinquency litera- ture includes only males (see Harris 1977; and Loeber and Dishian 1983~. Harris (1977) argues that we cannot begin to understand the origins of delinquency and crime until we understand why males and females differ so much in delinquent behavior. However, few studies have focused their attention on this sex difference. These studies suggest that, in general, sex differences in fre- quency of these transition behaviors has been declining in recent years, but that the patterns of these behaviors and their meanings may still be quite different for males and females. Social Class The study of the relationship of social class to adolescent be- havior differs somewhat from the study of social class to adult behaviors. First, for children and adolescents social class is an ascribed rather than an achieved characteristic as it is for adults-- measures of social class are based on the child/adolescents' parents rather than on their own characteristics. Second, the basic organi- zation for children/adolescents is based on age; while parental social class certainly influences which school a child attends, age is much more important in determining placement for children than for adults. For adults, social class seems to influence how they feel about them selves to a greater extent than it does for children or adolescents. Rosenberg and Pearlin (1978) compared the relationship of social class to self-esteem for children and adults. They found no relationship for children, a modest association for adolescents and a moderate associa- tion for adults. They suggest that adults perceive their social class standing as a reflection of their efforts and achievements while children do not. Social class may influence adolescent behavior, then, by way of different expectations or practices of parents, differences between schools in lower class neighborhoods vs. schools in middle or upper class neighborhoods, or the adolescents' realistic expectations about their possibilities of college education. Third, since most studies do not survey parents as well as adoles- cents, the measures of social class are often based on the child's report of the parents' education, income, or occupation. The validity of such measures has not been established. As reported by Hofferth, relationships between measures of parental social class and teenage sexual activity often disappears once certain factors are controlled, especially educational aspirations. However, parental social class may influence adolescent sexual behavior through its effect on aspirations--adolescents from lower class backgrounds have lower educational aspirations and adolescents with lower education aspirations are more likely to be sexually active.
52 Miller and Simon (1974) conceptualized measures of parents' social class as social origin indicators and educational aspirations as an- ticipated social class. They found that social origins were not related to teenage sexual behavior, but that anticipated social class was for both males and females. Hogan and Kitagawa (1983) found in a study of black teenage girls living in a poverty area in Chicago that these girls who were from poor areas had a higher rate of initial sexual intercourse than did those living in better off neighborhoods. Data from the National Longitudinal Survey of Youth presented in Volume 1 of the panel's report also indicates that teenagers from lower class families (as indicated by mothers' education) are more likely to be sexually active than are teenagers from middle or upper class families. These findings are stronger for females than males. The relationship of parental social class to adolescent substance use has not been strong. In a survey of a nationally representative sample of high school seniors, a mean of father's and mother's educa- tional attainment shows little association with drug use (Bachman, Johnston, and O'Malley 1981~. Kandel (1980) in a review of studies of drinking and drug use, reports that alcohol and marijuana use do not vary according -to socioeconomic status. Whether lower social class is associated with higher rates of delinquency has been an issue of some controversy. Some suggest that while poor youths have higher rates of official delinquency than middle class youths, they have similar self-reports of delinquent activities, indicating discrimination by juvenile justice authorities (Gould, 1969, 1981~. Others argue that self-report measures that tap more serious forms of delinquency do show more involvement by lower class youth {Elliott and Ageton 1980; Elliott and Huizinga 1983~. In a review of studies that examine the relationship between crime or delinquency and social class, Tuttle, Villemez and Smith (1978) con- clude that the relationship is weak. However, in a review of pre- diction studies of delinquency Loeber and Dishian (1984) report that the parents' socioeconomic status significantly improved the pre- diction of both delinquency, and recidivism over chance, even though it was one of the lowest ranking predictors. Race and Ethnic Differences In surveys of drug use and sexual exper fence race has of ten been measured and soc ial class has not, mak ing it hard to control on sac ial class when studying effects of race. This is a problem because of the strong relationship between poverty and being black. In the most recent national surveys of adolescent drug use, blacks have slightly lower reports of drug use than whites {Bachman, Johnston and O'Malley 1981~. However, blacks report higher rates of adolescent sexual experience (Zelnik and Kantner 1980~.
53 Conclusions regarding the relationships of social class and race to the patterns of adolescent behaviors must be regarded as tentative. The purpose of this review is to examine whether these major social structural variables relate to the transition behaviors in similar ways. The studies reviewed here that have included measures of mul- tiple transition behaviors have not been done on populations with enough variation in ethnic group and race or social class to carefully examine these interrelationships. So while we know that adolescent blacks tend to report less drug use and more sexual activity than whites, these findings come from different studies; and we cannot examine how patterns of these activities vary by race of social class. For example, a normative sequence of transition behaviors may exist, but it also may vary by these social structural divisions. Current data either does not exist or has not been analyzed to answer these questions. DISCUSSION An important issue underlying this review has been whether a unique set of factors explains each adolescent behavior that have been con- sidered here or whether a common set of factors precedes them all. The research reviewed here strongly suggests that the initiation of sexual activity is correlated with alcohol and drug use and delinquency and that at least some of the social and psychological factors are the same. However, not enough is known yet about the sequence of these behaviors, how subgroups may differ in the sequence or in the social and psychological antecedents, or how the relationships among the behaviors and their antecedents may vary by subgroup. Very few studies have focused on examining the interrelationships among these behaviors, and even fewer have examined the similarity or differences in antecedents or conditions that are associated with the behaviors. The work of Robins and Wish strongly suggests that there may be a patterning to the initiation of these behaviors, but since their sample was composed of black men born in the 1930s, we cannot assume that the sequence is the same or even similar in other popula- tions. Clearly, we need much more study of how these behaviors are interrelated for more recent cohorts of adolescents. We also know very little about whether or how the interrelationships vary for different groups of adolescents. Gender, social and family origins, race and ethnic group membership, and schooling all influence the experience that young people have had. Yet, we know very little about how these variations may affect the patterning of these behaviors. The function of these behaviors for the individual have not been adequately conceptualized either. While problem behavior theory sug- gests that adolescent sexual behavior, substance use and delinquency may be attempts by adolescents to assert their independence, few investigators have attempted to differentiate those adolescents whose involvements in these behaviors endanger their successful completion
54 of schooling or establishment of other adult social responsibilities from those whose involvements do not entail such risks. Certain adolescents may engage in "too early. sexual behavior and excessive alcohol or drug use as a way of avoiding school or the development of other social skills, while other teenagers, in contrast, may be en- gaging in such behaviors as an attempt to establish adult status. We need better conceptualization and more examination of the function of these behaviors. A second area in which both research and theory-building is lacking concerns the relationship of these adolescent behaviors to societal norms and expectations. One common factor underlying these behaviors is that they are all activities that are considered inappropriate or deviant for young people to engage in. Yet, we know very little about what the expectations are about when is the appropriate timing or how these expectations may vary across society. For example, much research has indicated that parental expectations for their adolescent chil- dren's future education are very important in influencing their children and that these expectations vary by the education and income level of the parents. However, we do not know what age parents expect their adolescent children to begin sexual activity, or decide whether to drink alcohol or smoke cigarettes. We also do not know if these expectations are related to each other. So even though these be- haviors all represent deviations from adult societal sanctioned behaviors, we know very little about what timing and patterning is appropriate or approved. More attention needs to be paid to what societal expectations are for the initiation of these behaviors and how these expectations vary across society. A separate issue concerns the placement of adolescent sexual behavior in a similar framework with misbehaviors such as drug use or delinquency. Sexual activity may have positive contributions for certain adolescents such as enhanced self-esteem, increased indepen- dence and improved social competence. If examined only within a problem behavior framework, any possible positive contributions would not be observed. In summary, the research findings reviewed here suggest that sexual activity is generally not an isolated behavior. Adolescents who are sexually active also are more likely to be involved in other behaviors that are not considered appropriate for adolescents such as smoking, alcohol and drug use and delinquency. Jessor suggests that these be- haviors may reflect a lifestyle. Clearly more work is needed in this area to better understand these interrelationships. Additional study of these interrelationships should further our understanding of early sexual behavior of adolescents and whether different adolescent be- haviors, including early sexual activity, may be serving similar psychological and social functions. The answer to these questions may have very important implications for prevention and intervention pro- grams. Prevention programs aimed at widening teenagers aspirations for the future, increasing social and interpersonal skills, and
55 involving them in more school and community activities may reduce their participation in a variety of deviant activities. Conversely, preven- tion or intervention attempts aimed at one specific outcome may have the unintended result of increased involvement in other undesired behaviors. These types of prevention/intervention efforts would have theoretical as well as practical significance and could be used to test whether different adolescent behaviors stem from the same social and psychological causes. We still have much to learn about how these adolescent behaviors interrelate and what functions they serve in adolescent development.
