Sex, contraception and childbearing among high-risk youth: do different factors influence males and females?

During adolescence, many young people begin to experiment with new roles. One important area of exploration is sexual activity, which involves a certain amount of risk-taking.

Sex, Contraception and Childbearing Among High-Risk Youth: Do Different Factors Influence Males and Females? By Lori Kowaleski-Jones and Frank L. Mott Context: The likelihood that adolescents will engage in sexual activity, use contraceptives or become parents is influenced by a range of attitudes and behaviors. These factors may differ for males and females. Methods: Data on female respondents to the 1979-1992 waves of the National Longitudinal Survey of Youth and the linked 1994 young adult data file on their children provided background information on 959 adolescents who had been born to young mothers. Partial correlation analy­sis was used to examine the factors related to sexual behavior, contraceptive use and childbirth, controlling for maternal and familial characteristics, in this relatively disadvantaged sample. Results: Youth who are inclined toward risk-taking and those who have run away from home are more likely than others to be sexually active. For young women, having intercourse at an early age, not using contraceptives and having a child are linked with depression, low self­esteem and little sense of control over their lives. The results for young men are less consistent and often in the opposite direction. Young people who have become parents evidence greater maturity than their childless peers; women are less likely to consume alcohol or to spend time with friends who drink, and men are more likely to participate in socially productive work. Conclusions: Although sexual behavior is tied to risk-taking in both adolescent males and fe­males, some noticeable psychological differences are evidenced early. Behaviorally, there is room for optimism, in that young parents appear to adopt more mature traits. Family Planning Perspectives, 1998, 30(4):163-169 During adolescence, many young people begin to experiment with new roles. One important area of exploration is sexual activity, which in­volves a certain amount of risk-taking. The association of sexual activity and risk-tak­ing is firmly established both in conven­tional wisdom and in the research litera­ture. Many scholars view adolescent sexual activity as a problem behavior be­cause it often departs from what is socially defined as appropriate for an age or for a stage of the life cycle.1 Adolescent sexual activity is also viewed as a problem in that teenagers often use contraceptives inef­fectively and have early or otherwise mis­timed pregnancies. Sometimes, adolescent sexual behavior is accompanied by other problem behav­iors, such as alcohol and drug use, a phe­nomenon often referred to as the "prob­lem behavior syndrome."2 However, the birth of a child can alter this problem be­havior path. Faced with the responsibili­ty of raising a child, many teenagers may change their views toward risk-taking. Using data on mothers participating in the National Longitudinal Survey of Youth (NLSY) and their young adult chil­dren, we previously found that there were significant links between maternal and adolescent sexual debut, that prior risk- taking behaviors also affected early initi­ation of sex and that these associations varied somewhat between male and fe­male adolescents.3 In this article, a follow- up to that analysis, we use more recent data for the young adult children of fe­male NLSY respondents to more fully in­vestigate early sexual activity. Addition­ally, we examine the correlates of ineffective contraceptive use and early childbearing. Our major aim is to explore whether common attitudinal and behav­ioral covariates are associated with these behaviors, and in particular whether there are gender differences in this area. The adolescents in our sample represent a relatively disadvantaged group in that they were bom to women who were most­ly in late adolescence or young adulthood at the time of the birth. Therefore, results from this sample may not represent patterns among adolescents bom to a more diverse cross section of mothers. However, this sample is uniquely suited to our research goals-to identify patterns of influence among those at the highest risk of early sex­ual behavior and early childbearing. Background Currently, 76% of young women and 80% of young men in the United States have sexual intercourse by age 20.4 Teenage sex­ual activity varies by gender in several im­portant ways. Typically, male adolescents become sexually active at earlier ages than females. Even though this gender gap is narrowing, males consistently have high­er levels of sexual activity at earlier ages.5 Rates of sexual activity also vary by race and ethnicity. A larger proportion of black adolescents than of white or Hispanic teenagers have sex at early ages. More than half of non-Hispanic black males have sex by age 15, but Hispanic and non- Hispanic white males do not attain this level of sexual activity until age 17. A sim­ilar pattern is seen for females.6 A number of important behavioral and attitudinal correlates influence adolescent sexual behavior. Low self-esteem is an im­portant predictor of sexual activity, and it is sensitive to the gender of the adolescent. Early sexual debut is a more negative and problematic phenomenon for women than for men.7 Whereas males typically re­port that their first sexual experience was exciting and satisfying, women are likely to report feelings of fear, guilt, anxiety and embarrassment associated with early sex­ual activity.8 More generally, depressive affect and disorder show marked in­creases in adolescence, especially for fe­males. 9 Teenage women's self-esteem de­clines as they make the transition into junior high school; adolescent males' per­ception of their own well-being and com­petence declines during college and rises during the following decade. Whether an adolescent attributes suc­cess in life to his or her own ability and ef­fort (internal locus of control) or to cir­cumstances and fate (external locus) is related to sexual activity, but the direction of the effect depends on the youth's age. Lori Kowaleski-Jones is a postdoctoral fellow, Institute for Policy Research, Northwestern University, Evanston, IL; and Frank L. Mott is senior research scientist, Center for Human Resource Research, and professor, Depart­ment of Sociology, The Ohio State University, Columbus. A previous version of this article was delivered at the an­nual meeting of the Population Association of America, New Orleans, May 9-11,1996. The authors thank Eliza­beth C. Cooksey for very helpful comments. The research on which this article is based was supported with funds from grant APR01 RA 000961 from the Office of Popu­lation Affairs, Department of Health and Human Ser­vices. The views expressed are those of the authors and are not meant to represent those of the funding agency. Volume 30, Number 4, July/August 1998 163 Sex, Contraception and Childbearing Among High-Risk Youth Younger adolescents with an external locus and older adolescents with an internal locus are at increased risk of engaging in coitus.10 Even though the link between an individual's internal state and sexual in­tercourse is complicated, prior research in­dicates the importance of assessing it.11 Traditional attitudes may lower ado­lescents' propensity to become sexually active. And among sexually experienced adolescents, such attitudes may reduce the likelihood of effective contraceptive use and increase the odds of carrying a preg­nancy to term.12 Adolescents' attitudes toward risk-tak­ing and their nonsexual risk-taking be­havior also are linked with decisions about sexual activity. Often, high-risk behaviors overlap. Teenagers who abuse drugs, for example, are much more likely than oth­ers to drink and smoke heavily/drop out of school, have sex at early ages and expe­rience early childbearing.13 Teenagers who engage in risky behavior may have more difficulty in, and less support for, handling the stresses and tasks of adolescence, and may be more depressed or may have a greater tolerance or need for risk-taking.14 The overlap of early sexual activity and other problem behaviors has attracted con­siderable research attention over the past 25 years.15 Associations b&tween sexual ac­tivity and drinking, drug use, school prob­lems and a wide variety of deviant be­haviors have been a consistent research finding.16 One analysis of NLSY data found that sexually experienced 15-17- year-olds were more likely than their sex­ually inexperienced peers to have been suspended from school or involved in such behaviors as theft, violence and drug use.17 Another documented that drug involve­ment is a risk factor for early onset of sex­ual activity.18 Early tobacco and alcohol use have been associated with permissive sex­ual attitudes and peer relationships that, in turn, predicted sexual behavior.19 Some evidence indicates that the co­occurrence of problem behaviors ^s more common among males than among fe­males. The available research suggests that the sequence of these behaviors typically runs from using substances and having deviant peer group associations to en­gaging in adolescent sexual intercourse, rather than in other directions.20 However, adolescents often learn from experience, and their experiences may af­fect their subsequent behavior. This process may vary for male and female adolescents, since society often places different mean­ings on sexual activity, contraceptive use and early childbearing for males and fe­164 males. Historically, stronger social and emotional sanctions have been associated with sexual activity and childbearing for females than for males.21 Attitudes sur­rounding the permissibility of sexual be­havior for males and females have con­verged in recent decades, but sharp differences persist.22 Males tend to receive more social support from their peers for premarital sexual activity than do females. Given that sexual activity and child­bearing have different implications for males and females, what young people learn from experience will vary as well. Because females invariably assume the bulk of childrearing responsibilities, par­enthood will likely result in more marked behavioral and attitudinal changes among them than among young men. Methodology Data and Analysis Our data are from the 1979-1992 annual waves of the NLSY and the linked 1994 young adult data file. The NLSY is an om­nibus national longitudinal sample, which initially included 12,686 men and women who were aged 14-21 on January 1,197923 In the original sample, military personnel and Hispanic, black and economically dis­advantaged white youth were overrepre­sented to permit statistically meaningful racial, ethnic and socioeconomic compar­isons. For funding reasons, the military personnel and economically disadvan­taged white respondents have been dropped from the sample. As of 1994,90% of the original sample who have remained eligible were still par­ticipating. Data on the maternal and fa­milial characteristics we include as con­trols were for the most part collected from the annual interviews and thus are not as subject to recall errors and biases as in­formation collected retrospectively. Avail­able weights permit one to present valid national statistics for the sample of women and children examined in this study. Our sample consists of 959 youth, most­ly between the ages of 14 and 18 as of the 1994 survey, who are the oldest children of the original female respondents. Our analy­sis examining the correlates of early sexu­al activity is based on the entire sample; the analyses of the predictors of contraceptive use and childbearing are limited to the 483 youth known to be sexually active. The three outcomes of interest are whether the adolescents have ever had sex and, among those who are sexually experi­enced, whether they used any form of birth control the last time they had intercourse and whether they have had a live birth (or, for males, fathered a child). We examine the independent associations between these outcomes and a variety of theoretically rel­evant, temporally proximate explanatory variables, controlling for a range of familial and maternal characteristics that are asso­ciated with early sexual activity 24 The ex­planatory and control variables are defined in the appendix (page 168). We must be cautious in making causal inferences, because data for many of the explanatory variables are available only for 1994, the same year for which we are evaluating the outcomes, and because some of the variables may be either a cause or a function of the outcomes. (For example, depression may result from early sex and childbearing, but it may also pre­dispose an adolescent to seek intimacy by engaging in sex.) Therefore, we conduct­ed correlation analyses to examine the re­lationships between the explanatory vari­ables and the outcomes. We present zero-order correlations and, to show the effects of controlling for familial and ma­ternal characteristics, partial correlations. Hypotheses For the most part, we anticipate that our key attitude and behavior variables will be predictive of having early sex and not using contraceptives. However, we antic­ipate different patterns of association be­tween the inputs and becoming a parent. Consistent with our views regarding the learning of new behaviors, we expect that having an early birth may alter prior atti­tudes and some types of socially nonpro­ductive behaviors. We expect different as­sociations for females than for males. This is because of the differential impact of child­bearing on the mother and the father, but also because empirical evidence confirms that the teenage female's immediate fami­ly- which here typically includes her par­ents- is more directly affected than the teenage male's. Parents' awareness of a young person's involvement in a pregnancy is likely to be much greater than their awareness of the youth's sexual activity or contraceptive behavior. Males' and females' attitudes and behaviors will probably at least partly reflect their concern about how their parents will respond to the birth. In preliminary analyses, we examined the independent associations of interest by entering the control variables in a sequen­tial manner, beginning with demographic characteristics. We first added just the child's age. This typically altered the asso­ciations for variables that are known to be very age-sensitive (such as the items on substance use). We then added all of the Family Planning Perspectives Table 1. Mean characteristics of adolescents aged 14 and older in 1994 who had been born to teenage mothers, by sexual experience and gen­der, National Longitudinal Survey of Youth Characteristic All Sexually inexperienced Sexually experienced Total (N=959) Male (N=496) Female (N=463) Total (N=413) Male (N=218) Female (N=195) Total (N=483) Male (N=236) Female (N=247) Sexual outcomes % ever had sex 54 52 56 na na na na na na % ever used birth control 69 66 72 na na na 69 66 72 % ever had a child 7 3 11 na na na 13 6 20 Demographic characteristics % male 52 na na 53 na na 49 na na Mean age 16.26 16.23 16.30 15.39 15.34 15.44 16.93 16.98 16.88 % black 32 32 32 24 20 28 39 42 35 % Hispanic 10 10 10 12 10 13 9 10 8 Attitudes Planning takes the fun out of things 2.20 2.30 2.11 2.20 2.25 2.16 2.20 2.33 2.08 Have to use self-control to stay out of trouble 2.54 2.65 2.44 2.51 2.62 2.39 2.56 2.65 2.47 Life without danger is dull 2.48 2.57 2.39 2.52 2.62 2.41 2.47 237 ' '2.37 Enjoy taking risks 2.51 2.61 2.41 2.50 2.61 2.39 2.54 2.66 2.41 Feel depressed at times 0.55 0.33 0.79 0.47 0.32 0.63 0.64 0.32 0.95 Depression score 0.73 0.66 0.79 0.65 0.64 0.67 0.80 0.69 0.90 Feel that 1 am a failure 1.58 1.60 1.54 1.57 1.64 1.49 1.59 1.56 1.61 Self-esteem 3.19 3.22 3.16 3.17 3.17 3.17 3.21 3.29 3.13 Locus of control 3.09 3.10 3.08 3.10 3.06 3.14 3.10 3.15 3.04 Traditional roles scale 2.00 2.08 1.91 2.03 2.12 1.93 1.97 2.06 1.89 Behaviors Frequency of drinking 3.01 3.28 2.76 1.95 1.97 1.94 3.69 4.12 3.26 Peers drink 1.99 2.02 1.96 1.56 1.67 1.45 2.36 2.40 2.33 Ever used marijuana 1.86 1.97 1.75 1.19 1.26 1.12 2.41 2.56 2.26 Problems because of substance use 6.39 6.49 6.28 6.17 6.27 6.06 6.59 6.72 6.45 No. of times ran away 0.23 0.20 0.26 0.11 0.10 0.12 0.34 0.30 0.38 % who got help for emotional problems 9 10 9 7 8 5 11 11 12 % who volunteered 34 30 37 42 35 48 27 25 29 Current maternal and family characteristics Mother's mean years of schoolingt 11.13 11.20 11.06 11.51 11.53 11.50 10.88 10.98 10.77 No. of children in homef 2.55 2.52 2.57 2.54 2.56 2.52 2.55 2.50 2.61 % urbanf 74 75 74 71 69 73 77 78 76 Mother's mean no. of weeks worked per yearj 24.35 24.67 23.99 24.11 24.54 23.62 24.94 25.50 24.32 % of surveys in which mother's partner was present^ 72 69 74 79 79 79 66 61 71 % of surveys in which grandparents were present^ 10 9 10 7 6 8 11 11 11 Mean family income-to-needs ration 1.91 1.87 1.95 2.06 2.07 2.05 1.79 1.68 1.89 Maternal characteristics Mean age at first sex 15.96 16.03 15.89 16.55 16.62 16.47 15.45 15.42 15.48 Mean age at first menarche 12.53 12.59 12.46 12.68 12.65 12.71 12.41 12.50 12.31 Early substance use 0.60 0.59 0.62 0.47 0.47 0.47 0.71 0.68 0.73 Early deviant behavior score 0.90 0.96 0.84 0.87 0.87 0.85 0.93 1.02 0.85 f1988. 979-1988. Notes: For definition of all variables and scores, see appendix (page 168). na=not applicable. other characteristics except the maternal characteristics, and finally included those items. Typically, the addition of each set of variables modestly reduced the coefficients. Results Sample Characteristics The adolescents' characteristics show clearly that many of these young people and their families are disadvantaged (Table 1). For example, two-fifths of these adolescents are members of racial or eth­nic minority groups. On average, their mothers had had only about 11 years of schooling and had become sexually active by their 16th birthday. And their income- to-needs ratio (1.9) suggests that the av­erage income for these families is not quite twice the poverty level. Volume 30, Number 4, July/August 1998 Because the young men and women are drawn from similar backgrounds, their family and maternal attributes are very similar. However, some gender distinc­tions are evident in attitudes and behav­iors that we expect to be associated with the outcomes. For example, the data sug­gest that males are more likely than fe­males to feel that "planning takes the fun out of things" or that they need to "use self­control to stay out of trouble," and they are more likely to be risk-takers and are less likely to be depressed. They also are more likely to use various substances. Thus, everything else being equal, one might an­ticipate gender differences in the three out­comes strictly on the basis of these factors. The sexually active adolescents were markedly more disadvantaged than their sexually inexperienced counterparts. On average, their mothers had less education, and their family economic well-being (as measured by the long-term ratio of income to needs) was lower. Perhaps even more im­portant, the mothers of the sexually active youth had become sexually active at a sub­stantially younger age than the mothers of the sexually inexperienced teenagers (15.5 vs. 16.6 years) and were substantially more likely to have been involved in troublesome adolescent behaviors. Additionally, the sex­ually active adolescents have used sub­stances to a greater extent and have evi­denced more problem behaviors. In virtually all instances, the differences between the sexually active and sexually inexperienced youth appear for both males and females, although the level of 165 Sex, Contraception and Childbearing Among High-Risk Youth Table 2. Correlations between the likelihood of having had sex, having used birth control or having had a child and selected attitudes and behaviors, by gender, among adolescents aged 14-18 in 1994 who had been born to teenage mothers Attitude or behavior Uncontrolled Controlled Total Male Female Total Male Female HAD SEX Attitudes Planning takes the fun out of things -.00 .05 -.05 -.00 .04 -.03 Have to use self-control to stay out of trouble .03 .02 .04 .01 -.01 .05 Life without danger is dull -.03 -.02 -.02 .03 .04 .03 Enjoy taking risks .02 .03 .02 .08*** .07* .11" Feel depressed at times .10*** -.01 .16*** .03 -.08" .07* Depression score .11*** .04 .16*** .06" -.01 .12"* Feel that 1 am a failure .01 -.07 .10" .00 -.07* .07* Self-esteem .04 .13*** -.04 .00 .05 -.02 Locus of control .00 .10** -.09 .01 .07* -.03 Traditional roles scale -.06 -.07 -.05 -.03 -.02 -.03 Behaviors Frequency of drinking .38*** .44*** .33*** .28*" .32*** .23*** Peers drink .29*** .26*** .32*** .20*** .17*" .24*** Ever used marijuana .36*** .36*** .36*** .30*** .29*** .31*** Problems because of substance use .14*** .12** .21*" .11*** .09" .17*** No. of times ran away .19*** .16*** .21*** 17**. .17*" .17*" % who got help for emotional problems .07** .04 .11** eg... .06 .11" % who volunteered -.15*** -.10** -.20*** -.06** -.03 -.11" USED BIRTH CONTROL Attitudes Planning takes the fun out of things -.04 -.06 .00 -.04 -.06 -.01 Have to use self-control to stay out of trouble -.07 -.01 -.11 -.05 -.00 -.10" Life without danger is dull .01 .07 -.03 .01 .08 -.06 Enjoy taking risks .05 -.01 .11 .03 -.00 .05 Feel depressed at times -.12*** -.11 -.19*** -.09** -.10* -.14" Depression score -.05 -.06 -.06 -.03 -.07 -.04 Feel that 1 am a failure -.12*** -.07 -.18*** -.09** -.04 -.13** Self-esteem .06 .04 .11 .04 -.02 .08 Locus of control .07 .07 .09 .05 .03 .08 Traditional roles scale -.16*** -.29*** -.03 -.11** -.25*** .04 Behaviors Frequency of drinking -.04 .07 -.15** -.06 .02 -.14" Peers drink -.09 .01 -.19*** -.10" -.03 -.17*** Ever used marijuana -.06 -.05 -.05 -.08* -.08 -.07 Problems because of substance use -.04 -.07 .03 -.03 -.06 .03 No. of times ran away .03 -.02 .07 .04 .01 .08 % who got help for emotional problems .02 .05 -.00 .02 .05 .01 % who volunteered .08 .07 .08 .06* .08 .04 HAD A CHILD Attitudes Planning takes the fun out of things -.10" -.13 -.04 -.09** -.08 -.01 Have to use self-control to stay out of trouble .05 .00 .10 .00 -.02 .05 Life without danger is dull -.17*** -.14** -.16** -.11" -.09* -.05 Enjoy taking risks -.13*** -.12 -.10 -.07* -.07 -.00 Feel depressed at times .18*** .12 .13** .17" .10* .08 Depression score .08 -.04 .10 .09" -.01 .08 Feel that 1 am a failure .09** -.02 .15** .08* -.02 .13" Self-esteem -.08 .05 -.11 -.11" .01 -.10* Locus of control -.05 .08 -.09 -.07* .10* -.12" Traditional roles scale .03 .00 .10 .00 .00 .10* Behaviors Frequency of drinking -.07 .07 -.12 -.11* .07 -.16" Peers drink -.00 .08 -.05 -.05 .07 -.12" Ever used marijuana -.02 .09 -.06 -.03 .09 -.07 Problems because of substance use -.05 -.02 -.06 -.05 .03 -.05 No. of times ran away .06 -.06 .12 .13*" .03 .13" % who got help for emotional problems .02 -.03 .05 .07* .06 .07 % who volunteered -.10** .13 -.23*** -.06 18*.. -.19*** *p<.10. **p <.05. ***p<.01. Note: Controlled analyses adjust for adolescent demographic, current maternal and family, and maternal characteristics. For definition of all explanatory variables and scores, see appendix, page 168. negative adolescent behaviors is system­atically higher for males. The only clear instance of a significant gender differen­tial by sexual activity is for depression: Sexually active young women appear sub­stantially more depressed than their sex­ually inexperienced counterparts, a pat­tern not evidenced for the young men. Attitude Variables •Risk-taking. The variables regarding plan­ning, staying out of trouble, danger and enjoying risks are intended to tap dimen­sions of risk-taking behavior. While they could be combined into a scale, we feel that the individual items reflect distinct dimensions. For example, the first of these items is likely linked to notions of delayed gratification, and the second might reflect an adolescent's impulsiveness. The last two items are directly related to risk-tak­ing likelihoods. In the uncontrolled analysis, for the overall sample, none of these items are as­sociated with either being sexually active or using contraceptives; when the analy­sis is controlled for background charac­teristics, however, young people who enjoy taking risks also are likely to be sex­ually experienced (Table 2). We were sur­prised to find no overall association be­tween an unwillingness t& plan and the use of contraceptives. Our hypothesis regarding associations between these attitude items and child­bearing is supported: Sexually active youth who have had a child are marginally less likely than their childless peers to express a willingness to take risks (p<.10) and are less likely to agree that planning takes the fun out of things. The contrast between these results and the signs (if not always the significance) of the coefficients for sex­ual activity suggests that young people's experiences can alter their attitudes. Do the overall associations mask im­portant gender differences? With respect to sexual activity, the answer clearly is no. The coefficients are typically not signifi­cant for either young men or young women. The only exception is that in the controlled analysis, the "enjoy taking risks" item is significant for females and marginally significant for males. With regard to contraceptive use, the controlled results indicate that young women have a significant likelihood of disagreeing with the statement that they need to exercise self-control to stay out of trouble; for young men, there is no statis­tically significant association between the self-control item and contraceptive use. One plausible implication of these find­ings is that women might be less likely than men to be impulsive when deciding whether to use contraceptives because the consequences of not using a method can have more direct and immediate reper­cussions for them. Finally, with respect to childbirth, we find no important associations related to risk-tak­ing for either gender in the partial correla­tion analysis. The significant associations noted for the total sample apparently are due to its size and smaller standard error. 166 Family Planning Perspectives •Depression. We use two measures of de­pression: an individual item ("feel de­pressed at times") and a composite score. Without any controls, being sexually active is strongly correlated with being depressed for the sample as a whole, regardless of which measure is used. However, this is largely a female phenomenon, as may be seen by comparing the coefficients for males and females. When controls are added, de­pression continues to be associated with above-average levels of sexual activity for women. However, according to the indi­vidual item, sexually active young men are less likely than their sexually inexperienced peers to be depressed. Thus, we find some evidence of differential psychological man­ifestations for young men and women who become sexually active. With respect to birth control use, we find significant associations, albeit only for the individual depression item. Even with all controls, depressed youth are signifi­cantly less likely than other young people to be using contraceptives. While depres­sion could arise from not having used a method appropriately, it is much more likely that ineffective contraceptive use is a result of depression. The magnitude of the associations is similar for both gen­ders, but the coefficient for males achieves only marginal significance. Finally, for both genders, depression is moderately associated with having had a child. The overall partial coefficient is strong and significant. The separate gender coef­ficients are somewhat smaller, and only among males is one of the measures mar­ginally significant. Given that the linkage between depression and childbearing is sim­ilar for males and females, the association may reflect more than just a postpartum physiological-psychological connection. • Self-esteem. We explore three partially re­lated factors: an individual item that asks the youth whether they "feel they are a fail­ure," as well as well-validated scales of self-esteem and locus of control. It is diffi­cult to predict the nature of the associations between these variables and our outcomes, because of the differing male and female norms relating to the appropriateness of the outcomes. Depending on a youth's other personal values, these explanatory measures could be associated with either higher or lower levels of sexual activity. Overall, in both the uncontrolled and the controlled analyses, these measures are not associated with early sexual activity. How­ever, in the uncontrolled associations, sex­ually active young men have higher self-es­teem and a stronger locus of control than their sexually inexperienced peers, while Volume 30, Number 4, July/August 1998 sexually active young women are more like­ly than others to perceive themselves as fail­ures. For the partial correlations, the results are similar, but much less robust. More con­trol and less feeling of failure are margin­ally linked with sexual activity for males, whereas feeling a failure is marginally as­sociated with sexual experience for females. These results, although modest in strength, suggest that these young men and women may be viewing sex quite differently. Regarding contraceptive use, only one association is significant, but it is quite ro­bust and is consistent with our findings regarding sexual activity. For young women, nonuse of birth control is strong­ly and independently associated with per­sonal feelings of failure. With regard to a history of childbearing, our results are generally consistent with prevailing norms within traditional ado­lescent culture. For the total sample, we find marginally significant associations that are consistent with the female results. That is, young mothers are more likely to feel they are a failure, have lower self-es­teem and feel they have less control over their environment than their childless counterparts. For young men, the only re­sult even approaching significance sug­gests that becoming a father is associated with a stronger locus of control. • Women's roles. The traditional roles scale consists of a well-established series of items regarding youth's ideas of the ap­propriate roles for women. When the re­lationships are controlled, only limited as­sociations are evident. Adolescent males with more traditional values are much less likely to use contraceptives than are those with less-traditional views. Teenage women with traditional values are mar­ginally more likely than others to have had a child. A traditional view of gender roles was not independently associated with sexual activity for either gender. Behavior Variables •Substance use. For both genders, sexual activity is strongly associated with drink­ing alcohol, using marijuana, having peers who drink and having substance use prob­lems. These associations remain statisti­cally significant in the controlled analysis. Flowever, the results for using birth con­trol and for having had a child tell a dif­ferent story. For females, drinking alcohol and spending time with peers who drink are independently linked with not using birth control. For males, there are no as­sociations. Thus, teenage women may be more susceptible than young men to neg­ative peer influences. On the other hand, young mothers drink less than their child­less peers and are less likely to spend time with friends who drink. •Other negative behaviors. One obvious manifestation of a difficult adolescence is a tendency for youth to run away from home. The motivations for leaving home can be inextricably linked with other at­titudes and behaviors that are associated with our outcomes. They can also be a consequence of the outcomes, and one would expect this to be particularly true with respect to giving birth. The pattern that emerges is reasonably consistent with our expectations. The par­tial correlations between running away and being sexually active are quite strong for teenagers of both genders. This is consis­tent with the notion that youth who are en­countering a difficult adolescence are more likely than other young people to leave home. It may also suggest that youth who wish to form sexual relationships find it easier to do so outside their parents' home. We had no theoretical rationale for ex­pecting a relationship between running away from home and using contracep­tives, and no associations were found. However, the linkage between running away and being a parent shows an im­portant gender distinction that is consis­tent with many of our other findings. Ado­lescent mothers are significantly more likely than their childless peers to have run away from home. This likely reflects that the presence of an infant may result in an awkward, if not impossible, residential sit­uation for some young women. As is well documented, a substantial proportion of young fathers have never resided with the mother and child.25 Thus, the child often becomes the responsibility solely of the mother. In many instances, her parents may wish no involvement, or she may avoid family conflict by leaving home. This highlights once again how female and male behavior paths may fundamentally differ, reflecting physiological variations and, perhaps more importantly, differences in perceived societal norms regarding ado­lescent sexual behavior. Finally, sexually active females were more likely than their sexually inexperi­enced peers to have received help for emo­tional problems within the past year. • Socially productive behavior. In addition to asking these youth about "problem" be­haviors, we queried them about their in­volvement in positive activities, such as volunteering in their community. Teenage women who are more involved in positive activities are less likely to be sexually ac­tive. In the controlled results, adolescent 167 Sex, Contraception and Childbearing Among High-Risk Youth mothers are less likely than their childless peers to volunteer, whereas young fathers are much more likely than childless young men to volunteer. Not surprisingly, teenage women with a young child have more time constraints than others, and they surely have less time for outside involvement. For fathers, the results may suggest the begin­ning of a move toward maturity, or they may indicate that young men who ac­knowledge fatherhood are the more re­sponsible of the young fathers, and thus are more likely than other fathers to be in­volved in positive activities. However, our methods do not permit us to draw con­clusions about the direction of causation. Discussion Our findings suggest that in certain re­spects, the factors associated with ado­lescent sexual behavior are the same for males and females. The young people in our sample who considered themselves risk-takers and who used or were in­volved with substances were somewhat more likely than others to become sexu­ally active at an early age. Additionally, those who became sexually active early were more likely than other youth to have run away from home. More typically, however, the linkages between the explanatory variables and the outcomes differed by gender. Some vari­ables had significant associations with an outcome for adolescents of both genders, but the associations were in opposite di­rections. For example, above-average lev­els of depression were associated with en­gaging in sexual activity, not using birth control and having an early birth for the sample overall. However, whereas sexu­ally active females had elevated levels of depression, sexually active males had re­duced depression levels. Other variables had a significant relationship with an out­come for men or women but not both. (Additionally, however, in a number of in­stances, "the- relationship between ex­planatory items and the outcomeSwas not significant for either gender.) Thus, a number of gender distinctions may apply to this population, although we cannot claim strong statistical support for any single association. Consistent with tra­ditional notions regarding gender behav­ior, youth who are inclined toward risk-tak­ing appear more likely than others to be sexually active. However, among sexual­ly active young women, those who use con­traceptives can be differentiated from nonusers by their acknowledgment that they do not need to use self-control to stay out of trouble. Additionally, it may well be 168 that the implications of not using contra­ceptives are so transparent and the conse­quences so immediate that for at least some of these young women, little conscious thought is required. However, this suppo­sition is applicable only to the subset of young women who are sexually active. We found fairly persuasive and sys­tematic evidence that for teenage women, having intercourse early, not using con­traceptives and having a child are linked with being depressed, feeling a failure and (with respect to having a child) having low self-esteem and little sense of control over one's life. For young men, the results are much more erratic. Males are similar to fe­males in showing associations between evidence of depression and both ineffec­tive contraceptive use and early childbirth. However, more typically, the correlation coefficients show either no relevant asso­ciations or weak associations in the op­posite directions. Behaviorally, gender distinctions are modest. The most pronounced differences are evidenced early: Teenage men who are sexually active express low levels of de­pression and feel in control of their envi­ronment. Furthermore, this feeling of con­trol does not appear to be transient, as those who go on to become young fathers main­tain this belief. By contrast, young women show significant associations between hav­ing sex and feeling depressed. Addition­ally, young women who have had a child report feeling less control over their life. After young people become parents, tentative evidence of maturity begins to appear for both genders; again, however, our methods do not disentangle causali­ty. Young mothers' below-average level of alcohol consumption and reduced ten­dency to spend time with peers who drink may suggest that widespread media mes­sages publicizing the potential dangers of alcohol abuse have had an effect, or that new mothers have different lifestyles than they had before they gave birth. Additional analyses using methods that more direct­ly test for causality might shed light on whether some of their other psychologi­cal traits-feelings of failure and lack of control over their world-could alter their likelihood of quickly having another birth. Young fathers exhibit noticeably lower levels of willingness to take risks than their peers, higher levels of depression and, particularly, greater involvement in socially productive activities. These young men may have learned something, al­though they paid a high price, and one is left with a feeling that they may act more responsibly in the future. In summary, we have found that on av­erage, the adolescent males in our sample approached sexual behavior in a more ca­sual manner, and from a different per­spective, than the young women. One might conclude that sexually active youth are following behavior paths for the wrong reasons, although those reasons are not easy to remedy. To some extent, the essential problem could perhaps be mediated if the value­laden baggage that many young men and women carry with them toward adult­hood could be dropped. In many in­stances, personalized counseling may be required. The strong and independent as­sociation between conservative notions re­garding appropriate roles for women and men's unwillingness to use contraceptives may partly reflect that young men ratio­nalize their behavior. It surely is an im­portant symptom of the problem. This life-event perspective leaves some room for optimism, given the various ways in which the young fathers and mothers may have altered their attitudes and behaviors in association with their new roles as parents. This may be predic­tive of changes in their likelihood of con­tinuing early childbearing. Subsequent waves of data from this same sample may shed light on this outcome. Appendix: Coding of Variables Sexual Outcomes All variables are coded 1 if the characteristic is present. Demographic Characteristics All variables are coded 1 if the characteristic is present, except age, which is continuous. Attitudes Planning takes the fun out of things, have to use self-control, life without danger is dull, enjoy tak­ing risks, feel a failure and feel depressed at times are coded on a four-point scale indicating strength of agreement (l=strongly disagree, 4=strongly agree). Self-esteem is measured by the nine-item Rosenberg index (Cronbach's alpha=.85). Each Likert-scaled item has a range of 1-4, with high­er scores representing greater self-esteem. Locus of control is measured by a seven-item mean index (Cronbach's alpha=.72). Each Likert- scaled item has a range of 1-4, with higher scores indicating that adolescents feel they have greater control over their actions and environment. The depression score is an abbreviated version of the Centers for Epidemiologic Studies De­pression Inventory. (The full scale is a self-report measure of adult depressive symptoms. It has high internal consistency and adequate test-retest reliability.26) The score used here is a seven-item mean index (Cronbach's alpha=.71) that ascertains how often mood has affected an adolescent's en­ergy level, sleeping habits and ability to concen­trate. Each item has a range of 1-3, with higher scores representing greater depression. Family Planning Perspectives The traditional roles scale is measured by an eight-item mean index (Cronbach's alpha=.70). Each Likert-scaled item has a range of 1-4, with higher scores indicating more traditional attitudes toward women's roles in the home and workplace. Behaviors Frequency of drinking is coded on a nine-point scale (l=never in last year, 9=daily). Having peers who drink is a measure of how many of the re­spondent's friends drink at least sometimes (0=none, 4=all). Marijuana use indicates how many times the respondent has ever used it (l=never, 7=100 times or more). Problems because of substance abuse is a summed index of six items (Cronbach's alpha=.84) about whether the ado­lescent has experienced difficulties as a result of using drugs, such as having had grades suffer, driven a car while high or done things uninten­tionally; each item has a four-point range, with higher scores representing more problems. Number of times ran away is coded on a four- point scale (0=never, 3=six or more). Got help for emotional problems is coded 1 if the respondent had received help for an emotional, behavioral or family problem in the past 12 months. Volunteered is coded 1 if the respondent had performed any vol­unteer or community work in the past two years. Current Maternal and Family Characteristics Urban residence is coded 1. All other measures are continuous variables. Average ratio of fami­ly income to needs is based on the income level that the federal government uses to define mini­mum acceptable living standards, taking into ac­count such factors as the state of residence and family size. Maternal Characteristics Age at first sex and at menarche are continuous variables. Early substance use is a composite of scores on cigarette and alcohol items asked in 1984. Alcohol use was coded 1 if the mother had had six or more drinks on at least two occasions in the previous month; cigarette use was coded 1 if she had had one or more cigarettes per day in the previous month. The deviant behavior score ranges from 0 to 8, reflecting how many of the following behaviors the mother reported engaging in during the year prior to the 1980 interview: taking a car without the owner's permission; breaking into a building to steal or just look around; destroying property more than once; shoplifting; smoking marijuana more than once; using force to obtain money or items; attacking someone with the intent to injure; and try­ing to get something from someone by lying. References 1. Jessor R and Jessor S, Problem Behavior and Psychoso­cial Development: A Longitudinal Study of Youth, New York: Academic Press, 1977. 2. Ibid.; Costa FM et al., Early initiation of sexual inter­course: the influence of psychosocial unconventionali­ty, Journal of Research on Adolescence, 1995,5(1):93-121; and Donovan JE and Jessor R, Structure of problem behav­ior in adolescence and young adulthood, Journal of Con­sulting & Clinical Psychology, 1985,53(6):890-904. 3. Mott FL et al., The determinants of first sex by age 14 in a high-risk adolescent population, Family Planning Per­spectives, 1996,28(1):13-18. 4. Moore KA et al., Adolescent Sex, Contraception, and Childbearing: A Review of Recent Research, Washington, DC: Child Trends, 1995. 5. Ibid. 6. Ibid. 7. Waterman C and Nevid J, Sex differences in the reso­lution of the identity crisis, Journal of Youth and Adoles­cence, 1977,6(4):337-342. 8. Weis D, Affective reactions of women to their initial experience of coitus, Journal of Sex Research, 1983, 19(2):209-237. 9. Mortimer J, Adolescence, in: Levinson D, ed., Ency­clopedia of Marriage and the Family, New York: Macmil­lan, 1996, pp. 12-14. 10. Day RD, The transition to first intercourse among racially and culturally diverse youth, Journal of Marriage and the Family, 1992,54(4):749-762. 11. Udry JR, Kovenock J and Morris NM, Early predic­tors of nonmarital first pregnancy and abortion, Family Planning Perspectives, 1996,28(3):113-116. 12. Herzog AR and Bachman JG, Sex Role Attitudes Among High School Seniors: Views About Work and Family Roles, Ann Arbor, MI: Institute for Social Research, Uni­versity of Michigan, 1982. 13. The Alan Guttmacher Institute (AGI), Sex and Amer­ica's Teenagers, New York: AGI, 1994; Dryfoos J, Adoles­cents at Risk: Prevalence and Prevention, New York: Oxford University Press, 1990; and Mott FL and Haurin RJ, Link­ages between sexual activity and alcohol and drug use among American adolescents, Family Planning Perspec­tives, 1988,20(3):128-136. 14. AGI, 1994, op. cit. (see reference 13). 15. Moore KA et al., 1995, op. cit. (see reference 4). 16. Mott FL and Haurin RJ, 1988, op. cit. (see reference 13); Rodgers JL and Rowe DC, Adolescent sexual activ­ity and mildly deviant behavior: sibling and friendship effects, Journal of Family Issues, 1990, ll(3):274-293; Ya- maguchi K and Kandel D, Drug use and other determi­nants of premarital pregnancy and its outcome: a dy­namic analysis of competing life events, Journal of Marriage and the Family, 1987,49(l):257-270; and Zabin LS et al., Substance use and its relations to sexual activ­ity among inner-city adolescents, Journal of Adolescent Health Care, 1986,7(4):320-331. 17. Ketterlinus RM et al., Adolescent non-sexual and sex- related problem behaviors, Journal of Research on Adoles­cence, 1992, 7(4):431-456. 18. Rosenbaum E and Kandel D, Early onset of adoles­cent sexual behavior and drug involvement, Journal of Marriage and the Family, 1990,52(3):783-798. 19. Whitb^ck LB et al., Minor deviant behaviors and ado­lescent sexual activity, Youth and Society, 1993, 24(2):166-177. - ~ 20. Moore KA et al., 1995, op. cit. (see reference 4). 21. Reiss I, Family Systems in America, second ed., Hins­dale, IL: Dryden Press, 1976. 22. Chilman CS, Adolescent Sexuality in a Changing Amer­ican Society: Social and Psychological Perspectives, Wash­ington, DC: U.S. Department of Health, Education, and Welfare, 1983; Chilman CS, Some psychosocial aspects of female sexuality, The Family Coordinator, 1974, 23(2):123-131; and HofferthSL, Influences on early sex­ual and fertility behavior, in: Hofferth SL and Hayes CD, eds., Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Washington DC: National Academy Press, 1987, Vol. 2, pp. 7-35. 23. Mott FL, The NLSY Children, 1992: Description and Evaluation, Columbus, OH: Center for Human Resource Research, The Ohio State University, 1995; and Baker PC et al., NLSY Child Handbook, rev. ed., Columbus, OH: Cen­ter for Human Resource Research, The Ohio State Uni­versity, 1993. - 24. Mott FL et al., 1996, op. cit. (see reference 3). 25. Marsiglio W, Adolescent fathers in the United States: their initial living arrangements, marital experience and educational outcomes, Family Planning Perspectives, 1987, 19(6):240-251; Marsiglio W, Young nonresident biolog­ical fathers, Marriage and Family Review, 1995, 20(3^):325-348; and Pirog-Good MA and Good DH, Child support enforcement for teenage fathers: problems and prospects, Journal of Policy Analysis and Management, 1995,14(l):25-42. 26. Radloff LS, The CES-D scale: a self-report depression scale for research in the general population, Applied Psy­chological Measurement, 1977,1(3):385-401. Volume 30, Number 4, July/August 1998 169