Association of sexual abstinence in adolescence with mental health in adulthood.

We examined whether adolescent sexual abstinence predicts better adult mental health. A total of 1,917 adolescents, recruited from middle schools at age 13, were surveyed at ages 13, 18, 23, and 29. In bivariate analyses, adolescent sexual abstinence was associated with better mental health at age 29 for females, but not for males," three adolescent factors, educational prospects, family bonding, and unconventionality, were investigated as explanatory variables of this relationship. The abstinence-mental health relationship was nonsignificant when educational prospects were included in multivariate models, and was marginally significant when family bonding and unconventionality were included; all three explanatory factors accounted for significant proportions of the variance in adult mental health. Girls who are uninvolved in school, have weak family backgrounds, and exhibit unconventionality may have poor adult mental health, whether or not they abstain from sex in adolescence. Interventions that strengthen adolescents' connections to families and schools may reduce risk for long-term mental health problems.

Policymakers are increasingly promoting sexual abstinence as an antidote to problems among society's youth. This belief in the benefits of abstinence was codified in Section 510 of Title V of the 1998 Social Security Act, which established an appropriations program for abstinence education (U.S. Department of Health & Human Services, 1998). To receive federal funds for abstinence education, such programs are required, for example, to teach youth that the benefits of abstinence include "social, psychological, and health gains," and that "sexual activity outside of the context of marriage is likely to have harmful psychological and physical effects."

Although assumptions regarding the beneficial psychological effects of sexual abstinence underlie Department of Health & Human Services [DHHS] policies, only a handful of studies have examined the relationship between adolescent sexual abstinence and long-term mental health. The majority of research on this topic has been cross sectional, finding significant associations between sexual abstinence and better mental health in adolescence (e.g., fewer symptoms of depression, stress, conduct disorder, and substance abuse or dependence; Brooks, Harris, Thrall, & Woods, 2002; Hallfors, Waller, Bauer, Ford, & Halpern, 2005; Rector, Johnson, & Noyes, 2003; Stiffman, Earls, Robins, Jung et al., 1987; Waller et al., 2006). In such cross-sectional research, however, sexual activity and mental health were measured at the same point in time, making the direction of the relationship unclear.

A few studies have investigated the relationship between abstinence and mental health longitudinally, with adolescent abstinence measured prior to mental health symptoms. Using data from the Bureau of Labor Statistics' National Longitudinal Survey of Youth, Finger and colleagues (Finger, Thelen, Vessey, Mohm, & Mann, 2004) found that women who were virgins at age 18, versus those who were sexually active as teenagers, were less likely to have an emotional illness at age 40; the corresponding effect for men was not significant. The effect for women persisted when the analysis was conducted without those reporting early pregnancy (a risk factor for later mental health problems). A population-based study of male and female twins demonstrated that sexual abstinence prior to and including age 17, compared with sexual activity, was related to a lower prevalence of psychopathology at age 17 (McGue & Iacono, 2005); a general problem behavior factor including sexual activity, substance use, and deviance prior to age 18 predicted psychopathology at age 20. An analysis of the data of children in grades 7-11 in the National Longitudinal Study of Adolescent Health (Add Health; Hallfors et al., 2005) indicated that engaging in sexual behavior and using substances (combined into one risk factor), versus abstaining from sex and drug use, predicted a higher number of depressive symptoms one year later, controlling for age, race/ethnicity, parent education, family structure, and physical maturity. Depressive symptoms did not predict sexual behavior and substance use in longitudinal analyses, however, suggesting that sexual activity precedes mental health problems. Another longitudinal analysis of data from Add Health (Sabia, 2006) found that sexual activity at ages 13-14 was associated with a higher number of depressive symptoms at ages 15-16 among girls, but not among boys; this relationship was reduced to nonsignificance after controlling for time-varying covariates related to mental health (i.e., romantic relationship status, attempted suicide of family members and friends, perception of parents' caring, college aspirations, weight perceptions, religiosity, alcohol use, grade point average, physical health, and loneliness).

In sum, the limited body of longitudinal research has indicated a significant association between teenage sexual abstinence and mental health. Although potential confounding variables such as early pregnancy, family background, and physical maturity have been examined, no research to date has comprehensively placed the relationship between abstinence and mental health in the context of a range of confounding variables that may covary with both abstinence and mental health. Moreover, two of these prior studies lack specificity, creating confounds by combining sexual activity and substance use to create predictors. Significant associations in prior research may have been due to a lack of a range of appropriate statistical controls, or to these nonspecific measures, rather than a causal effect of sexual abstinence.

In the current research, we investigated whether demonstrated correlates of both abstinence and mental health in prior research--including family bonding, educational prospects, and unconventional behaviors and attitudes--functioned as confounding variables that would explain any observed relationship between abstinence and mental health. Prior research indicates that stable and nuclear family structures (Capaldi, Crosby, & Stoolmiller, 1996; Lammers, Ireland, Resnick, & Blum, 2000; Newcomer & Udry, 1987; Upchurch, Aneshensel, Sucoff, & Levy-Storms, 1999), good parental communication, and strong parentchild relationships (Fasula & Miller, 2006; Kowaleski-Jones & Mott, 1998; Lammers et al., 2000; Maguen & Armistead, 2006), are robust correlates of sexual abstinence in adolescence. Strong family environments may inhibit sexual activity in adolescence, possibly because nuclear families offer greater resources, in terms of parental supervision, support, and control (Lammers et al., 2000). Family factors also predict better mental health in adulthood, including greater optimism and emotional well-being, and fewer depressive symptoms (Barrett & Turner, 2005; Bell & Bell, 2005; Ek, Remes, & Sovio, 2004; Luo & Waite, 2005; Spruijt, DeGoede, & Vandervalk, 2001).

In the United States, strong educational prospects in adolescence may similarly confound the relationship between adolescent sexual abstinence and adult mental health, especially for girls. Adolescent girls who have better life prospects--in terms of their parents' higher educational attainment and their own high academic achievement and aspirations--tend to refrain from early sexual activity, childbearing, and marriage (Astone & Upchurch, 1994; Capaldi et al., 1996; Fergusson & Woodward, 2000; Handler, 1990; Raine et al., 1999; Resnick et al., 1997; Schvaneveldt, Miller, Berry, & Lee, 2001; Scott-Jones & White, 1990; Small & Luster, 1994), possibly because they have consciously decided to postpone such activities until they achieve their educational and vocational goals. Higher socioeconomic status (SES) is associated with better mental health and adjustment in both adolescence and adulthood (Brown & Adler, 1998; Daniels, 1986; Evans & English, 2002; Goodman, McEwen, Dolan, Schafer-Kalkhoff, & Adler, 2005; Ostrove, Feldman, & Adler, 1999).

The relationship between abstinence and mental health also may be confounded by conventionality and substance use. Adolescents who use substances (Blinn-Pike, Berger, Hewett, & Oleson, 2004; Harvey & Spigner, 1995; Lammers et al., 2000; Mott, Fondell, Hu, Kowaleski-Jones, & Menaghan, 1996; Raine et al., 1999; Santelli, Kaiser, Hirsch, Radosh, Simkin, & Middlestadt, 2004), and who exhibit other types of unconventional behaviors and attitudes, including deviance and rebelliousness (Fergusson & Woodward, 2000; Jessor & Jessor, 1977), are more likely to start having sex or to be sexually active than are those who do not use substances or who are more conventional, respectively. Teenage substance use and deviance also predict worse mental health in adulthood, including higher depression, emotional distress, and anxiety (Fergusson & Woodward, 2000; Newcomb & Bentler, 1988).

In the current study, we examined the association between sexual abstinence until the end of the high school years and mental health at age 29. We investigated whether any observed relationship between adolescent sexual abstinence and adult mental health was attributable to demonstrated correlates of both abstinence and mental health, including family environment factors, educational prospects, and unconventional behaviors and attitudes. We examined this association separately by gender, based on prior research demonstrating gender differences in sexual behavior patterns and attitudes that may have different implications for mental health. For example, men are more likely to have sex before age 18 (De Gaston, Weed, & Jensen, 1996; Laumann, Gagnon, Michael, & Michaels, 1994); hold more permissive attitudes about premarital sex (Oliver & Hyde, 1993); and report different reasons for abstinence in adolescence (Loewenson, Ireland, & Resnick, 2004); and are less likely to experience negative mental health effects from early sexual activity (Finger et al., 2004).

Method

Participants

Participants were recruited at age 13 from 30 California and Oregon middle schools for the RAND Adolescent/ Young Adult Panel, a longitudinal cohort first surveyed during grade 7 (age 13) and followed until age 23. The schools were selected to represent a broad range of community environments with respect to urbanicity (urban, suburban, rural), ethnic composition, and SES. Nine schools had minority populations of 50% or greater, and 18 schools drew from neighborhoods with household incomes below the median for their state (Ellickson & Bell, 1990).

A total of 6,527 adolescents participated at age 13. Of the 2,172 individuals who participated in the baseline assessment and both follow-up surveys of interest to the current analysis (at ages 23 and 29), 1,985 participants had complete and consistent data to compute the sexual abstinence variable. Of those 1,985 participants, 1,963 were not missing data on the primary outcome, mental health at age 29. Because we aimed to examine whether any of the potential confounding variables were predisposing or covarying factors, rather than consequences, of sexual abstinence in adolescence, we further restricted the baseline sample to the 1,871 participants who had never had sex by age 13 (i.e., 1,107 females, 764 males).

Male participants, African Americans, and those who exhibited high-risk behavior at an early age were more likely to drop out of the study or be lost to follow-up than those who were retained. To reduce bias associated with sample attrition, we developed nonresponse weights to provide estimates consistent with the original 1985 sample. We used logistic regression to create the weights, obtaining predicted probabilities of responding to the 1995 survey based on information collected in the 1985 survey, including race, gender, family structure, deviance, substance use, and grades. These weights removed 90% or more of the observed bias on several key measures for the obtained young adult sample (see Table 1).

Construction of Sexual Abstinence Measure

We constructed a variable indicating whether participants abstained from sexual activity through the end of high school (ages 17-18), using data from the baseline (age 13), age 23, and age 29 follow-up surveys. At ages 23 and 29, participants were asked when they first had sexual intercourse (vaginal or anal), if ever. Response options were grouped into multiple-year intervals (e.g., for the age 29 survey, the categories were: age 10 or younger; 11-12; 13-14; 15-16; 17-18; 19-20; 21-22; age 23 or older). Participants were classified as abstinent through high school if they reported their age at first sex being 19 or older on both surveys or they reported never having had sex. For the 19 cases in which responses to the sexual intercourse question were missing for the age 23 survey, we used responses from the age 29 survey; for the 31 cases in which responses to the sexual intercourse question were missing for the age 29 survey, we used responses from the age 23 survey. In addition, 186 respondents' data were set to missing on the abstinence variable because their responses between surveys were inconsistent (e.g., they indicated that they had sex before age 19 on one survey and after age 19 on the other survey, or they indicated they never had sex on the age 23 survey and that they had sex before age 23 on the age 29 survey).

Age 29 Mental Health

Mental health in the past 4 weeks was assessed with the well-validated MHI-5 (alpha = .83; Stewart & Ware, 1992). The average of the 5 items was computed. A sample item is, "How much of the time, during the past month, have you felt calm and peaceful?" Response options ranged from 0, all of the time to 5, none of the time. The MHI-5 has been used in numerous studies and has shown excellent reliability and validity in population studies (Berwick, Murphy, Goldman, Ware, & et al., 1991; McCabe, Thomas, Brazier, & Coleman, 1996; McHorney & Ware, 1995; Rumpf, Meyer, Hapke, & John, 2001).

Baseline (Age 13) Covariates

Race/ethnicity and gender. Race/ethnicity and gender were asked at age 13.

Family bonding. Students indicated whether they were living in a nuclear family (i.e., with both the mother and father). Parental communication was assessed by asking respondents whether they would talk to their parents if they had a personal problem (yes/no).

Educational prospects. Highest educational attainment of the respondents' two parents was used to indicate family SES (0 = no high school degree to 3 = high school degree plus some college). Academic orientation was measured in terms of grades and academic intentions. Grades in the 6 months prior to the seventh-grade survey were self-reported by the respondent as mostly As, mostly Bs, mostly Cs, mostly Ds, or mostly Fs. Academic intentions were assessed with one item, "What is the highest level of school you plan to finish?" with response options 1, I may not finish high schoo; 2, I plan to finish high school; 3, I plan to attend vocational or trade school after high school graduation; 4, I plan to go to college; or 5, I plan to go to graduate or professional school, such as medical or law school.

Unconventionality. Three measures were used to assess unconventional behaviors and attitudes that represented participants' disengagement from conventional social institutions: substance use, deviance, and rebelliousness. Frequency of cigarette and marijuana use by grade 7 were each measured with substance-specific scales that ranged from 0, never used, to 11, used 20 or more days in the past month. Binge drinking at grade 7 was assessed by asking participants if they drank any alcohol in the past year, and, if so, whether they had three or more alcoholic drinks in the last month; participants were classified as "nondrinker," "drinker but not a binge drinker," and "binge drinker." Because hard drug use was infrequent at age 13, it was not included in analyses.

Deviance at age 13 was measured as the average of four items: number of times in the past year the student skipped school without a good excuse, cheated on a test, took something from a store, and was sent out of the classroom for causing trouble. Response options ranged from 0 (never) to 3 (more than 3 times). Internal consistency was adequate (alpha = .64).

Two items were standardized and averaged for a measure of rebelliousness: "Sometimes I enjoy doing things I shouldn't do just for the fun of it" (from 1, strongly agree, to 4, strongly disagree) and "Do you ever do things you're not supposed to do just to bother your parents?" (yes/no). The standardized scores ranged from -3 (not rebellious) to + 3 (rebellious). Although the internal consistency of the scale was low (alpha = .42), we chose to keep the scale in the analysis due to the conceptual importance of this variable.

Statistical Analysis

Bivariate analyses. We first used Fisher's Exact test to examine gender differences in sexual abstinence. We then used linear regression to examine the bivariate association between adolescent sexual abstinence and age 29 mental health, in separate models for females and males. When the bivariate association between abstinence and mental health was significant, we used logistic regression to examine the bivariate relationships between each of the proposed explanatory variables (i.e., educational prospects, family bonding, and unconventionality), and both adolescent abstinence and adult mental health.

Multivariate analyses. To test for factors explaining the relationship between abstinence and mental health, we constructed four multivariate models predicting adult mental health with adolescent sexual abstinence, using explanatory variables that were significantly (p < .05) related to both adolescent sexual abstinence and adult mental health at the bivariate level. The first three models controlled for each set of explanatory variables (family bonding, educational prospects, and unconventionality) in turn. For each multivariate model, we examined whether the coefficient for the relationship between abstinence and mental health became nonsignificant, which would suggest an explanatory role for the set of variables included in the model. We also tested whether each set of explanatory variables (i.e., family bonding, educational prospects, or unconventionality) was significant as a group using an adjusted Wald test. Each adjusted Wald test assessed whether the different conceptually related sets of explanatory variables significantly improved the bivariate model with abstinence as the sole predictor of mental health (i.e., whether each set of variables explained a significant amount of variation in mental health at age 29 that was not explained by adolescent abstinence).

Following the separate tests of each set of explanatory variables, we constructed one full logistic regression model that included all three sets of potential explanatory variables as covariates. The adjusted Wald test for this model assessed whether the entire block of explanatory variables significantly improved the bivariate model that included abstinence as the sole predictor.

Handling dependence and missing data. To account for dependence of observations within schools, Huber variance estimates (Huber, 1967) were calculated. Least squares regression (Little & Rubin, 1987) was used to impute missing values, using 60 variables (e.g., age, race, gender, religiosity, income, household composition, deviance, education) as predictors; predicted values were substituted for the small amount of missing data (i.e., 0 to 2.3%) on each of the predictors.

Results

Bivariate Models for Mental Health, Sexual Abstinence, and Potential Explanatory Variables

More than a fifth (21%) of women, versus more than a quarter (26%) of men, reported sexual abstinence in adolescence, p < .05. Adolescent abstinence was a significant bivariate predictor of adult mental health for women (b = .15, SE=.05, p < .01), but not for men (b = .13, SE= .10, p > .10). Women who had abstained from sexual activity during high school had more positive mental health at age 29. Because the corresponding effect was not significant for men, only the effect for women was considered further.

In bivariate analyses for females only, all of the proposed explanatory variables except marijuana use and binge drinking were significantly associated with abstinence, and all of the proposed explanatory variables except binge drinking, cigarette use, and marijuana use were significantly associated with mental health (Table 2). Thus, we included all potential explanatory variables in the multivariate models, with the exception of substance use.

Multivariate Associations Between Sexual Abstinence and Mental Health Among Women

The multivariate models are shown in Table 3. In the unconventionality model (Model I), the coefficient for abstinence was reduced to marginal significance, and the adjusted Wald test was significant for the set of unconventionality indicators, F(2,28) = 6.71, p < .01; within this unconventionality set, the coefficient for deviance was significant and negative, indicating that deviant girls had worse mental health in adulthood. Similarly, in the family bonding model, the abstinence coefficient was marginally significant and the adjusted Wald test was significant (Model II), F (2, 28) = 6.95, p < .01. The coefficients for both variables within the family bonding set--parent communication and nuclear family--were significant and positive, indicating that girls from dual-parent homes who have good relationships with their parents are likely to have better mental health in adulthood. In the educational prospects model (Model III), the coefficient for abstinence was reduced to nonsignificance, and the adjusted Wald test was significant, F (2, 28) = 12.06, p < .001; the coefficient for low academic orientation was significant and negative, and the coefficient for parent education was marginally significant and positive. Thus, poor educational prospects in adolescence lead to a higher risk for worse adult mental health and appear to account for the relationship between abstinence and mental health.

To control for any intercorrelations among the explanatory variables, we tested a final model in which all explanatory variables were included (Table 3, Model IV). The coefficient for abstinence was not significant in this full model. The adjusted Wald test indicated that the full set of explanatory variables explained a significant amount of the variance in mental health, F (6, 24) = 6.11, p < .001. Only low academic orientation was a significant predictor of mental health; the coefficients for parent education and communication were marginally significant, and none of the unconventionality variables were significant. These findings suggest that the set of explanatory variables fully accounted for the relationship between abstinence and mental health.

Discussion

The results of this comprehensive longitudinal study indicate that adolescent sexual abstinence is associated with better adult mental health among girls, prior to controlling for explanatory variables related to educational prospects, family bonding, and unconventionality. Some prior work has shown that, when asked about their first sexual encounter, girls report more negative emotions, including regret, than do boys, and boys report greater feelings of pleasure than do girls (Karofsky, Zeng, & Kosorok, 2000). Our findings suggest that these are only short-term psychological effects. In our multivariate models, there was a diminished and nonsignificant relationship between abstinence and mental health, once adolescent girls' educational prospects, family bonding, and unconventionality factors were taken into account.

One educational prospects variable--academic orientation-was a particularly robust predictor in explaining the abstinence-mental health relationship, remaining significant in the full model including all explanatory variables. Girls who had higher grades and educational aspirations at age 13 were more likely both to abstain from sexual intercourse throughout adolescence and to report better mental health in adulthood. Girls who aspire to go to college avoid sexual intercourse in order to avoid the risk of pregnancy (Teti & Lamb, 1989; Teti, Lamb, & Elster, 1987), and girls who aspire to go to college may eventually attain higher SES, which is associated with better mental health (Brown & Adler, 1998; Daniels, 1986; Evans & English, 2002; Goodman et al., 2005; Ostrove, et al., 1999). This latter effect, however appears to be independent of early sexual abstinence.

Our results indicate a smaller, albeit significant, explanatory role for family bonding factors in the relationship between abstinence and mental health. These findings are consistent with prior work, which has demonstrated that better parent-child relationships lead to later and less risky sexual activity in adolescence (Henrich, Brookmeyer, Shrier, & Shahar, 2006; Hutchinson, Jemmott, Jemmott, Braverman, & Fong, 2003; Miller, Norton, Fan, & Christopherson, 1998; Ream & Savin-Williams, 2005; Resnick et al., 1997) as well as better mental health in childhood, adolescence (Conger, Ge, Elder, Lorenz, & Simons, 1994; Conger, et al., 2002; Gutman, McLoyd, & Tokoyawa, 2005), and adulthood (Bell & Bell, 2005). Moreover, youth from stable, dual-parent families are less likely to be sexually active (Capaldi et al., 1996; Lammers et al., 2000; Newcomer & Udry, 1987; Upchurch et al., 1999) and to experience negative mental health outcomes (Spruijt et al., 2001). Stable families and strong parent-child relationships may set up an environment in which the child has the resources, support, and motivation to avoid sexual activity and to achieve high educational levels. Nevertheless, the effect of abstinence was marginally significant in the family bonding model, and the final multivariate model indicated diminished effects of parent communication once educational prospects were taken into account. Thus, although family bonding variables may contribute to the abstinence mental health relationship, they do not fully account for it.

Unconventionality variables similarly contributed to the relationship between abstinence and mental health, although the effect of unconventionality was eclipsed by the more robust effect of educational prospects in the full model. Moreover, adolescent substance use, one indicator of unconventionality, was not consistently associated with either abstinence or mental health at the bivariate level. This suggests that associations observed among abstinence, substance use, and mental health in some prior studies may be attributable to differences in, and a failure to control for, educational prospects.

Taken together, our findings suggest that the observed associations between abstinence and mental health are not causal. This finding is consistent with other recently published findings showing no significant association between early abstinence per se and adult life satisfaction (Else-Quest, Hyde, & Delamater, 2005). Low educational prospects, weak family factors, and unconventionality likely set the stage for early sexual activity, and these characteristics also independently contribute to worse emotional well-being in adulthood. Promoting sexual abstinence, by itself, is therefore unlikely to result in fewer mental health problems over the long term. Whether or not they initiate sexual behavior, adolescents who come from certain family backgrounds or who are less involved in school are more likely to exhibit poorer mental health in adulthood. These conclusions are bolstered by our analysis, which controlled for educational prospects, family bonding, and unconventionality at age 13, and included sexual abstinence after that age. Nevertheless, because abstinence at age 13 co-occurred with our control variables, we cannot be certain that stronger educational prospects, higher-quality parent relationships, and more conventional behavior preceded adolescent sexual decisions up to age 13. In addition, prior longitudinal research suggests a dynamic relationship, whereby adolescents who feel less connected to their parents are more likely to initiate sexual activity, and adolescents who continue to be sexually active feel less close to their parents (Ream & Savin-Williams, 2005). Future longitudinal research measuring academic, family, unconventional, and sexual factors at multiple points in time before and after adolescence would further help to clarify whether this relationship is a dynamic one that continues after adolescents begin to be sexually active.

Because our results indicate that families and schools play an essential role in promoting teenagers' health, interventions that strengthen adolescents' connections to each are likely to reduce risk for a range of short- and long-term health problems. Due to the robust linkages among all types of problem behaviors, including early sexual initiation, interventions that target adolescents with poor school achievement, low motivation, and deviant behaviors, and that engage parents and children in effective communication are likely to have beneficial effects across a number of health risk domains.

Our test of the association between sexual abstinence and mental health is one of the most comprehensive to date. The rich dataset from the RAND Adolescent/ Young Adult Panel allowed us to take into account several different possible confounding factors. Nevertheless, a large percentage of the sample was lost to attrition. Although we statistically corrected for most of the potential bias due to attrition between baseline and age 29, we could not control for unmeasured factors that may have contributed to bias. Use of weights, however, corrected for more than 90% of the differences between those who attrited versus those who did not with respect to gender, race/ethnicity, SES, and family status. In addition, sexual behavior measures may be subject to recall and social desirability bias, both of which may have contributed to some of the observed inconsistent reports between survey assessments at ages 23 and 29. Furthermore, we did not assess biological or sexual coercion factors. Pubertal timing and testosterone levels may affect the timing of first intercourse in adolescence (Halpern, Udry, & Suchindran, 1997; Miller et al., 1998), and first intercourse that occurs at age 14 or younger, versus over age 14, is more likely to be involuntary (Abma, Martinez, Mosher, & Dawson, 2004). Any future comprehensive study of this issue would need to include a full range of biological, family background, psychosocial, and sexual coercion factors.

Conclusion

We set out to examine support for the basic assumptions of a national policy--abstinence-only education--that sexual abstinence in adolescence promotes long-term mental health. We did not find support for this contention among boys, and among girls, the association was largely explained by adolescents' educational prospects, with some contribution from family bonding and unconventional attitudes and behaviors. Although the context surrounding some adolescent sexual behavior may be harmful (Else-Quest et al., 2005), adolescent sexual activity by itself is not likely a causal factor in long-term risk for poorer mental health. Adolescents' low academic aspirations, resource-poor family environments, and unconventionality may create risk for sexual initiation in adolescence and worse emotional well-being in adulthood. We caution against narrow interventions that focus on abstinence exclusively, without taking into account adolescents' family, school, and neighborhood context. Instead, we recommend the development of broader interventions that address a full range of risk behaviors simultaneously, with a focus on increasing adolescents' self-efficacy for academic achievement and strengthening parental communication.

References

Abma, J. C., Martinez, G. M., Mosher, W. D., & Dawson, B. S. (2004). Teenagers in the United States: Sexual activity, contraceptive use, and childbearing, 2002. National Center for Health Statistics, Vital Health Stat, 23, 1-48.

Astone, N. M. & Upchurch, D. M. (1994). Forming a family, leaving school early, and earning a GED: A racial and cohort comparison. Journal of Marriage & the Family, 56, 759-771.

Barrett, A. E. & Turner, R. J. (2005). Family structure and mental health: The mediating effects of socioeconomic status, family process, and social stress. Journal of Health and Social Behavior, 46, 156-169.

Bell, L. G. & Bell, D. C. (2005). Family dynamics in adolescence affect midlife well-being. Journal of Family Psychology, 19, 198-207.

Berwick, D. M., Murphy, J. M., Goldman, P. A., Ware, J. E., et al. (1991). Performance of a five-item mental health screening test. Medical Care, 29, 169-176.

Blinn-Pike, L., Berger, T. J., Hewett, J., & Oleson, J. (2004). Sexually abstinent adolescents: An 18-month follow-up. Journal of Adolescent Research, 19, 495-511.

Brooks, T. L., Harris, S. K., Thrall, J. S., & Woods, E. R. (2002). Association of adolescent risk behaviors with mental health symptoms in high school students. Journal of Adolescent Health, 31, 240-246.

Brown, H. D. & Adler, N. E. (1998). Socioeconomic status. In H. S. Friedman (Ed.), Encyclopedia of mental health (Vol. 1). San Diego: Academy Press.

Capaldi, D. M., Crosby, L., & Stoolmiller, M. (1996). Predicting the timing of first sexual intercourse for at-risk adolescent males. Child Development, 67, 344-359.

Conger, R. D., Ge, X., Elder, G. H. Jr., Lorenz, F. O., & Simons, R. L. (1994). Economic stress, coercive family process, and developmental problems of adolescents. Child Development, 65, 541-561.

Conger, R. D., Wallace, L. E., Sun, Y., Simons, R. L., McLoyd, V. C., & Brody, G. H. (2002). Economic pressure in African American families: A replication and extension of the family stress model. Developmental Psychology, 38, 179-193.

Daniels, S. (1986). Relationship of employment status to mental health and family variables in black men from single-parent families. Journal of Applied Psychology, 71, 386-391.

De Gaston, J. F., Weed, S., & Jensen, L. (1996). Understanding gender differences in adolescent sexuality. Adolescence, 31, 217-231.

Ek, E., Remes, J., & Sovio, U. (2004). Social and developmental predictors of optimism from infancy to early adulthood. Social Indicators Research, 69, 219-242.

Ellickson, P. L. & Bell, R. M. (1990). Drug prevention in junior high: A multi-site longitudinal test. Science, 247, 1299-1305.

Else-Quest, N. M., Hyde, J. S., & Delamater, J. D. (2005). Context counts: Long-term sequelae of premarital intercourse or abstinence. Journal of Sex Research, 42, 102-112.

Evans, G. W. & English, K. (2002). The environment of poverty: Multiple stressor exposure, psychophysiological stress, and socioemotional adjustment. Child Development, 73, 1238-1248.

Fasula, A. M. & Miller, K. S. (2006). African-American and Hispanic adolescents' intentions to delay first intercourse: Parental communication as a buffer for sexually active peers. Journal of Adolescent Health, 38, 193-200.

Fergusson, D. M. & Woodward, L. J. (2000). Educational, psychosocial, and sexual outcomes of girls with conduct problems in early adolescence. Journal of Child Psychology and Psychiatry, 41, 779-792.

Finger, R., Thelen, T., Vessey, J. T., Mohm, J. K., & Mann, J. R. (2004). Association of virginity at age 18 with educational, economic, social, and health outcomes in middle adulthood. Adolescent and Family Health, 3, 164-170.

Goodman, E., McEwen, B. S., Dolan, L. M., Schafer-Kalkhoff, T., & Adler, N. E. (2005). Social disadvantage and adolescent stress. Journal of Adolescent Health, 37, 484-492.

Gutman, L. M., McLoyd, V. C., & Tokoyawa, T. (2005). Financial strain, neighborhood stress, parenting behaviors, and adolescent adjustment in urban African American families. Journal of Research on Adolescence, 15, 425-449.

Hallfors, D. D., Waller, M. W., Bauer, D., Ford, C. A., & Halpern, C. T. (2005). Which comes first in adolescence--sex and drugs or depression? American Journal of Preventive Medicine, 29, 163-170.

Halpern, C. T., Udry, J. R., & Suchindran, C. (1997). Testosterone predicts initiation of coitus in adolescent females. Psychosomatic Medicine, 59, 161-171.

Handler, A. (1990). The correlates of the initiation of sexual intercourse among young urban black females. Journal of Youth and Adolescence, 19, 159-170.

Harvey, S. M. & Spigner, C. (1995). Factors associated with sexual behavior among adolescents: A multivariate analysis. Adolescence, 30, 253 264.

Henrich, C. C., Brookmeyer, K. A., Shrier, L. A., & Shahar, G. (2006). Supportive relationships and sexual risk behavior in adolescence: An ecological transactional approach. Journal of Pediatric Psychology, 31, 286-297.

Hutchinson, M., Jemmott, J., Jemmott, L., Braverman, P., & Fong, G. (2003). The role of mother-daughter sexual risk communication in reducing sexual risk behaviors among urban adolescent females: A prospective study. Journal of Adolescent Health, 33, 98-107.

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

Karofsky, P. S., Zeng, L., & Kosorok, M. R. (2000). Relationship between adolescent-parental communication and initiation of first intercourse by adolescents. Journal of Adolescent Health, 28, 41-45.

Kowaleski-Jones, L. & Mott, F. L. (1998). Sex, contraception and childbearing among high-risk youth: Do different factors influence males and females? Family Planning Perspectives, 30, 163-169.

Lammers, C., Ireland, M., Resnick, M., & Blum, R. (2000). Influences on adolescents' decision to postpone onset of sexual intercourse: A survival analysis of virginity among youths aged 13 to 18 years. Journal of Adolescent Health, 26, 42-48.

Laumann, E. O., Gagnon, J. H., Michael, R. T., & Michaels, S. (1994). The social organization of sexuality: Sexual practices in the United States. Chicago: University of Chicago Press.

Loewenson, P. R., Ireland, M., & Resnick, M. D. (2004). Primary and secondary sexual abstinence in high school students. Journal of Adolescent Health, 34, 209-215.

Luo, Y. & Waite, L. J. (2005). The impact of childhood and adult SES on physical, mental, and cognitive well-being in later life. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 60, S93-S101.

Maguen, S. & Armistead, L. (2006). Abstinence among female adolescents: Do parents matter above and beyond the influence of peers? American Journal of Orthopsychiatry 76, 260-264.

McCabe, C. J., Thomas, K. J., Brazier, J. E., & Coleman, P. (1996). Measuring the mental health status of a population: A comparison of the GHQ-12 and the SF-36 (MHI-5). British Journal of Psychiatry, 169, 517-521.

McGue, M. & Iacono, W. G. (2005). The association of early adolescent problem behavior with adult psychopathology. American Journal of Psychiatry, 162, 1118-1124.

McHorney, C. A. & Ware, J. E. (1995). Construction and validation of an alternate form general mental health scale for the medical outcomes study short-form 36-item health survey. Medical Care, 33, 15-28.

Miller, B. C., Norton, M. C., Fan, X., & Christopherson, C. R. (1998). Pubertal development, parental communication, and sexual values in relation to adolescent sexual behaviors. Journal of Early Adolescence, 18, 27-52.

Mott, F. L., Fondell, M. M., Hu, P. N., Kowaleski-Jones, L., & Menaghan, E. G. (1996). The determinants of first sex by age 14 in a high-risk adolescent population. Family Planning Perspectives, 28, 13-18.

Newcomb, M. D. & Bentler, P. M. (1988). Consequences of adolescent drug use: Impact on the lives of young adults. Beverly Hills, CA: Sage.

Newcomer, S. & Udry, J. (1987). Parental marital status effects on adolescent sexual behavior. Journal of Marriage and the Family, 49, 235-240.

Oliver, M. B. & Hyde, J. S. (1993). Gender differences in sexuality: A meta-analysis. Psychological Bulletin, 114, 29-51.

Ostrove, J. M., Feldman, P., & Adler, N. E. (1999). Relations among socioeconomic status indicators and health for African-Americans and Whites. Journal of Health Psychology, 4, 451-463.

Raine, T. R., Jenkins, R., Aarons, S. J., Woodward, K., Fairfax, J. L., El-Khorazaty, M. N., & Herman, A. (1999). Sociodemographic correlates of virginity in seventh-grade black and Latino students. Journal of Adolescent Health, 24, 304-312.

Ream, G. L. & Savin-Williams, R. C. (2005). Reciprocal associations between adolescent sexual activity and quality of youth-parent interactions. Journal of Family Psychology, 19, 171-179.

Rector, R. E., Johnson, K. A., & Noyes, L. R. (2003). Sexually active teenagers are more likely to be depressed and to attempt suicide. A report of the Heritage Center for Data Analysis. Washington, DC: The Heritage Foundation.

Resnick, M. D., Bearman, P. S., Blum, R. W., Bauman, K. E., Harris, K. M., Jones, J., Tabor, J., Beuhring, T., Sieving, R. E., Shew, M., Ireland, M., Bearinger, L. H., & Udry, J. R. (1997). Protecting adolescents from harm: Findings from the national longitudinal study on adolescent health. JAMA: Journal of the American Medical Association, 278, 823-832.

Rumpf, H.-J., Meyer, C., Hapke, U., & John, U. (2001). Screening for mental health: Validity of the MHI-5 using DSM-IV Axis I psychiatric disorders as gold standard. Psychiatry Research, 105, 243-253.

Sabia, J. J. (2006). Does early adolescent sex cause depressive symptoms? Journal of Policy Analysis and Management, 25, 803-825.

Santelli, J. S., Kaiser, J., Hirsch, L., Radosh, A., Simkin, L., & Middlestadt, S. (2004). Initiation of sexual intercourse among middle school adolescents: The influence of psychosocial factors. Journal of Adolescent Health, 34, 200-208.

Schvaneveldt, P. L., Miller, B. C., Berry, E. H., & Lee, T. R. (2001). Academic goals, achievement, and age at first sexual intercourse: Longitudinal, bidirectional influences. Adolescence, 36, 767-787.

Scott-Jones, D. & White, A. B. (1990). Correlates of sexual activity in early adolescence. Journal of Early Adolescence, 10, 221-238.

Small, S. A. & Luster, T. (1994). Adolescent sexual activity: An ecological, risk-factor approach. Journal of Marriage & the Family, 56, 181-192.

Spruijt, E., DeGoede, M., & Vandervalk, I. (2001). The well-being of youngsters coming from six different family types. Patient Education and Counseling, 45, 285-294.

Stewart, A. L. & Ware, J. E. (1992). Measuring functioning and well being: The medical outcomes study approach. Durham, NC: Duke University Press.

Stiffman, A. R., Earls, F., Robins, L. N., Jung, K. G., & et al. (1987). Adolescent sexual activity and pregnancy: Socioenvironmental problems, physical health, and mental health. Journal of Youth and Adolescence, 16, 497-509.

Teti, D. M. & Lamb, M. E. (1989). Socioeconomic and marital outcomes of adolescent marriage, adolescent childbirth, and thier co-occurrence. Journal of Marriage and the Family, 51, 203-212.

Teti, D. M., Lamb, M. E., & Elster, A. B. (1987). Long-range socioeconomic and marital consequences of adolescent marriage in three cohorts of adult males. Journal of Marriage and the Family, 49, 499-506.

U.S. Department of Heath & Human Services. (1998). Understanding title V of the Social Security Act. Retrieved August 14, 2006, from ftp://ftp.hrsa.gov/mchb/titlevtoday/UnderstandingTitleV.pdf

Upchurch, D., Aneshensel, C., Sucoff, C., & Levy-Storms, L. (1999). Neighborhood and family contexts of adolescent sexual activity. Journal of Marriage and the Family, 61, 920-933.

Waller, M. W., Hallfors, D. D., Halpern, C. T., Iritani, B. J., Ford, C. A., & Guo, G. (2006). Gender differences in associations between depressive symptoms and patterns of substance use and risky sexual behavior among a nationally representative sample of U.S. adolescents. Archives Women's Mental Health, 9, 139-150.

Laura M. Bogart, Rebecca L. Collins, Phyllis L. Ellickson, and David J. Klein

RAND Corporation 

This research was funded by grant #R01 DA 13515 from the National Institute on Drug Abuse. Support for preparation of this manuscript was partially provided by CDC Center Grant #448/ D8000056.

Correspondence should be addressed to Laura M. Bogart, Ph.D., RAND Corporation, 1776 Main Street, P.O. Box 2138, Santa Monica, CA 90407-2138. E-mail: [e-mail]

Table 1. Weighted and Unweighted Percentages of the RAND
Panel's Socio demographic, Substance Use, and Problem
Behaviors and Attitudes

                                               Young Adult Sample
                           Baseline Sample        (N = 2,172)
                             (N = 6,527)
                             Unweighted      Unweighted   Weighted

Gender
  Male                          52.0            42.6        50.7
  Female                        48.0            57.4        49.3
Race/Ethnicity
  White                         67.3            74.4        67.3
  Black                         10.3             5.9         9.8
  Hispanic                      10.3             7.7        10.8
  Asian                          8.2             9.6         8.5
  Other                          4.0             2.4         3.7
Substance use (lifetime)
  Ever used marijuana           21.5            12.8        20.4
  Ever used alcohol             77.2            74.8        77.0
  Ever used cigarettes          54.4            44.7        54.1
C average or worse              30.1            20.2        30.0
  Ever shoplifted               23.3            17.1        22.7

Note: The weights were constructed for the subset of 2,172 young
adults who responded to the age 13 baseline survey and age 23 and
age 29 follow-up surveys.

Table 2. Bivariate Relationships of Potential Explanatory Variables
at Age 13 With Adolescent Abstinence and Age 29 Mental Health
(Women Only, n = 1,107)

                             Bivariate Relationship
                               With Abstinence
                                 Until Age 19
Predictors at Age 13               b (95%CI)

Unconventionality
  Binge drinking              -1.05 (-2.18,.07) (+)
  Cigarette use               -1.29 (-2.19, -.39) **
  Marijuana use               -1.19 (-3.09, .71)
  Deviance                     -.61 (-.86, -.35) ***
  Rebelliousness              -1.10 (-1.83, -.37) **
Family bonding
  Parent communication          .64 (.42, .86) ***
  Nuclear family                .60 (.23, .98) **
Educational prospects
  Parent education              .29 (.15, .43) ***
  Low academic orientation     -.62 (-.90, -.33) ***

                             Bivariate Relationship
                              With Mental Health
                                  at Age 29
Predictors at Age 13               b (95%CI)

Unconventionality
  Binge drinking               -.25 (-.50, .01) (+)
  Cigarette use                -.10 (-.29, .09)
  Marijuana use                -.27 (-.59, .04) (+)
  Deviance                     -.09 (-.15, -.02) *
  Rebelliousness               -.19 (-.32, -.06) **
Family bonding
  Parent communication          .18 (.08, .27) **
  Nuclear family                .17 (.04, .30) *
Educational prospects
  Parent education              .12 (.06, .18) ***
  Low academic orientation     -.21 (-.30, -.12) ***

(+) p <.10, * p<.05, ** p<.01, *** p<.001.