What predicts adolescent sexual activity? This question has been of central interest to researchers for more than two decades.^1–5 Among the factors that have been investigated, family structure has received a considerable amount of attention. For instance, research has consistently demonstrated that adolescents from intact two-parent families are more likely to delay sexual activity and have fewer sexual partners than adolescents in singleparent and stepfamily homes.^5–8 Relatedly, Kapinus and Gorman⁹ found that both males and females in singlemother and mother-stepfather families have more positive views about becoming teenage parents than do adolescents who are adopted or living with both biological parents, and Trent¹⁰ reported that white adolescents who live with a mother and a stepfather are more likely than those living with both biological parents to anticipate becoming teenage parents.

This focus on family structure, although meaningful, is problematic, because it ignores the potential influence that family relationships—or family context—may have on adolescents’ sexual attitudes and practices. As Davis and Friel¹¹ noted, developing an understanding of adolescent sexual activity is contingent upon an examination of "the total family environment," rather than family structure alone. Some studies have examined the link between family context and teenage sexuality, but most have focused on the general effects of the biological parents^8,12 or the specific effects of the relationship between the biological mother and the child.^1,11 A more complete understanding of the influence of family context requires an exploration of the extent to which relationships with nonbiological family members, such as stepparents or stepsiblings, affect teenage sexuality. For instance, close stepparent-stepchild relationships may lessen the likelihood of teenagers’ having sex. However, the existing literature contains very little about how the stepfamily environment is related to adolescents’ attitudes toward sex. We attempt to address this gap by focusing on the stepfather-stepchild relationship and its potential impact on adolescents’ attitudes toward sexual behavior.

Why stepfathers? The number of children living in a stepfamily household exceeded five million in 2001, up from approximately 4.5 million in 1991; the vast majority of these children were living with a biological mother and stepfather.¹³ Moreover, about one-third of all children in the United States will live with a remarried or cohabiting parent before they reach adulthood.¹⁴ For a large number of children, then, understanding family context necessitates understanding the impact of stepfathers in their lives. We do this by considering the degree to which the stepfather-stepchild relationship is associated with adolescents’ attitudes toward sex.

We focus on attitudes, rather than actual behaviors, because both research evidence^15–18 and a general theoretical argument¹⁹ support the notion that adolescents’ attitudes are a significant predictor of engagement in sexual activity. Furthermore, family context influences sexual behavior through teenagers’ attitudes,^20–22 although this link has not been adequately explored in the stepfamily literature.

Understanding behaviors is certainly important, given that the consequences associated with adolescent sexual activity, from early childbearing and STDs^23,24 to diminished educational aspirations,²⁵ have the potential to dramatically alter one’s life chances. Moreover, adolescents living with single parents or stepparents are more likely to expose themselves to these risks than are teenagers living with both biological parents.^3,5,8 Future investigations should attempt to uncover direct or indirect links between relationships with stepfathers and teenagers’ sexual behavior.

BACKGROUND

Perspectives on Teenage Sexuality

Existing research in this area can generally be classified within one of three frameworks: family control, family culture or family involvement.¹¹ Each of these frameworks may also be used to explain the link between family structure and adolescents’ attitudes toward sexuality. Scholars who subscribe to the family control argument assert that two-parent families offer the greatest stability and supervision, with the result that adolescents delay sexual activity.^4,21,26 One would expect that if the family control argument holds, greater regulation and supervision would correlate with less permissive sexual attitudes among adolescents. However, stepparents may have only a limited ability to provide control and stability, because they are not viewed, at least by their stepchildren,^27–29 as having the same authority as biological parents.

The family culture explanation asserts that family culture (which includes parents’ attitudes and behaviors) is the conduit that links family structure and adolescent sexual behavior. More precisely, different family structures reflect different cultural norms regarding sexuality.^3,5,30 Parents in single-parent families and reconstituted families have relatively permissive sexual attitudes, which they model through their actions.^21,31 Parents’ attitudes are important because adolescents are more likely to engage in sexual activity when they believe their mothers approve of this behavior than otherwise.^2,32,33 Thus, according to this explanation, teenagers in reconstituted families may have increased motivations to engage in sexual activity because they share their parents’ more permissive sexual attitudes.

Both the family control and the family culture frameworks are grounded in an understanding of teenage sexual attitudes and behaviors as stemming from differences in family structure. Although both do, at some level, address family environment, they assume that the influence of environmental factors works through family structure. The family control model suggests that the two-parent biological family is inherently better at providing control and stability for children, while the family culture model suggests that almost by default, the dating and courtship activities of single or remarried parents leads to permissive attitudes toward sexual engagement.

Scholars who subscribe to the family involvement framework, however, assert that interactions, communication and the quality of the relationship between a parent and child influence adolescent sexual behavior. From this perspective, environmental influences are not tied to a particular family structure; rather, they hinge on the individual-level relationships within the family. For instance, Davis and Friel¹¹ found that relationships between mothers and adolescents that are defined as close, warm and loving are associated with decreases in early sexual initiation among teenagers. Likewise, Miller, Forehand and Kotchick²⁰ reported that adolescents who talk with their mothers about general life issues or sexual behaviors are less inclined than others to engage in sexual activity. From a sample of black, inner-city adolescents, Jaccard, Dittus and Gordan³³ reported that even after communication with mothers and mothers’ attitudes are controlled for, adolescents who are less satisfied with their relationship with their mothers have an increased likelihood of having had sexual intercourse and of engaging in frequent coitus, and a reduced likelihood of using contraceptives. In contrast, adolescents who feel close to their mothers have elevated odds of delaying sex and of using birth control when they do have sex, and reduced odds of becoming pregnant.³² Thus, one would expect that closer parent-child relationships are associated with a reduced disposition toward having sex.

Although the focus on context provided by the family involvement framework is a welcome addition to the literature on teenage sexuality, most of these studies are limited in that they focus only on the mother-child relationship. Moreover, although scholars have speculated that adolescents’ attitudes may be an intervening variable linking family context and teenage sexual behavior,³⁴ few have investigated factors that influence teenagers’ attitudes. Therefore, in this study, we examine the family involvement framework by exploring the link between adolescents’ perceptions of closeness to stepfathers and attitudes toward sexual behavior.

Stepfather Involvement

Existing literature on stepfather-stepchild relationships suggests two possible—and contradictory—ways that stepfather relationships may be associated with adolescents’ attitudes toward sex. Because children perceive their stepparents as friends or advisors who hold less authority and decision-making rights than their biological parents,^27,28 stepfathers may have only minimal impact on their attitudes toward sexual activity. Relatedly, society’s generally negative view of stepfamilies^14,35 and the stigma attached to nonbiological families (those formed through adoption or remarriage)^27,35–39may lessen any potential benefits that a stepparent might bring to the life of a stepchild. For example, biological parents are often referred to as "real" parents or "natural" parents, thereby suggesting that stepparents and adoptive parents are somehow less real and even abnormal in children’s lives. Members of stepfamilies may internalize these negative views, which may adversely affect their perceptions of and experiences with their own families. This kind of an internalization could limit a stepfather’s ability to influence his stepchild’s attitudes toward sex.

Yet, a stepfather’s presence may be influential on a stepchild. Some studies have found that with respect to levels of cohesion, support, control and emotional closeness, stepfathers often function much as biological fathers.^14,40–42Furthermore, children who feel close to their residential father (whether a biological father or a stepfather) exhibit significantly fewer behavioral problems than children who report feeling estranged from their residential father.^43,44 Likewise, children with highly involved stepparents tend to experience the same kinds of behavioral problems as children raised in traditional nuclear families.⁴⁵ Thus, when the relationship is close, the stepfather may influence his stepchild’s attitudes toward sex.

We expect that a stepfather’s influence will depend on how the stepchild experiences the relationship. Children who perceive the relationship to be close and who are more engaged with their stepfather may benefit from his presence, or at least be protected from the negative outcomes associated with living outside of a two-parent biological home.

The Importance of Gender

The effects of closeness with stepfathers may also depend on the child’s gender. The quality of family relationships and discussions with parents about sexual values and activity have stronger effects on daughters’ sexual attitudes than on sons’.^4,46,47 However, social learning theorists have asserted that children’s behavior is likely to reflect the behavior of the same-sex parent.⁴⁸ Consistent with this notion, some studies have suggested that fathers are more likely to influence the development of sons than of daughters^49,50 and that males benefit more from the presence of a stepfather than do females.^51,52 Research by Risch, Jodl and Eccles⁵³ has indicated that whereas males who feel close to their biological fathers and stepfathers are more likely than others to report negative attitudes toward divorce, closeness does not matter for females.

The varied and sometimes contradictory findings regarding the importance of gender on adolescent outcomes make it difficult to hypothesize how gender will influence the association between stepfather involvement and teenagers’ sexual attitudes. It is clear, however, that the teenager’s gender will likely matter in some respect. Thus, we examine effects for males and females separately.

METHODS

Data

Data for this study come from the first wave of the National Longitudinal Study of Adolescent Health (Add Health), which surveyed adolescents who were in grades 7–12 in 1994–1995. The survey collected extensive information on measures of adolescent health, well-being, daily activities and parental involvement, as well as school and community contextual characteristics. A resident parent questionnaire was also administered for this wave of the survey and contains additional household and parental data.

Our sample was restricted to adolescents whose biological mothers were currently in a marriage or marriage-like relationship, whose biological fathers did not live in the household and for whom information was available for independent and dependent variables. Adolescents who were younger than 15 or were married did not answer questions that were used to construct the dependent variable and are omitted from the analyses. Because of the questionnaire’s skip pattern, data are not available for adolescents who did not identify a father figure on the household roster (i.e., adolescents who did not list a father or who identified the potential father figure as a "mother’s partner" or "mother’s husband"). Of the remaining 923 eligible adolescents, several were missing data on independent and control variables. Survey-adjusted averages were substituted for missing values when more than 1% of eligible respondents were missing data on a control measure; the proportion of adolescents to which mean substitution applies is noted in the variable descriptions. Further details regarding compensation for survey design effects in Add Health appear elsewhere.⁵⁴

Measures

The dependent variable, disposition toward having sex, is a scale composed of the average of eight Likert items measuring the adolescent’s motivations to have sex. Specifically, adolescents were asked how strongly they agreed that having sex would cause their friends to respect them more, cause a partner to lose respect for them, make them feel guilty, upset their mother, give them a great deal of physical pleasure, relax them, make them more attractive to members of the opposite sex and make them feel less lonely. Items were collapsed into one scale (with the second, third and fourth items entered negatively), on which higher scores indicated a more positive disposition toward having sex (alpha=0.76).

The main independent variable, stepfather’s involvement, is a standardized composite* measure composed of 11 items—the adolescent’s perception of closeness and of the stepfather’s caring and nine items measuring whether the adolescent had participated in specific activities with his or her stepfather in the past four weeks (e.g., gone shopping, played a sport, gone to a religious service or church-related event, or talked about schoolwork or grades). Higher scores on the scale indicate greater involvement, and a score of zero equates to an average score on all items (alpha=0.71).

Controls are included for adolescent’s sex (1=male), black or Latino origin (1=yes), age (in years), average grades (1=D or lower, 4=A; mean substitution used in 4% of cases) and religiosity. The last was measured on a standardized scale of six items (having a religious affiliation, agreeing that sacred scriptures are the literal word of God, frequency of attending religious services, importance of religion to oneself, frequency of prayer, frequency of attending teenage religious activities; alpha=0.69). Higher scores on the scale indicate greater religiosity, and a score of zero equates to an average score on all items.

Two measures of the adolescent’s romantic experience are used—whether the adolescent had ever had sex and whether he or she had been in a romantic relationship at any time during the 18 months prior to the interview (1=yes).† Measures of the resident mother’s involvement (identical to the scale for father’s involvement; alpha=0.61) and mother’s educational attainment (0=less than high school, 4=more than a college degree; mean substitution used in 4% of cases) are also included. Several items drawn from the resident mother’s questionnaire are also used: her reluctance to talk about sex (five-item scale with higher scores indicating greater reluctance; alpha=0.80), the frequency with which she discussed sex in a cautionary way with the adolescent (six-item scale with higher scores indicating greater frequency; alpha=0.90), her disapproval of the adolescent’s having sex at this point in life (five-point Likert scale with higher score indicating disapproval), whether she has recommended birth control to the adolescent (five-point Likert scale with higher score indicating greater agreement that she has made such a recommendation; mean substitution used in 1% of cases) and her level of agreement that it would be acceptable for the adolescent to have sex within a relationship (five-point Likert scale with higher score indicating greater agreement). Our models also include controls for the amount of child support received in a typical month (0=none, 4=more than $500; mean substitution used in 1% of cases) and nonresident biological fathers’ involvement (similar to the involvement scales for mothers and step-fathers, but with the inclusion of frequency of overnight visits and frequency of communication; alpha=0.81). Adolescents who had no contact with their biological fathers received the minimum possible score on this scale.

Several stepfather-related controls, based on mothers’ reports, are also included. The duration (in years) of the mother’s marriage to or marriage-like relationship with the stepfather, the stepfather’s educational level (0=less than high school, 4=more than a college degree; mean substitution used in 2% of cases), and whether the stepfather is employed full-time (1=yes; mean substitution used in 12% of cases) appear in models that include stepfather involvement.

Analysis

We examined gender differences at the bivariate level, using two-tailed tests. All linear regression models are adjusted to account for sample design characteristics and include binary marker variables to indicate mean substitution. The gender-specific linear regression models also incorporate results from difference-of-slopes tests on the combined sample (i.e., within-gender interactions with all independent and control variables in the model) to indicate significant differences in effect sizes for males and females. Where results from supplemental analyses are noted, survey-adjusted linear regression was used.

RESULTS

Although males and females in the sample are very similar in many respects (e.g., the amount of child support they receive, their mothers’ levels of education, the proportion who have had sex and their levels of involvement with their stepfathers), they also differ in some ways (Table 1, page 85). For example, males view sex more positively, females get slightly better grades, and a higher proportion of females report romantic involvement. In addition, mothers approach their sons and daughters differently about sex: They are more reluctant to talk about sex with their sons than with their daughters, warn their daughters about sex more than their sons and more adamantly disapprove of their daughters’ having sex than their sons.’ Although stepfather involvement does not differ by child’s gender, biological fathers are slightly more involved with their sons than with their daughters.

Is stepfathers’ involvement associated with adolescents’ disposition toward having sex? In our initial model for the full sample, which serves as a baseline for further comparison, males are more positively disposed toward sex than females, and sexually experienced adolescents are more positively disposed than are those who have not had sex (Table 2). In contrast, the more religious adolescents are and the greater their mothers’ disapproval of adolescents’ having sex view, the less positive the view. No other variables are significant.

The model that includes stepfather involvement and stepfather-related controls supports the first hypothesis: Adolescents whose stepfathers are more involved with them are less positively disposed toward having sex. The results remain largely unchanged when stepfather-related variables are included, and the addition of these variables does little to improve model fit in the full sample.

However, the same associations may not hold for males and for females. In baseline analyses, the higher males’ grades in school, the less positive their disposition toward sex, but this association is not found among females (Table 3). Religiosity, mother’s warnings about sex and mother’s disapproval of sex are negatively associated with sexual attitudes for daughters but not for sons. (For both genders, sexual experience is positively associated with disposition toward sex.)

Gender differences persist with the inclusion of stepfather-related measures, which modestly increase the amount of variance accounted for in the males’ model but do little to improve the fit of the females’ model. Stepfather involvement and the length of time the stepfather has lived in the household are associated with a significantly less positive disposition toward sex among males, but not among females. The magnitude of this association can be better understood by predicting the amount of change in the dependent variable when the independent variable is allowed to vary while all other measures are held constant. Males who are about one standard deviation above the average in involvement with their stepfathers score about one-third of a standard deviation lower in their disposition toward sex than those whose involvement is one standard deviation below average. The difference in involvement between one standard deviation above average and one standard deviation below average equates to the adolescent’s engaging in three fewer activities per month with his stepfather, feeling a bit less close to his stepfather and thinking that his stepfather cares "somewhat" about him instead of "very much." Furthermore, the longer the stepfather has lived in the household, the less positively disposed a male adolescent is toward sex, but this association is relatively small. Other results remain similar to those in the previous model, except that males’ grades are no longer significant.

Similar results for the stepfather-related variables are not found among females, which indicates that the link between disposition toward sex and relationships with stepfathers is largely gender-specific. However, a difference-of-slopes test reveals that very few slope magnitudes differ significantly by gender. It is therefore possible that a larger sample with less variance might find significant associations for both males and females.

Supplemental analyses (not shown) were run to investigate whether particular aspects of stepfathers’ involvement (i.e., individual activities, closeness or caring) were especially strong indicators of sexual attitudes.

Analyses of the combined sample suggest that the more a stepfather cares about his stepchild, the less the adolescent’s disposition toward having sex may be, whereas the more the two attend public entertainment events together (e.g., movies or sporting events), the more positive the disposition. In separate analyses by gender, however, no specific aspect of involvement significantly predicted disposition toward sex among males, and participation in religious activity with stepfathers was associated with negative disposition among females. Thus, few specific activities are especially important predictors of the association between stepfather involvement and sexual attitudes. However, in conjunction with the prior analyses, these results suggest that the combination of these activities and the underlying level of involvement they represent is important for males. In contrast, for females overall, stepfather involvement is unrelated to attitudes toward sex, but the specific act of attending a religious service or event with their stepfather is related to their sexual attitudes. These results, then, seem to underscore the importance of religiosity for females.

DISCUSSION

Our results show that the aspects of family context that are central to adolescents’ sexual attitudes differ by gender. For males, greater stepfather involvement is associated with less motivation to have sex, but for females, the association is nonsignificant. Females’ attitudes toward sexual activity are related to religiosity and mothers’ attitudes toward teenage sex. The role of religiosity is so important for females that the only significant association between stepfather involvement and females’ sexual attitudes is for participation in religious activities with one’s stepfather.

Although we have not explicitly tested the family control, family culture and family involvement schemes against one another, the results suggest that family involvement may be more closely associated with teenagers’ sexual attitudes than family structure. There is an important exception to this assertion: One potential proxy for parental control—the duration of the stepfather’s presence in the household—was significant among male adolescents, and the coefficient estimate indicates that it had a modest association with the outcome. Nonetheless, when the findings regarding stepfather involvement are considered in context with other work,^27,28,40,41 they suggest that the association between parental authority and teenage sexuality may be determined less by family structures and the duration of such structures than by intrahousehold, interpersonal relationship dynamics that vary according to gender.

The strength of our research, therefore, lies in its focus on both gender and the nonbiological family context. That is, research often treats the stepfamily household as an issue of structure, rather than one of context. Here we do just the opposite; in so doing, we address calls for greater attention to family processes⁵⁵ while expanding upon the studies that have focused on the biological parent-child relationship.^1,8,11,12 We illustrate that context not only matters, but matters in gender-specific ways, corroborating the literature that indicates males benefit more from the presence of a stepfather than do females.^51,52

This study was limited to examining the association between stepfather-stepchild relationships and the stepchild’s attitudes toward engaging in sex. Because the influence of the stepfamily context on teenage sexuality is understudied, an examination of the effect of stepfather involvement on adolescents’ attitudes was a plausible starting place. Now that we have established a link between involvement and attitudes, future research may seek to disentangle any direct or indirect relationships among stepfather involvement, teenagers’ attitudes and a variety of sexual behaviors (e.g., initiation, number of partners, use of birth control and protection against disease transmission). Such work would build upon the findings presented here as well as those studies that have focused on the link between attitudes and behaviors.^15–19

Additional research should also conduct parallel analyses among households with stepmothers. More broadly, similar research on adolescent outcomes will benefit from investigating the relative merits of the family control, family culture and family involvement perspectives by comparing associations with involvement across a variety of family structures and cultures. More attention should be paid to variations within family type and the factors that influence females’ and males’ motivations to engage in sexual behavior.

Different factors influence males’ and females’ motivations to engage in sex, and intervention programs need to be structured with these gender differences in mind. Our results suggest that for males, programs that work toward cultivating close relationships between stepfathers and stepchildren would be beneficial in influencing attitudes that promote healthy decisions regarding sexual activity. More specifically, programs that provide stepfathers opportunities to spend time in meaningful ways and foster feelings of closeness with their stepsons are likely to have the most impact. In contrast, programs that highlight the important role that mothers play in making clear their beliefs about the acceptability of teenage sexual activity are likely to be the most successful in developing attitudes that promote healthy decision-making about sex by female adolescents. Additionally, programs that strengthen stepfamily ties and promote attendance at religious services together would benefit both males and females, but especially females.