Although most teenagers believe that 16 is too young an age at which to have sex, one in four have had intercourse by that age.¹ And among sexually experienced teenagers, the youngest are the most likely to report having had an older first sexual partner.² Sex at a young age and sex with an older partner have been linked with increased risk of unintended teenage pregnancy, childbearing and STDs.³ Furthermore, the combination of early sexual activity and having an older partner may be associated with poor reproductive health outcomes because of teenagers' undeveloped decision- making and negotiation skills and the potential for power imbalances and sexual exploitation by older partners.⁴

In the 1990s, research that found that a substantial proportion of children born to teenage mothers were fathered by adult males highlighted the issue of sex between young teenagers and older partners. According to this research, the youngest teenage mothers were the most likely to have had a much older male partner.⁵ State-level statutory rape laws prohibiting sex between minors and adults were strengthened in response to these findings.⁶ Yet, many young teenagers engage in risky sexual relationships with older partners that do not result in births.⁷ Thus, it is important to examine correlates and consequences of all teenagers'—not just teenage mothers'—early sexual relationships.

In this article, we use data from males and females interviewed in the 2002 National Survey of Family Growth (NSFG) to examine teenagers whose sexual debut occurred at a young age with an older partner. We ask the following research questions: What is the prevalence of first sex at a young age with an older partner? What family and individual characteristics are associated with having this type of relationship? Is having sex at a young age, having sex with an older partner or the combination of these behaviors associated with reduced contraceptive use and increased risk of a teenage birth? And are associations between these relationships and reproductive health outcomes due, in part, to characteristics of the teenage-partner relationship? Because partner characteristics differ by gender (and most sexual relationships involve an older male and a younger female),⁸ we conducted all analyses separately for males and females. A better understanding of the prevalence, correlates and consequences of early sexual activity with older partners will help inform teenage pregnancy and STD prevention efforts.

Age and Age Difference: Correlates and Outcomes

Extensive research has documented the associations between family and individual factors and early first sexual intercourse. Teenagers whose parents have a low income or low level of education, whose mother gave birth as a teenager, who belong to racial or ethnic minority groups and who have relatively low cognitive ability are more likely than others to have sex at an early age; teenagers who attend religious services relatively frequently are less likely than others to have early sex.⁹ In addition, females who experience menarche at an early age, those with an older dating partner and those with low educational attainment are more likely than others to engage in sex at an early age.¹⁰ The few multivariate studies of factors associated with having an older sexual partner suggest that females who grow up without either biological parent or are Roman Catholic are at increased risk of having sex with an older partner.¹¹

Early age at first sex is associated with poor reproductive health outcomes: Teenagers who have sex at an early age have reduced odds of using contraceptives and increased odds of having multiple sexual partners, and thus are at increased risk of a teenage birth or an STD.¹² In addition, having an older sexual partner is associated with reduced and inconsistent use of contraceptives,¹³ including the condom,¹⁴ and a greater risk of teenage pregnancy.¹⁵

The poor reproductive health outcomes associated with having first sex at a young age with an older partner may occur partly because these sexual experiences often are coercive or take place within casual relationships. Teenage females who have first sex at an early age and those who have an older sexual partner are more likely than other women to report that these sexual experiences were nonvoluntary and unwanted,¹⁶ and these coercive sexual experiences are associated with reduced contraceptive use and greater risk of a teenage birth.¹⁷ Also, early sexual debut and having an older partner are associated with casual relationships,¹⁸ which are themselves associated with reduced contraceptive use.¹⁹

Hypotheses

In this article, we examine four hypotheses. First, we hypothesize that both males and females have sexual experiences at a young age with older partners, although a larger proportion of females than of males do so because males are generally older than females in sexual relationships.²⁰ Second, we expect that family and individual socioeconomic disadvantage, racial or ethnic minority status and growing up without two biological parents are associated with having sex at an early age with an older partner. Our third hypothesis is that among both females and males, having early sex and having an older partner are independently associated with reduced contraceptive use and increased risk of a teenage birth, and the combination of the two factors is associated with especially poor reproductive health outcomes. Finally, we expect that young teenagers with older partners experience poor reproductive health outcomes partly because these relationships are often casual or unwanted.

METHODS

Data and Sample

We used data from the 2002 NSFG, conducted by the National Center for Health Statistics, which interviewed 7,643 females and 4,928 males aged 15–44; Hispanics, blacks and teenagers were oversampled. We restricted our sample to unmarried females and males aged 18–24 who responded to questions on their first sexual relationship before age 18 to assess the role of older partners in early sexual relationships and to provide a comparison group of teenagers who had had sex during their high school years. Individuals aged 25 or older were excluded to reduce recall bias.

Our overall sample included 1,838 females and 1,426 males aged 18–24. For analyses examining contraceptive use at first sex, we restricted our sample to the 1,113 females and 869 males who had had intercourse before age 18. To provide complete information on whether respondents had experienced a teenage birth, we further restricted the sample to 808 females and 576 males aged 20–24.

Measures

•Dependent variables. We examined three dependent variables: respondents' age at first sex and age difference with their first sexual partner, contraceptive use at first sex and having or fathering a birth before age 20. Because we classified teenagers across first sexual relationships, we created discrete categories of age and age difference with partners. We based our decisions on state-level statutory rape laws, which determine when it is illegal to have sex with a minor, often on the basis of the age of the teenager and the age of the partner. We chose to examine female and male teenagers who had had sexual intercourse before or after age 16—the modal age of consent at which teenagers are legally allowed to have sexual intercourse, regardless of their partner's age.*²¹

For the first dependent variable, we divided the sample of females into six categories: those who had had sex at an early age (before age 16) with a partner five or more years older, with a partner 3–4 years older and with a partner of similar age (less than three years older); those who had had sex at age 16–17 with a partner three or more years older^† and with a partner of similar age; and those who had not had sex before age 18. Because of small sample sizes, we divided males into four categories: those who had had sex before age 16 with a partner three or more years older, those who had done so with a partner of similar age, those who had had sex at age 16–17 and those who had not had sex before age 18.^‡ We examined two categories of older partners: those 3–4 years older and those five or more years older. Because males were less likely than females to report older partners, we combined the categories for males.

•Family and individual characteristics. Family background factors included were mother's education (less than high school, high school, some college and bachelor's degree or higher), family structure at age 14 (two biological or adoptive parents, one biological and one adoptive parent or stepparent, single biological parent and other) and a dichotomous measure of whether the respondent's mother had had her first birth as a teenager. Individual characteristics included were race and ethnicity,^§ immigrant status, early menarche (before age 12) and religious attendance at age 14 (never, less than once a month, 1–3 times per month, once a week and more than once a week).

•First sexual relationship. We included measures of voluntary status and wantedness of respondents' first sexual experience, as well as a measure of type of relationship with partner at first sex. To address voluntary status, females (but not males) were asked "Would you say that this first vaginal intercourse was voluntary or not voluntary, that is, did you choose to have sex of your own free will or not?" To address wantedness of first sex, females and males were asked "Think back to the very first time you had vaginal intercourse with a [male/female]. Which would you say comes closest to describing how much you wanted that first vaginal intercourse to happen?" The three possible responses were "I really didn't want it to happen at the time" (categorized as not wanted), "I had mixed feelings—part of me wanted it to happen at the time and part of me didn't" (categorized as mixed) and "I really wanted it to happen at the time" (categorized as wanted). Relationship type at first sex was categorized as cohabiting or engaged; going steady; or going out once in a while, just friends, just met or other. Because of small sample sizes, the second and third categories were combined for males.

Analytic Methods

We conducted bivariate chi-square analyses to assess the associations between the selected characteristics and different combinations of age at first sex and partner age difference at first sex. Next, we conducted multinomial logistic analyses to test the associations between the selected characteristics and having first sex at an early age with an older partner, rather than having first sex at age 16–17. We then conducted logistic regression analyses, including sexual history measures, to examine associations of age at first sex and partners' age difference at first sex with contraceptive use at first sex and involvement in a teenage birth.

Finally, we conducted contrast analyses to test whether the combination of early first sex and an older partner was associated with poor health outcomes. Thus, we tested whether early first sex with an older partner was associated with lower odds of contraceptive use at first sex and higher odds of a teenage birth than were early first sex with a similarly aged partner, first sex at age 16–17 with a similarly aged partner and first sex at age 16–17 with an older partner. Contrast estimates were obtained by repeating analyses and changing the reference category of the variable. Contrasts for females were based on separate analyses with a six-level independent variable measuring age at first sex and partners' age difference; contrasts for males were based on separate analyses with a similar three-level independent variable.

All analyses were weighted, run separately by gender and conducted in Stata to control for the complex sampling design of the NSFG.²²

RESULTS

The mean ages of female and male respondents were 21.1 and 20.8 years, respectively. Of females, 63% were white, 15% black, 16% Hispanic and 6% of other racial or ethnic groups; the proportions among males were 62%, 13%, 19% and 6%. Fifty-two percent of respondents reported that their mother had at least some college, and 17% said that she had not graduated from high school.

Bivariate Analyses

•Females. More than half (57%) of women reported having had sexual intercourse before age 18 (Table 1, page 199). Fourteen percent had had sex before age 16 with an older partner (5% with a partner five or more years older, and 9% with a partner 3–4 years older), and 13% had had sex before age 16 with a partner of similar age. The remaining 30% had had sex at age 16–17 (7% with a partner three or more years older, and 23% with a partner of similar age). Among females who had had sex before age 16, 75% of older male partners had been teenagers, including all of those 3–4 years older and 35% of those five or more years older (not shown); in comparison, 81% of the partners of women who had first sex at age 16–17 had been in their teens.

In general, women who had had early sex with an older partner had the most disadvantaged family backgrounds, whereas those who had not had sex before age 18 had the most advantaged backgrounds. For example, 60% of females not sexually experienced at age 18 reported that their mothers had attended or graduated from college, compared with 36–38% of those who had had sex before age 16 with an older partner and 45–49% of other sexually experienced females. Also, women who had had early sex with an older partner were the group with the lowest proportion residing with both biological parents at age 14 (51–53%), and women who had had early sex with a partner at least five years older had the highest proportion whose mother had had a teenage birth (53%). In contrast, females who had not had sex before age 18 were the group with the highest level of religious service attendance and the lowest proportion reporting early menarche.

Among sexually experienced females, greater proportions of those who had had sex before age 16 with an older partner than of others reported their first sex had been nonvoluntary (16–25% vs. 7–11%), unwanted (20–36% vs. 9–17%) and with a casual partner (35–46% vs. 16–26%). A smaller proportion of females who had had early sex with a partner five or more years older than of others had used contraceptives at first sex (45% vs. 62–74%), and a greater proportion of females who had had sex before age 16 with a partner three or more years older than of others had had a teenage birth (41–46% vs. 19–34%).

•Males. Fifty-eight percent of males had had first sex before age 18 (Table 2). Six percent reported that their sexual debut had occurred before age 16 with a partner three or more years older, 22% before age 16 with a partner less than three years older and 31% at age 16–17; about two-thirds of those who had had sex before age 16 reported their first sex had been wanted. Males represented 27% of all teenagers who reported a first sexual experience before age 16 with a partner three or more years older (not shown). Among males who had had early sex with a partner three or more years older, 74% had had teenage partners; the proportion among males aged 16–17 at first sex was 93%.

The proportions of males who had had first sex before age 16, at age 16–17 and later were similar to those of females; however, a smaller proportion of males than of females reported early sex with a partner three or more years older (6% vs. 14%). A smaller proportion of males who had had sex before age 16 with an older partner than of others reported that their mother had attended college (34% vs. 46–56%), said that they had lived with both biological parents at age 14 (54% vs. 61–81%), were white (38% vs. 49–71%) and had gone steady with or been engaged to their first sexual partner (15% vs. 53–59%). A greater proportion of those who had had sex before age 16 with an older partner than of others were Hispanic (40% vs. 12–23%) and had a mother who had had a teenage birth (50% vs. 21–35%). Furthermore, regardless of age difference with their partner, a smaller proportion of males who had had sex at an early age than of others reported having used a contraceptive at first sex (61–67% vs. 77%), and greater proportions of males who had had sex at an early age than of others had fathered a child as a teenager (16–19% vs. 6%).

Multivariate Analyses

•Early sex with an older partner. Among females, those who had lived with neither biological parent at age 14 and those who had experienced early menarche had elevated odds of having had first sex before age 16 with a partner 3–4 years older (odds ratios, 3.1 and 1.9, respectively) or five or more years older (3.5 and 2.7, respectively), relative to those who had had first sex at age 16–17 (Table 3). In addition, being Hispanic was associated with increased odds of having had sex before age 16 with a partner five or more years older (2.1), whereas a one-step increase in religious attendance at age 14 (e.g., from never to less than once a month) was associated with a 20% reduction in the odds of having had early sex with a partner 3–4 years older (0.8).**

Several family and individual characteristics were associated with women's remaining abstinent until age 18. Having lived with one biological parent and one stepparent or adoptive parent and being black were associated with reduced odds of not having had sex by age 18 (odds ratios, 0.4 and 0.6, respectively); having a mother who had had her first birth as a teenager was marginally significant. Being foreign-born and more frequent religious attendance at age 14 were associated with increased odds of having been abstinent until age 18 (2.2 and 1.2, respectively).

Among males, those who had lived with neither biological parent at age 14 and Hispanics had significantly elevated odds of having had first sex before age 16 with a partner three or more years older (odds ratios, 3.7 and 2.4, respectively); having lived with a single biological parent and being black were associated with marginally elevated odds of this outcome.^†† Males who had lived with one biological parent and one adoptive parent or stepparent, Hispanics and blacks had reduced odds of postponing sex until age 18 (0.5–0.6), whereas those who had attended religious services more frequently had marginally increased odds of remaining abstinent (1.1).

•Contraceptive use at first sex and teenage childbearing. In multivariate analyses controlling for family and individual characteristics, having been younger than 16 at first sex, having had a first partner five or more years older, being Hispanic and being foreign-born were associated with reduced odds of females' having used contraceptives at first sex (odds ratios, 0.4–0.7—Table 4). After sexual history characteristics were added to the model, age difference between partners, being Hispanic and being foreign-born remained negatively associated with contraceptive use (0.4–0.6); age at first sex became nonsignificant. In addition, women whose first sexual experience had been unwanted and those whose first relationship had been casual had reduced odds of contraceptive use (0.6 and 0.7, respectively), although the latter finding was only marginally significant.

Women who had first had sex before age 16 had greater odds than those who had done so at age 16–17 of having a teenage birth (odds ratio, 1.6); having a partner who was 3–4 years older was marginally associated with an increased risk of teenage birth (1.6). Having a mother who had had a teenage birth and being Hispanic or black were positively associated with having a teenage birth (1.8–2.2), whereas a one-step increase in maternal education (e.g., from high school to some college) was associated with a 40% reduction in the odds of having had a teenage birth (0.6). After characteristics of first sexual experience were added to the model, age at first sex remained positively associated with teenage birth (1.7); the association between age difference at first sex and teenage birth became nonsignificant. One first sexual relationship characteristic was associated with teenage birth: Women who had been cohabiting with or engaged to their first partner had nearly three times the odds of those who had been going steady with their first partner of having a teenage birth (2.9).

Among males, being younger than 16 at first sex was marginally associated with decreased odds of contraceptive use (odds ratio, 0.7—Table 5). Partners' age difference at first sex was not significantly associated with the outcome. Being black was positively associated with contraceptive use at first sex (1.7), and having lived with neither biological parent at age 14 was negatively associated with contraceptive use (0.4). When first sexual relationship characteristics were added to the model, the association between age at first sex and contraceptive use remained marginally significant (0.7), and the associations of black race and family structure with the outcome remained significant (1.8 and 0.5, respectively). In addition, men who had had a casual first sexual relationship had reduced odds of having practiced contraception at first sex (0.7)

Compared with males who had had sex at age 16–17, those who had had sex before age 16 had greater odds of having fathered a child as a teenager (odds ratio, 2.9); Hispanics, those whose mothers had had a teenage birth and those who had lived with neither biological parent at age 14 also had increased odds of this outcome (2.6–4.0). When first sexual relationship characteristics were added to the model, all the significant associations from the first model remained significant. In addition, males whose first sexual experience had been unwanted had more than four times the odds of those who had mixed feelings of having fathered a child as a teenager (4.3).

•Contrasts of associations with age and partner age difference. In contrast analyses controlling for family and individual characteristics, females who had had first sex before age 16 with a partner five or more years older had lower odds of practicing contraception at first sex compared with those who had had first sex before age 16 with a similarly aged partner or at age 16–17 with a similarly aged partner (odds ratios, 0.4 and 0.5, respectively—Table 6, page 204). Women who had had first sex before age 16 with a partner 3–4 years older had marginally higher odds of contraceptive use than those who had had sex before age 16 but with a similarly aged partner (1.5). When we added first sexual relationship characteristics to the model, the difference in contraceptive use between women who had early sex with a partner five or more years older and those who had early sex with a similarly aged partner remained significant (0.4), whereas the difference with those who had had sex at age 16–17 with a similarly aged partner became marginally significant. The marginal difference in contraceptive use between women who had early sex with a partner 3–4 years older and those who had early sex with a similarly aged partner became significant (1.6).

In analyses of teenage births among females, those who had had early sex with a partner five or more years older did not differ significantly from any of the comparison categories. Women who had had early sex with a partner 3–4 years older had increased odds of teenage birth compared with those who first had had sex at age 16–17 with a similarly aged partner (odds ratios, 3.0) or a partner 3–4 years older (2.5). In addition, women who had had early sex with a partner 3–4 years older had marginally higher odds of teenage birth than those who had had early sex with a similarly aged partner (1.7). When we added first sexual relationship characteristics to the model, the marginal difference in teenage birth between those who had had early sex with a partner 3–4 years older and those who had had early sex with a similarly aged partner became nonsignificant. All other significant findings remained significant.

Among males, the only significant comparison was that those who had had sex before age 16 with a partner three or more years older had marginally higher odds of having fathered a child as a teenager than those who had delayed first sex until ages 16–17 (odds ratio, 2.6). This association became significant when first sexual relationship characteristics were added to the model.

DISCUSSION

This article provides an exploratory view of first sexual relationships between young male and female teenagers and older partners. We found that among 18–24-year-olds, 14% of females and 6% of males had had their first sexual relationship before age 16 with a partner at least three years older, and that 5% of females had had first sex before age 16 with a partner five or more years older. Therefore, on the basis of our weighted estimates, almost two million females and 782,000 males in this cohort have an increased risk of poor reproductive health outcomes.

Prevalence and Partner Information

These are the first available nationally representative analyses that provide gender-specific information on sexual relationships between young teenagers and older partners. As hypothesized, and consistent with age differences in relationships among adults,²³ a smaller proportion of males than of females reported having sex at a young age with an older partner. Yet, more than one-quarter of sexual experiences occurring before age 16 with a partner three or more years older were among males. This proportion is higher than estimates from the National Center for Juvenile Justice (NCJJ), which found that only 5% of statutory rape reports to law enforcement were made by males.²⁴

One primary difference between the studies is that the NSFG estimates were classified solely by the ages of the teenagers and their sexual partners, whereas the NCJJ estimates were based on subjective decisions by teenagers and their families to report cases to law enforcement. Also, the NSFG estimates show that the majority of young males with older partners classified their sexual experience as wanted, which may be associated with less reporting to law enforcement than occurs among females, who were more likely to classify these sexual relationships as unwanted and nonvoluntary. Lower reporting of statutory rape and higher wantedness of these sexual relationships among males may reflect cultural norms or peer encouragement for males to initiate sexual experiences. Alternatively, greater reporting among females may be because of larger age differences with older male partners, which may be more likely to qualify as statutory rape under state laws.²⁵

Family and Individual Correlates

As hypothesized, family and individual factors were associated with having had sex at a young age with an older partner. Hispanic males and females, black males and those who lived at age 14 with neither biological parent had increased odds of having had sex at a young age with an older partner. The finding on family structure suggests that teenagers with the fewest family supports have the highest odds of engaging in risky sexual relationships, possibly because of limited parental monitoring or because teenagers from disadvantaged families tend to have more problem behaviors and riskier peers than others.²⁶ The association between Hispanic ethnicity and early sex with an older partner confirms prior research suggesting that Hispanics—especially those with greater ties to their families' countries of origin—may have different cultural norms than whites about appropriate age differences between dating and sexual partners.²⁷ Also, we found that early menarche was associated with women's elevated odds of having sex at a young age with an older partner, whereas greater religious attendance was associated with reduced odds. The result concerning menarche supports previous findings that females with an early age of development are more likely than others to date older partners²⁸ and to engage in sex at an early age.²⁹ The finding on religious attendance supports research linking greater attendance to reduced sexual activity.³⁰

Reproductive Health Outcomes

As hypothesized, early age at first sex and a large age difference with a first partner were independently associated with the odds of contraceptive use and a teenage birth among females; however, only having had a much older partner was associated with reduced contraceptive use, whereas having had a partner 3–4 years older was associated with marginally increased odds of a teenage birth. The result concerning contraceptive use complements other research and suggests that females may have reduced negotiating power in their relationships with older males.³¹ One study supports these findings by showing that female teenagers fear more negative reactions from older partners than from similarly aged partners if they bring up using a condom.³²

In addition, the combination of young age at first sex and having had an older partner was associated with especially high odds of teenage birth. Specifically, females who had had early sex with a partner 3–4 years older had greater odds of having a teenage birth than those who had had early sex with a similarly aged partner and those who had had sex at age 16–17. These findings correspond with estimates suggesting that teenage births occur disproportionately among young women who have had older partners;³³ such teenage births, however, may not involve women's first sexual partners.

Females who had had sex before age 16 with a partner five or more years older had lower odds of contraceptive use at first sex than those who had had first sex before age 16 or at age 16–17 with a similarly aged partner, but not lower odds than those who had had sex at age 16–17 with an older partner. Thus, we cannot say that those who had had early sex with an older partner had especially low odds of contraceptive use, although both age and age difference were independently associated with this outcome. Interestingly, females with partners 3–4 years older had greater odds of using contraceptives than those with similarly aged partners. Although we do not know who initiated contraceptive use in these relationships, these findings suggest that some older male teenagers (in this case, still high school–age) may be more motivated than younger males to use contraceptives. Alternatively, females may be more comfortable negotiating positive contraceptive outcomes with males who are 3–4 years their senior than with those who are even older.

Among males, early age at first sex, but not age difference, was associated with reduced contraceptive use and elevated odds of fathering a child as a teenager, suggesting that no additional barriers to contraceptive use are associated with having an older sexual partner in this sample. This finding contrasts with research that has shown reduced contraceptive use among males with older partners.³⁴ Also, males who had had sex before age 16 with an older partner had more than twice the odds of fathering a child during their teenage years compared with those who had delayed sex until age 16–17. This finding suggests that delaying sexual activity among males will help reduce early fatherhood, and thus increase the likelihood that these fathers can support their families in a marital relationship.³⁵

Unwanted and Casual Sex

We hypothesized that unwanted sexual experiences and more casual relationships may help explain the association between early sex with an older partner and risky reproductive health outcomes. According to our bivariate analyses, greater proportions of males and females who had had first sex at a young age with an older partner than of others categorized their first sexual relationship as casual, and a greater proportion of those females than of others reported that their first sexual relationship had been unwanted. In fact, one in four first sexual relationships between females younger than 16 and partners five or more years older were classified as nonvoluntary. Thus, the majority of females reported that their sexual relationships at a young age with an older partner had been voluntary, and more than half had considered these steady relationships. These findings highlight the differences between rape and statutory rape, which may be prosecuted even if both parties consider the sexual relationship to be voluntary.³⁶

Consistent with other research, our multivariate analyses indicated that sex in a casual relationship was associated with reduced contraceptive use at first sex for males and females.³⁷ Unwanted first sexual experiences were associated with reduced odds of contraceptive use among females and greater odds of involvement in a teenage birth among males; these findings add to the literature on the associations between poor reproductive health outcomes and coercive relationships.³⁸ And for females, the association between having sex before age 16 and low contraceptive use was attenuated after we included relationship type and wantedness, providing some support for our hypothesis.

Limitations

This study has several limitations. First, because states' statutory rape laws differ, we were unable to create a national measure of sexual relationships between young teenagers and older sexual partners. Second, ideally, we would have classified teenagers by age and partner's age difference for all of teenagers' sexual relationships, rather than only for their first. Third, we relied on retrospective accounts of age at first sex and partner's age, which may have been biased. Fourth, because of sample size issues, we were not able to test whether very young teenagers (such as those younger than 14) with an older first partner experience especially poor reproductive health outcomes. Finally, because of small sample sizes for multinomial outcomes, our reference group for analyses of characteristics associated with sex at a young age with an older partner combined all teenagers who delayed sex until age 16–17, regardless of partners' age; ideally, we would have used separate comparison categories for similarly aged and older partners. Despite these limitations, our analyses provide a new examination of sex between young teenagers and older partners and the link between these relationships and poor reproductive health outcomes.

Implications

Our findings highlight the importance of preventing early sex, regardless of partners' age difference, and of preventing sex with an older partner, especially for females. An emerging body of research suggests a role for parents, service providers and communities in helping define norms of appropriate age differences between teenagers and their dating and sexual partners; however, no consensus exists on what age is too young for sex and how large an age difference is too large for sexual partners.³⁹ And a qualitative study of teenage mothers' accounts of their relationships with much older partners suggests that some teenagers may perceive that their families, peers and communities consider older partners to be normal.⁴⁰ According to the extensive research literature, parents may have an influence on adolescent sexual behaviors through communication and monitoring of their teenagers' activities.⁴¹ Our research suggests that parents should monitor teenagers' partners and intervene if a partner is much older or younger. Also, our findings highlight the importance of culturally appropriate interventions for Hispanic and black families and communities to address potential risks associated with sex at a young age and with older partners.⁴²

Effective pregnancy prevention programs incorporate role-playing exercises to help teenagers negotiate decisions about whether to have sex and practice contraception with sexual partners.⁴³ Our research suggests that it may be appropriate to incorporate exercises on how to deal with the pressures for risky sex from an older, higher status partner. And in light of our finding that many teenagers engage in sex with older teenagers, prevention efforts should attempt to reach teenagers and adults, to educate them about the legal and ethical implications of age differences within sexual relationships.