Sexual intercourse during adolescence is a significant and complex event. It leads to a radically new understanding of one's identity as a physically developing person, and it is a major psychological milestone on the road to adulthood.1

Adolescent sexual activity in the United States is associated with a broad range of health and social problems.2 Sexual activity places youths at risk for STDs3 and pregnancies—every year about three million cases of STDs are reported among 10–19-year-olds,4 and 870,000 pregnancies occur among 15–19-year-old women.5 Early sexual activity is associated with alcohol and drug use, intimate partner violence, pregnancy and inconsistent condom use, as well as multiple sex partners.6 STDs and adolescent pregnancies place an economic burden on the youth involved and on society.7 Therefore, it is critical to understand the factors that are associated with the first sexual intercourse of young adolescents.

There is a dearth of information about Asian American adolescents' sexual behavior, even though Asian Americans, whose number grew by 52% between 1990 and 2000, are the fastest growing ethnic minority in the United States.8 Most studies that have included Asian American youth have used only small samples, often convenience samples, and have been cross-sectional.9 In addition, many have focused exclusively on college students.10

The cultural heritage of Asian Americans makes the systematic study of sexual activity and attitudes about sexuality complex, because in Asian cultures, there is a sharp distinction between public and private selves, and sexuality is kept within the realm of the private self. Unmarried people, especially women, are expected to abstain from expressing sexual desire because they may bring shame or dishonor to the family. This sexual conservatism is viewed as vital to the maintenance of family unity.11

Although Asian American adolescents have lower rates of sexual experience than white, black and Latino adolescents,12 the overall rate is almost 30% by age 18. However, the stereotype of Asian Americans as a "model minority" does not hold for adolescents in the area of sexual activity, particularly with HIV-related risk behaviors.13 A nationally representative survey and studies in southern California have found that once Asian American adolescents and young adults engage in sexual intercourse, they are similar to other ethnic groups in terms of their number of lifetime partners and sexual behaviors that may transmit HIV.14 Furthermore, young adults and adolescents of Asian descent who were surveyed in British Columbia and Massachusetts had less knowledge about the risk of HIV and other STDs than respondents of other ethnicities.15 Among Asian American students in several southern California colleges, only 11% reported consistent condom use, with no significant differences between men and women.16 In a national study of HIV-positive adults, Asian Americans had the highest rate of late intervention.17

Many factors are associated with adolescents' initiation of sexual intercourse: age, ethnicity, gender, parental attachment, substance use and socioeconomic status.18 A higher level of acculturation also is associated with higher rates of risky sexual activities among ethnic minority youth.19 Acculturation is the complex psychological process of adaptation to a different culture, by which members of an ethnic group gradually change their behaviors and attitudes to be more like those of the host society. High acculturation suggests adopting the values, language and beliefs of the new culture, while low acculturation suggests retaining the values, languages and beliefs of the original culture.20 A critical aspect of acculturation is renegotiating or redefining gender roles in the new cultural context. Gender role acculturation occurs when the host culture's gender role norms influence the individual's perceptions of masculinity, femininity and sexuality.21

Asian American adolescents, particularly women, often live in a bicultural world where they experience contradictions between their heritage and American culture. They observe American notions of gender roles through interactions with peers and at schools, and Eastern notions of gender roles through family and ethnic community socialization. Highly acculturated Asian American females may adopt gender roles shaped by American values, which emphasize an egalitarian distribution of social status, power and sexual freedom. However, less acculturated young women may internalize traditional gender-specific norms, which are characterized by passivity and submissiveness during interactions with men.22 For Asian men, sexual behavior may not be strongly influenced by level of acculturation, because both cultures value young men's independence, freedom and sexual accomplishment.23

Associations between parental communication and delayed adolescent sexual activity have been clearly established.24 Discussion with parents about sexuality, romantic relationships, HIV and AIDS, and strategies to reduce risky behaviors are linked to reductions in sexual activity, pregnancies and unprotected intercourse.25 However, because of gaps in language and culture, the advantages of parental communication may not be available to highly acculturated adolescents whose parents are immigrants.26 Highly acculturated adolescents may interpret a lack of communication as a lack of parental interest and may therefore feel that their peers are their only sources of information and emotional support.27

Sexual identity develops differently for young men and women.28 In the overall U.S. population, men tend to have their first sexual experience at a younger age and hold more permissive attitudes than women.29 However, findings from studies of Asian American youth suggest some differences. In a study of Chinese American college students, a higher proportion of women than of men had experienced sexual intercourse by age 21.30 However, in a study of Asian college students in Canada,31 men had higher rates of sexual experience. In other samples of Asian American adolescents32 and young adults,33 no significant gender differences were found in the prevalence of sexual intercourse.

In this study, we investigate the association between levels of acculturation and rates of sexual experience for Asian American adolescents, using data from the National Longitudinal Study of Adolescent Health (Add Health). Drawing from the literature, we hypothesize that factors associated with sexual experience will differ by gender. More specifically, we expect that when compared with young women who are less acculturated, young women who are highly acculturated will have higher odds of having had sexual intercourse. We do not expect that acculturation level is associated with sexual experience for the young men.

METHODS

Data

Add Health is a longitudinal study of a nationally representative sample of youth who were in grades 7–12 in 1995.34 The sample was selected from the enrollment rosters of 132 schools, derived from a sampling frame that comprised all school districts in the United States with a high school. To ensure a representative sample of adolescents, the Add Health study designers stratified schools by region, urbanicity, type (public, parochial, private) and racial composition. At Wave 1 (April–December 1995), roughly 50,000 adolescents completed in-school questionnaires, and 20,745 of them participated in confidential in-home interviews; more than 14,700 students were reinterviewed at home for Wave 2 (January–December 1996). The response rate for Wave 1 was 79%; for Wave 2 it was 88%.

The present study included the 1,048 Asian American respondents who participated in the in-school interview and in both the Wave 1 and the Wave 2 in-home surveys. However, the final sample was reduced to 689 participants (323 female, 366 male) because of missing values. On average, individuals for whom complete data were available were half a year younger, and had slightly higher socio- economic status and self-esteem, than those with missing data. There was no difference in the proportion of men and women for whom data were missing.

Variables

Participants were divided into four groups according to level of acculturation: U.S.-born adolescents who reported using English at home, foreign-born adolescents who reported using English at home, U.S.-born adolescents who reported exclusive use of another language at home and foreign-born adolescents who reported exclusive use of another language at home. Research on Latinos has demonstrated that these groups lie on a continuum from the most acculturated to the least.35 Since language may be a stronger component of acculturation than place of birth, those who are not U.S.-born but use English at home are considered more acculturated than those who are U.S.-born but do not use English at home.36 Because the differences between the ordered groups are not equal, we analyzed the data categorically, with the least acculturated group as the reference group.

All covariates were measured at the time of the Wave 1 in-home interview. Age was measured in years. Gender was coded 0 for male and 1 for female. Family socioeconomic status was determined by combining parental educational attainment and occupational status. Parental educational attainment was coded on a scale of 1–6 (eighth grade or less; some high school, or business, trade or vocational school; complete high school or general equivalency diploma; some postsecondary; four-year college or university graduate; and postgraduate professional training). Occupations were coded on a scale of 1–5, ranging from blue-collar (low) to white-collar professionals (high): construction workers, mechanics, factory workers, laborers, bus drivers, security officers, farmers or fishermen; office workers or salespeople; technical workers, computer specialists, teachers, librarians or nurses; managers, including executives or directors; and doctors, lawyers or scientists. The scores for education and occupation were summed; for respondents reporting information on two parents, the higher score was used (possible range, 2–11; alpha, 0.73). Observed values ranged from 2 to 11; the mean socioeconomic status score was 6.93 (standard deviation, 2.4). Similar measures have been used by Ford et al. 37 and Guterman et al.38

Parental attachment was assessed by 13 items that focused on the participant's perception of his or her relationship with each parent. For example, the adolescents were asked if they are satisfied with their relationships and communication with their mothers and with their fathers, if their families understand them and if their parents care about them. Responses were on a scale of 1–5, ranging from "not at all" to "very much" (possible range, 13–65; alpha, 0.90). The observed values ranged from 21 to 65 (mean, 55.1; standard deviation, 7.5).

The self-esteem score was determined by the responses to 11 items: "I have a lot of energy"; "I seldom get sick"; "when I do get sick, I get better quickly"; "I am well coordinated"; "I have a lot of good qualities"; "I am physically fit"; "I have a lot to be proud of"; "I like myself just the way I am"; "I feel like I am doing everything just about right"; "I feel socially accepted"; and "I feel loved and wanted." Responses were on a scale of 1–5, ranging from "never" to "every day" (possible range, 11–55; alpha, 0.84). The observed values were 22–55, and the mean self-esteem score was 43.3 (standard deviation, 5.9).

The school attachment scale was the sum of the following six items: "I feel close to people at my school"; "I feel like I am a part of my school"; "students at my school are prejudiced"; "I am happy to be at my school"; "the teachers at my school treat students fairly"; and "I feel safe in my school." Responses ranged from "strongly disagree" to "strongly agree," on a scale of 1–5 (possible range, 6–30; alpha, 0.70). The observed values ranged from 9 to 29; the mean school attachment score was 21.8 (standard deviation, 3.2).

For grade point average, the average of the grades in English, math, history and science from the most recent grade period was used. The possible values on this scale were 1–4 (alpha, 0.73); the observed values ranged from 1 to 4, and the mean grade point average was 3.1 (standard deviation, 0.75).

After the socioeconomic status, parental attachment, school attachment, self-esteem and grade point average scales were created, they were recoded into categorical variables. Values more than one standard deviation below the average were coded as low, within one standard deviation of the average as medium and more than one standard deviation above the average as high.

Binge drinking and tobacco use were coded as dichotomous variables. Those who reported one or more days in the last 12 months on which they drank five or more drinks in a row were coded as binge drinkers. Those who reported ever smoking cigarettes in the past 30 days were coded as tobacco users.

The question "Have you ever had sexual intercourse?" from the Wave 2 interview was used as the outcome variable. A response of yes was coded as 1; a response of no was coded as 0.

Statistical Analysis

The distribution of specific Asian ethnic groups among respondents for whom complete data were available differed from the distribution among the U.S. population. Therefore, specific weights were estimated and applied to the entire analysis, to adjust the population estimates to match the population ethnic distribution. Data from the 2000 census were used for Asian subgroups, and the proportion of each group aged 5–24. Stata 8.0 was used for all analyses.

In bivariate analyses of relationships between selected variables and rates of sexual experience, the usual chi-square test statistic was adjusted to account for the weighting; the adjusted test statistics have an approximate F distribution. Explanatory variables that had a p value less than .10 in the bivariate analyses were included in the multiple logistic regression models to assess the association between acculturation and sexual intercourse for young women and young men. The standard errors and test statistics in the logistic regression are also adjusted to account for the weighting. Chi-square testing was performed to assess differences between females and males in rates of sexual intercourse.

RESULTS

At Wave 1, the mean age of all Asian American adolescents included in this study was 16.1 years (standard deviation, 1.5). The sample included adolescents from different Asian ethnic backgrounds (Table 1). Similar proportions of young women were Chinese and Filipina, and the highest proportion of young men were Chinese. The vast majority (96%) of adolescents had parents who were both of an Asian ethnicity (not shown). Those who spoke English at home and were U.S.-born represented the highest proportion of both females and males (46% of each). The majority of women (89%) and men (90%) indicated that they had either medium or high parental attachment. Most women and men reported medium or high school attachment (91% and 83%, respectively). Thirteen percent of women and 15% of men reported binge drinking in the past year; 11% of females and 15% of males reported tobacco use.

A total of 24% of the young women and 20% of the young men reported in Wave 2 that they had had sexual intercourse (Table 2, page 31). In the bivariate analyses, foreign-born adolescents who spoke English at home had the highest rates of sexual intercourse for both women and men (37% and 34%, respectively). However, for the most acculturated group, U.S.-born adolescents who spoke English at home, a much higher proportion of women (31%) than of men (18%) reported having had intercourse (p=.01—not shown).

For both males and females, a higher proportion of older adolescents than of younger adolescents had had sexual intercourse.

More than half of young women with low parental attachment had had sexual intercourse by Wave 2. By contrast, among those with high parental attachment, fewer than one in 10 reported having had intercourse. There is a stunning gender difference in results for young people with low parental attachment—57% of such women had had sexual intercourse, compared with 28% of men (p=.01—not shown).

Among young women who reported low school attachment in Wave 1, more than half had had sexual intercourse by Wave 2. The proportion of young women who had had sexual intercourse varied by level of school attachment and was less than one-fifth at the highest level. However, differences in rates of sexual intercourse according to school attachment were not statistically significant for young men.

Among women, 72% who reported binge drinking in the past 12 months had had sexual intercourse, compared with 17% of those who did not. Among men, 39% who reported binge drinking had had sexual intercourse, compared with 16% who did not. For the young women, 60% of those who reported tobacco use had had sexual intercourse, compared with 20% of those who did not. For the young men, 47% of those who reported tobacco use had had sexual intercourse, compared with 14% who did not.

Differences in rates of sexual experience by ethnicity, socioeconomic status, grade point average and self-esteem were not statistically significant for either women or men.

The multiple logistic regression analyses adjusted for age, parental attachment, socioeconomic status, self-esteem, school attachment and substance use (Table 3). For the women, those who spoke English at home and were either U.S.-born or foreign-born had significantly higher odds of having had sexual intercourse by Wave 2 than those who did not speak English at home and were foreign-born (odds ratios, 4.9 and 4.3, respectively). For the men, there is no evidence of association between acculturation and sexual experience.

Other factors that were associated with significantly increased odds of sexual experience for women were older age (odds ratio, 1.9) and binge drinking in the past 12 months (6.4). Factors associated with decreased odds of having had sexual intercourse for young women were medium socioeconomic status, compared with low (odds ratio, 0.4); high and medium parental attachment (0.2 and 0.4, respectively); and high and medium school attachment (0.3 and 0.4, respectively).

For men, only older age and tobacco use were associated with increased odds of sexual experience (odds ratios, 1.7 and 3.0, respectively). In contrast to the multivariate results for women, sexual intercourse was not associated with parental attachment or school attachment.

DISCUSSION

Overall, the weighted population estimates for the proportion of Asian American adolescents who were sexually experienced were 24% for females and 20% for males. By contrast, a national longitudinal study by Grunbaum et al. showed a 28% prevalence rate,39 and a California study by Schuster et al. showed a 27% prevalence rate.40 Our results may differ because we used weights calculated specifically for Asians rather than weights calculated for all ethnic groups, as in Grunbaum et al., or no weights, as in Schuster et al.

Acculturation was not associated with sexual intercourse for young men. For young women, however, high levels of acculturation were associated with significantly increased odds of sexual experience. In other words, the influence of American cultural values on young Asian women's sexuality is greater than their influence on the sexuality of young Asian men.

Why is level of acculturation associated with elevated odds of sexual intercourse for young Asian American women? From the sociocultural perspective, there are three possible explanations. First, Asian cultures convey different expectations about independence and sexual activity for young men and young women. Sexual activity is accepted and even subtly encouraged for young men.41 As young women acculturate, they may perceive sex as an assertion of independence and gender equality.

In addition, the more acculturated young women become, the less importance they may place on the cultural values of their families. For instance, in a comparison of university students, Chinese Canadian women held less traditional views of gender and family than did Chinese Canadian men. Also, a greater proportion of women than of men reported that their views differed from those of their parents.42

In a study of Asian American college students, Cochran and colleagues found that the most important reason for refraining from sexual activity for women was the maintenance of cultural, family and religious values and harmony. In contrast, the main reasons given by men were partner's pregnancy, AIDS and the lack of opportunity.43

Second, differences in dating partners may help explain gender differences in sexual experience. Asian American adolescent women may be dating older men.44 If young Asian American women are dating and having sex with non-Asian young men, their dating experiences are part of the acculturation process as well as a consequence. Young Asian American women who are more acculturated are more likely to adopt American attitudes about sex and are more likely to go out with whites.45 Also, young Asian American women may become more acculturated by dating young white men, and consequently adopt American attitudes toward sex.

These assertions are supported by studies of Asian American adults. A higher proportion of Chinese American women who date whites than of those who date Chinese men are sexually active.46 When compared with whites, blacks and Latinos, Asian American adult women have the highest rate of interethnic dating and marriage, particularly with white partners.47 While women of all ethnic groups have higher rates of interethnic dating than men, the difference is greatest for Asian Americans.48

The findings of these studies are consistent with American cultural stereotypes about Asian Americans, as seen in fiction, popular films and advertisements. Asian women are seen as "exotic, sensual and submissive," whereas Asian men are often viewed as "asexual, passive or sexually less desirable."49 Adolescents are sensitive to such messages, both explicit and implicit. Media play a prominent role in acceptance of sexual stereotypes and shaping sexual attitudes and behaviors among American youth.50

Another possible explanation for the elevated odds of sexual intercourse for highly acculturated females is that the longer immigrant families live in the United States, the more family support declines.51 Decreases in family support as a result of acculturation may affect adolescent females and males differently. More acculturated women may engage in sexual activity because of a desire for intimacy or as a way of being accepted by their peers. As indicated in our multivariate model for women, parental attachment was associated with decreased odds of sexual experience for young women.

Our finding that low parental attachment was associated with greater odds of sexual intercourse for young women implies that increasing parental attachment may lead to closer identification with and acceptance of parental values, possibly reducing the risk of sexual intercourse. Communication is an important component of parental attachment. Holtzman and Rubinson found that parent-daughter communication about sexual issues, including AIDS, is associated with reduced numbers of sex partners and more protected sexual intercourse among young women in general.52 Clear and explicit messages from parents about the importance of delaying sex or using protection and the prevention of HIV and AIDS are crucial.53 For example, for African American adolescents, maternal disapproval of young women's sexual activity and having a satisfying maternal relationship are associated with abstinence and less frequent intercourse.54

Parental attachment was significant for young men in the bivariate analyses, but after controlling for the other covariates, it became insignificant. This implies that parental attachment may be associated with other significant variables, and thus may have an indirect association with sexual intercourse. Statistical exploration confirmed that higher parental attachment is associated with lower levels of tobacco use, which is significant for the young men.

Studies that have found significant parental influence on young women also have noted that young men are more strongly influenced by their peers than by their parents.55 Adolescents' sexual intercourse is also significantly associated with best friends' mildly deviant behaviors, such as smoking, drinking and cheating on tests.56

Binge drinking was not associated with sexual intercourse for the young men in our study, although it was strongly associated with sexual intercourse among the young women. Previous work with the Add Health data found that moderate and high parental attachment protects highly acculturated adolescents from alcohol use,57 and acculturation leads to binge drinking via social interaction with substance-using peers.58 Further investigation into the characteristics of peer groups and substance use may lead to a better understanding of the gender difference we found in the association between substance use and sexual intercourse. Nevertheless, available evidence suggests that school-based substance abuse and sex education programs must be integrated.59 Teachers, counselors and mental health professionals need to be aware that adolescents who have histories of binge drinking and tobacco use are at increased risk for engaging in sexual intercourse.

The association of school attachment with sexual intercourse among young women may be a further indicator of the importance of sociocultural experiences. School attachment may lead to adherence to school messages in a similar manner that parental attachment leads to adherence to parental values. Schools are an important venue for sex education, though the potential role that feelings of attachment to one's school have in influencing sexual behavior remains unexplored.

Limitations

These findings should be considered in light of the limitations of this data set. First, sexual experience and the covariates were self-reported. Add Health was administered using both computer-assisted self-interviewing (CASI) and audio-CASI systems, in which adolescents listened to the questions through headphones and replied via computer. Though CASI and audio-CASI are effective in eliciting answers to highly sensitive questions, such as ones about drug and alcohol use,60 approximately 30% of Add Health respondents of every ethnic group (Asian, white, black, Latino and American Indian) did not answer the question regarding sex. This indicates a strong need to improve the survey in this area of study by devising new ways of eliciting answers to questions about sex. This could involve changes in wording, the use of multiple questions or modifications of the survey method.

Second, the measure of acculturation was based on use of English at home and place of birth, which have been validated as indicators of acculturation using adolescent samples.61 However, Add Health did not ask adolescents if they were bilingual or if their parents were bilingual. Future research on ethnic and acculturation effects on sexuality should incorporate a richer measure of acculturation, including measures of exposure to Asian and American cultures, specific languages used at home, motivation for acculturation and attitudes about acculturation.62

Third, this study used a dichotomous measure of sexual experience, ever versus never having had intercourse. The study of the frequency and exact nature of sexual activities would lead to a more complete understanding of the sexual development of Asian American adolescents, provided accurate and detailed data could be collected.

Fourth, the sample sizes for most ethnic groups were not large enough to allow for comparisons among them. Since Asian Americans are diverse, the findings for associations with acculturation and other characteristics should be investigated for each major ethnic group.

Finally, more comprehensive longitudinal studies are needed to investigate changes in acculturation over time and how these changes influence sexual norms, family support and parent-adolescent communication, as well as their relationships with sexual activity.

Conclusions

Despite these limitations, our findings have important implications for practice. Since a higher level of acculturation was associated with young women's sexual intercourse, cli-nicians need to give particular attention to assessing the level of acculturation among Asian American young women.

Innovative preventive programs must incorporate culture, gender and family relationships. To engage and interest adolescents, interventions need to be culturally specific, in terms of not only ethnic culture but teenage culture as well. A sex education intervention that used theater to present real-life scenarios tailored to the issues and culture of its Latino audience resulted in positive changes in the knowledge and intentions of the adolescents.63 Such an approach has potential for Asian American adolescents as well, by recruiting ethnically representative, young adult actors whom the audience can relate to as peers.

Interventions should acknowledge the importance and influence of family by containing a component for parents. Since low parental attachment was associated with greater likelihood of young women's sexual intercourse, interventions that improve parent-adolescent communication, parental involvement and support are urgently needed. Parents could attend workshops and be given take-home videos and instructional workbooks on parent-child communication.64

All parents need assistance in initiating discussion of intimate subjects such as premarital sexual activity, STDs and pregnancy. However, Asian American parents need extra assistance because of their cultural heritage. Values and expectations need to be communicated more explicitly in an American context than in an Asian environment, where the culture reinforces restraint and family cohesion.65 A motivation for parents to overcome their reticence and to develop their communication skills is their desire to create the conditions in which their adolescents can delay sexual intercourse and be responsible when they become sexually active. Parents need particular assistance in understanding the peer culture that their children encounter at school. For Asian American adolescents, there is a crucial need for preventive programs that are culturally sensitive, inclusive of family and gender-specific.