The Influence of Sexual Risk Communication between Parents and Daughters on Sexual Risk Behaviors
Hutchinson, M. Katherine, Family Relations
The Influence of Sexual Risk Communication Between Parents and Daughters on Sexual Risk Behaviors*
Two hundred thirty-four 19- to 21-year-old women completed interviews that assessed parent-adolescent relations, sexual communication, and selected sexual risk behaviors and outcomes. Overall, Hispanic-Latina respondents reported less parent-adolescent sexual communication than others. Early parent-adolescent sexual communication was associated with later age of sexual initiation, consistent condom use and, indirectly, less likelihood of sexually transmitted diseases. Mother-daughter communication about condoms was associated with consistent condom use. Recommendations for family-based HIV-STD prevention are presented.
Key Words: adolescents, family processes, parent, sexual behavior.
Although all sexually active persons are at some risk for negative sexual outcomes or "sexual risks" (Hutchinson, 1999), adolescents are a group at great risk. Every year, nearly one million unintended pregnancies and more than three million new cases of sexually transmitted diseases (STDs) occur among American adolescents (Alan Guttmacher Institute, 1994). Adolescents account for one-fourth of all STDs (American Social Health Association [ASHA], 1998) and a significant proportion of new HIV infections (Centers for Disease Control and Prevention [CDC], 1999).
In response to high rates of sexual risk behaviors and negative outcomes among adolescents, delaying the initiation of sexual intercourse and increasing condom use among adolescents are identified as two public health priorities in Healthy People 2000 (Department of Health and Human Services [DHHS], 1991). Recent studies have challenged popular beliefs that parents have no influence over the sexual behavior of the adolescent children. Several well-designed studies found parents' sexual values and sexual communication with their children exert significant influences on adolescents' attitudes toward sexuality, their initiation and participation in sexual activity, and their use of contraceptives and condoms (e.g., Jaccard & Dittus, 2000; Jaccard, Dittus, & Gordon, 1996; Miller, Norton, Fan, & Christopherson, 1998). Although mothers have consistently been found to be the primary sexual communicators with children, particularly daughters (e.g., Dilorio, Kelley, & Hockenberry-Eaton, 1999; Dutra, Miller & Forehand, 1999; Hutchinson, 2000a; Hutchinson & Cooney, 1998), fathers may play an important role through the discussion of sociosexual issues with daughters (Hepburn, 1981). Fathers may be able to provide daughters with unique perspectives, enhance their overall understanding of men, and provide opportunities to role-play communication strategies with men (Hutchinson, 1999, 2000b). However, much of the recent research on parent-adolescent sexual communication and family influences of adolescent sexual risk behaviors has been limited to communication and relationships with mothers. The purpose of the present study was to examine the influence of sexual communication with mothers and fathers on the sexual risk behaviors of adolescent daughters.
Sexual Risk Behavior of Adolescent Females
Female adolescents face even greater risk for STD transmission than their male peers and older adult women (Donovan, 1993; Hatcher et al., 1994: Panchaud, Singh, Feivelson, & Darroch, 2000). Teenage girls are more likely to be unmarried, have multiple sexual partners, and have unprotected sex (CDC, 2000; Panchaud et al.). Compared to adult women, adolescent females are also at greater biological risk to contract an STD, if exposed (CDC, 2000; Donovan, 1993, 1997; Santelli, DiClemente, Miller, & Kirby, 1999). Women face more severe long term sequelae from STD infections than men. Undiagnosed and untreated STDs in women may result in pelvic inflammatory disease, increased risk for ectopic pregnancy, infertility, devastating congenital infections in infants born to infected women (Hutchinson & Sandall, 1995; Institute of Medicine [IOM], 1997), and cervical cancer (Donovan, 1997). In addition, many STDs can increase the likelihood of HIV transmission if an individual is exposed to an infected partner.
As a result of such social and biological vulnerability, adolescent females exhibit some of the highest STD rates of any age group (CDC, 2000). Forty percent of chlamydia cases are reported among adolescents. Prevalence rates among adolescent girls often exceed 5-10%. Although chlamydia is common among all races and ethnicities, its prevalence is higher among African Americans and Hispanics. Gonorrhea rates are high and increasing among all groups. However, rates are highest among adolescent females and males in their 20s, particularly African Americans. Rates for 15- to 19-year-old females range from approximately 200/100,000 for Whites to almost 3,700/100,000 for African American teen females.
Adolescents also face significant risk for sexually transmitted HIV infection. Although this age group has represented a small proportion of reported cases of AIDS, their low numbers primarily may be attributable to the long incubation period that precedes the development of symptoms and AIDS-defining conditions (Rotheram-Borus, Jemmott, & Jemmott, 1995). Of the 6,387 women newly diagnosed with HIV in the U.S. (July 1998 to June 1999), 22% were among 13- to 24-year-old females (CDC, 1999). The number of new HIV cases among girls (13-- 19 years old), exceeded the number for boys, accounting for 61% of new HIV cases among this age group. Among 20-24 year olds, females accounted for 43.7% of new HIV cases (CDC, 1999). Because only 4% of cases among 13- to 19-year-old women and 7% among 20- to 24-year-old women were attributed to injection drug use (CDC, 1999), sexual activity appears to be the primary mode of transmission.
Significant risk for STDs and HIV among adolescents is not surprising, given their sexual risk behaviors. Although there have been small declines recently, more than half of all high school students (53%) have had sexual intercourse (Kann et al., 2000). Rates of sexual activity vary by age, ethnicity, and urbanicity. Only 32% of 9th grade females and 41% of 9th grade males have had sexual intercourse whereas more than 65% of 12th grade students have (Kann et al., 2000). Urban and minority youth report earlier ages at sexual initiation and higher overall rates of sexual activity than other adolescents (Kann et al., 2000; Santelli et al., 1999). African American high school students are significantly more likely to have had sexual intercourse than White or Hispanic students, 73.4%, 48.9%, and 57.6%, respectively (Kann et al., 2000). These differences hold for both male and female adolescents. Urban youth initiate sexual activity at even younger ages than their nonurban peers. More than 25% of inner-city 6th and 7th graders in Philadelphia reported having had sexual intercourse at least once (Jemmott, Jemmott, & Fong, 1998). Mean age at first sex for these young adolescents was 11.8 years. Thus, urban teens who become sexually active at young ages are at particularly high risk for STDs and HIV (Jemmott, Jemmott, & Fong, 1992; Miller et al., 1997).
Adolescents who initiate sexual activity at younger ages are more likely to have multiple sexual partners (e.g., Miller et al., 1997) and more likely to become infected with an STD (Aral & Wasserheit, 1995; Miller et al., 1997). Despite having multiple sexual partners, many adolescents do not recognize their risk for STDs including HIV and do not use condoms. Many do not equate serial monogamy with multiple sexual partners and so may not view their sexual behavior as "risky" (Alan Guttmacher Institute, 1994). This lack of personal risk recognition may contribute to why adolescents fail to use condoms consistently (Hutchinson, 1998, 1999). In lieu of condom use, adolescent females may implement other, less effective, sexual protective strategies, such as choosing low-risk partners (Hutchinson, Sosa, & Thompson, 2001).
Studies have repeatedly documented inconsistent condom use among teens and youth. The proportion of teens who report using condoms during their last episode of sexual intercourse varies across studies and ranges from about 40% to 70% (Jemmott et al., 1992; Kann et al., 2000; Whitaker, Miller, May, & Levin, 1999). African American students are more likely than White or Hispanic students to report condom use, 66%, 52.5% and 44.4%, respectively (Kann et al., 2000), and even fewer teens report consistent condom use (Kann, Warren, Harris, Collins, & Williams, 1996). This unprotected sexual activity places young people, especially young women, at risk for both HIV and other STDs (Aral & Wasserheit, 1995; Jemmott, Catan, Nyamathi, & Anastasia, 1995).
Sexual Risk Communication
Recently, researchers have begun to look beyond the individual to identify family variables that may influence adolescent sexual risk behavior. Several family processes are identified as potentially important. Girls who talk to their mothers about sexual topics are more likely to have conservative sexual values and less likely to have initiated sex, compared to girls who mostly talked to their friends about sex (Dilorio et al., 1999). Parents also appear to serve as buffers for adolescents, moderating the effects of peer pressure and environmental influences on sexual activity (Santelli et al., 1999; Whitaker & Miller, 2000). Girls who talk to parents about when they should have sex were less influenced by whether they thought their peers had initiated sex early or later. In contrast, if parents had never discussed sexual initiation with them, age at sexual debut was significantly lower in girls who thought their peers started early (Whitaker & Miller).
Open and receptive sexual communication between adolescents and their parents is associated with less adolescent sexual activity. Parent-adolescent sexual communication is linked with later onset of sexual initiation (e.g., Dilorio et al., 1999; Whitaker & Miller, 2000). Among adolescents who are already sexually active, parent-adolescent sexual communication is associated with greater condom use self-efficacy (Hutchinson & Cooney, 1998), sexual communication with the male partner (Hutchinson & Cooney; Whitaker et al., 1999), and increased condom use (Kotchick, Dorsey, Miller, & Forehand, 1999; Whitaker & Miller). Further, the timing of parent-adolescent sexual communication may be as important as the occurrence of the communication itself. Adolescents who report discussing condoms with their mothers before becoming sexually active are almost three times more likely to use condoms (a) the first time they had sex, (b) consistently, and (c) the last time they had sex, compared to teens who do not have such discussions (Miller, Levin, Whitaker, & Xu, 1998).
Communication and relationships with fathers are also potentially important influences of daughters' sexual attitudes and behaviors (Hepburn, 1981). Although mothers provide children, particularly daughters, with more sexual information than fathers (Dilorio et al., 1999; Dutra et al., 1999; Hutchinson & Cooney, 1998), fathers may be important sources for the discussion of more general moral and sexual issues (Hepburn). They also may provide daughters with a better general understanding of men and men's perspectives in heterosexual relationships (Hutchinson, 1999, 2000b.)
Unfortunately, our present understanding of parental influences of adolescent sexual behavior is somewhat limited. Past studies of parent-adolescent sexual communication are limited by their use of selective samples (e.g., sexually active teens; STD clinic attendees; young, urban adolescents; college students), inconsistent operationalization of parent-adolescent sexual communication, and limited measures of parent-adolescent sexual communication (Fisher, 1993; Hutchinson & Cooney, 1998). Although most studies rely upon dichotomous questions regarding whether a sexual topic has ever been discussed with the parent, greater sensitivity is found by using 5-point response scales to indicate the extent of communication on a given topic (Hutchinson, 2000a; Jaccard, Dittus, & Gordon, 1998). Inconsistencies in methodology and restrictive samples limit our ability to compare parent-adolescent sexual communication across populations. Crude or global measures of parent-adolescent sexual communication limit our ability to discriminate between patterns of sexual communication and may limit our power to detect differences in the influence of parent-adolescent sexual communication on adolescent sexual risk behaviors. Finally, the general omission of fathers from studies of parent-adolescent sexual communication affects our understanding regarding the role and influence of fathers.
Study Aims and Hypotheses
This study extended previous work in four important ways. First, we included ethnically diverse statewide samples of young women to identify differences in families' patterns of sexual communication by race and ethnicity. Second, we included respondents from urban and nonurban areas to extend the literature that has been largely limited to urban populations and identify differences that may need to be considered in the design of family based prevention interventions for nonurban populations. Third, we included father-daughter sexual communication and general communication as potentially important family influences of adolescent sexual risk behaviors, although earlier studies show low overall levels of father-daughter sexual communication (Hutchinson & Cooney, 1998). Fourth, we focused exclusively on young women's reports of parent-adolescent sexual communication and sexual risk behaviors to avoid confounding results by gender differences in such communication.
The specific research questions included: Are there differences in the amount and timing of parent-adolescent sexual risk communication with mothers and fathers reported by Hispanic-- Latina, African American and White female adolescents? Is there a relationship between the timing of parent-adolescent sexual risk communication and female adolescents' sexual risk behaviors (initiating sexual intercourse, consistent condom use, and self-reported STDs)? Is there a relationship between amounts of parent-teen sexual risk communication with mothers and fathers and female adolescents' sexual risk behaviors (initiating sexual intercourse, consistent condom use, and self-reported STDs)? Are the relationships between parent-adolescent sexual risk communication and female adolescents' sexual risk behaviors and outcomes moderated by race and ethnicity and urbanicity?
Lists of all female licensed drivers born in 1976, presorted by race, were obtained from the Department of Motor Vehicles (DMV) of a Mid-Atlantic state. Initially, contact letters were sent to random samples of 250 African American and White women. Home telephone numbers were obtained through SelectPhone (ProCD, 1996) using the addresses on DMV records. Women who could be located and contacted were invited to participate. Because greater difficulty was encountered reaching young African American women, particularly those in urban areas, a supplemental sample of 100 young African American women was randomly drawn from the DMV list and included in the potential subject pool. Telephone interviews were scheduled with women who agreed to participate.
One year after the original interviews, supplemental funds allowed extending the study to include young Hispanic-Latina women from the same birth cohort. Questionnaires were translated and back-translated. The same procedures outlined were followed to recruit these young women and again we encountered difficulty. Two supplemental random samples were drawn and included in the potential subject pool. Bilingual interviewers were hired and trained. Only four (6.2%) Hispanic-Latina participants completed interviews in Spanish. Hispanic-Latina subjects were interviewed 1-2 years later and, thus, were 1-2 years older than other study participants.
A total of 234 young women participated in the study-65 Hispanic-Latinas, 78 African Americans, and 91 Whites. Participation rates for those who could be located and contacted were 43%, 52%, and 56% for Hispanic-Latina, African American, and White participants, respectively. Difficulty in locating and recruiting Hispanic-Latina and African American women resulted in smaller numbers than anticipated. Analyses were limited to African Americans, Hispanic-Latinas, and Whites. Respondents who indicated multiple races and ethnicities (n = 2) and those who self-identified their race and ethnicity as "other" (n = 2) were excluded. All study participants were paid $10.
Respondents answered questions about demographics (race and ethnicity, community type), family structure (parents' levels of education, parental divorce), family processes (closeness with each parent, general communication and sexual risk communication with each parent, whether parents discussed sexual risk topics before sexual initiation), and adolescent sexual risk behaviors (age at sexual initiation, numbers of partners, frequency of condom use, and whether they had ever had an STD). Qualitative data regarding parent-adolescent sexual communication and sexual protective strategies also were collected but are not reported here. (Daughters' qualitative reports of the adequacy and importance of father-daughter sexual communication are discussed elsewhere; see Hutchinson, 2000b).
Demographic characteristics. Respondents described their ethnicity or race, and race and ethnicity was coded using two dummy variables (n categories - 1). The first variable was coded 1 for Black and 0 for non-Black. The second dummy variable was coded 1 for Hispanic-Latina and 0 for non-Hispanic.
Also, respondents were asked to describe the type of community they grew up in as either urban, suburban, or rural. Because our primary interest was comparing urban with nonurban, community type was coded as 1 for urban or 0 for nonurban.
Family structure. Parental education was measured separately for mothers and fathers. Eight response choices, ranging from 1 = less than a high school diploma to 8 = graduate or professional degree, were available. Parental marital status was assessed for current status as well as when the respondent was 16 years old and was coded as married, separated, divorced, or never married. The analyses reported here are based on marital status when respondents were age 16 and coded as 1 = married or 0 = not married.
Parent-adolescent general communication. General communication was measured separately for each parent using the single item, "When you were in high school, how well could you and your mother/father share ideas or talk about things that really mattered to you?" Responses ranged from: 1 = not well at all to 4 = very well, with higher scores indicating better general communication.
Parent-adolescent sexual risk communication. The PTSRC-- III scale is an 8-item measure of parent-adolescent sexual communication about specific sexual risk-related topics. Sexual risk communication is assessed separately for mothers and fathers. Respondents were asked, "When you were growing up, how much information did your mother/father share with you about ... contraception/birth control, STDs, HIV, how to protect against STDs and HIV, condoms, postponing or not having sex, pressure from boys to have sex, and how to resist pressure from boys to have sex." Responses ranged from I = nothing/none to 5 = extensive/everything, and possible scores ranged from 8-40.
The development and psychometric properties of the PTSRC-III are discussed elsewhere (Hutchinson & Cooney, 1998). The scale shows high levels of internal reliability (alpha > .90 for both mothers and fathers in earlier studies; .92 and .95 for current study) and test-retest reliability, r = .80 (Hutchinson, 2000c). Data from the African American and White participants in this study showed that mother-teen PTSRC-III scores were correlated with daughters' reports of mother-adolescent closeness (r = .38), mother's comfort with sexual communication (r = .61) and a global measure of sexual communication (r = .73) (Hutchinson & Cooney). Father-teen PTSRC-III scores were correlated with daughters' reports of closeness with the father (r = .30), father's comfort with PTSRC (r = .61) and a single-item measure of father-teen sexual communication (r = .80).
Timing of parent-adolescent sexual risk communication. Timing of sexual communication was assessed with one yes/no question, "Did your parents talk to you about sex before you started having sex?" It should be noted that, prior to this item, "having sex" had been defined as "having sexual intercourse" or "going all the way" with a boy or man.
Adolescent sexual risk behavior. Respondents were asked three questions regarding their sexual risk behaviors and outcomes: age at first episode of sexual intercourse, ever had an STD (e.g., chlamydia, syphilis, gonorrhea, herpes, human papilloma virus [HPV]), and consistent condom use before age 18. Responses for the condom use item ranged from 1 = never to 5 = almost every time/every time, and later coded as 1 = consistent condom use [almost every time/every time] or 0 = inconsistent.
Analyses and Results
Characteristics of the study sample and differences between groups are summarized in Table 1. Forty of the 234 (17.1 %) young women reported never having sexual intercourse, and 194 (82.9%) reported having had vaginal intercourse with a man at least once. The majority of women (60.7%) reported having sex prior to age 18, and 11% (13% of those who had had sex) reported having had at least one STD during their lifetime. There were no significant differences between African American, White, and Hispanic-Latina women in the proportion who made their sexual debut prior to age 18, consistent condom use, or STD occurrence. Among those who initiated sex prior to age 18, there were no differences between groups in age at first sex. Because 17% of respondents had not yet initiated sexual intercourse, there was significant right censoring of the data (Allison, 1995). Because right censoring can be biasing and because Hispanic respondents were older and had a longer period in which to initiate sex, intergroup comparisons of mean age at first sex were limited to those who had initiated prior to age 18, a consistent time frame across the three groups. Associations between race and ethnicity and the likelihood of initiating sexual intercourse were analyzed using proportional hazards models.
Intergroup differences in demographic characteristics and descriptive data were assessed using chi-square analyses for categorical variables and analysis of variance (ANOVA) for interval and higher level data. Because these young women, who were from the same birth cohort, were interviewed 1-2 years later than other respondents, Hispanic-Latinas were significantly older than other study women, F(2, 232) = 649.9, p < .0001. Age and Hispanic ethnicity were highly correlated (p = .74). Hispanic-Latina young women were more likely to report having had sexual intercourse at least once, X2(2, N = 234) = 8.0, p < .05; more than 93% of Hispanic respondents had initiated sex, compared to 77% and 80% of Whites and African Americans, respectively. This difference was not surprising given that Hispanic-Latina respondents were older than other respondents. African American participants were more likely to report being raised in an urban area, X2(2, N = 234) = 31.5, p < .01, and less likely to report their mothers were married when they were growing up, X2(2, N = 234) = 8.0, p < .05. Because of these intergroup differences, dummy variables representing urban area of residence and race and ethnicity were included in all subsequent multivariate analyses.
Measures and bivariate associations. Although a number of moderate interitem correlations were found (e.g., mother-daughter sexual communication and general communication, r = .50), the degree of association was such that it would not preclude the inclusion of both variables in a given model (see Table 2). Only age and Hispanic-Latina ethnicity were highly intercorrelated (r = .92, p = .74). Because of concerns regarding multicollinearity, the procedures outlined by Lewis-Beck (1989) were followed, and each demographic independent variable was regressed on all the others (African American, Hispanic-Latina, urbanicity, and age). A maximum R2 of .85 was obtained when age was regressed on the other independent variables. Therefore, age was not included in multivariate analyses. Table 2 presents the correlation matrix for the study variables.
PTSRC scores were moderately correlated with measures of general communication with each parent (r = .50 for communication with mothers; r = .44 for fathers). Mean scores were 21.4 (SD = 8.7) and 14.3 (SD = 7.9) for sexual risk communication with mothers and fathers, respectively. Scores for mothers were fairly normally distributed, whereas fathers' scores were skewed (skewness = 1.41, kurtosis = 1.26), with nearly onethird of responses (33%) reporting no sexual communication with their fathers (PTSRC = 8).
Analyses Related to Study Hypotheses
Patterns of parent-adolescent sexual risk communication. Intergroup differences in patterns of parent-adolescent sexual risk communication were assessed by examining both the timing and amount of such sexual communication that occurred. Intergroup differences in the timing of sexual communication (whether sexual communication with parents occurred prior to the initiation of sexual activity) were analyzed using chi-square (see Table 1). Nearly 78% of African American young women reported their parents discussed sexual topics with them before they became sexually active, compared to 53% of Hispanic-Latinas and 50% of Whites, X2(2, N =193) = 12.3, p < .01.
Differences in amounts of parent-adolescent sexual communication between African American and White respondents from the original study sample are discussed elsewhere in greater detail (Hutchinson & Cooney, 1998). In general, African Americans reported significantly more communication with both mothers and fathers, compared to their White peers. After expanding the sample to include Hispanic-Latinas, these findings held. Two separate one-way ANOVAs were used to test for differences in sexual communication with mothers and fathers between African American, Hispanic-Latina, and White daughters. Both were significant (see Table 1). T-tests were used to test for differences in communication with mothers and fathers between urban and nonurban women. Urban respondents reported more communication with both mothers, t(231) = 2.90, p < .005, and fathers, t(232) = 2.15, p < .05, than nonurban respondents.
Multivariate OLS regression models (Table 3) were used to verify ANOVA findings of racial and ethnic differences and identify other influences of communication. Parental education was nonsignificant in both models and deleted. Parental marital status was nonsignificant for both parents, although it approached significance for fathers. Hispanic-Latina ethnicity was associated with significantly less sexual risk communication with both parents. Urban residence was significantly associated with more sexual communication with fathers, but only marginally so for mothers. Young African American women reported more communication with both mothers and fathers. Overall, the two models yielded similar findings. By far, the single greatest predictor of parent-adolescent sexual communication was the quality of parent-adolescent general communication.
Potential moderator effects were assessed by including all possible two-way interaction terms in models using the procedures outlined by Baron and Kenny (1986). Interaction terms were created using product terms (e.g., African American x urban) and included in models with main effects. No significant two-way interaction effects were found for either sexual communication with mothers or fathers.
Parental influence of age at first sexual intercourse. The influence of family structure and other process variables on adolescents' age at first sexual intercourse was estimated using partial likelihood estimates based on proportional hazards modeling (Cox, 1972). These data could not be analyzed using more traditional techniques because 40 of the 234 study participants had not experienced sexual intercourse, resulting in right censoring of the data (Allison, 1995). Estimates would be biased because young women who had never had sex would have missing values for the variable. Partial likelihood and proportional hazards models take into consideration that the event of interest has not occurred in some of the participants when calculating estimates of effects (see Allison, 1995; Cox, 1972).
Demographic variables (African American, Hispanic, urban) were entered into the model first. Family structural characteristics (parents' marital status and education) were entered next, followed by family process variables (parent-adolescent general communication, sexual communication that occurred prior to sexual initiation). Participants' reports of the amount of parent-- adolescent sexual communication were excluded from the model because of the retrospective nature of the reporting and the likelihood that parent-adolescent sexual communication would have continued well after the initiation of sexual intercourse by the adolescent. Other than demographic and control variables, only two family process variables were significant and retained in the final model. Urban residence showed a marginal association with increased hazard of sexual initiation but this association was not significant (p > .05). The final model is shown in Table 4.
Discussing sex with the parents prior to becoming sexually active exerted the greatest influence on the likelihood of initiating sexual intercourse. The hazard ratio indicates that young women who discussed sex with their parents before becoming sexually active were much less likely to initiate sexual intercourse than those who did not discuss sex with their parents. In addition, those who reported better general communication with their fathers were less likely to initiate sexual intercourse. For each one-unit increase in the quality of father-daughter communication, the risk for initiating sex at a given age decreases by 19% (100*[.81-1.00]; Allison, 1995). In sum, young women with higher quality of general communication with fathers and those who had discussed sex with their parents had lower rates of sexual initiation than other young women.
Family influence of consistent teen condom use. Logistic regression (Hosmer & Lemeshow, 1989) was used to determine each parent's influence on the outcome of consistent condom use during adolescence. Young women's reports of condom use frequency during adolescence were coded for using condoms as 1 = always or almost always and 0 = less frequent use. Demographic variables were entered into the model first. Of these, only urban residence was significant. Family structure variables (parents divorced, parents' educations) were entered next, and only father's education was significant and retained in the final model. Higher levels of father's education were associated with an increased likelihood of consistent condom use.
Family process variables were added next. General mother-- daughter communication, whether mother discussed condoms (gave at least some information on condoms), and early parentadolescent sexual communication that occurred before first sex were all significant predictors of consistent condom use. Total mother-daughter sexual communication was only weakly associated with consistent adolescent condom use (r = 10) and highly intercorrelated with discussing condoms (r = .87, p < .0001) and discussing sex before first sex (r = .61, p < .0001). Therefore, mother-daughter sexual communication scores were eliminated from the model because of potential multicollinearity. Father-adolescent variables were added to the model next. All were nonsignificant and subsequently eliminated. The reduced final model (shown in Table 5) was highly significant, X2(7, N = 134) = 42.671, p = .0001, and the model explained 27.3% of the variance.
General communication with the mother, communication with the mother about condoms (single item), and sexual communication prior to sexual debut were significant predictors of consistent adolescent condom use. Young women who reported that their parents talked to them about sex before they became sexually active were nearly seven times more likely to report consistent condom use during adolescence. Young women who reported more communication with their mothers about condoms and those who reported good general communication with mothers were more than 60% more likely to report consistent condom use compared to other young women. In addition, urban young women were almost eight times more likely than their nonurban peers to report consistent adolescent condom use.
Potential moderation of effects was evaluated using interaction terms. Included in the models were both main effects and two-way interaction terms (e.g., African American x urban). Models were then evaluated for significant coefficients and model improvement (Allison, 1999). None of the interaction terms were significant.
Family influence of STD occurrence. Logistic regression also was used to assess respondents' reports of ever having had a sexually transmitted disease (STD) and to estimate the effects of family structure and process variables on the likelihood of STDs. The final model is shown in Table 5.
Once again, African American, Hispanic, and urban were entered first in the model. Urban residence was significantly associated with STD incidence, although race and ethnicity were not. Family structure variables were added next, contributed little to the model, and were subsequently deleted. Family process variables were added last. As was the case in the previous model, father-daughter variables were not significant and were deleted from the model. Mother-daughter general communication also was nonsignificant and deleted. Mother-daughter communication about condoms and early sexual communication were added to the model last. Their inclusion raised the covariate X2(5, N = 191) = 22.82, p = .0004, although only early sexual communication was significant. However, when adolescent risk behaviors (age at first sex and consistent adolescent condom use) were included in the model, early sexual communication no longer exerted a significant main effect.
These results indicate no direct effect from early parent-- adolescent sexual communication on STD likelihood, although indirect effects may be present. Early sexual communication was associated with both later age at sexual initiation and consistent condom use. Both of these variables were, in turn, associated with reduced likelihood of STD occurrence. Urban residence was significant in all models. In the final model, after controlling for other factors including adolescent condom use and age at first sex, urban young women were 18 times more likely to report having had an STD. Once again, no significant interaction effects were found.
Discussion and Practice Implications
The findings from this study extend the extant literature by examining daughters' reports of sexual communication with mothers and fathers by incorporating measures of both the timing and amount of sexual risk communication into models of adolescent sexual risk behavior and by including ethnically diverse families from urban, suburban, and rural areas. However, the study findings should be interpreted with caution. Because the study sample was drawn from all counties of a single Mid-Atlantic state, it is diverse but not necessarily representative of the entire population of 19- to 21-year-old women. Difficulty in contacting potential study participants may have led to the underrepresentation of some groups. Further, the use of driver's license records and telephone contacts may under-represent urban young women and the very poor. Small subsample sizes may result in undetected differences in some of the analyses. In addition, the use of retrospective reporting raises the possibility of intentional or unintentional recall errors, or the reinterpretation of past events.
Finally, because Hispanic-Latina respondents were 1-2 years older than participants from other racial and ethnic groups, findings of differences in parent-adolescent sexual communication related to Hispanic-Latina ethnicity may be confounded by age. High levels of multicollinearity were found between age and Hispanic ethnicity (r = .91, p = .74, R2 = .85), that precluded including both variables in the regression models. Hispanic ethnicity was retained and age was dropped from the analyses reported. In order to address this potential confounding of the results, Hispanic ethnicity was subsequently residualized by age (in months) and analyses were rerun using the residualized variable. No major changes in the results were found. The residualized measure of Hispanic ethnicity continued to be a significant predictor of mother-daughter sexual risk communication (p < .01), although only a marginal predictor of father-daughter sexual communication (p < .09). As was found in the original models, residualized measures of Hispanic ethnicity were not significant predictors of sexual outcomes-age at first sexual intercourse, consistent adolescent condom use, and self-reported occurrence of sexually transmitted infections.
Parent-Adolescent Sexual Risk Communication
We found that young Hispanic-Latina women reported less parent-adolescent sexual risk communication with both parents compared to their non-Hispanic peers, although differences with fathers were only marginal when examined using ethnicity measures that were residualized on age. Findings of ethnic differences are in contrast to Miller et al.'s (1997) findings of no differences between Hispanic and African American inner-city families. We believe that differences in study findings are likely attributable to sampling differences and sensitivity of measures of parent-adolescent sexual communication. The Miller et al. (1997) study included inner-city male and female adolescents, whereas our sample included only females from urban, suburban, and rural areas. In addition, our results indicate a small but significant effect of urban residence on levels of sexual risk communication. Urban Hispanic-Latina families may be more similar to other families in patterns of parent-child communication than those in suburban and rural areas. Although we found no significant race X urban interaction effects, our sample may have been too small to detect such differences. Another possible explanation for our finding of differences could lie in our assessment of how much sexual risk communication occurred rather than whether or not any communication occurred. The younger ages of Miller et al.'s (1997) subjects also may contribute to different findings.
On the basis of our findings, we question whether young Hispanic-Latina women receive less parent-adolescent sexual communication than female adolescents of other ethnic groups. Qualitative data from the same study (Hutchinson, 2000b) indicate that the family and religious norms of some Hispanic-- Latina families limits parent-adolescent sexual communication. We speculate that these patterns may vary by the parent's country or region of origin, religiosity, and gender of parent and child. Communication also may vary by other contextual factors, such as urbanicity, community norms, and level of acculturation. Some Hispanic-Latina families, those with conservative religious values and those who hold strong beliefs that any type of premarital sexual intercourse is morally wrong, may be uncomfortable and avoid sexual communication. Although the findings from the current study raise interesting questions about ethnic variations in parent-adolescent sexual communication, they represent a beginning. Future research needs to more fully explore variation within ethnic groups to refine our understanding and enhance our ability to develop programs that are tailored to the needs of individual families.
We also found lower levels of sexual communication with fathers than with mothers. The special communication needs of fathers should be addressed in family based interventions to promote father-daughter sexual communication. Although these fathers provided daughters with little information regarding sexuality, daughters reported that fathers were important sources of support (Hutchinson, 2000b). Many of the young women here reported a desire for more sexual communication with their fathers, particularly in the areas of "understanding men" and "resisting pressure from men to have sex" (Hutchinson, 2000b). Family interventions should include components that are specifically targeted to fathers and their unique roles and needs regarding parent-adolescent sexual communication.
Curricula for fathers that are theory-based, culturally sensitive, and tailored to fathers in the target population based on elicitation research with men themselves are warranted. Father should be inclusively defined to include both biological fathers, stepfathers, male relatives, guardians, and other father figures.
We recommend that the goals for programs to enhance father-adolescent sexual communication include: (a) clarifying men's personal goals for their children's sexual socialization; (b) exploring personal feelings regarding their roles as fathers; (c) providing information about the role of fathers in sexual socialization; (d) assisting men to identify cultural, family, and personal barriers to father-child sexual communication; (e) providing opportunities to learn and practice active listening and open communication skills; and (f) providing opportunities to role play and practice initiating sexual conversations and responding to sexual queries from children.
Parents' Influences on Adolescent Sexual Risk Behaviors
As in other studies (e.g., Miller, Forehand, & Kotchick, 1999), parental influences varied somewhat across outcomes. Early parent-adolescent sexual communication (discussions that occurred before the onset of adolescent sexual activity) was directly related to age at sexual initiation and consistent condom use during adolescence. Mother-daughter communication about condoms also was associated with consistent condom use during adolescence. Parental influences on age at sexual initiation and condom use are particularly important as both early age at sexual debut and inconsistent condom use have been linked to STD risk (Aral & Wasserheit, 1995; Miller et al., 1997; Santelli et al., 1998). There are, of course, alternative explanations for these findings. Given the cross-sectional retrospective nature of the data, it is not possible to determine the direction of effects with certainty. It may be that young women who initiated sex at later ages had more time for parents to discuss sexual topics before their children began having sex.
Sexual communication with fathers was not a significant influence of daughters' sexual risk behaviors in this study. Although we and others failed to demonstrate a significant effect for sexual communication with fathers, we do not believe that fathers are unimportant. We concur with Hepburn's (1981) assertion that the paternal role in sexual socialization of children, particularly daughters, is qualitatively distinct from that of mothers and not adequately measured. We recommend that qualitative studies with children and fathers be undertaken to further our understanding of this potentially important aspect of fatherhood. Qualitative studies could yield a deeper understanding of fatherchild sexual communication and illuminate critical father-child processes that have not been identified as important in studies of mother-child interactions and sexual socialization.
We were somewhat surprised at the absence of significant interaction effects on parental influences of sexual risk behaviors. Neither race and ethnicity nor urbanicity qualified the effect of parental variables on reports of adolescent sexual risk behaviors. However, small subsample sizes, particularly among urban White and Hispanic-Latina women, may have resulted in undetected differences. Although urban young women reported significantly more parent-adolescent sexual communication than nonurban young women, the effects of sexual communication on adolescent sexual risk behaviors were not conditioned by urbanicity. Despite higher levels of parent-adolescent sexual communication among urban families, urban young women were more likely to report STDs than their nonurban peers. Thus, any protective effects provided by such communication are inadequate compared to the higher prevalence and relative risk for STDs present in an urban environment.
Also of great concern are differences in sexual communication among urban and nonurban families. Nonurban (suburban and rural) respondents reported less parent-adolescent communication than urban young women. We question whether non-- urban parents are less aware than urban parents of the sexual risks facing adolescents today. Parents may not appreciate the number of sexual risks facing today's youth, if they grew up in the 1960s and 1970s when the primary sexual risk was pregnancy. Even parents who are aware of the extent of STDs and HIV in the United States may not apply this knowledge to their own children. Parents also may be unaware that their adolescents are sexually active. Family professionals need to raise awareness among all parents that STDs and HIV are pervasive and that all adolescents are at risk. Accurate information about the prevalence of STDs and HIV across groups, heterosexual risk for STDs and HIV, and current rates and patterns of adolescent sexual behavior may help parents to recognize the relevance of the problem and address it within their own families. Even those adolescents who are not currently sexual active will presumably become sexually active at some point, thus making the sexual risk information and communication skills relevant for all parents.
Promoting Parent-Child Sexual Communication
In addition to a lack of risk recognition, parents may lack the communication skills and self-efficacy necessary to effectively communicate about sexual topics with their children. In focus groups with suburban parents (unpublished data), we found that many recognize the importance of communication and want to communicate with their children but they lacked good sexual communication role models in their own lives and are unaware of how and when to initiate sexual conversations. Parents need both information and opportunities to practice these communication skills, identify teachable moments, plan age-appropriate conversations, develop open communication and listening skills, and examine their own fears and attitudes toward sexuality in order to create a family environment that normalizes sexuality and promotes communication.
The findings of this and another study (Miller, Levin, Whitaker, & Xu, 1998) lend support to the contention that sexual communication may be most effective when it begins prior to the adolescent becoming sexually active. Thus, family based interventions targeted towards preadolescent parent-child sexual communication are needed. Many family based programs designed to promote parent-child sexual communication target early adolescents and their parents (Meschke, Bartholomae, & Zentall, 2000) to establish sexual communication prior to the adolescent's sexual debut. However, we believe that parents should begin laying the foundation for positive and open sexual communication even earlier. Parents of toddlers and preschool-age children can do so by promoting the child's self-esteem and pride in his or her body, conveying an accepting attitude toward the young child's self-exploration and masturbation. Also, parents can begin to share family mores and values regarding sexuality. For example, parents can be accepting of masturbation, yet provide the context that this behavior is done in the privacy of one's room. Parents of school-age children can continue by being approachable, identifying and capitalizing on teachable moments in everyday life (e.g., discussing moral issues or sexual situations raised in a t.v. program or popular song), and initiating conversations with the child who does not ask. Reminiscing about one's own experiences at a similar age and asking the child about his or her concerns and questions also can elicit questions and stimulate conversation.
Resources for Parents and Adolescents
In addition to providing parents with individual guidance and communication skills, family professionals can help parents to identify and obtain resources for themselves and their children. There are hundreds of books on sexuality related topics for children of all ages and for parents. Annotated bibliographies of sexuality related books are available through the Sexuality Information and Education Council of the U.S. (SIECUS) website at www.siecus.org. SIECUS provides up-to-date information on a variety of sexuality topics, as well as useful communication tips for parents in a how-to format.
The Planned Parenthood Federation of America (810 7th Ave., New York, NY; 1-800-669-0156; www.plannedparenthood.org) also distributes educational materials for parents, children, and adolescents on a wide variety of sexuality related topics. Their pamphlets are well-written and provide thorough, accurate information in brief form. Pamphlets are available on topics ranging from "Sexually Transmitted Infections: The Facts," to "Human Sexuality: What Children Should Know and When They Should Know It." Planned Parenthood also distributes a kit on parent-child sexual communication that includes a video, parent's guide, and activity workbook for 10- to 14-year-olds. Many local chapters have lending libraries through which community groups and parents can borrow videos and other materials. Although books, pamphlets, and videos can serve as valuable supplemental resources, they are not substitutes for active parental involvement and should not be used in lieu of ongoing, age-- appropriate, parent-child sexual communication and guidance.
*This work was supported by grants from the American Nurses Foundation, the Rutgers University Research Council and the Research Institute of the New Jersey State Nurses Association. Support was also provided by an institutional postdoctoral fellowship award to the University of Pennsylvania Center for Health Outcomes and Policy Research (T32 NR07104, Linda Aiken, PI.). An earlier version of this article was presented at the National Institute for Mental Health (NIMH) research conference, "The Role of Families in Preventing and Adapting to HIV/AIDS," in July, 2000. The author would like to thank Dr Adeline Nyamathi for her invaluable assistance in planning the Hispanic-Latina component of this project.
Alan Guttmacher Institute. (1994). Sex and America's teenagers. New York: Author.
Allison, P. (1995). Survival analysis using the SAS system. Cary, NC: SAS Institute
Allison, P (1999). Logistic regression using the SAS system. Cary, NC: SAS Institute.
American Social Health Association (ASHA). (1998). Sexually transmitted diseases in America: How many cases and at what cost? Menlo Park, CA: Kaiser Family Foundation.
Aral, S., & Wasserheit, J. (1995). Interactions among HIV, other sexually transmitted diseases, socioeconomic status, and poverty in women. In A. O'Leary & L. S. Jemmott (Eds.), Women at risk: Issues in the primary prevention of AIDS (pp. 13-42). New York: Plenum.
Baron, R., & Kenny, D. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182.
Centers for Disease Control and Prevention (CDC). (1999). HIV/AIDS Surveillance Report, 11(1), 1-42. Atlanta, GA: CDC.
Centers for Disease Control and Prevention (CDC). (2000). Tracking the hidden epidemics: Trends in STDs in the United States. Atlanta, GA: CDC.
Cox, D. (1972). Regression models and life tables (with discussion). Journal of the Royal Statistical Society, B34, 187-220.
Department of Health and Human Services [DHHS]. (1991). Healthy people 2000. Washington, DC: Author.
Dilorio, C., Kelley, M., & Hockenberry-Eaton, M. (1999). Communication about sexual issues: Mothers, fathers, and friends. Journal of Adolescent Health, 24, 181-189.
Donovan, P. (1993). Testing positive: Sexually transmitted diseases and the public health response. New York: The Alan Guttmacher Institute.
Donovan, P. (1997). Confronting a hidden epidemic: The Institute of Medicine's report on sexually transmitted diseases. Family Planning Perspectives, 29(2), 87-89.
Dutra, R., Miller, K., & Forehand, R. (1999). The process and content of sexual communication with adolescents in two-parent families: Associations with sexual risk-taking behavior. AIDS and Behavior, 3, 59-66.
Fisher, T (1993). A comparison of various measures of family sexual communication: Psychometric properties, validity, and behavioral correlates. The Journal of Sex Research, 24, 305-311.
Hatcher, R., Trussel, J., Stewart, E, Kowal, D., Guest, E, Cates, W., & Policar, M. (1994). Contraceptive technology (16th ed.). New York: Irvington Publishers.
Hepburn, E. (1981). The father's role in sexual socialization of adolescent females. Journal of Early Adolescence, 1, 53-59.
Hosmer. D., & Lemeshow, S. (1989). Applied logistic regression. New York: Wiley.
Hutchinson, M. K. (1998). Something to talk about: Sexual risk communication between young women and their sexual partners. Journal of Obstetric, Gynecologic, and Neonatal Nursing, 27, 127-133.
Hutchinson, M. K. (1999). Individual, family, and relationship predictors of young women's sexual risk perceptions. Journal of Obstetric, Gynecologic. and Neonatal Nursing, 28, 60-67.
Hutchinson, M. K. (2000a, July). Influence of family structure and communication variables on the sexual risk behaviors of late adolescent females. Paper presented at the National Institute of Mental Health Conference on the Role of Families in Preventing and Adapting to HIV/AIDS, Chicago, IL.
Hutchinson, M. K. (2000b. July). Talking to Daddy's little girl about sex: Daughters' qualitative reports of sexual communication and support from fathers. Poster session presented at the National Institute of Mental Health Conference on the Role of Families in Preventing and Adapting to HIV/AIDS, Chicago, IL.
Hutchinson, M. K. (2000c). [Test-retest reliability of PTSRC-III scale]. Unpublished raw data.
Hutchinson, M. K., & Cooney, T. (1998). Parent-teen sexual risk communication: Implications for intervention. Family Relations, 47, 185-194.
Hutchinson, M. K., & Sandall, S. (1995). Congenital TORCH infections in infants and young children: Neurodevelopmental sequelae and implications for intervention. Topics in Early Childhood Special Education, 15, 65-82.
Hutchinson, M. K., Sosa, D., & Thompson, A. (2001). Sexual protective strategies of late adolescent heterosexual females: More than just condom use. Journal of Obstetric, Gynecologic, and Neonatal Nursing, 30, 429-438.
Institute of Medicine (IOM). (1997). The hidden epidemic: Confronting sexually transmitted diseases. Washington, DC: National Academy Press.
Jaccard, J., & Dittus, P (2000). Adolescent perceptions of maternal approval of birth control and sexual risk behavior. American Journal of Public Health, 90, 1426-1430.
Jaccard, J., Dittus, P, & Gordon, V. (1996). Maternal correlates of adolescent sexual and contraceptive behavior. Family Planning Perspectives, 28, 159165, 185.
Jaccard, J., Dittus, P, & Gordon, V. (1998). Parent-adolescent congruency in reports of adolescent sexual behavior and in communication about sexual behavior. Child Development, 69, 247-261.
Jemmott, J., Jemmott, L., & Fong, G. (1992). Reductions in HIV risk-associated sexual behaviors among Black male adolescents: Effects of an AIDS prevention intervention. American Journal of Public Health, 82, 372 377.
Jemmott, J., Jemmott, L., & Fong, G. (1998). Abstinence and safer sex HIV risk-reduction interventions for African American adolescents: A randomized control trial. Journal of the American Medical Association, 279, 1529-1536.
Jemmott, L. S., Catan, V., Nyamathi, A., & Anastasia, J. (1995). African Amer
ican women and HIV-risk-reduction issues. In A. O'Leary & L. S. Jemmott (Eds.), Women at risk: Issues in the primary prevention of AIDS (pp. 131158). New York: Plenum Press.
Kann, L., Kinchen, S., Williams, B., Ross, J., Lowery, R., Grunbaum, J., & Kolbe, L. (2000). Youth risk behavior surveillance-U.S., 1999. Morbidity and Mortality Weekly Report, 49, 5505, 1-96.
Kann, L., Warren, C., Harris, W., Collins, J., & Williams, B. (1996). Youth risk behavior surveillance-United States, 1995. Morbidity and Mortality Weekly Report, 45, SS-4, 1-83.
Kotchick, B., Dorsey, S., Miller, K., & Forehand, R. (1999). Adolescent sexual risk-taking behavior in single-parent ethnic minority families. Journal of Family Psychology, 13, 93-102.
Lewis-Beck, M. (1989). Applied regression: An introduction. Newbury Park, CA: Sage.
Meschke, L., Bartholomae, S., & Zentall, S. (2000). Adolescent sexuality and parent-adolescent processes: Promoting healthy teen choices. Family Relations, 49, 143-154.
Miller, B. C., Norton, M. C., Fan, I., & Christopherson, C. R. (1998). Pubertal development, parental communication, and sexual values in relation to adolescent sexual behavior. Journal of Early Adolescence, 18, 27-52.
Miller, K., Clark, L., Wendell, D., Levin, M., Gray-Ray, P, Velez, C., & Webber, M. (1997). Adolescent heterosexual experience: A new typology. Journal of Adolescent Health, 20, 179-186.
Miller, K., Forehand, R., & Kotchick, B. (1999). Adolescent sexual behavior in two ethnic minority samples: The role of family variables. Journal of Marriage and the Family, 61, 85-98.
Miller, K., Levin, M., Whitaker, D., & Xu, X. (1998). Patterns of condom use among adolescents: The impact of mother-adolescent communication. American Journal of Public Health, 88, 1542-1544.
Panchaud, C., Singh, S., Feivelson, D.. & Darroch, J. (2000). Sexually transmitted diseases among adolescents in developed countries. Family Planning Perspectives, 32(I),.
Rotheram-Borus, M., Jemmott, L. S., & Jemmott, J. (1995). Preventing AIDS in female adolescents. In A. O'Leary & L. S. Jemmott (Eds.), Women at risk: Issues in the primary prevention of AIDS (pp. 103-130). New York: Plenum.
Santelli, J., Brener, N., & Lowry, R. (1998). Risk factors for multiple sexual partners among U.S. adolescents. Family Planning Perspectives, 30, 271-275. Santelli, J., DiClemente, R., Miller, K., & Kirby, D. (1999). Sexually transmitted
diseases, unintended pregnancy, and adolescent health promotion. Adolescent Medicine: Prevention Issues in Adolescent Health Care, 10, 87-108.
Whitaker, D., & Miller, K. (2000). Parent-adolescent discussions about sex and condoms: Impact on peer influences of sexual risk behavior. Journal of Adolescent Research, 15, 251-273.
Whitaker, D., Miller, K., May, D., & Levin, M. (1999). Teenage partners' communication about sexual risk and condom use: The importance of parent-- teenager discussions. Family Planning Perspectives, 31(3), 117-121.
M. Katherine Hutchinson**
**Center for Urban Health Research, University of Pennsylvania, 420 Guardian Drive, Philadelphia, PA 19104-6096 (email@example.com).…