Wayne State University, Detroit, USA

Introduction

Sex education is no doubt an intense debate, and issues of who should teach it and when are central to the argument. An appropriate theoretical model with which to conceptualize sex education is an ecological approach (e.g. Bronfenbrenner, 1989).

In an ecological framework, multiple contexts of influence on a child’s/adolescent’s life are considered when examining their development. The most immediate set of contexts (microsystem) that may have an influence on sexual outcomes (defined here as sexual knowledge, attitudes and behavior) are peers, family, schools and other adults in adolescents’ lives. The relation between these contexts are the next

level in this approach (mesosystem), and collaboration of sex education sources could be reflected at this level. At the third level (exosystem), contexts influence children though they do not directly involve children. Media’s portrayal of sexuality can serve as an example of a potential influence on sexual development. The final level (macrosystem) involves overarching beliefs, behavior, institutions, etc. that

influence development, and in this case, American society’s general discomfort with sex education and its attitudes toward sex education are likely to play a significant, though probably unmeasurable, role in adolescents’ sexual outcomes. The purpose of the current study was to investigate sex education from multiple contexts in an effort to understand combined effects.

Impact of schools

Research on the impact of formal school sex education on sexual outcomes has been mixed. For example, studies have shown that school-based sex education programs have demonstrated positive effects on adolescent sexual knowledge (Finkel & Finkel, 1985; Melchert & Burnett, 1990). However, they have revealed that adolescents’ sexual attitudes (Finkel & Finkel, 1985) or behavior (Maslach & Kerr, 1983) were not influenced. Contraceptive use has similarly been found to be inconsistently related to school-based sex education, with some studies finding a positive influence (e.g. Kvalem et al., 1996) and others finding no influence at all (e.g. Taylor et al., 1989). This sampling of findings appears to be representative, and suggests that consensus has not yet been achieved regarding the impact of school-based sex education in general.

Impact of parents

There are also likely to be other sources of sex education that influence adolescents’ sexual development. Research has also reported that adolescents perceived that their preferred sources of information were not always the most accessible (de Castilla et al., 2001). Indeed, some studies have examined the independent impact of sources such as parents, peers and media. For example, teenagers’ sexual attitudes and

behavior have been shown to be related to greater communication from their parents (i.e. Fisher, 1986; Miller & Fox, 1987; Leland & Barth, 1993; Whitaker & Miller, 2000). DiIorio et al. (1999) reported that early adolescence was a period of more discussions about sexuality with mothers than other developmental times, and that conversations with fathers about sex were generally limited.

In a study that explored the sources of first sexual discussion with children, Moore and Davidson (1999) revealed that in a group of 717 unmarried college women, those who talked about sex with their mother first were more likely to use contraception at first sexual intercourse than those who learned from peers or teachers.

However, findings by O’Sullivan et al. (2001) suggest that for some daughters, sexual development causes conflict, which impairs communication with their mothers and necessitates obtaining sex education from other sources. Another study showed that females are more strongly influenced by their mothers than are males (Ballard & Morris, 1998), and yet others have demonstrated the importance of

fathers above and beyond mothers in influencing sexual behavior (Dittus et al., 1997). Mostly, however, it is unclear which sexuality outcomes are influenced by various sources of sex education, and which sources are most related to sexual outcomes. Again, sexual outcomes are defined here as sexual knowledge, attitudes and behavior.

Impact of peers

The literature provides a wealth of data supporting the influence of peers on adolescents’ sexual behavior. For example, when adolescents perceive that their friends are sexually active, they report higher levels of sexual activity (Brooks-Gunn & Furstenberg, 1989). Adolescents have also reported greater sexual activity when they had older siblings modeling sexually active behavior (Rodgers & Rowe, 1988;

East et al, 1993). Regardless of the actual effect of this peer communication, same-age peers have been found to be more influential than older peers (Ballard & Morris, 1998). Apart from the frequency of sexual behavior, other effects of sexual information from peers have been explored. In the same study as noted earlier, Moore and Davidson (1999) found that women who first learned about sex from peers were less likely to consider love a prerequisite for sex than the parent-educated group.

Impact of media

Research has also explored the impact of media on adolescent sexual attitudes. In a study that analyzed teenagers’ top 10 programs, Ward (1995) found that more than a quarter of the shows contained interactions of sexual content. In these shows, a majority of interactions involved men seeing women as sex objects, sex as a competition, sex as a defining aspect of masculinity and sex as fun and exciting. In

another example of media influence, Ward (2000) reported a positive relationship between time spent viewing videos and prime-time television, and the tendency for females to view dating as a game, to perceive women as sex objects and to believe that the behavior of males is motivated by sex. Research by Steele (2000) found that adolescents are often exposed to sexual scripts and value systems in media that may be contrary to the values and beliefs of their parents. The specific way in which media influences adolescent cognition and development has been theorized to be through identity-based needs which determine which messages are selected for attention, how the information is interactively interpreted and processed, and how the messages are incorporated or omitted in application in daily life (Steele, 2000).

Source and timing of sex education

Clearly, education about sexuality does not occur from a single source. From an ecological perspective, this assumption is not even possible. Therefore, an analysis of the combined effects of multiple sources, or a comparison between sources, is warranted. To date, only a few studies have attempted this. For example, relative to school-based sex education, sex education administered by parents has been found to

have a more direct behavioral impact in terms of decreasing sexual behavior (Fisher, 1986; Warren & Neer, 1986). A later study expanded on this research by examining additional sources of sex education (Somers & Gleason, 2001). They surveyed high-school male and female adolescents and asked whether or not they had learned from any or all of the following sources: media, peers, professionals, school personnel and family. Results indicated that more frequent sexual behavior was significantly explained by having experienced less education from school and more education from non-sibling family members. Research by Wellings et al. (1995) found no relationships between school-based sex education and the hastening of the onset of sexual activity.

The age at which various sexual topics are best introduced is also an area of both research and practical concern. However, Somers and Gleason (2001) did not consider the ages at which the education had occurred. The focus on the timing of the first sexual intercourse is especially important because early sexual activity is associated with more risky sexual behavior such as having multiple partners (Seidman

et al., 1994). Also, even if sexual activity occurs at earlier ages, it assumed that cognitive maturity would not develop any earlier than is typical. Thus, this may increase the length of time at risk for adolescent pregnancy and sexually transmitted infections. However, relatively few studies have examined the relation between the age at which sex education is taught and the occurrence of sexual behavior during

adolescence. For example, Miller et al. (1998) showed that teens were more likely to use condoms during initial and subsequent sexual experiences if mothers and adolescents discussed the use of condoms before the adolescent had sexual intercourse for the first time. One of the problems, as suggested by Somers and Paulson (2000) and others (Lindberg et al., 2000), though, is that parents often wait to initiate communication until after their adolescents have already started having sexual intercourse. Somers and Eaves (2002) attempted to explore the relation between the timing of sex education and sexual outcomes. Results indicated that, overall, earlier education was shown to have several desirable effects, and in no case was it related to more frequent sexual behavior (i.e. oral sex, intercourse) or earlier onset of sexual

intercourse. Because a larger percentage of teenagers are having sexual intercourse and these teens are initiating sexual intercourse at younger ages (Haffner, 1996), the question of when sex education should begin is clearly important.

In fact, the issue of timing of the provision of sex education appears to be a significant variable which itself could be conceptualized as a protective intervention. It also seems clear from prior research that the most externally generalizable study is likely to be one that includes both source of education and timing of education variables.

Source and timing of sex education are likely to be related to one another and have differing influences on outcomes. It is important to determine which is more powerful in explaining sexual outcomes, or whether the two in combination explain the most desirable outcomes. Thus, the purpose of the current study was to expand the existing body of literature by exploring the comparative contribution that multiple sources of education about sexual topics (peers, media, school and other adults), and the timing of this sex education, make in explaining adolescents’ sexual attitudes and behavior choices. The variables of gender and race were also explored for potential differences.

Method

Participants

Participants were 672 adolescents (231 males, 413 females, 28 unreported) in the 9th through 12th grades (even distribution of students across grade levels) of one suburban and two urban high schools. The average age was 15.97

Measures

Demographic questions yielded information about age, grade, race and gender. Additional measures of sources of sex education, timing of sex education, and sexual attitude and behavior outcomes were also collected.

Sources of sex education. A measure of sources of sex education was created for the purposes of this research. Ten items reflecting various sexual topics were included: reproductive system, menstruation (answered by females)/nocturnal emissions (answered by boys; the term ‘wet dreams’ was used on the actual surveys), petting, oral sex, sexual intercourse, the importance of using birth control, consequences of teen pregnancy, sexually transmitted infections (STIs), love and/or marriage, and whether premarital sex is right or wrong. Adolescents were asked to indicate the source(s) from which they learned about each topic. The following four source categories emerged: ‘Peers’, ‘Media’ (all forms), ‘School’ sex education programs and ‘Adults’ other than sex education teachers. For each source and for each of the

10 topics, the responses were coded as 1 if they identified ever having learned from that source, and 0 if they never learned from that source.

Timing of sex education:

For each of the 10 sex education topics, adolescents were asked to report in which school grade they remember being when they first learned about each topic.

Sexual attitudes and personal sexual values:

A five-item scale was used to assess adolescents’ attitudes toward premarital sexual intercourse (i.e. ‘Unmarried people should not have sex’) and an additional five-item scale was used to assess clarity of personal sexual values (i.e. ‘I know what is right and wrong for me sexually’). Both were taken from the Mathtech Attitude and Value Scales (Kirby, 1990). For all

items, adolescents responded using the same five-point scale (1_Strongly Disagree, 5_Strongly Agree). Kirby (1990) yielded Cronbach’s alpha coefficients of .94 for boys and .73 for girls. In the current study, alphas were .88 and .62, respectively.

The Pearson correlation coefficient shrinks artificially due to a small number of items and, therefore, the Spearman-Brown Prophecy formula correction procedure was used for the alpha of girls’ personal sexual values. This formula is used when the number of items in a subscale is relatively small. Based on the idea that reliability is increased with test length, this formula demonstrates what the reliability of the instrument would be if additional similar quality items were added to the scale. In this case, projecting a more typical number of items for subscales of this type of 10, an alpha of .77 resulted. This is a projected figure to help interpret the original alpha that was obtained, and not the actual alpha to be reported.

Sexual behavior:

The original 18-item instrument used to derive the measures in this study was the SKAT-A (Sex Knowledge and Attitudes Test for Adolescents; Leif et al., 1990).

This scale included a variety of general sexual behavior, including reading pornography and talking with peers and parents about sex. This was modified in subsequent studies to 10 items and yielded internal consistency reliability coefficients of .86 (Somers & Paulson, 2000) and .88 (Somers & Fahlman, 2001). In the current study, this measure was further reduced to five overt sexual behaviors (Dating, Kissing, Petting/Fondling, Oral Sex, Sexual Intercourse). Adolescents responded on a five-point scale to reveal how frequently they have engaged in each behavior in the past year (ranging from ‘never’ to ‘daily’). These were used as individual items and also summed to create a total sexual behavior score. Higher total scores represented greater amounts of general sexual behavior. In the current sample, the total sexual

behavior scale resulted in an overall alpha of .83. Age of first sexual intercourse and number of sexual partners were also used as individual dependent measures.

Self-report, while some argue is questionable, is also argued to be a more valid method of obtaining such personal information than other methods (e.g. Stone et al., 1999) like interview or observation.

Procedure:

Students who were permitted by their parents to participate were then also asked for their own assent to participate. Participation was voluntary. Students completed the survey in their classrooms under supervision of the researcher and/or the teachers.

It is not believed that the students who participated were significantly different than those who did not participate. This is primarily because rates of behavior for the samples were similar to American national averages, and thus the sample is believed to be representative of the larger adolescent population (e.g. Grunbaum et al., 2002). The University Institutional Review Board approved all procedures.

Discussion

The purpose of this study was to explore the roles of multiple sources of education about sexual topics (peers, media, school and other adults) and the timing of this sex education in explaining adolescents’ sexual attitudes and behavior choices. Overall, both timing and source of education were important in explaining variance in outcomes. Several themes emerged in the results. Generally, the pattern was

consistent that later learning from any source and less learning from school sex education about various sexual topics (including all sexual behavior, importance of using birth control and consequences of teen pregnancy) was statistically significantly predictive of more frequent sexual activity, including oral sex and sexual intercourse. Other sources of sex education (peers, media, other adults) seemed to be less consistently important in their relations with outcomes in a typically preferred direction. This is all discussed in detail as follows. ‘Desired outcomes’ is defined in this interpretation as less risk-taking and more conservative attitudes, as both are related to lowered pregnancy and STI rates (e.g. Somers & Paulson, 2000).

It is apparent that school is an important setting in which to learn about these topics in a way that has desired relations with teens’ sexual behavior. Earlier introduction of topics also appears to be an important contributing factor to desired outcomes. Past literature suggests that teenagers are often educated after it is too late (Somers & Paulson, 2000). Research by Clawson and Reese-Weber (2001)

found that it was the timing of the discussion between mothers and adolescents that predicted sexual behavior; that is, when communication occurred before first intercourse it predicted age of first intercourse and number of pregnancies. Based on others’ research and the current study results, it appears that a preventive approach to sexual risk-taking would benefit adolescents. Specifically, rather than waiting until post-puberty (e.g. high-school grade/ages 15–18) to educate about these issues, it

would be useful to include developmentally appropriate sex education in our curricula earlier on. The issue of developmental appropriateness would be based on children’s biological, cognitive and social development and capacities at each age/ grade level, and the topics taught would be tailored appropriately. However, owing to many factors, efforts to educate about sexuality are often resisted. In one study, 46% of 1789 teachers noted pressure from the community, parents or school administrators as a primary problem in the teaching of information about sexuality (Landry et al., 2000).

Another theme observed in the current study was that adolescents consistently reported clearer personal sexual values (knowing what is right and wrong sexually for them) when more education about various topics occurred from other adults, and they were less clear about their personal sexual values when education came from peers and media. Implications for adolescents’ decision-making are significant.

Research has discussed the importance of adolescents identifying the consequences of sexual intercourse and unprotected intercourse (Duerst et al., 1997). When in risk-taking situations, those adolescents who are clear about their own values and intentions beforehand have been found in at least two studies to make healthier and safer decisions than those who are unclear about these issues (e.g. Miller et al., 1998;

Clawson & Reese-Weber, 2001). The data from the current study suggest that other adults in adolescents’ lives, and not peers and media, may be more successful at helping adolescents develop clear values and intentions regarding their sexual lives.

However, such an interpretation would require causal inferences from the data, and at this point, interpretations are correlational.

In addition to these consistent themes observed across groups, noteworthy subgroup patterns also surfaced. Among males, timing of sex education did not appear to be a significant contributor to outcomes. However, more school-based education about a few key behavioral topics (sexual intercourse, oral sex) was related to less sexual behavior in adolescence and less adolescent approval of premarital sex.

School may help to provide important information and perhaps alternatives for boys. For girls, it was consistently clear that schools provide an important source of information to them. For example, more learning from school about teen pregnancy and at any earlier time combined to significantly explain less frequent sexual behavior, and more education from other adults about love/marriage and whether

premarital sex is right or wrong significantly explained (above other variables) less adolescent approval of premarital sex. Perhaps it provides girls with social support for resisting sexual pressure, particularly when the education occurs in the realistic setting in which the real or perceived pressure occurs. For girls, earlier learning about overt activities like sexual intercourse and the importance of using birth

control was also related to less frequent sexual behavior. This suggests that girls may benefit from earlier learning about these issues in order to establish their values and develop an identity around resisting sexual pressure instead of succumbing to it.

Racial/ethnic group differences were important to explore given demonstrated differences found in prior literature. An understanding of what these differences are will likely facilitate more effective prevention and intervention efforts. In this study, correlations for African-American adolescents indicated that school-based education is more related to desired outcomes for these adolescents than is peer-based

education. However, more work needs to be conducted in order to better understand the relative lack of effects, including identifying which other variables may be making contributions to development. It remains unclear why African-American adolescents do not demonstrate similar patterns of influence of source or timing as did other racial/ethnic groups. It is also intriguing that there is a relative lack of

findings for this group, which may suggest that the variables tapped in this study are not the most salient factors involved in the development of sexual behavior for this subgroup.

Learning from school and earlier learning about some topics were generally related to desirable sexual outcomes among Caucasian adolescents as well. Unique to this group, though, earlier education about love and/or marriage and the consequences of teen pregnancy was related to less frequent sexual behavior. Perhaps these adolescents hear about realistic consequences for behavior and they are taught

about concepts of love and marriage earlier on, and then they learn to use that education to engage in alternatives to the sexual behavior that they see all around them. However, this requires specific testing, and the proposed directionality of relations would require formal testing through research.

The results for the Hispanic adolescents in this sample were somewhat puzzling.

Few significant regression analysis results emerged for this group, though those that did were some of the strongest of all subgroups in the sample, but there were many significant correlation results. With few exceptions, more learning from various sources and learning earlier were related to more risky sexual behavior, which is contrary to expectation. Why might more education be related to more behavior

among these adolescents? This is discussed in more detail below, but also, again, directionality of relations must be considered, and longitudinal research is needed in order to tease this out.

Cultural idiosyncrasies should also not be overlooked. As a whole, this group tends to practice religious and sexual conservatism through families and churches (e.g. McEachern & Kenny, 2002; Romo et al., 2002). Thus, adolescents may not be obtaining information needed to help them understand their own biological, cognitive and emotional changes, particularly those associated with sexual development.

Because of cultural restrictiveness around sex education and behavior, it is likely, therefore, that behaviors are coming first and then sex education follows either out of adolescents’ own curiosity or out of other adults who recognize the necessity to educate. School is perhaps an ideal context in which to provide basic information to these adolescents that they may not be getting in their relatively restrictive home environments.

Some of the significant findings for several subgroups were actually counterintu-itive, i.e. in some cases more education was related to more sexual behavior.

However, because these findings were accompanied with positive findings as well, this may best be explained by pubertal timing and maturation. Earlier maturation generally prompts males to seek education earlier than others may be ready to seek it. Thus, the issue of directionality of relations must again be considered.

Specifically, the role of maturation, and simply age, has been show to be significant in the development of sexuality (e.g. Somers & Paulson, 2000), and it is probable that there is a bidirectional relationship between physical maturation (becoming ready for and interested in sexual relations) and seeking out information about sex if it is not already being provided.

Interestingly, media provided little explanation of variance in outcomes, and it was inconsistent in whether it was related to more or less of certain variables.

Clearly, more investigation into the role of various forms of media as contexts of sex education is warranted. However, it is also possible that the exosystem level is not as strong or direct an influence as more proximal contexts such as family, peers and schools. Although not intentionally part of the research purpose, it was also interesting in and of itself that there were no race by gender interactions in any of the

outcome variables, and that there were no racial or gender differences in adolescents’ clarity of personal sexual values. Male and females were different in their outcomes, consistent with past research, but it did not also vary by race. Females, regardless of race, were similar to each other, and males, regardless of race, were similar to each other. Conversely, racial/ethnic differences appeared between the three groups sampled in this study, but within each race, gender differences did not appear.

These results must also be considered in light of several limitations, the greatest of which is the need for a longitudinal design. In some subgroups the sample sizes could be increased. The measurement of sex education sources and timing may be improved in future research. Inclusion of other variables may help to explain larger proportions of variance, as well as more inclusive identification of the mechanisms

at work for all racial/ethnic groups. It also needs replication not only through similar methodology but also, and most ideally, through experimental methodology, so that directionality and causality among relations can be established. This would likely be a methodologically challenging task, and thus may explain why so little detailed research has been conducted on sources and timing of sex education. Although the data-collection method used here could be improved, it nevertheless provides a solid

place in which to begin investigations of this topic.

Despite limitations, this study provides a more detailed investigation than has been done before of the relation between multiple sources of sex education and the timing of sex education and adolescents’ sexual attitudes and behavior. In general, the pattern was consistent that school-based sex education, when it occurs, is related to typically desired outcomes. Other sources of sex education (peers, media, other adults) seemed to be less consistently related to outcomes, and some, peers in

particular, may even have generally undesired relations with outcomes (i.e. in some cases, more education from peers was related to more behavior and more liberal attitudes). However, though not as prevalently significant a contributor as schools, adult influence appears to be related to desired outcomes and peer influence to undesired outcomes. Earlier education about various topics was also more consistently

related to desired outcomes. Important gender and race patterns were also observed, suggesting that prevention and intervention approaches may require some customization. However, this must be balanced with an avoidance of stereotyping.

In summary, this study provides important information applicable to research, theory and practice. Future research is needed to confirm and expand the findings.