Can Sex Education Delay Early Sexual Debut?

Abstract

In this study, we examine whether a nine-lesson sex education intervention, “Get Real: Comprehensive Sex Education That Works,” implemented in sixth grade, can reduce the number of adolescents who might otherwise become “early starters” of sexual activity (defined as heterosexual intercourse) by seventh grade. Participants were 548 boys and 675 girls who completed surveys in both sixth grade (baseline) and seventh grade (follow-up). The sample was 35% Latino, 32% Black, 24% White, 3% Asian, and 6% biracial. Students randomly assigned to the control condition were 30% more likely to initiate sex by follow-up when controlling for having had sex by sixth grade, demographic variables, and a tendency to give socially desirable responses. This finding is noteworthy because previous research has identified early starters to be prone to poor outcomes in sexual health, family formation, economic security, and incarceration and few middle school interventions have shown an effect on behavioral outcomes.

Keywords

curriculum health promotion intervention/prevention sexual behavior (including pregnancy)

In this brief report, we examine whether a comprehensive sex education intervention implemented in sixth grade can reduce the number of early adolescents who might otherwise become “early starters” of sexual activity by seventh grade. Early initiation of sex is a key health risk indicator because it increases a young person’s cumulative exposure to sexually transmitted infections (STIs) and likelihood of unintended pregnancy (von Ranson, Rosenthal, Biro, & Succop, 2000). Even more importantly, early sexual activity has been associated with a riskier developmental trajectory into adulthood because of its association with other health risk behaviors such as substance use (Coker et al., 1994; O’Donnell, O’Donnell, & Stueve, 2001), particularly cigarette smoking (Duncan, Strycker, & Duncan, 1999), and poor outcomes for adult sexual functioning and health (Magnusson & Trost, 2006; Sandfort, Orr, Hirsch, & Santelli, 2008), with increased risk of family dysfunction, poverty, substance abuse, and incarceration (Pergamit, Huang, & Lane, 2001). Consequently, reducing rates of early sexual initiation through interventions can have important health and well-being benefits for individuals as well as public health and social welfare benefits for society across the life course.

Abstinence-only versus comprehensive sex education interventions

Public discourse in the 1990s and the first decade of the 21st century has been witness to debates on the merits of comprehensive compared with abstinence-only sex education. Some interventions use comprehensive sex education curricula, which stress delaying sexual debut as a healthy choice and also provide information on the correct use of contraceptives for students who become sexually active, whereas others are abstinence-based, which advocate delaying sexual debut until marriage and do not provide information on contraception.

The state-of-the-art approach to evaluating sex education programs is a randomized control experimental design that does not allow self-selection into the intervention or comparison group. The U.S. Department of Health and Human Services’ sponsored review of programs to reduce teen pregnancy, STIs, and associated sexual risk behaviors reserves the quality rating of “high” for randomized controlled trials (RCT) that randomly assigned subjects to the study’s groups because this approach provides the strongest evidence that differences in the outcomes between the intervention and control groups can be attributed to the intervention (Mathematica Policy Research & Child Trends, 2010). To date, RCT evaluations of middle school sex education programs have produced only modest evidence for the effectiveness of comprehensive or abstinence-based interventions to delay early sex initiation. Among comprehensive sex education programs evaluated using RCT, Tortolero and her colleagues (2010) showed a general effect for their evaluation of It’s Your Game—Keep It Real. An evaluation of Draw the Line/Respect the Line with a racially mixed but predominantly Latino sample found that boys in the intervention group were less likely to have initiated sexual intercourse or to have had sex in the past 12 months compared with boys in the control group, but there were no program effects for girls (Coyle, Kirby, Marin, Gomez, & Gregorich, 2004). The reverse was found for an evaluation of The McMaster Teen Program with a predominantly African American sample, which showed delay of sexual behavior only for girls (Aarons, Jenkins, Raine, & El-Khorazaty, 2000). Another evaluation with an African American sample showed no behavioral effects (Koo et al., 2011).

Behavioral findings for the effectiveness of abstinence-based programs are similarly mixed. Jemmott, Jemmott, and Fong (2010) studied an African American sample and found effectiveness in changing students’ sexual behaviors, with reduced sex among students in the abstinence intervention condition. This finding contrasts with a larger-scale evaluation of four abstinence-only programs, which showed no behavioral differences between intervention and control groups (Trenholm et al., 2007). The inconsistent evidence of effectiveness of sex education programs aimed at middle school youth underscores the importance of identifying sex education curricula that reduce sexual risk-taking behavior.

The Present Study

In this article, we assess the behavioral impact at the beginning of seventh-grade relative to baseline data obtained at the beginning of sixth grade of a comprehensive sex education intervention, Get Real: Comprehensive Sex Education That Works (Planned Parenthood League of Massachusetts, 2010). The impact evaluation involved random assignment of 24 middle schools in the greater Boston area into intervention and control conditions. Get Real was implemented in intervention classrooms through nine 45-min lessons in the sixth grade by trained educators. The focus of the present study is on whether a sixth-grade comprehensive sex educational intervention can delay early sexual debut.

The Curriculum

Get Real: Comprehensive Sex Education That Works is designed to deliver factually accurate, culturally sensitive, and age-appropriate information on a range of topics pertaining to sexual health and relationships. The development of Get Real, its content, and implementation have adhered to the 17 characteristics of effective curriculum-based interventions proposed by Kirby (2007, pp. 127-135) based on his review of the extant research on sex and HIV/STI education programs. The development of Get Real received input from multiple experts in theory and research on sex and HIV/STI education. The curriculum was created using a logic model approach that specified the health goals, the types of behavior affecting the goals, the risk and protective factors affecting these behaviors, and activities to change relevant risk and protective factors. The activities incorporated into the curriculum are consistent with community values and resources available in a typical middle school in the greater Boston area. The curriculum was pilot tested in three middle schools and subsequently subjected to a formative evaluation in five middle schools before being evaluated for its impact.

The long-term goal of Get Real is to promote sexual health through delaying sexual activity and increasing the correct and consistent use of protection among those who are sexually active. Get Real addresses sexual psychosocial risk and protective factors that affect sexual behavior (e.g., knowledge, perceived risks, values, attitudes, perceived peer norms, and self-efficacy). The curriculum’s change model is consistent with the tenets of the widely used theory of planned behavior (Ajzen, 1991, 2006), namely, attitudes, beliefs, self-efficacy, and behavioral intentions through teaching positive attitudes toward delaying sexual onset and belief in one’s efficacy to act on those attitudes that can be formulated into intentions. These causal relationships identified by the theory of planned behavior are viewed to take place within multiple ecological contexts (Bronfenbrenner, 1979) in which an adolescent’s sexual development is imbedded. Among the different contexts which bear on the developing adolescent’s sexual behavior, the immediate context, where relationships that have the potential to turn into sexual situations take place, is identified as a pivotal one. Therefore, the curriculum teaches skills such as perspective taking, refusal skills, and assessment of long-term consequences of one’s actions that can be applied in potentially sexual situations. The change model postulates that the curriculum’s emphasis on relational skill building will serve adolescents well in translating to action their intentions to achieve sexual health as taught in the curriculum. It is expected that adolescents who have enhanced relationship skills will make healthy choices regarding sexual relationships.

The lesson plans include multiple activities to change targeted risk and protective factors that actively involve students and are appropriate to middle school students’ developmental age and sexual experience. The lessons cover topics in a logical sequence, as vetted by experts in theory and research on sex education, moving from creating a safe classroom climate in the first lesson to communication and refusal skills in the second, followed by relationships and personal boundaries, male anatomy and reproduction, female anatomy and reproduction, puberty, abstinence, decision making and values, and a review in the ninth and last lesson. There is also a family activity component associated with the first eight lessons designed to increase parent–adolescent comfort in communication so that parents or adult guardians can pass on their family values about relationships and sex and adolescents can go to a trusted family member with their questions.

Method

Participants

Participants were 1,223 adolescents who had valid surveys at both baseline and follow-up. The sample was 55% female, with intervention participants comprising 57.6% of the sample. Average age of participants at the time of the follow-up survey in seventh grade was 12.88 (SD = 0.59) and ranged from 10.33 years to 15.92 years. The sample was 34.5% Latino, 32.3% Black, 23.5% White, 3.2% Asian, and 5.6% biracial.

Procedure

For inclusion in the study, schools had to be within a 25-mile radius of Boston, include sixth-, seventh-, and eighth-grade classrooms, and have at least two and at most 10 sixth-grade classrooms. To be eligible, schools could not be involved in a formal sex education curriculum evaluation or already implementing Get Real and had to respond to the request for inclusion in the evaluation within the academic year. Approximately 140 middle schools met the screening criteria. Superintendents of the public school systems and school heads of charter and private schools were contacted to invite middle schools to participate in the impact evaluation. Initial information meetings were held with 80 schools willing to consider participation in the evaluation. Of the 24 schools that signed on to participate, one school withdrew prior to the follow-up data collection and was not included in the analyses reported here. The final sample included 12 public schools, nine charter schools, and two private schools. Except for one public, one charter, and one private school, the student body in each school was composed primarily of minority students. All but two schools were located in an urban area. Randomization took place at the school level; schools were randomly assigned to intervention or control conditions immediately after their chief school administrator signed a memorandum of understanding to participate in the evaluation.

Parent/guardian informed consent forms were distributed in multiple languages and obtained from participating students prior to survey administration. Schools determined whether active or passive consent was used. For each wave of data collection, in active consent schools, a study description and consent form was sent home with the student to be read, signed, and returned by a parent or guardian (forms not returned were considered consent not given). In schools using passive parental consent, the study description and consent form was sent home with the student and parents only returned the signed form in the event that they did not give consent. In passive consent schools 23 parents/guardians refused consent, equally distributed between control and intervention conditions. None of the intervention schools elected active parental consent, whereas in the control condition 167 students did not return signed parental permission slips when active consent was requested. Participating students provided written assent prior to completing the survey, after receiving written and oral information about the study, with only students who completed assent forms invited to take the survey. All but eight students agreed to participate in the evaluation across both conditions. Students without parental consent or their own assent were given supervised time outside of class while the survey was being administered. All procedures were approved by Wellesley College’s Institutional Review Board and the Review Board of one of the larger school districts involved and adhered to all human subjects protection.

In the intervention schools, sex education was taught by educators trained by the developer of the curriculum. The sex education class was typically taught during health or physical education classes, with one school using an advisory period for this purpose. In intervention schools the baseline and follow-up surveys were administered before Get Real lessons for that grade were taught. In control schools, sex education was taught as it had been in previous years. There, the survey was administered during a class period or home room period of the school’s choosing, although a few schools chose to group classes together in a cafeteria or auditorium to complete the survey. Members of the evaluation team administered the surveys. The average amount of time between assessments was 1 year, with a range of 9 to 13 months. Twenty-one students took the survey in Spanish at baseline and 25 at follow-up; all other students filled out an English version. Similar to the baseline data collection, the number of parents who opted out and students who did not give assent in follow-up was low (~1%) and only 33 students did not return parental consent slips in control schools with active consent—all intervention schools continued to employ passive consent.

Measures

Sexual behavior

In both baseline and follow-up surveys students were asked whether they had ever engaged in heterosexual intercourse: “Have you ever had sex?” and were provided a definition of sex that stated “Having sex means when a boy puts his penis inside a girl’s vagina. Some people call this ‘making love’ or ‘doing it.’”

Age

Seventh-grade age was calculated by subtracting participants’ birth date from the date of data collection.

Race/ethnicity

Race and ethnicity were based on follow-up survey data. Participants were given a wide range of race and ethnicity choices, asked to choose all that they felt applied to them and were also given the option of writing in the race-ethnicity category that they felt best represented them. For the present analyses, students who chose any Black ancestry such as “African American” or wrote in, for example, “Ghanaian” were coded as Black. Students who chose any European American ancestry such as “White” or “Irish American” were coded as White. Students who chose any Asian ancestry such as “Chinese American” or “Cambodian” were coded as Asian. Students who reported being biracial or chose more than one race were coded as biracial/multiracial. To be consistent with Census definitions that Hispanics can be of any race (Ennis, Rios-Vargas, & Albert, 2011), students who noted any Latino ethnicity were coded as Latino. However, when students checked biracial as well as Latino and/or one or more other races, they were coded as biracial.

Two-parent family

Family structure was based on follow-up survey responses to the question “Who do you live with?” Students who chose the response “two parents in one place” were coded as having a two-parent family structure as opposed to those who chose other responses including “one parent,” “two parents in different places,” “grandparents or other family members,” or “other.”

Social desirability

An abbreviated measure of social desirability from Baxter and colleagues (2004) was incorporated into the survey and included five items. Examples are “I have never hated another person,” and “I am always polite even to people who are not nice.” Items were scored 0 for a nonsocially desirable response and 1 for a socially desirable response, with possible scores from 0 to 5. Item scores were summed, with higher scores representing more socially desirable responses. This sum score was included as a control variable in all analyses.

Analytic Plan

Logistic regression with listwise deletion on the outcome variable was used to test the hypothesis that intervention status predicts lower probability of having had sex by follow-up. The standard errors of all parameters were adjusted for the complex sampling design at baseline (students nested within classrooms). For intervention students, this was the class in which they received the intervention. For control students, this was the class in which they were surveyed. Male gender and older age are known risk factors for being sexually active, whereas living with two parents is a protective factor (Abma, Martinez, Mosher, & Dawson, 2004). At baseline these characteristics differed significantly across intervention and control groups. In addition, there were significantly more Black and Latino students in the control group compared with the intervention group (see Table 1). Therefore, gender, age, family structure, and race were included as controls in the models, in addition to social desirability and reports of having had sex at baseline. Self-report of having had sex at baseline was used as a control variable to preserve the integrity of the randomization design. Omitting from these models students who had sex by baseline, while intuitively appealing, would bias the sample.

Table 1.

Descriptive Statistics and Tests for Mean Differences Between Intervention and Control Samples.

	Intervention (n = 708: 57.6%); M (SD)	Control (n = 525: 43.4%); M (SD)	t test
Age	12.81 (57)	12.99 (60)	5.40**
Social desirability (0 to 5 range)	1.43 (.27)	1.51 (.28)	0.74

	(%)	(%)	χ²

Female sex	59.7	40.3	3.23
Black	27.4	38.9	16.89**
White	32.7	11.2	71.74**
Latino	29.5	41.2	16.81**
Asian	3.8	2.3	2.36
Biracial	5.5	5.6	0.00
2-parent family	57.1	48.0	9.42**
Had sex at baseline	5.1	4.6	0.09
Had sex at seventh grade (but not at baseline)	5.3	8.5	4.26**

p < .05. **p < .01.

We tested five distinct models. The first model included only baseline sexual activity, social desirability, age, gender, and intervention status as predictors. The second model included these variables and also added in two-parent versus non-two-parent family status. The third model included all the variables from the second model and added Latino ethnicity. The fourth model removed Latino ethnicity and included Black race. The fifth model included all the predictors, including both Latino ethnicity and Black race simultaneously. The purpose of testing and building separate models was because some of the variables (e.g., family status and race/ethnicity) are typically overlapping risks, and we wanted to demonstrate the unique effect each predictor had on the outcome variable, in conjunction with the intervention.

Results

Characteristics of the sample are reported in Table 1. Table 2 presents the odds ratios and 95% confidence intervals for the models. In the first model all variables were significant. Participants who reported having had sex at baseline were more than four times as likely as other participants to report having had sex at follow-up. This is to be expected as the sexual behavior variable captures ever having had sex and contains redundant information about a student’s sexual activity among early starters. Participants reporting higher social desirability scores were less likely to report having had sex. Older participants were about 40% more likely and girls were about a third less likely to report having had sex. In this model, with a statistically significant odds ratio of 1.31, participants in the control group were about 30% more likely to report having had sex than those in the intervention group.

Table 2.

Logistic Regression Models Predicting Sexual Initiation by Seventh Grade.

	Model 1 (N = 1,072)	Model 2 (N = 1,062)	Model 3 (N = 1,055)	Model 4 (N = 1,055)	Model 5 (N = 1,055)
	OR [95% CI]	OR [95% CI]	OR [95% CI]	OR [95% CI]	OR [95% CI]
Had sex at baseline	4.14** [2.74, 6.24]	3.94** [2.56, 6.06]	4.01** [2.61, 6.18]	4.06** [2.69, 6.12]	3.94** [2.56, 6.06]
Social desirability	0.38** [0.22, 0.68]	0.39** [0.21, 0.70]	0.38** [0.22, 0.66]	0.38** [0.21, 0.69]	0.39** [0.22, 0.70]
Age	1.42** [1.21, 1.66]	1.40** [1.18, 1.68]	1.42** [1.18, 1.69]	1.40** [1.18, 1.68]	1.39** [1.17, 1.66]
Gender	0.68** [0.51, 0.89]	0.66** [0.49, 0.88]	0.65** [0.48, 0.87]	0.66** [0.49, 0.88]	0.66** [0.49, 0.88]
2-parent family	—	1.25 [0.98, 1.57]	1.25 [0.96, 1.61]	1.23 [0.98, 1.56]	1.17 [0.86, 1.61]
Latino	—	—	1.26 [0.98, 1.62]	—	1.62** [1.14, 2.30]
Black	—	—	—	1.07 [0.83, 1.38]	1.48* [1.04, 2.10]
Intervention status^a	1.31* [1.02, 1.69]	1.34* [1.04, 1.72]	1.30 [1.00, 1.66]	1.32* [1.03, 1.70]	1.25 [0.97, 1.61]

Note: OR = odds ratio, CI = confidence interval, “—” = variable not included in model. Sample sizes vary because of listwise deletion with missing items.

Intervention = 0, control = 1.

p < .05. **p < .01.

In the second model, two-parent family was added as a predictor variable. Living in a two-parent family did not reach significance, all other variables retained statistical significance, and the intervention effect was similar to Model 1 (OR = 1.34, 95% CI [1.04, 1.72], p < .05). In Model 3, Latino ethnicity was added as a predictor; it was not a significant predictor and other variables remained approximately the same in terms of significance and magnitude of effect. However, the intervention effect on having had sex, while at a similar order of magnitude, failed to reach the conventional level of statistical significance (OR = 1.30, 95% CI [1.00, 1.66], p = .06). Model 4 removed Latino ethnicity and added Black race. Black race was not significant, and other variables remained greatly the same, although in this model, intervention status was significant, such that participants in the control group were about 30% more likely to report having had sex in seventh grade (OR = 1.32, 95% CI [1.03, 1.70] p < .05) than those in the intervention group. In Model 5 both Black race and Latino ethnicity were added and both emerged as significant predictors (OR = 1.48 and OR = 1.62, respectively) and the intervention effect was reduced to a trend level (OR = 1.25, 95% CI [0.97, 1.61], p = .09).

Discussion

Our results show that students receiving sex education as usually taught in their schools were about 30% more likely to have initiated sex by seventh grade than students who were exposed to nine lessons of Get Real in sixth grade. The effect size of 30% difference in sexual initiation is comparable to effect sizes obtained in other comprehensive sex education interventions that target middle school students (e.g., Coyle et al., 2004; Tortolero et al., 2010).

When Latino ethnicity was added to the predictor variables, the intervention effect, reduced to 25%, lost its statistical significance, becoming the level of a statistical trend. This suggests that the difference in sexual initiation between the intervention and control groups coincides with a difference in sexual initiation between Latinos and non-Latinos. Because Latinos are less likely to have had sex by seventh grade, statistically controlling for the overrepresentation of Latinos in the control group reduces the difference in sexual initiation between the intervention and control group. The question remains as to whether this is an effect of sample selection (main effect) or whether the intervention operates differently for Latinos and non-Latinos (interaction effect), a distinction that should be addressed in future research.

The results of the present evaluation suggest that an early stage intervention can reduce the number of early starters of sexual behavior through exposure to a school-based program whose development, content, and implementation closely adhere to the recommended best practices for sex education curricula. Further research is needed to investigate which aspects of the curriculum made it effective and whether the same factors are effective for both boys and girls and across different racial/ethnic groups.

The intervention reported here was implemented by educators employed, trained, and closely supervised by the developers of the curriculum. As such, the current evaluation assesses the effectiveness of the curriculum under conditions of high fidelity of implementation. Its effectiveness when taught by classroom teachers awaits a scale-up evaluation.

A question regarding the generalizability of the results obtained in this evaluation concerns whether Get Real would be equally effective with student populations that show higher levels of sexual activity. In this study, 6% of students reported being sexually active at follow-up (average age of 12.88). This level of sexual activity places our sample in the mid-range of sexual activity rates reported in the 2009 Youth Risk Behavior Surveillance System (Centers for Disease Control and Prevention, 2010), where prevalence of having had sexual intercourse before 13 years of age ranged from 3.4% to 13.4% across state surveys (median: 5.7%). Although there is reason to believe that what worked in Massachusetts can work in many other states with similar early sexual activity rates, demonstration of this assumption is needed. The effectiveness of the Get Real curriculum for populations of students with higher rates of sexual initiation should be explored.

The limitations of this evaluation include the self-report nature of the measures. We acknowledge that, although students were assured of the confidentiality of their responses and the study used a measure of social desirability to control for the influence of a tendency to give socially desirable responses, it is possible that some youth may not have been truthful in their answers. The present evaluation shares this limitation with all other evaluations that rely on self-reports. Another limitation is the inability to block randomize by school characteristics, which may have contributed to the imbalance in our intervention and control samples. The imbalance occurred because schools were recruited over a period of approximately 12 months, and individual schools needed to know their intervention or control status upon entering the study to adjust their class schedules. Consequently, randomization took place on a rolling basis, rather than after the full complement of schools was ascertained. However, characteristics identified as significantly different between groups were used as predictors in the models we tested and the intervention retained significance.

A further limitation of this study is the large confidence intervals in the logistic regression, which indicate a wide range of possible intervention effects. This is primarily because of low statistical power, stemming from the rarity of the event (participants having sex in sixth or seventh grade). Despite this, we were able to find a significant, although small effect, when controlling for such known risk factors for sexual debut as older age and male gender.

The focus on vaginal intercourse did not address sexual risk taking among students who engage in nonvaginal types of intercourse and did not fully assess sexual activity among lesbian, gay, or bisexual students. Therefore, we do not know whether the intervention was effective in delaying nonvaginal intercourse.

In sum, the findings indicate that theory-based and age-appropriate comprehensive sex education programs for early adolescents can be effective in delaying sexual debut. Sex education in the sixth grade may be particularly critical for potential early starters, who are at increased risk for ongoing health and social adjustment difficulties that persist into adulthood (Magnusson & Trost, 2006; Pergamit et al., 2001; Sandfort et al., 2008; von Ranson et al., 2000). At this juncture of the evaluation, it is not possible to know whether current seventh-grade students in the intervention group will show long-term delays in sexual debut. Yet being able to reduce the number of early starters to sexual activity with 1 year of intervention is noteworthy because of the particularly dire consequences of early sexual debut.

Footnotes

Acknowledgements

The authors thank the schools for giving them access, the students for participating, and the PPLM sexual health educators for administering the intervention.

Authors’ Note

This research was presented at the annual meeting of the 2011 American Public Health Association in Washington, D.C.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the Planned Parenthood League of Massachusetts for the impact evaluation of their comprehensive middle school sex education curriculum.

Author Biographies

Sumru Erkut, the director of the impact evaluation study on which this article is based, is an associate director and senior research scientist at the Wellesley Centers for Women. She received her doctorate in social psychology from Harvard University. Her studies of youth development and women’s leadership from the perspective of gender equity and racial and ethnic diversity have been supported by federal, foundation, corporate, and private funds.

Jennifer M. Grossman, the codirector of the impact evaluation study, is a research scientist at Wellesley Centers for Women. She investigates risk and resilience in adolescent development, with a focus on the influence of contextual factors, such as race, religion, and gender on individuals’ academic, sexual, social, and emotional development.

Alice A. Frye is a lecturer in the Department of Psychology at the University of Massachusetts, Lowell. She is a developmental psychopathologist and methodologist with a focus on adjustment and psychopathology in emerging adulthood and adolescence.

Ineke Ceder is a research associate at the Wellesley Centers for Women where she is the project manager of the evaluation study described in this article. She has worked on projects related to issues of racial/ethnic and sexual diversity, gender equity, and youth development.

Linda Charmaraman is a research scientist at the Wellesley Centers for Women and the National Institute of Out-of-School Time. Her research interests center around positive minority youth development, conducting studies that focus on adolescent identity, youth media, bullying, and sexual harassment. She received her doctorate from the Graduate School of Education at University of California, Berkeley.

Allison J. Tracy is a senior research scientist and methodologist at the Wellesley Centers for Women. She has provided methodological consultation and statistical analysis to researchers at the Wellesley Centers for Women and to researchers in academic, governmental, and nonprofit institutions for more than 13 years. She received her doctorate in human development and family studies with a minor in statistics from the Pennsylvania State University.

References

Aarons

Jenkins

Raine

El-Khorazaty

(2000). Postponing sexual intercourse among urban junior high school students—A randomized controlled evaluation. Journal of Adolescent Health, 27, 236-247. doi:10.1016/ S1054-139X(00)00102-6

Abma

J. C.

Martinez

G. M.

Mosher

W. D.

Dawson

B. S.

(2004). Teenagers in the United States: Sexual activity, contraceptive use, and childbearing, 2002. National Center for Health Statistics. Vital Health Statistics, 23. Retrieved from http://www.cdc.gov/nchs/data/series/sr_23/sr23_024.pdf

Ajzen

(1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179-211. doi:10.1016/0749-5978(91)90020-T

Ajzen

(2006). Behavioral interventions based on the theory of planned behavior. Retrieved from http://www.people.umass.edu/aizen/pdf/tpb.intervention.pdf

Baxter

S. D.

Smith

A. F.

Litaker

M. S.

Baglio

M. L.

Guinn

C. H.

Shaffer

N. M.

(2004). Children’s social desirability and dietary reports. Journal of Nutrition Education and Behavior, 36, 84-89. doi:10.1016/S1499-4046(06)60138-3

Bronfenbrenner

(1979). The ecology of human development. Cambridge, MA: Harvard University Press.

Centers for Disease Control and Prevention. (2010). Youth risk behavior surveillance—United States, 2009. Surveillance summaries. Morbidity and Mortality Weekly Report, 59(No. SS-5). Retrieved from http://www.cdc.gov/mmwr/pdf/ss/ss5905.pdf

Coker

A. L.

Richter

D. L.

Valois

R. F.

McKeown

R. E.

Garrison

C. Z.

Vincent

M. L.

(1994). Correlates and consequences of early initiation of sexual intercourse. Journal of School Health, 64, 372-377. doi:10.1111/j.1746-1561.1994.tb06208.x

Coyle

K. K.

Kirby

D. B.

Marin

B. V.

Gomez

C. A.

Gregorich

S. E.

(2004). Draw the line/respect the line: A randomized trial of a middle school intervention to reduce sexual risk behaviors. American Journal of Public Health, 94, 843-851. doi:10.2105/AJPH.94.5.843

10.

Duncan

S. C.

Strycker

L. A.

Duncan

T. E.

(1999). Exploring association in developmental trends of adolescent substance use and risky sexual behavior in a high-risk population. Journal of Behavioral Medicine, 22, 21-34. doi:10.1023/ A:1018795417956

11.

Ennis

S. R.

Rios-Vargas

Albert

N. G.

(2011). The Hispanic population, 2010 (2010 Census Brief). Retrieved from http://www.census.gov/prod/cen2010/briefs/c2010br-04.pdf

12.

Jemmott

J. B.

III Jemmott

L. S.

Fong

G. T.

(2010). Efficacy of a theory-based abstinence-only intervention over 24 months: A randomized controlled trial with young adolescents. Archives of Pediatric and Adolescent Medicine, 164, 152-159. Retrieved from http://archpedi.ama-assn.org/cgi/content/short/164/2/152

13.

Kirby

(2007). Emerging Answers 2007: Research findings on programs to reduce teen pregnancy and sexually transmitted diseases. Washington, DC: National Campaign to Prevent Teen and Unplanned Pregnancy. Retrieved from http://eric.ed.gov/ERICWebPortal/contentdelivery/servlet/ERICServlet?accno=ED456171

14.

Koo

H. P.

Rose

El-Khorazaty

M. N.

Yao

Jenkins

R. R.

Anderson

K. M.

. . .Walker

L. R.

(2011). Evaluation of a randomized intervention to delay sexual initiation among fifth-graders followed through the sixth grade. Sex Education, 11, 27-46. doi:10.1080/14681811.2011.538146

15.

Magnusson

Trost

(2006). Girls experiencing sexual intercourse early: Could it play a part in reproductive health in middle adulthood? Journal of Psychosomatic Obstetrics and Gynecology, 27, 237-244. doi:10.1080/ 01674820600869006

16.

Mathematica Policy Research & Child Trends. (2010). Identifying programs that impact teen pregnancy, sexually transmitted infections, and associated sexual risk behaviors. Retrieved from http://www.hhs.gov/ash/oah/oah-initiatives/tpp/eb-programs-review.pdf

17.

O’Donnell

B. L.

O’Donnell

C. R.

Stueve

(2001). Early initiation and subsequent sex-related risks among urban minority youth: The Reach for Health Study. Family Planning Perspectives, 33, 268-275. doi:10.2105/AJPH.2003.026567

18.

Pergamit

M. R.

Huang

Lane

(2001). The long term impact of adolescent risky behaviors and family environment. Washington, DC: U.S. Department of Health and Human Services. Retrieved from http://aspe.hhs.gov/hsp/riskybehav01

19.

Planned Parenthood League of Massachusetts. (2010). Get real: Comprehensive sex education that works. Retrieved from http://www.getrealeducation.org/

20.

Sandfort

T. G. M.

Orr

Hirsch

J. S.

Santelli

(2008). Long-term health correlates of timing of sexual debut: Results from a national U.S. study. American Journal of Public Health, 98, 155-161. doi:10.2105/AJPH.2006.097444

21.

Tortolero

S. R.

Markham

C. M.

Peskin

M. F.

Shegog

Addy

R. C.

Escobar-Chaves

S. L.

Baumler

E. R.

(2010). It’s your game: Keep it real: Delaying sexual behavior with an effective middle school program. Journal of Adolescent Health, 46, 169-179. doi:10.1016/j.jadohealth.2009.06.008

22.

Trenholm

Devaney

Fortson

Quay

Wheeler

Clark

(2007). Impacts of four Title V, Section 510 abstinence education programs: Final report. Princeton, NJ: Mathematica Policy Research, Inc. Retrieved from http://aspe.hhs.gov/hsp/abstinence07/

23.

von Ranson

Rosenthal

Biro

Lewis

Succop

(2000). Longitudinal risk of STD acquisition in adolescent girls using a generalized estimating equations model. Journal of Pediatric and Adolescent Gynecology, 13, 87-99. doi:10.1016/S1083-3188(00)00013-9