In addition to describing the outcomes of the different approaches to sex education, in this chapter we examine their foundational premises and assumptions, with the intent to clarify not only what does or does not work but also the reasons for success or failure.
Following is an excerpt from the chapter “Abstinence Education in Context:
The History, Evidence, Premises, and Comparison to Comprehensive Sexuality Education.”
This chapter provides a broad picture of the abstinence education movement in the U.S. and the historical, political, and theoretical context of its journey. It describes the problems it has targeted, the policies and programs it has designed to solve them, and the results of these various interventions. Solutions to the problems posed by adolescent sexual activity have fallen into three camps: risk reduction, risk avoidance, and a combination of those two approaches. This historical context permits a comparative analysis of the most common sex education strategies and provides a better understanding of what abstinence education actually looks like — what it is and what it is not. In addition to describing the outcomes of the different approaches to sex education, we examine their foundational premises and assumptions, with the intent to clarify not only what does or does not work but also the reasons for success or failure. As important as the effectiveness question is, the underlying rationale, theory, or premise on which any program strategy is based is equally important if we are to understand how to benefit from prior successes and failures. In addressing these issues, we propose and describe an empirical model that delineates key predictors of adolescent risk behavior and demonstrates important causal mechanisms that program developers can focus on for more effective interventions. ……
B. Program Success with Risk Avoidance (Abstinence) Strategies
Given the lack of compelling evidence for CSE strategies, the high rates of condom user error, the partial protection levels achieved even with correct and consistent use, and the emotional harm and sexual violence which condoms do not reduce, it seems appropriate to ask whether it is possible and desirable, as an alternative strategy, to promote abstinent behavior and thereby substantially increase the number of teens who abstain from sexual activity.
Promoting abstinence as a lifestyle is no easy task, especially given the cultural context in which today’s adolescents live. Movies, music, peers, Internet pornography, and other influences are constantly pushing a different message. Many teens succumb to those influences. Abstinence education faces an uphill battle. In spite of that, it is apparently easier to convince adolescents to abstain that it is to convince them to use condoms consistently.
For example, if we do a side-by-side comparison of program strategies from Kirby’s list of credible studies, and look at those primary prevention programs which have a common setting and population, and which have (1) reasonably comparable outcome measures (CCU and abstinence), (2) a similar time frame (one to two years), and (3) broad based impact, we find 38 studies of comprehensive programs that fit all three categories. From Kirby’s same list, we find seven abstinence education studies that meet those same criteria. From the list of 38 CSE studies, 10 reported a significant improvement in abstinence, but none reported an increase in CCU. And this was in programs where abstinence was not the central message. In the seven abstinence education evaluations, five reported a significant reduction is sexual initiation rates. Kirby’s (1991) statement that “it may actually be easier to delay the onset of intercourse than to increase contraceptive practice” (p.262) is being borne out.
According to data from both the National Survey of Family Growth, and the Youth Risk Behavior Surveys, the national trends in teen sexual activity have shown a steady decline from 1988 through 2010. (The 2014 data show a continuing decline.) Females 15-19 who were sexually experienced declined significantly from 51% in 1998 to 43% in 2006-2010. For males of the same age group, the decline was even larger: 60% in 1998 to 42% in 2006-2010 (Martinez, Copen & Abma, 2011, p6. Fig. 1). Apparently, teens’ sexual behavior is amenable to change. Contrary to the perception that all teens are “doing it,” more than 80% of teens under 15 and more than 70% of teens 15 to 17 say they have not had sex (Healthy People, 2013). This increase in teen abstinent behavior corresponds with the decline in teen pregnancy, teen births, and teen abortions — an encouraging trend by anyone’s standards.
The CDC’s Guidelines for Effective School Health Education to Prevent the Spread of AIDS recommended that schools provide programs that encourage abstinence from sex:
School systems should make programs available that will enable and encourage young people who have not engaged in sexual intercourse…to continue to abstain from sexual intercourse until they are ready to establish a mutually monogamous relationship within the context of marriage… For young people who have engaged in sexual intercourse…school programs should enable and encourage them to stop engaging in sexual intercourse until they are ready to establish a mutually monogamous relationship within the context of marriage (CDC, 1988, p.1).
The American College of Pediatricians’ policy is similar (American College of Pediatricians, 2013), and Healthy People 2020 challenges health advocates to increase abstinence among teens by 10% (Healthy People, 2013). To meet these goals, evidence-based abstinence education programs are needed.
We are seeing a pattern of evidence that well-designed and well-implemented programs can be effective. Of the seven abstinence programs examined in Kirby’s Emerging Answers review (2007), five reported a significant reduction in rates of sexual initiation (Clark et al., 2005; Denny & Young, 2006; Doniger et al., 2001; Howard & McCabe, 1990; Weed et al., 1992).
Let us review some additional examples of successful abstinence education programs that were not included in Kirby’s 2007 review:
- The Heritage Keepers Abstinence Education study used a large sample size (n=1,535), matched comparison group, and a 12-month follow-up (Weed, Ericksen & Birch, 2005). It found that program students were about half as likely to initiate sexual intercourse after one year as were the comparison students, after controlling for pretest differences (odds ratio=.539, p<.001). Program students also had significant improvement on cognitive factors that appeared to mediate their abstinent behavior (Weed, Ericksen, Lewis, et al., 2008). This study was replicated with a larger sample and more rigorous methods, and produced very similar results (Weed, Olsen, Ericksen, 2013).
- A 2008 evaluation of the Reasons of the Heart abstinence curriculum found that adolescent students who were virgins who received the program were about half as likely as the matched comparison group to initiate sexual activity after one year (odds ratio=.413, p<.05). This program also achieved impact on cognitive mediators that appeared to contribute to the program’s success (Weed, et al., 2008)
- In a 2006 study conducted by Dr. John Jemmott at Princeton University, African-American youth (ages 10-15) were randomly assigned to one of four interventions: abstinence, safer-sex, safer-sex and abstinence, and health-promotion control intervention. Adolescents who received the abstinence intervention were less likely to report ever having sexual intercourse at 24-month follow-up than were those in the health-promotion control intervention (p=.02), the safer-sex intervention (p=.007), or the safer-sex and abstinence intervention (p=.05). Youth in the abstinence program who did start having sex were no less likely than those in the other groups to use condoms (Jemmott, Jemmott, & Fong, 2006).
- A 1990 study in the Family Planning Perspectives journal of the Postponing Sexual Involvement (Abstinence Version) found that low-income, minority students in the 8th grade (in Atlanta, Georgia) who participated in the Postponing Sexual Involvement program were 5 times less likely to have had sex at the end of the 8th-grade than students who did not participate (4% vs. 20%). By the end of the 9th-grade, the difference between the two groups was still significant, with rates of 24% vs. 39% (Howard & McCabe, 1990).
- A 1992 evaluation by the U.S. Office of Population Affairs of the Sex Respect and Teen Aid programs found that the two programs together reduced the initiation of sexual activity among at-risk students by 25% compared with similar at-risk students who did not receive any abstinence education (Weed, Olsen, DeGaston, & Prigmore, 1992).
- A 2005 study funded by the Department of Health and Human Services evaluated the Choosing the Best program in Georgia and found that 11.5% of virgin students exposed to this abstinence education program had begun having sex one year later, compared with 21.6% of the students who received no abstinence education (Weed & Anderson, 2005).
- A 2001 study in the Journal of Health Communications found that after a 5-year county-wide mass communications program called Not Me, Not Now, there was a 32% reduction in the percent of teens under 16 who had experienced sex (p<.05). The adolescent pregnancy rate for Monroe County dropped from 63.4% in 1993 to 49.5% in 1996. Similar counties in New York not exposed to this campaign did not experience a comparable decline in the teen pregnancy rate (p <.01) (Doniger, Adams, Utter, & Riley, 2001).
- The 2010 Promoting Health Among Teens study compared behavioral outcomes using four approaches: abstinence-only, safer-sex (contraception focus), combined approach (abstinence and safer sex), and an untreated control group. Two years later, students in the abstinence-only cohort initiated sex at a significantly lower rate than in any of the other cohorts (Jemmott, Jemmott, & Fong, 2010).
Taken together, these studies provide more rigorous evidence than Kirby’s Emerging Answers review (2007) that abstinence education programs can be effective. A developing pattern of scientific evidence indicates that abstinence education programs, if properly designed and implemented, can cut rates of teen sexual activity by as much as half for significant periods of time, without reducing condom use by the sexually active. (Condom use was measured by the Jemmott et al. (2006) and Treholm et al. (2007) studies of abstinence programs, and no reduction in use by sexually active teens was found). This body of research suggests that teaching adolescents to avoid sexual activity is a viable primary prevention strategy, one that can fully prevent the harmful and costly consequences of teen sex.
Do all abstinence education programs work? Of course not. We have also evaluated several programs that do not work, do not work well, or do not work for all the program participants. Such findings were truer of abstinence programs in the early stages of development and implementation, when they did not have program evaluation data to give direction to program improvement. Findings of ineffectiveness will, of course, continue to be true of abstinence programs that do not provide adequate dosage to go beyond superficial impact, do not impact the cognitive mediators that bring about enduring behavioral change, are not delivered by teachers who themselves believe in the abstinence message they are presenting to teens, and so on. As with all educational interventions, those programs that are well-designed and well-implemented are the ones that will be effective.
Other Commentaries and Research Regarding Abstinence Education
Critics of abstinence education typically cite reviews of abstinence education studies that found no positive impact on teen sexual behavior (Kirby, 2007; Kohler, et al., 2008; Underhill, et al., 2007). We have also evaluated programs that don’t work well, along with those that do. Most of the ineffective programs we have evaluated, like the “no impact” programs reviewed by abstinence education critics, were developed and implemented during the first decade of federal abstinence funding at a time when most abstinence programs and their research evaluation were still in their infancy. Later, drawing from our evaluation research, we will discuss the program features that differentiate effective abstinence education interventions from ineffective ones.
Let us look more closely at two well-publicized research reports that concluded that the abstinence programs were ineffective in reducing teen sexual activity. The first such research report was the aforementioned Emerging Answers 2007, the exhaustive review of the best sex education evaluation studies of the preceding 15 years. Not included in Emerging Answers were subsequently published studies of abstinence interventions, several of which reported significant increases in rates of teen abstinence for at least 12 months after the program. For example, two of these studies found that teen participants were half as likely to become sexually active as non-participants after one year (Weed, et al., 2005, Weed, et al., 2008). The third study found that the program had increased teen abstinence significantly for a period of two years. Furthermore, students in this third program who did initiate sexual activity were no less likely to use condoms than those in the control group (Jemmott, et al., 2010). And, as previously noted, Emerging Answers 2007 found only one CSE program out of 115 studies that improved for at least 12 months teens’ CCU(the outcome that is, for risk reduction programs, what abstinence is for risk avoidance programs) (DiClemente, et al., 2004).
The second well-publicized research report often cited by critics of abstinence education was a federally funded evaluation by Mathematica Policy Research, Inc. It conducted a longitudinal study of four abstinence interventions and found no reduction in teen sexual activity (Trenholm, et al., 2007). The media reported this study as showing that abstinence education does not work, and that therefore CSE (condom-centered) programs would work. Once again, a closer examination of the actual study and its findings does not support that conclusion.
The design of the Mathematica study was rigorous in some ways — it investigated four different abstinence programs and followed students longitudinally — but it also suffered from a number of limitations:
- Inappropriate timing of program dose. The age group for the interventions in the Mathematica study was quite young — elementary and early middle school. Some were as young as 4th- and 5th-grade. The interventions did not continue or in any way reinforce the initial treatment during the key years (9th, 10th, 11th grade) when transition into sexual activity typically occurs. Thus, the treatment was not delivered or reinforced when it was most relevant and needed. As the Mathematica report points out, “The findings provide no information on the effects the programs might have if they were implemented for high school youth or begun at earlier ages but served youth through high school” (p. 61). At the outset, then, the evaluation started with interventions that, because of their inappropriate timing, had little hope of impacting behavior in the long run.
- Unusually long time between intervention and follow-up evaluation. The follow-up time frame employed in this study — 2½ to 5½ years after the program’s end — is too long for any type of sex education intervention to have a sustained effect on behavior without interim reinforcement of the program message. A myriad of negative influences operate in adolescents’ lives to overpower any initial program effect that may have occurred so far in the past. The follow-up interval for measuring behavioral outcomes was much longer than what is typical in evaluations of CSE programs. For example, Emerging Answers 2007 reported that only 7 out of 107 studies of CSE programs used a follow-up interval of 4 years or longer, and none of these 7 programs significantly increased teen condom use of any kind, including CCU, for that length of time. Thus, when held to the same standard of effectiveness used in the Mathematica study, no CSE program in the past 15 years would be called “effective.” In fact, we are not aware of any evaluations of school based CSE programs that have shown positive changes in teen condom use after three years, and are aware of only two that have shown impact after two years, and these were using the lower standard of success (not CCU).
- Cross-contamination of program effects. The benefits of a random assignment research design are best realized when the treatment and control groups can be kept separate and their integrity maintained. In this way, the treatment or “medicine” is not shared between the groups. However, in field experiments (which sex education evaluations typically are), this requirement of separate treatment and control groups is difficult to achieve, especially with teenagers, and particularly with an intervention that deals with a topic as highly charged and commonly discussed as sex. Students randomly assigned to the treatment and control groups don’t live in these groups — they interact with friends, siblings, and dating partners in the other group. Any new values or behaviors adopted by each group tend to be shared across the groups, and the longer that sharing lasts, the more likely it is that the differences between the two groups will disappear as their attitudes, values, beliefs and behaviors merge over time. This cross-group contamination is likely to be a more powerful “intervention” — a stronger influence on attitudes and behavior — than a typical one-hour-per-day, short-term classroom-based intervention. With almost six years for this spillover effect to operate, as was true in the Mathematica study, the cross-contamination would minimize the measurable differences between the groups, even if the program had successfully reduced the participants’ sexual activity. The Mathematica study did not address this problem in its design, nor did it take it into account when reporting its findings.
- Non-representative study sample. The high-risk population used in the Mathematica study does not represent the teen population in the U.S. The majority of the sample was African-American youth from poor, single-parent households; the fact that these abstinence programs produced no impact on this sample does not tell us whether the same programs would have had an impact on a different group of teens.
- Inadequate attention to mediator variables. Sex education research must identify and track over time the important causal mechanisms (cognitive mediators such as future orientation, personal efficacy, personal values, peer influence, etc.) that influence adolescent sexual risk behavior. Unless we identify and monitor how programs do (or do not) impact these causal mechanisms, program success or failure cannot be understood, intervention modifications cannot be made, and longer-term program potential cannot be identified. The Mathematica study had several shortcomings in this area. First, the study’s generic logic model did not take into account differences in how the four specific programs conceptualized mediating causal mechanisms, and therefore each program’s theory was not tested by the study. Second, Mathematica’s design did not assess a great enough variety of mediating variables; only two of the ones it chose to measure showed a significant relationship to the targeted sexual behavior, and neither of those showed significant pre-post change.
- Failure to share interim data with program designers. The Mathematica study did not share interim data on the causal mechanisms with the four programs to support their improvement. In Mathematica’s case, data on the mediating variables was not shared with the programs until four or five years later. Had we taken that approach with some of our own program evaluations (e.g., in Arkansas, Virginia, South Carolina, and Georgia), we would likely have seen the same “no impact” result when measuring behavior five years later. Instead, in part because these programs benefitted from interim data on how they were influencing the targeted mediating variables, they matured over time, and are now realizing up to 50% reduction in teenagers’ initiation of sexual activity.
Because of these design limitations, the findings of the Mathematica study regarding the four particular abstinence programs it evaluated cannot be generalized to represent the overall efficacy of abstinence education. But equally important to a correct understanding of the Mathematica study is the fact that it did not set out to compare abstinence education with CSE (risk reduction) programs. It did not study any “safer sex” programs, nor suggest that they are the obvious default if abstinence programs are not successful. We remind the reader that a substantial number of other studies during the past two decades have examined condom-based school interventions, and only one out of 50 reported an improvement in CCU after a period of at least one year. We believe that pattern of evidence is hardly grounds for embracing a condom-based sex education policy if a particular abstinence program appears not to be working.
A final point about the Mathematica study is a basic one that applies to interpreting all research: Any new study must be viewed not in isolation but in the context of the total body of relevant research that preceded it. Prior to the Mathematica study, as we have shown, a growing number of well-designed studies of abstinence programs had achieved substantial reductions in teen sexual activity (as much as half) for periods ranging from 12 to 24 months. That success stands in contrast to the failure of school-based CSE to achieve the goal of increasing CCU. When CSE and abstinence education are evaluated in terms of the outcomes that each approach aims for — CCU in the case of CSE and reduced sexual activity in the case of abstinence education–the evidence is better for the effectiveness of abstinence education.
C. Can the Risk Reduction and Risk Avoidance Approaches Be Combined?
Some would argue that abstinence education and condom (or CSE) education should both occur so as to not only give a risk avoidance message, but also to accommodate those students who are already sexually active. Advocates of this combination approach have sought support from public opinion polls that ask parents about their preferences regarding sexuality education. But on the whole, such polls have repeatedly found American parents to be very much pro-abstinence education; they want schools to teach their children not to have sex. For example, across three different national polls, 70% to 90% of parents said they want a strong abstinence message given to teens (NPR/Kaiser Foundation, 2004; (Zogby, Bonacci, Bruce, & Wittman, 2003); Zogby, 2004). More than 90% believe that adolescents should not become sexually active (Zogby, et al., 2003), and fully two-thirds (67%) say it is morally wrong for them to do so (Zogby, 2004). Only 8% believe that teaching adolescents about condom use is more important than teaching abstinence (Zogby, et al., 2003). And only 7% want sex education to convey the message that it’s okay for teen to engage in sexual intercourse as long as they use a condom (Zogby, et al., 2003).
As is often the case with polls, opinion on a specific matter has varied with how the question is asked. It is true that many parents respond favorably when asked whether teens should be given information about how to obtain and use condoms (39% and 58% in one poll (NPR/Kaiser Foundation, 2004), and 78% and 81% in another (Zogby, 2004). However, when asked to respond to the actual content of popular CSE curriculum materials, the large majority of parents (70% to 90%) opposed the explicit information about sexual practices, condom use, and masturbation that such materials contained. These polls also report that parents opposed (76%) to withholding from teens medically accurate information showing that condoms provide only partial protection against STI, and 70% do not want their own child to be given contraception in school or taught how to obtain contraception without their knowledge or approval.
Finally, while a majority of parents believe teens should have information about contraception, fewer than half (40%) think that abstinence and contraception should be taught in the same classroom. Most parents prefer that biological facts about contraception either be taught in a health curriculum separate from the abstinence program (56%), or not taught at all (22%)(Zogby, 2004).
Advocates of combining the risk reduction and risk avoidance approaches might argue that doing so is consistent with program evaluations that have measured both condom use and abstinence as outcomes. Evaluating both kinds of outcomes might seem to suggest that the different messages are somehow compatible and could be effective in combination. But we don’t think it follows logically that because it’s possible for the same study to evaluate two very different kinds of outcomes (increased condom use and decreased sexual activity), those outcomes represent educationally compatible program goals and underlying philosophies. The risk reduction and risk avoidance approaches are based on very different assumptions and premises about human sexuality, healthy relationships, and family formation (differences we address in greater detail in Section VII). It is difficult to see how these two different ideologies could be combined.
In evaluating whether a combination model approach makes sense in sex education, we think it’s helpful to consider what we do in other areas of health education. We normally seek to attain our program goals by transmitting a coherent, consistent message rather than mixed messages that undercut each other. In drug education, for example, we don’t say, “Avoid using illegal drugs and their harmful consequences to users and society,” and then add, “But if you decide to use them, here’s a way to reduce the risks.” Why, then, would we say to teens, “Avoid premarital sex and the risks of pregnancy, disease, and emotional hurt — but if you decide to have sex, here’s a way to partially reduce some of those risks”? Teenagers — and their parents — will see that for what it is: a weak endorsement of abstinence.
It’s also useful to recall that thus far, no condom education programs have been able to increase teenagers’ CCU. Why would abstinence programs that have succeeded in increasing teen abstinence want to add to their efforts an educational strategy — urging adolescents, “Always use a condom when you have sex” — that has for more than two decades been unable to achieve its goal? In education, program designers do not normally reason, “Let’s try to improve this strategy that’s working by adding one that isn’t.”
Two final arguments against the combination approach: (1) When it has been studied, it has been found to weaken the effectiveness of the abstinence component. Recall the 2010 Promoting Health Among Teens study that included a comparison of an abstinence-only program with one that combined abstinence with “safer sex” instruction. Teen sexual activity rates were higher in the combination approach; and (2) In practice, combination programs, rather than giving equal time to risk reduction and risk avoidance, typically focus on condom education and treat abstinence superficially. Abstinence in combination programs is, on average, given about 10% of the attention. That is a recipe for making abstinence education ineffective.
V. Using a Causal Model to Evaluate Abstinence Education: A Case Study
In recent decades, health and education program evaluation studies have emphasized the need to measure hypothesized mediating variables in order to strengthen causal explanations of program effects or their absence (Fitz-Gibbon & Morris, 1996; Reynolds, 1998; Worthen, 1996). Known as “Confirmatory Program Evaluation” (CPE), this approach provides a framework for conducting theory-driven outcome evaluations, ones that generate the data needed to test the theoretical constructs underlying a program’s design. Reynolds writes, “Of special interest is testing the causal mechanisms that may lead to longer-term program effects. In CPE, the evaluator investigates the empirical relationships among program, intervening, and outcome variables . . . If the identified causal pathways leading to the desire outcome are consistent with the theory and operation of the program, causal inference is strengthened and the coherence of the program outcome relationship is supported” (Reynolds, 1998, pp. 206, 209).
In this section we describe in more detail our evaluation of an abstinence education program, Heritage Keepers (2014), which incorporated the CPE emphasis on measuring mediating variables (Weed, Ericksen, & Birch, 2005). We believe this study met most of the challenges inherent in applied field research. It sought to examine not only the basic outcome of reduced sexual activity, but also the causal mechanisms operating to bring about that reduction. As Reynolds pointed out, research that examines the mediating variables that link program inputs to desired outcomes enables program designers to subsequently modify their intervention so as to impact those causal mechanisms even more effectively.
To identify causal mechanisms, it is helpful to look at social science literature that addresses mediators of behavior and how those mediators are influenced. Social learning theorists use cognitive theory to identify important cognitive mediators of social learning. Protection motivation theory focuses on behavioral intentions, self-efficacy, outcome expectancies, attitudes, and social norms (Ajzen, 1991; Armitage & Conner, 2000; Bandura, 2004; Conner & Armitage, 1998; Floyd, Prentice-Dunn, & Rogers, 2000). These and similar psychosocial constructs have been shown in previous research to be significantly related to adolescent sexual behavior (Kirby, et al., 2007b; Plotnik, 1992; Resnick, et al., 1997). Constructs similar to these were helpful to us in designing the abstinence program we will now describe.
The Heritage Keepers Abstinence Education curriculum had been reviewed by the federal government, selected as meeting federal criteria for “evidence based,” and thereby approved for federal funding. Designed for middle and/or high schools, Heritage Keepers was based on a set of psychosocial constructs posited to influence adolescent sexual behavior. Heritage Keeper’s 450-minute interactive curriculum can be presented in 45-minute class periods over 10 consecutive school days, or in 90-minute sessions over five days. Heritage Keeper’s curriculum and training encourages sensitivity to race, gender, sexual experience, sexual orientation, family of origin structure, the persons students are living with, and whether students already have a child. However, the program’s message does not vary based on those variables, since all students are assumed to have the capacity to, and are encouraged to, abstain from sexual activity.
Program content is consistent with Title V, Section 510 A-H standards and includes definitions of abstinence and recommitment to abstinence, reproduction and anatomy, STI information, determining and integrating personal values with behavior, goal setting, establishing protective boundaries, building healthy relationships without having sex, benefits of marriage, and developing skills to refute and refuse sexual initiation. A “whole person” approach is applied that takes into account how students see themselves, what and whom they value, how they relate to others, where they are going, and how they will react to stimuli introduced into their lives.
Psychosocial constructs believed to be predictive of teen sexual behavior provided the theoretical foundation guiding Heritage Keeper’s curriculum design, training, implementation, monitoring, and program improvement processes. These constructs included Behavioral Intention (Intention); Abstinence Values (Values), Future Impact of Sex (Future Impact), Abstinence Efficacy (Efficacy), and Justifications for Sex (Justifications). During annual 3-day trainings, program instructors learned how to engage students in active learning processes that addressed these targeted mediators. For example, the program addresses common “justifications for sex” by listing typical reasons teens give for initiating sex and by providing alternative arguments. Students practice these arguments in directed role-plays. They also take turns in role-plays in which they alternate playing someone engaging in sex outside marriage, someone effectively resisting those arguments, and a third person encouraging the resistance. These exercises are designed to increase “abstinence efficacy.” The program emphasizes the “future impact of sex” through interactive activities that help students make a personal connection between the possible consequences of sexual activity and the plans they have for their future. They are also given data about the benefits to the couple and any children they may have and about the benefits of forming and raising a family within a long-term legal and ethical commitment. This fosters the development of students’ “abstinence values” by promoting class discussions differentiating between short-term infatuations and lasting love.
To facilitate consistent delivery of the program, a fidelity-to-plan checklist helped instructors keep track of what they had taught to each class. Using the same checklist, we provided program administrators with feedback on each of the mediator variables as short-term outcomes variables in order to help administrators gauge instructors’ effectiveness and take steps to improve program delivery.
Our evaluation study of Heritage Keepers (Weed, et.al, 2005) had two primary purposes: (1) to test the program’s effectiveness at postponing adolescent sexual debut, and (2) to simultaneously test the program’s theoretical framework regarding the causal linkages between initiation of teen sexual intercourse and the psychological constructs hypothesized to be mediators of abstinence. Our study’s sample included 2,215 students, grades 7 to 9, of which 63% were African American and 42% were male. Program and comparison students were matched using propensity score analysis, a procedure that enabled us to establish strong baseline equivalence between these the program and comparison groups on demographic and mediator measures. The Heritage Keepers study appears to be the first in the U.S. to use propensity score matching in an outcome study of sex education. This matching procedure, by establishing baseline equivalence of the program and comparison groups on key factors, mitigated possible bias/mismatch in the original sample. Our assessment of program outcomes included pretest-posttest comparisons and follow-up tests after one year.
Using structural equation models and mediation analysis, several results stand out:
- After one year, the group that experienced the Heritage Keepers program exhibited substantially and statistically significant lower rates (67% lower) of sexual initiation than the comparison group, suggesting a strong program effect on sexual behavior.
- Program participants who reported sexual experience at the pre-test also reported significantly fewer sexual partners at the time of the one year follow-up (p=.035).
- At the one-year follow-up, there were clear and statistically significant differences between the program and comparison groups on four of the five program-targeted mediator variables (the one exception being personal efficacy).
Pertaining to the second purposes of the study, the observed program effects on sexual behavior were almost entirely explained or mediated by the four program-targeted cognitive constructs on which they program and comparison groups differed. This evidence provided clear empirical support for the program’s theoretical model, identifying important causal mechanisms that can influence adolescent sexual risk behavior. The statistical relationship demonstrated among three factors — the program intervention, the change in psychosocial mediators targeted by the program, and long-term sexual behavior — provides stronger evidence than has been previously available for a program’s causal impact on teen abstinence.
The Heritage Keepers study (Weed, et al., 2005) shows that abstinence education programs that influence key attitudes, values, and behaviors which are directly predictive of sexual risk behavior, can produce a long-term delay in initiation of sexual intercourse as well as a reduction in sexual partners. Future research should continue to identify and test important mediating factors, and designers of abstinence programs should use all such findings to strengthen program impact on the causal mechanisms that influence young people’s sexual behavior.
VI. What Are the Characteristics of Effective Abstinence Education Programs?
As mentioned earlier in this chapter, our research shows that some abstinence programs work and some don’t. The important questions to ask are, “Which ones work, and why?” Abstinence interventions are most effective when they incorporate what has been learned from research about how to reduce adolescent sexual risk behavior. Successful abstinence education programs tend to do that and to share a number of other attributes. Listed below are a dozen characteristics of effective programs we have observed over 20 years of evaluating more than a hundred abstinence interventions. We have not collected empirical data that enables us to rank or weight these characteristics; we encourage program designers to treat all these factors as worthy of attention as they seek to maximize overall program effectiveness. The more of them they incorporate, the greater likelihood of success.
Given the evidence regarding the importance of parents, some abstinence programs have involved parents in one or more ways: assigning homework that requires parent-teen discussion of sex-related issues; offering parents workshops on topics such as adolescent development and effective parent-teen communication; and providing workshops that parents and teens attend together. Parent involvement strategies such as these can lead parents to take advantage of opportunities in family life to reinforce the school’s abstinence message, thereby increasing the program’s dosage and encouraging a young person to attach a high value to refraining from sexual activity.
Abstinence education programs that incorporate a majority of these characteristics will have a high likelihood of producing a sustained reduction in teen sexual behavior among their participants. Well-designed and well-implemented abstinence education programs can reduce teen sexual activity by at least one half for periods of one to two years, substantially increasing the number of adolescents who avoid the full range of problems related to teen sexual activity.
Characteristics of Effective Abstinence Education Teachers
One of the most important characteristics listed above has to do with the classroom teacher or instructor. This feature deserves more attention. In addition to measuring the content of an educational program, an effective evaluation pays attention to the process by which that program is delivered, including the teacher’s critical role as an element in the educational paradigm. The typical objective of a classroom teacher of traditional subjects is the students’ acquisition of knowledge or the understanding of concepts. When attempting to reduce teen pregnancy and STIs, however, transmitting knowledge is not enough, because the ultimate objective is to change students’ values, their personal efficacy, resistance to negative peer influence, the decisions they make about sex in real-life situations, and ultimately their behavior. Teachers play an even more important role in the process of influencing values and behavior than they do in the transmission of knowledge. We have found that in addition to adequate competence in basic teaching skills, the teacher’s bond with the student, commitment to the program, modeling of the desired outcomes, and sense of self-efficacy in influencing youth are key factors influencing students’ cognitive, affective, and behavioral outcomes. Research on teacher effectiveness in both the traditional classroom as well as in risk behavior interventions supports this point of view.
Studies of teacher effectiveness have shown that the quality of the interpersonal relationship between the teacher and student contributes to higher credibility and greater student impact (Burke & Nierenberg, 1998; Wentzel, 1997, 1998, 2002; Peart & Campbell, 1999; Howard, 2002). Furthermore, a teacher’s high expectations (Ennis, 1998; McEwan, 2002; Peart & Campbell, 1999; Wentzel, 2002), and attitudes towards the program (Serow 1994; DeGaston, et al., 1994) increase positive change in students. Finally, program implementation and success are highly dependent on teacher qualities, including the key role played by teachers in achieving fidelity to program implementation (Dusenbury, et al., 2003), whether for academic results (see Abbott, 1998; Whitehurst, 1994), prevention of drug or alcohol abuse (Hansen, 1991; Rohrbach, 1993) or other aspects of school climate and success (Haynes, 1998)In summary, the effectiveness of a program intervention designed to influence adolescent values and behavior is dependent in many ways — especially in sex education — on the qualities of the teachers assigned to implement it. If abstinence education teachers strive to make personal connections with their students, feel confident in their ability to have an impact on teens (high self-efficacy), are enthusiastic about the program, and model the lifestyle they are teaching, they will have optimum success in influencing their students to avoid sexual activity and the problems that flow from it. Selection, training, monitoring, and feedback to teachers all contribute to effective teaching of abstinence education.
VII. Ideology vs. Evidence in the Sex Education Debate
In the debate about sex education, we typically debate the solutions but don’t adequately examine the ideological premises that underlie them. In reality, whether one favors a risk-reduction approach to sex education or a risk-avoidance, abstinence-until-marriage approach is very much influenced by one’s underlying ideology or philosophy regarding the role of sexuality in human relationships.
One of the revealing moments in the history of the sex education debate occurred during a House of Representatives Committee hearing in April 2008. Seven members of a panel testified before the Committee on behalf of groups such as the American Public Health Association (Georges Benjamin), the American Academy of Pediatrics (Margaret J. Blythe), the Halbren Department of Population and Family Health at Columbia University (John Santelli), the Institute of Medicine of the National Academies (Harvey Fineberg), and the Institute for Research & Evaluation (Stan E. Weed). In the course of this hearing, the panel members were asked a simple yes or no question by Representative Virginia Foxx (Republican, North Carolina). The question was, “If [you were] provided evidence that abstinence education programs are as [effective] as or more effective than comprehensive sex education, would you support optional federal funding for such programs? Five of the seven panel members voted “No.” Only two (Weed and Fineberg) voted “Yes.” [Hearing before the Committee on Oversight and Government Reform, House of Representatives, One Hundred Tenth Congress, Second Session, April 23, 2008].
If abstinence education were proven to work — to delay sexual involvement, reduce sexual partners, diminish STIs and pregnancies, and protect against the harmful psychological consequences of premature sexual involvement — what could be the possible reason to not support it? It’s significant that abstinence education not only lacks support from its opponents but also has been aggressively targeted for government defunding. The SIECUS organization, for example, states on its web site (www.siecus.org) that since its inception, “SIECUS has been tracking abstinence-only-until-marriage programs, advocating for an end to federal funding for these programs, and helping educators and parents keep these harmful programs out of their schools.”
Passionate statements of opposition to abstinence education from CSE and its allies have a familiar ring when we recall the history of sex education in America. Mary Calderone, Alfred Kinsey, Margaret Sanger, and other early twentieth century champions of sexual freedom expressed similar strong opposition to premarital abstinence and other traditional restraints on sex. The ideology of sexual freedom culminated in the sexual revolution of the 1960s and 1970s and helped to shape a sex education that carried a similar message to the young: They, too, should be free to enjoy the pleasures of sex with as few restrictions as possible. In 1988, Debra Hafner, then executive director of SIECUS, wrote in SIECUS Report that teens should
. . . explore the full range of safe sexual behavior. . . . a partial list of safe sex practices for teens could include talking, flirting, dancing, hugging, necking, massaging, caressing, undressing each other, masturbation alone, masturbation in front of a partner, and mutual masturbation (Hafner, 1998, p. 9)
The Philosophy of Abstinence Education
Contrast the ideology of maximizing sexual freedom, including sexual experimentation by the young, with the following statement of the philosophy of abstinence education, based on our work over four decades with abstinence education programs:
To summarize the fundamental ideological differences between the two major approaches to sex education: The philosophy that seems to us to underlie the CSE, risk reduction approach views sex as pleasure-seeking that does not necessarily involve commitment, love, or even emotional engagement. In this view, normal sexual development should include early experimentation and discovery. Sex education based on this view says, “Make sex safer, but don’t constrain it.”
By contrast, the philosophy underlying abstinence education views sex as much more than physical intimacy. From this perspective, sex also includes emotional, intellectual, and moral dimensions and therefore requires commitment, trust, maturity, and exclusivity in the relationship — the conditions most likely to occur within the context of a committed relationship historically known as marriage.
Where Does Evidence Come into the Picture?
What is the role of evidence in this clash of ideologies? In a rational world, of course, evidence should matter. Programs that can cite credible, rigorous evidence of effectiveness deserve the support of the relevant stakeholder groups, from legislators to academics, boards of education, principals, sex education teachers, parents, the media, and young people themselves. We have argued that an objective review of the best evidence to date shows that the risk reduction or CSE model has not achieved its stated goals of reducing teen pregnancy or STIs or increasing CCU, whereas a growing number of peer-reviewed studies find that well-designed abstinence education programs have been successful in achieving their goal of reducing teen sexual activity and maintaining that reduction for a year or more after the program.
In the light of this evidence, why is there such fierce opposition to abstinence education from the CSE camp and its allies? Some CSE partisans may be misled — by a pro-CSE media and by research reviews claiming positive results for CSE when the data show otherwise — into sincerely believing that teaching students to use condoms “works” and encouraging them to abstain doesn’t. But it seems to us that something deeper than beliefs about effectiveness is also operating, namely, an unwillingness to abandon a fundamental, less publically acknowledged agenda: breaking down restrictive barriers surrounding sex.
Our analysis leads us to conclude that at the end of the day, at least for the advocates of maximizing sexual freedom, the sex education debate is often less about evidence and more about ideology. That seems to be the most plausible explanation of why opponents of abstinence education say they wouldn’t support it even if the evidence showed that it works. We hope our chapter will encourage all concerned groups to take a fresh look at what the accumulated research on sex education really shows about the relative effectiveness of CSE and abstinence education, and to objectively consider the philosophy and rationale that undergirds these different approaches to teaching about human sexuality.
From its early beginnings, sex education in America has generated considerable controversy that continues unabated today. Our chapter has examined the goals and effectiveness of the three rival approaches — risk reduction (CSE), risk avoidance (abstinence education), and a combination of those two strategies — that currently compete for the support of educators and other stakeholder groups. We have argued that a close examination of the evidence shows that the risk reduction and combination approaches have thus far not achieved their professed goals of reducing teen pregnancies or STIs or even the more modest intermediate goal of getting sexually active teens to use condoms consistently. By contrast, well-designed abstinence education programs — those that provide adequate dosage, target important mediating causal mechanisms, utilize effective teachers, etc. — have achieved significant reductions in teen sexual activity that are still evident a year or more later. The evidence suggests that it may in fact be easier to get a teenager to abstain from sex than to use a condom consistently.
We have also sought to show that beneath this debate about evidence are fundamental philosophical differences among the competing approaches concerning questions of values and beliefs: What is the purpose of human sexuality? What is its role in human relationships and its connection to society’s stake in healthy families and communities? Is sex just about personal pleasure-seeking, or is it tied to higher values and a bigger vision that includes committed love and responsibility to one’s community? We think it is clear from mounting social science evidence that the goal of the sexual revolution — to break down all restrictions on sexual freedom — has carried a very high societal cost, including an epidemic of STIs, unwed pregnancies, fatherless families, and the many psychological repercussions of temporary sexual relationships.
Ultimately, in choosing an approach to sex education, the question we must answer is, “What is truly in the best interest of children, families, communities, and society as a whole?” We think the best answer lies in programs that deal with students as whole persons — that foster their development of a future orientation, their respect for self and others, their self-control, their capacity to delay gratification, and their concern for the health and happiness of those persons whose lives they impact, including those with whom they become romantically involved and the children they may someday bring into the world. Effective abstinence programs are designed to foster these very qualities and have in fact produced higher levels of sexual self-control in teens exposed to them. For that reason, we think they merit support from those seeking seek better solutions to teen pregnancy, single parenthood, STIs, and the emotional consequences of premature sexual activity.