A recent study by Rosenbaum in Pediatrics  concludes that the sexual behavior of virginity pledgers does not differ from that of closely matched nonpledgers, and pledgers are "less likely to protect themselves from pregnancy and disease before marriage." The media immediately pounced on the study results and concluded that abstinence education "does not work." All in spite of other reports that virginity pledges help delay sexual debut in adolescents.
The Rosenbaum article states that the study results show no significant differences in sexual behaviors or sexually transmitted diseases (STDs) in pledgers and nonpledgers. But general statements such as "virginity pledges are not effective" or "abstinence education programs do not work" are unwarranted conclusions from the results here. All the study results really show are that in a particular group of adolescents with religious, conservative attitudes and beliefs there were no significant differences amongst them in sexual behaviors and STDs; and pledges do not seem to have any added effect to the pre-existing beliefs and attitudes of that particular group.
Here are some points about the study that need to be considered carefully before drawing over-arching conclusions:
– According to the author, there were baseline differences between the pledgers and all nonpledgers, hence it was necessary to match pledgers with similar non pledgers. The pledgers were more religious, less sexually experienced, expected more negative and fewer positive effects of sex and birth control use, and had lower knowledge of birth control than non pledgers. Due to these baseline differences, the Rosenbaum study looked at only a sub-group of adolescents.
– The author matched pledgers and nonpledgers so there would be no baseline differences between the pledgers and nonpledgers in the study group. In fact, the author says that there were no significant differences seen when comparing 128 characteristics in the pledgers and nonpledgers. However, the author also says that about 5% significant differences are seen on average in any comparison. These two facts taken together suggest that there may have been some "over-adjustment" during the matching process – making the groups so similar that there may be no real differences at all.
– The biggest concern is the representativeness of the sample. The sample under study is not representative of all teens in the country – simply because the author looked at a very small group with specific characteristics. So the study results cannot be generalized to all US teens. This is also one of the study limitations the author acknowledges in the article.
– Pledgers are reported as less likely to use condoms/birth control as compared to nonpledgers. Although pledgers were less likely to use condoms over the last year, their condom use was not significantly different during last sex. The author also states that abstinence programs and pledges do not prepare the youth to "protect their health." Apparently those who did not pledge were also unable to "protect their health" since there was no difference in STDs in both groups. So how are virginity pledges specifically to blame for the health status of the study group?
– The author is quick to draw a causal inference from the data analysis results. The article states that pledgers may be less likely to use condoms/contraceptives because abstinence programs cause participants to develop negative attitudes about their effectiveness. Since the study does not look at youth who had participated in abstinence education programs vs. youth who did not participate in such programs, it is unclear how the author can say abstinence programs cause less condom/contraceptive use from her data and results.
– The article states that the study results are biased in favor of showing a "pledge effect" since the pledgers would be less likely to report sexual activity. It is also important to note that 82% of the pledgers did not remember they had pledged. If most of the youth do not remember having pledged at all, the sexual decisions or reporting of most of the pledgers could not possibly be driven by the pledge. And under-reporting of sexual activity is common in all adolescents due to social desirability bias (the desire to report what is perceived to be more socially acceptable). Religious and conservative teens would be as likely, if not more so, to under-report their sexual activity as well. In this case, under-reporting in both groups would not be likely to swing the results in either direction.
In fact, it may be that the results of the study were due to the underlying beliefs of the study group. Having taken a virginity pledge probably did not add anything to the sexual decision making of the youth in the study. It is thus a little far-fetched to conclude that abstinence education does not work if virginity pledges do not show an effect. First, not all abstinence education programs require virginity pledges. Second, a count of the number of virginity pledges is not a measure of the success or effectiveness of the program. Just like a count of the condoms distributed cannot be a measure of the success or effectiveness of a sex education program. Any program needs to be evaluated continually to obtain feedback on how the program can be improved. This means more than just providing a count of youth in the program or those involved in an event (eg, pledging) in the program.
The underlying belief system of the youth (possibly from their families) had a protective effect as compared to youth outside the study group. These youth had fewer sexual partners and later age of sexual debut. As reinforced by these results, families and parents play an important role in helping adolescents avoid risky sexual behaviors.
– It is important to ensure rigorous implementation and evaluation of any sex education (including abstinence education) program to see how a prevention message can be delivered effectively.
– Parents need to talk to their children about sexual health issues and guide them toward the healthiest sexual decision making.
1. Rosenbaum JE. Patient teenagers: A comparison of the sexual behavior of virginity pledgers and matched nonpledgers. Pediatrics. 2009 Jan;123(1):e110-20.
2. Martino SC, Elliott MN, Collins RL, Kanouse DE, Berry SH. Virginity pledges among the willing: delays in first intercourse and consistency of condom use. J Adolesc Health. 2008 Oct;43(4):341-8. Epub 2008 Jun 5.
[12 February, 2009, The Medical Institute Responds to the Virginity Pledge Report, www.medinstitute.org]