First published online:

| DOI: https://doi.org/10.1363/4113209

Fathers' Pregnancy Intentions

We read with great interest "The Link Between Couples' Pregnancy Intentions and Behavior: Does It Matter Who Is Asked?" [2008, 40(4):194–201], in which Maureen R. Waller and Marianne P. Bitler discuss the importance of evaluating both parents' pregnancy intentions in measuring the effect of unintended childbearing on parental behaviors during pregnancy. Knowing that pregnancies that were unintended by at least one parent are associated with mothers' receipt of prenatal care and with fathers' provision of material support during the pregnancy is important. Late initiation of prenatal care and lack of financial support from nonresident fathers are associated with negative pregnancy and childhood outcomes.1, 2 We heartily agree with the authors' conclusion that future research on pregnancy intentions should include fathers. Further, we believe that the impact of a father's unintended pregnancy involvement extends beyond the duration of pregnancy and is associated with long-term child outcomes.

To evaluate our hypothesis, we used data from the 2002 National Survey of Family Growth to examine associations between resident fathers' self-reported involvement with a child younger than five and their pregnancy intentions at the time of that child's conception. We focused on two conceptual domains of fathers' involvement: caregiving (including such activities as feeding and bathing the child) and optional social interaction (including such activities as going on outings with the child and playing with or reading to the child). We hypothesized that fathers who had not intended to make their partner pregnant would be less involved than others, and that differences would be particularly pronounced in the optional social interaction domain.

Our data suggest that fathers' reported involvement in caregiving activities does not differ by intention status. The proportion who were highly involved in caregiving was 42% (95% confidence interval, 32–52%) among fathers who reported the pregnancy to have been unwanted or mistimed, and 46% (95% confidence interval, 39–54%) among fathers who reported that the pregnancy had been wanted. This slight difference was explained entirely by father's educational attainment. For optional social interaction, the proportion reporting high involvement was 53% (95% confidence interval, 41–65%) among fathers who reported that the pregnancy had been unwanted or mistimed, and 63% (95% confidence interval, 56–70%) among those who said that it had been wanted. Unfortunately, the small size of the available sample results in a lack of precision of these estimates.

As Waller and Bitler discuss, the link between pregnancy intentions and health and developmental outcomes in infants is well established, but little research has examined the impact of fathers' pregnancy intentions on child outcomes either in infancy or later in life. The literature clearly demonstrates that fathers are important in children's cognitive, emotional and educational development.2, 3 Our analysis suggests that fathers may be less involved with children whose conceptions were unintended than with children whose conceptions were planned, even if the fathers and children reside together—a finding that corroborates the suggestion that a father's pregnancy intentions have long-term effects on children. Larger studies are needed to confirm these suggestive data.

That 3.1 million unintended pregnancies occur each year in the United States4 is unacceptable. The Institute of Medicine's 1995 report Best Intentions5 stated that little was known about the impact of unintended childbearing on fathers and that what information is available was based on adolescent fathers. This is still true today. Clearly, more work needs to be done to reduce the number of unintended pregnancies. This endeavor must include research on fathers' pregnancy intentions and must involve men in efforts to reduce unintended pregnancy. We applaud Waller and Bitler for giving much-needed attention to this area.

Brianna Magnusson and Kate Lapane
Department of Epidemiology and Community Health
School of Medicine
Virginia Commonwealth University
Richmond, VA

1. Kotelchuck M. The adequacy of prenatal care utilization index: its U.S. distribution and association with low birthweight, American Journal of Public Health, 1994, 84(9):1486–1489.

2. Cabrera N et al., Fatherhood in the twenty-first century, Child Development, 2000, 71(1):127–136.

3. Harris KM, Furstenberg FF, Jr., and Marmer JK, Paternal involvement with adolescents in intact families: the influence of fathers over the life course, Demography, 1998, 35(2):201–216.

4. Finer LB and Henshaw SK, Disparities in rates of unintended pregnancy in the United States, 1994 and 2001, Perspectives on Sexual and Reproductive Health, 2006, 38(2):90–96.

5. Brown S and Eisenberg L, eds. The Best of Intentions: Unintended Pregnancy and the Well-Being of Children and Families, Washington, DC: National Academy of Sciences, 1995.

Virginity Pledges May Work For Some Adolescents

The March 2009 issue includes a digest ["Virginity Pledgers Are Just as Likely as Matched Nonpledgers to Report Premarital Intercourse," 2009, 41(1):63] of an article by Janet Rosenbaum that appeared in the January 2009 issue of Pediatrics.1 Her study used data from the National Longitudinal Study of Adolescent Health (Add Health) to compare the sexual behavior, contraceptive use and STD diagnoses of adolescents who reported taking a virginity pledge with those of a matched sample of nonpledgers. Similar proportions of pledgers and nonpledgers were found to have initiated sexual intercourse by the follow-up; rates of STD diagnoses were also comparable in the two groups.

Rosenbaum's findings appear to be in direct conflict with those of a similar study that I led at the RAND Corporation, which was published in the October 2008 issue of the Journal of Adolescent Health.2 Like Rosenbaum, we used propensity score modeling to compare the rate of sexual initiation among adolescents who had taken a virginity pledge with that among a closely matched sample of nonpledgers. Unlike Rosenbaum, however, we found that pledging was associated with delayed sexual initiation.

One possible explanation for the two studies' disparate results is the difference in age of surveyed individuals: Rosenbaum investigated the effectiveness of pledges taken at age 15 or older, whereas the RAND study focused on pledges reported by youth aged 12–17. Virginity pledges may be effective mostly (perhaps solely) among younger adolescents. This possibility is supported by findings from an earlier study of Add Health youth that compared the effectiveness of virginity pledges among younger and older adolescents, and found that pledges were effective only in the younger group.3

Another possible explanation for the disparate findings is that pledges may work for only a limited period of time or at a certain stage of life. The RAND study followed youth for three years, until they were 15–20 years old, and found that 42% of pledgers had remained virgins, while only 33% of similar nonpledgers had done so. Rosenbaum's study followed participants for five years after they reported having taken a pledge, at which point they were 20 or older, and found that the mean age at first sex for pledgers and nonpledgers was 21. By that time, both groups appeared equally likely to have had sex.

Rosenbaum concludes that virginity pledges are ineffective. Our study suggests that they have some utility. Considering all the evidence, the most prudent course of action may be to offer virginity pledges as part of a comprehensive program of sex education that includes information on birth control methods and condom use. After all, most young people do not take virginity pledges, and most (65% of males and 70% of females4) have sex as teenagers. Even many virginity pledgers will have sex at a young age, and they need to know how to protect themselves from unintended pregnancy and STDs when they do.

This is not to say that all adolescents should be encouraged to make virginity pledges. The RAND study showed that pledges worked for teenagers with strong religious backgrounds or with less positive attitudes toward sex, and for those whose parents keep close track of them. Other research has indicated that pledges must be freely undertaken and that pledges are ineffective if all youth in a school or community take them.3 Instead, sex education programs should encourage virginity pledging for those adolescents who sincerely wish to make a commitment to abstinence, and should provide all young people with the knowledge and skills they need to protect themselves from unintended pregnancy and STDs once they become sexually active.

Steven C. Martino RAND Corporation Pittsburgh, PA

1. Rosenbaum JE, Patient teenagers? a comparison of the sexual behavior of virginity pledgers and matched nonpledgers, Pediatrics, 2009, 123(1):e110–e120.

2. Martino SC et al., Virginity pledges among the willing: delays in first intercourse and consistency of condom use, Journal of Adolescent Health, 2008, 43(4):341–348.

3. Bearman PS and Brückner H, Promising the future: virginity pledges and first intercourse, American Journal of Sociology, 2001, 106(4):859–912.

4. Abma JC et al., Teenagers in the United States: sexual activity, contraceptive use, and childbearing, 2002, Vital and Health Statistics, 2004, Series 23, No. 24.

Rosenbaum replies:

I welcome Dr. Martino's contrasting of our respective virginity pledge studies, because they were published simultaneously and so any contrast could not appear in the papers themselves. However, I believe the RAND results can be largely explained by uncontrolled religious and attitudinal differences between pledgers and nonpledgers.

Both studies used matching in an attempt to minimize or even remove confounding. The RAND study matched pledgers and nonpledgers on 24 prepledge factors; my study matched them on 112. Not all of these additional factors were necessary to achieve a good match, but some differed substantially between pledgers and nonpledgers, and thus were potential confounders.

For example, the RAND survey measured religiosity with a single item—agreement that "religion is very important in my life"—and neglected religious differences that might be more substantial confounding factors. The virginity pledge movement originated in conservative evangelical churches, membership in which also is an independent predictor of pledging, but the RAND survey did not assess which respondents came from conservative religious backgrounds. Before writing this response, I looked at the Add Health data from Wave 1 for respondents who had said that religion was "very important" to them. Among these teenagers, pledgers were still much more likely than nonpledgers to report behaviors that encourage abstinence independent of a virginity pledge—for example, being born-again Christians (66% vs. 42%), attending church weekly (78% vs. 61%) or a youth group weekly (57% vs. 34%), and praying daily (79% vs. 65%). Pledgers in the RAND study may have likewise been more religiously conservative and religiously involved than nonpledgers.

The RAND study also did not consider whether respondents had negative expectations about sex—feelings that may motivate teenagers both to take virginity pledges and to abstain. When I looked at the Wave 1 data for all adolescents, pledgers were more likely than nonpledgers to say that if they had sex, they would feel "very guilty" (24% vs. 8%) and their partner would lose respect for them (19% vs. 9%), and were more likely to worry that their friends would think they were "looking for sex" (30% vs. 20%).

I have collaborated with three of the five coauthors of the RAND study, and I know they do research of the highest caliber. The survey on which they based their virginity pledge study, however, did not measure a sufficient variety of religious and attitudinal variables to control for the substantial differences between pledgers and nonpledgers. Hence, it is impossible to say whether the pledge or these other factors were responsible for the observed difference in sexual activity.

Janet E. Rosenbaum
Johns Hopkins Bloomberg School of Public Health
Baltimore, MD