The purpose of this research is to improve understanding of the effectiveness with which American adolescents practice contraception, as well as of those factors which contribute to differences in effective use and to estimate the extent to which unintended pregnancy could be reduced if the contraceptive effectiveness of sexually active teens in the U.S. could be improved. The project will calculate contraceptive use-effectiveness rate of adolescents in the U.S. by investigating and making appropriate adjustments for the substantial underreporting of unintended pregnancies ending in abortion that occurred in Cycle III of the National Survey of Family Growth (NSFG), the only nationally representative data set available that contains the information necessary to calculate life-table measures of contraceptive effectiveness. To obtain the data needed to estimate the extent of such underreporting and to estimate accurate rates of contraceptive use-effectiveness, a nationally representative sample of women will be surveyed regarding their contraceptive use prior to conception. Data from Cycle IV of the NSFG will extend analysis to cover contraceptive use intervals from 1979 through early 1987. Using the NSFG data adjusted for underreporting, factors that might affect contraceptive effectiveness will be analyzed, using hazards-model analysis to test the hypothesis that there are no significant differences across subgroups of adolescents in important factors. Comparisons will be made between teens and adult women of the levels of contraceptive use-effectiveness and factors influencing effectiveness rates to test for differences in the rates and the hypothesis that factors influencing differences in effectiveness rates are the same for teens and adults. The current contribution of contraceptive failure to the levels of unintended pregnancy among U.S. adolescents will be calculated. The reduction in unintended pregnancy possible through improving contraceptive effectiveness to levels observed in some subgroups of the adolescent population, or in older women, will be estimated along with the reductions possible through changes in levels of contraceptive use among adolescents and through changes in the distribution of methods among users. Finally, levels of underreporting of unintended pregnancies ending as abortions in other national surveys - Cycle III of the NSFG, the 1976 and 1979 Johns Hopkins Surveys of Young Women and the National Longitudinal Survey - will be calculated by population subgroup. The effects of different survey designs and question wording on the level of underreporting will be evaluated and ways to improve pregnancy reporting will be identified.
