Between 1971 and 1981 the race difference in illegitimacy rates among unmarried adolescents was reduced by half. This reduction resulted from an increasing rate among whites and a declining rate among blacks. The proposed research will integrate aggregate and individual level analyses in order to develop a general explanation of this historically unique trend. Specifically, the proposed research will first use a log-decomposition procedure to identify the relative contributions of four major components of illegitimacy to both racial differences in the adolescent illegitimacy rate, and the convergence in those rates between 1971 and 1981. These components, or proximate determinants, are: the proportion sexually active; the proportion of the sexually active who become pregnant; the proportion who have a live birth among those who conceive; and the proportion who marry before the birth occurs. Second, the research will determine the contribution of selected individual characteristics to the race differences (and their convergence) in each of the proximate determinants. These contributions will be allocated, using a regression decomposition procedure, to compositional (mean) and effect (slope) differences in the characteristics. Finally, the results from the first two steps will be integrated and the ways in which the individual level characteristics operate via the proximate determinants to affect the race differences and convergence in illegitimacy rates will be determined. In this way the process by which the individual characteristics operate to affect trends and differences in illegitimacy can be identified. The above tasks will be accomplished using two data sets, the 1971 National Survey of Young Women and the National Survey of Familly Growth - Cycle III. The former data set will be used to calculate the levels of the proximate determinants at year 1 (1971) and to estimate the individual level regression equations at year 1. All year 2 (1981) analyses will be performed with the NSFG-III.
