Effects of family structure on the risk of first premarital birth in the presence of correlated unmeasured family effects

This paper assesses the effects of family structure on the risk of a first premarital birth for a sample of women from the 1979 National Longitudinal Survey of Youth. The sample reflects the family structure and family formation experiences of a cohort of women who were at risk of out-of-wedlock childbearing during the 1980s and early 1990s. We focus on assessing the effects of family structure in the presence of correlated unmeasured family effects, which are identified through the use of sibling data. The availability of multiple sibling respondents per family permits identification of family-level unobserved heterogeneity in a multi-level context of individuals nested within families. Our models account for family-specific sources of unobserved heterogeneity in the processes generating family structure and nonmarital childbearing, and provide estimates of the association between these sources of unobserved heterogeneity along with the effects of family structure and other covariates. We find that accounting for the correlation between unobserved family-level effects in processes generating family structure and first premarital birth leads to attenuated estimates of the effects family structure on the risk of first premarital birth. This suggests that other family-level factors may play a mediating role in generating both family structure and nonmarital childbearing.