Abstract

PIP: Using data (the initial sample was 5533 women aged 14 to 24) from the National Longitudinal Surveys conducted by Ohio State University's Center for Human Resource Research, and the log linear probability technic, the impact of family background variables and personal characteristics on the timing of 1st birth are examined. These factors are expected to affect age at entry into motherhood directly, and indirectly, through education. A variable which increases the education attainment of either father or mother will indirectly delay childbearing. The study attempts to disentangle direct and indirect influences, and determine whether some effects may be nonlinear. Regarding other variables: 1) maternal employment only slightly increases the likelihood of an early 1st birth; 2) a broken home is associated with more than a 100% increase in the probability of an early birth; and 3) planning to be employed at 35 is associated with a higher likelihood of bearing the 1st child at an early age, confirming the human capital theory that the optimal amount of investment in human capital is a positive function of how long one plans to use it. Total effects of race on both age at transition to motherhood and school are great, being almost double for nonwhites for probabilities of an early 1st birth; the probability of having dropped out of school by the date of the 1st birth is 0.3185 for nonwhites, but only 0.2260 for whites. In conclusion, parental education, race, and whether the young woman grew up in an intact family are significant variables. Number of siblings and plans for age 35 have opposite direct and indirect influences. Nonlinearities in the impact of some explanatory variables on age at 1st birth and education are shown. This study illustrates the potential usefulness of log linear probability models, and may provide a framework for the statistical analysis of a wide range of economic issues.