The Impact of Time-Varying Exposure to a Two-Parent Household on Black-White Skill Gaps at School Entry

Jill Bowdon, University of Pennsylvania

In this study, I employed marginal structural models with inverse probability of treatment weighting to estimate the effects of time-varying exposure to an intact family structure on teacher ratings of academic engagement in kindergarten. This method of estimating time-varying treatment effects is an improvement over conventional regression models because it does not “control away” the effect of prior exposure to father presence operating indirectly through the time-varying characteristics of the child or family. I found that father presence in both infancy and the year prior to kindergarten has a positive effect on approaches to learning that is comparable in size to its effect on reading ability and externalizing behavior. Given the fact that only one-third of black children had a biological father present in the household in both infancy and at age four, these estimated effects of father presence can help explain black-white skill gaps at kindergarten entry.