This Magic Moment: Predictors of Non-Resident Father Involvement in Parenting
As a result of these shifts in both the make up of the family and in norms about the role of fathers in parenting, the large existing literature on paternal financial support (Furstenberg, Sherwood & Sullivan 1992; McLanahan, Garfinkel & Robins 1994) has been joined by a growing interest in father involvement in direct care activities.
A large body of evidence demonstrates that children in non-married families are more likely to experience behavioral problems than children living in two-parent married families (McLanahan & Sandefeur 1994). Growing up in a household with non-married parents has been shown to be associated with lower levels of school engagement, as well as behavioral and emotional problems (Brown 2004). Similar work has found that children who spend time in cohabiting families are at risk for poor cognitive and behavioral outcomes, while children brought up in single-parent homes face the greatest levels of risk (Carlson & Corcaran 2001; Hofferth 2006; Amato 1994; Wu, Cherlin & Bumpass, 1996; Magnuson & Berger 2009).
This paper explores whether the norms of active fathering have been adopted by low-income, unmarried, non-resident fathers, and whether these men are active parents one and three years after birth. I draw on data from the Fragile Families and Child Wellbeing Survey, a longitudinal birth cohort study (Reichman, Teitler, Garfinkel & McLanahan 2001).
I draw on a range of father and mother characteristics from the baseline (birth) survey that may be associated with father involvement. Specifically, I identify 5 overarching categories of potentially important predictors of father involvement: (1) fathers’ financial and other support during pregnancy (2) fathers’ presence and actions in the hospital at the time of birth (3) fathers’ perceptions of the paternal role (4) expectations about the likelihood of marriage (5) parental relationship quality.
I employ a series of progressively controlled OLS and logistic regressions models to estimate the extent to which baseline characteristics predict paternal parenting activities when the child is one and three years old. Parenting activities include: play activities; direct care; in-kind support; time spent with child; and cash support. I find that paternal and maternal conceptions of the paternal role, as well as father/mother family background are largely not predictive of later involvement. Surprisingly, I find that fathers presence in the hospital at the time of birth or soon after is strongly predictive of paternal parenting one and three years later; these results remain even when controlling for a host of father and mother characteristics, birth histories, and relationship quality. Last, I attempt a series of instrumental variable analyses to further parse this association.