The Fragile Families and Child Well-Being Study estimates that 30 percent of unmarried births in large urban areas are to non-cohabiting, but romantically involved parents, termed couples in a visiting relationship. If the proportion of births to visiting couples in the Fragile Family sample were applicable nationally, parents in visiting relationships would account for 12 percent of all births. As researchers give more attention to non-marital births through visiting relationships, there is a similar interest in what happens to father involvement as non-marital children mature.

A previous study based on four waves of the Fragile Families and Child Wellbeing Study (FFCW) by Tach, Mincy and Edin estimated the trajectory of father involvement after visiting unions dissolved, e.g., after romantic involvement had ended. Because fathers in visiting unions are nevertheless non-resident fathers, Mincy, Pouncy and Zilanawala in a previous APPAM paper used five waves of the FFCW to estimate trajectories of father involvement for waves in which couples were still romantically and after such couples broke up.  The odds that a non-resident father still in a romantic relationship visited his child at least once per survey wave were 20 times greater than a peer non-resident father who had been in a cohabiting relationship.  After his romantic relationship ended, there was no significant difference in the odds of visitation between the two such fathers.  

The present paper offers two innovations. First, it examines an additional father involvement measure, sleepover nights in the last year. Second, it estimates Generalized Estimating Equations to account for the non-normal distributions of the three father involvement measures used thus far (any involvement since the last wave/year, days of involvement in the previous month, and sleepover nights). The models control for economic, behavioral, and subsequent partnership and fertility information collected from both the mother and father.

The paper’s main finding occurs at the intensive margin of father involvement, that is, days of father involvement last month, for visiting couples. The race of the mother is highly correlated with the variable measuring whether she is currently involved in a romantic relationship with the father. This means that studies examining non-resident father involvement that cannot distinguish between visiting couples and other non-resident couples are conflating the effect of race and romantic involvement.

The paper also estimates the effect of maternal and paternal re-partnering and confirms “perturbation superiority’ for mother re-partnering at the extensive margin, any visits since the last survey wave.  It finds ‘perturbation parity,’ meaning equal effects of mother or father re-partnering on visitation in days last month, and ‘perturbation superiority’ for father re-partnering when the dependent variable is sleepover visits.  

The paper suggests that the national census should follow the lead of Caribbean governments, the Millennium study of Great Britain and the Family Families Survey and include a ‘births by visiting relationship’ category in its non-marital births supplemental reports.