Does Fathers' Residency Matter? Paternal Influence on Children's Health and Development

Abstract

This dissertation focuses on the effects of fathers' residency status and father involvement on children's health and development in the United States. The data for this dissertation were obtained from the Fragile Families and Child Wellbeing Study (Fragile Families) and the Early Childhood Longitudinal Study- Birth Cohort (ECLS-B). Chapter 1 examined the association between fathers' residency status and child BMI. We performed a series of cross-sectional linear regression and propensity score matching analyses using three waves (Years 3, 5, and 9) of the Fragile Families data. We did not find a significant difference in BMI between children who had residential fathers and those with nonresidential fathers at any age. Chapter 2 assessed item bias in the father involvement scale used on the 24-month ECLS-B questionnaire for resident and nonresident fathers. We used Differential Item Functioning (DIF) to detect whether the 17-item scale measured father involvement fairly for nonresidential and residential fathers. We found that almost half of the items (8 of the 17 items) showed signs of uniform DIF. That is, the eight items were biased in favor of residential fathers. One item showed signs of nonuniform DIF, which indicated that item was consistently bias towards one residential type over the other. The remaining eight items did not show signs of bias, and therefore could be used as anchor items on the father involvement scale. Chapter 3 examined the intergenerational effects of father involvement on children's social development. We used structural equation modeling (SEM) to model father-child interactions across three generations. We found positive 1st generation father-son interactions to be significantly associated with more frequent 2nd generation father involvement in cognitively-stimulating activities (e.g. reading books, singing songs and telling stories to children), which was also found to be associated with higher teacher reports of externalizing behaviors in children. We stratified the data by child sex and found differential effects for boys and girls. In boys, we found that positive 1st generation father- son interactions was associated with more frequent 2nd generation father involvement in cognitively-stimulating activities, which was also significantly associated with higher teacher reports of internalizing behaviors. However, there were no significant associations between 1st generation father-son interactions and externalizing and internalizing behaviors in girls.