A significant amount of literature has examined negative outcomes for children associated with divorce. However, most research has treated divorce as a static event rather a dynamic event that may include years of conflict preceding the divorce. Thus, little research has examined the temporal effects of divorce-i.e., how the divorce process affects children in the years leading up to the divorce, at the time of the divorce, and over time after the divorce. Our objectives in the proposed work are to (1) examine these temporal effects of divorce on children's problem behavior (e.g., hyperactivity) and on teenagers' outcomes and risky behaviors (substance use, delinquency, sexual behavior, and fertility-related outcomes); and (2) examine whether gender, age, and pre-divorce parental conflict moderate the negative effects of divorce. The latter moderating factor can be informative on whether children from high-conflict families actually benefit from divorce. We will examine children from the Child and Young Adult Survey that accompanied the NLSY-79, and we examine teenagers from the NLSY of 1997 (NLSY-97). Most studies on the effects of parental divorce on children rely on the assumption that divorce is random across families. But, much research has shown that families that will divorce years later are different from families that will remain intact. A few studies use fixed-effects models to address this problem of heterogeneity possibly confounding the relationship between divorce and children's outcomes. However, these fixed-effect studies are not capturing the full effect of divorce, as the baseline measure may already capture much of the effect of the divorce process (e.g., the family conflict leading to the divorce). The few studies on the temporal effects of divorce are closer to capturing the full effects of divorce, but, they do not control for unobserved heterogeneity determining which families divorce. We will combine the positive features of these two types of models. In particular, we will address the unobserved heterogeneity issue by restricting the sample to children of families that experience a divorce (as the fixed-effect studies do). This changes the necessary assumption from """"""""divorce is random across families"""""""" to """"""""the timing of the divorce is random across families that experience a divorce."""""""" We can even eliminate the need for this new assumption by taking advantage of the panel surveys to estimate models with individual fixed effects. And to capture the full effects of divorce, we use a reference periods that is 3 to 5 years prior to the divorce for the analysis of teenagers and one that is 10 to 16 years prior to the divorce for the children's analysis. The proposed work should help the research community understand the temporal effects of parental divorce. In particular, the research will indicate how early before divorces kids may be impacted, how much the negative effects increase around the time of the divorce, and whether the divorce effects subside or persist over time. In addition, the research may be informative for whether children in high-conflict families do better with their parents divorcing or staying together. ? ? ?
| Arkes, Jeremy (2015) The Temporal Effects of Divorces and Separations on Children's Academic Achievement and Problem Behavior. J Divorce Remarriage 56:25-42 |
| Arkes, Jeremy (2012) Longitudinal association between marital disruption and child BMI and obesity. Obesity (Silver Spring) 20:1696-702 |
