Interrelationships between life course events in the United States

Bohyun Joy Jang, Ohio State University
John B. Casterline, Ohio State University
Anastasia R. Snyder, Ohio State University

Family events are closely related to residential changes (Kulu & Milewski 2007; Clak & Withers 2007). People consider residential moves in response to changes in family size or in anticipation of new family members. While previous research has demonstrated interrelationships between mobility and fertility (Kulu & Steele 2013), relatively little is known about the association between mobility and union transitions. Union formation (i.e., marriage and cohabitation) may trigger residential changes because of a need for additional space. In addition, mobility may influence union transitions. For example, researchers have found that family formation behaviors are related to housing career and homeownership (Murphy & Sullivan 1985; Mulder & Billari 2010), which mostly require residential changes. Using data from the National Longitudinal Survey of Youth 1997, we examine how the two life course transitions are linked to each other. We separate union types into marriage and cohabitation and mobility into migration (a between-county move) and residential mobility (a within-county move). As individuals experience multiple union transitions and residential changes over the life course, we use a multi-level, multi-process, competing-risks model allowing for person specific characteristics. In the NLSY97, about 14% of the sample has experienced marriage without cohabitation and 57% have cohabited. Regarding mobility, about 18% have moved within the same county and 59% have changed residence to different county. Preliminary findings from separate estimation of each transition suggest that migration and residential mobility are the most critical determinants for marriage and cohabitation, and vice versa. Moreover, we found significant person specific random effects in each equation. We will estimate two multi-level competing risks models simultaneously for a full model.