The importance of the common family background for the similarity of divorce risks of siblings: a multi-level event history analysis

Abstract

This study examines the effects of a family's and individual children's characteristics on the probability of having a divorce. Current research shows a clear indication of increased divorce risks if an individual's parents or siblings have experienced a divorce. Explanations include both shared family characteristics (including genetic effects) and common characteristics of the individual children involved. This study analyzes the effects of shared family background characteristics on the divorce risk of individuals. By analyzing siblings within families and including individual children's characteristics in the analysis, it is possible to separate individual-level and family-level effects.

In addition to employing a multi-level structure of individual siblings nested within families, the data cited here are censored. For all individuals, the length of the marriage and the divorce status are known, but the divorce status is interpreted differently for individuals who have or have not experienced divorce. For divorced individuals, the final divorce status is known; for individuals who have not experienced divorce, the final marriage status is unknown or censored. The proper analysis model for such data is event history (also called survival) analysis. This study therefore employs a multi-level event history model.

Our results show that there is a similarity in the divorce risks of siblings from the same family, which is not explained away by the available child and family characteristics. This finding suggests that shared genetic and social heritage play an important role in the intergenerational transmission of divorce risks.

Citation

Dronkers, J. and Hox, J. (2006), "The importance of the common family background for the similarity of divorce risks of siblings: a multi-level event history analysis", Yammarino, F.J. and Dansereau, F. (Ed.) Multi-Level Issues in Social Systems (Research in Multi-Level Issues, Vol. 5), Emerald Group Publishing Limited, Leeds, pp. 217-238. https://doi.org/10.1016/S1475-9144(06)05010-7