Empirical Work in Family Law

Until fairly recently, researchers have not done much theoretical work on the subject of family law. Although the move towards theoretical work is a positive one, unfortunately, most of the latest reforms in family law have been uninformed by empirical studies. Furthermore, the few empirical studies that have been conducted are replete with intractable problems. In this essay, Margaret Brinig discusses some of the problems researchers have encountered in their attempts to conduct empirical work in the area of family law. For example, most researchers have used state cross-sectional data for their experiments. Reliance on this type of data can be misleading because a population’s characteristics can vary even within state lines. This casts doubt on whether a researcher’s conclusions are generalizable over any particular segment of the state’s population. Additionally, researchers are often forced to conduct their experiments using incomplete data. Because family law involves many highly sensitive issues, much of the data that could be used for empirical studies is not available to the public. Establishing a true control group has been problematic for the very same reasons. The author further points out that much of the empirical work has been conducted by individuals with a stake in the outcome of their research. For example, a researcher who was a single father conducted a well-known study on the effects of father custody on children. The author implores researchers and legislators to stop reacting so quickly to a particular study. First, she argues that many experiments are not generalizable over other populations. An experiment in Minnesota suggested that mandatory arrest of domestic abusers led to a low rate of recidivism. Courts and legislators reacted to these figures with swift reform. Future experiments, however, uncovered that mandatory arrest only works when the abuser has a stake in the community or employer reputation. Otherwise, mandatory arrest actually increases recidivism. Second, there is evidence that individuals do not care about what a particular law says. The author cautions us that we should not credit the existence of a particular law for any behavior in a population. Future research should try to eliminate other possible causes. The author concludes this essay by suggesting that future research should focus on topics such as the nonfinancial costs of divorce on the family, the effects of various custody patterns on children, the relative costs of divorce and staying together, and, in the case of adoption, whether good or bad parents can be identified in advance.

The full text of this Symposium is available to download as a PDF.