The Association Between Wife Assault and the Socioeconomic Characteristics of Women and Their Families

Abstract

This thesis is composed of three essays on the topic of wife assault and is based on the 1993 Violence Against Women survey (VAWS) data. In the first two essays, we estimate the impact of economic status on the frequency of wife assault using several models and large random samples. Tauchen, Witte and Long (1991) (TWL) find that both female and male income have a significant impact on the frequency of domestic abuse. Their study is based on a small, self-selected sample of battered women (n=l25) . In the first essay, we find similar results to TWL when we use a sample and a model much like theirs. However, when we use a Zero Inflated Poisson model and a less self-selected sample of physically abused women (n=877), we find there is less evidence of a significant relationship between either female or male income and the frequency of abuse. In the second essay, we expand the sample to include all currently married women (n=5596) . Again we estimate several models and find little evidence of a significant relationship between either female or male income and the frequency of abuse. In the third essay, we investigate the source of the positive association between the frequency of wife assault in a first marriage and the probability of that marriage ending. We find that the association between abuse and the decision to separate remains significant and largely unaffected when all VAWS variables thought to influence the decision to separate are added to the probit. Most of the variables that are not available in the VAWS do not appear to be capable of generating a spurious positive relationship between the frequency of abuse and marital dissolution. We also find no evidence of systematic reporting differences that might generate a spurious association between abuse and dissolution.