In examining the relation between violence against women and women's economic dependence, existing literature treats the incidents of violence either as a binary or as a continuous variable. However, the incidents of violence is a count variable and, quite often, data on the number of violent incidents is categorized. This paper estimates the relation between violence against women and economic dependence by using a categorized negative binomial regression model. The model is suitable for categorized count data and thus provides a more accurate estimation of the relation than what is provided in the literature. Data analyses in this paper show that less economic dependence of women is associated with less violence.