This document summarizes statistical analysis techniques for assessing the validity of survey results on violence against women. It discusses how stratified analysis and multivariate logistic regression can be used to control for confounding variables. An example is provided showing how multivariate analysis revealed that poverty, urban residence, and high parity independently increase women's risk of violence. The document also describes how more advanced techniques like life table analysis can provide additional insights, such as that high parity may result from violence rather than be a risk factor for it.