ABSTRACT
Many persons analyzing survey data are interested in learning about associations between variables measured in the survey. Are members of some demographic groups more likely to be crime victims than others, or are persons with health insurance less likely to have high blood pressure? Chi-square tests are often used to explore relationships among variables, but clustering and stratification affect the properties of the test statistics. This chapter tells how to carry out chi-square tests and fit loglinear models with data from a complex survey.
