A statistical modelling approach to community prevalence data

Sample survey techniques are often used to assess the prevalence of illness in a community and to determine any variation with environmental, social and demographic factors. Analysis of survey data is often carried out using several elementary statistical procedures. The formulation of a statistical model is an effective way of conducting a unified analysis. The model provides a concise description of the study population and is a most effective way of summarizing community prevalence data. The testing of statistical hypotheses is equivalent to model simplification and is conveniently performed using general procedures. Two alternative statistical models are given for an investigation into the extent of minor psychiatric morbidity in Perth, Western Australia.