Comparison Of Variances Due To Design Types In With And Without Replacement Using Two-Stage Sampling

In some applications, it is cost efficient to sample data in two or more stages. In the first stage, a simple random sample is drawn and then stratified accordingly to some easily measured attributes. For each subsequent stage, a random subset of previous selected units is sampled for more detailed observation, with a unit\'s sampling probability determined by its attributes as observed in the previous stages. This paper described two-stage sampling designs and estimating equations based on the resulting data. It also compared the resulting variances in each sampling design. The designs, the estimates and the corresponding variances are illustrated with data on diabetic patients in Niger State. JORIND Vol. 5 (1) 2007: pp. 78-85