Comparative study on the complex samples design features using SPSS Complex Samples, SAS Complex Samples and WesVarPc

Abstract

Unlike simple random sampling, complex sample designs involve additional considerations such as multistage sampling, stratification, and unequal probability of selection. A basic problem with complex surveys is in variance estimation which requires the use of approximate methods. Generally, such methods are based on either the Taylor series linearization or the replication techniques. Statistical software that use standard packages usually assume that simple random sampling of elements is inadequate for data analysis from complex surveys, especially for purpose of variance estimation. This study compares the complex sample design features produced by three statistical software packages designed to handle complex surveys (SPSS 16.0 Complex Samples, SAS 9.0 Complex Samples, and WesVarPc 5.1). Comparisons among the software are made based on the types of sample design, sampling error estimates, method of variance estimation and cost of software packages. The results of the finding show that WesVarPc can be downloaded for free from Web and offers complete basic of descriptive analyses. Although expensive, SPSS 16.0 Complex Samples and SAS 9.0 Complex Samples have been dominant in the field of data management and data analysis.