Abstract
The multilevel modeling-scaled weights (MM-SW) technique and the standard regression-robust (bootstrap) variance estimation (SR-RV) technique take into account the complexities such as stratification, clustering and unequal probability of the selection of individuals of multistage complex survey data but the ways are different. The performance of these two statistical techniques were examined based on the analysis of real life cross-sectional complex survey data and the Monte Carlo simulation study using cross-sectional complex survey data. The results obtained from the Monte Carlo simulation study indicated that the performance of the MM-SW technique and the SR-RV technique were comparable to analyze the cross-sectional complex survey data although results obtained from the analysis Canadian Heart Health Survey (CHHS) data were inconsistent. Our conclusion based on this study was that both statistical techniques offered similar results when used to analyze the cross-sectional complex survey data; however SR-RV technique might be preferred because it fully accounts for stratification, clustering and unequal probability of selection.