The advantage and disadvantage of implicitly stratified sampling

Abstract

Explicit stratified sampling (ESS) and implicit stratified sampling (ISS) are alternative methods for controlling the distribution of a survey sample, thereby potentially improving the precision of survey estimates. With ESS, unbiased estimation of standard errors is possible, whereas with ISS it is not. Instead, usual practice is to invoke an approximation that tends to result in systematic over-estimation. This can be perceived as a disadvantage of ISS. However, this article demonstrates that true standard errors are smaller with ISS and argues that this advantage may be more important than the ability to obtain unbiased estimates of the standard errors.