A Note on Sample Design

Abstract

As a most scientific approach to sampling, probability sampling has been a common tool for conducting general household survey. Sample design generally refers to the way in which population elements are included in the sample. In this regard, two contrasting approaches to sample design - element and cluster - have been evolved. At the same time, sample design also has to take into account the various population sub-groups through stratification process. When the concept of stratification involves in sampling, four basic forms of sample designs applicable to general household survey can be envisaged. With this, sampling and estimation process tends to be more complex and a single method to sample selection, evaluation of sample design and estimation do not work. In this context, based on the available literatures and author’s own work experience on sampling, this article describes various types of sample design for large-scale general household survey, methods of evaluating their efficiency, and the process of sample selection and estimation.