Stratified and Cluster Sampling
The random sampling paradigm, typically introduced in basic statistics courses, ensures that a sample of data is, loosely speaking, ‘representative’ of the underlying population. When the population parameters are identified, many common estimation techniques, including least squares, maximum likelihood, and instrumental variables, have desirable statistical properties under random sampling. Unfortunately, while random sampling is convenient, it can be, and often intentionally is, violated when cross-sectional data and panel data are collected. Two important deviations from random sampling are stratified sampling and cluster sampling, or perhaps a combination.
KeywordsCluster correlation Cluster sampling Exogenous sampling Heteroskedasticity Multinomial sampling Probability sampling Sampling Stratified sampling Survey sampling Two-stage sampling Unbiased estimators Variable probability sampling Variance Weighted least squares
- Cosslett, S.R. 1993. Estimation from endogenously stratified samples. In Handbook of statistics, ed. G.S. Maddala, C.R. Rao, and H.D. Vinod, Vol. 11. Amsterdam: Elsevier.Google Scholar
- Wooldridge, J.M. 2002. Econometric analysis of cross section and panel data. Cambridge, MA: MIT Press.Google Scholar