서베이앤헬스폴리시리서치센터

Kim, S. W., Heeringa, S. G., and Solenberger, P. W. (2003). “A probability sampling approach for variance minimization,” in Proceedings of the Survey Research Methods Section, American Statistical Association, 2168-2173, San Francisco, California: Joint Statistical Meetings.

Abstract

A number of techniques for probability sampling without replacement (SWOR) have been introduced, although it is not clear which method is superior in terms of statistical efficiency.  Jessen (1969) suggested one such sampling technique. The Jessen method 4 shows high efficiency in comparisons of variances of estimates relative to those of alternative SWOR selection schemes. However, Jessen’s method may be difficult to employ in practical problems due to the arbitrariness and complexities of trials to determine the joint inclusion probabilities that are required for the variance estimation formula suggested by Yates and Grundy (1953). In this paper, we suggest a non-linear programming (NLP) approach to overcome some computational disadvantages of Jessen’s method 4. We illustrate the practicality and statistical efficiency of our method through the application of the method to several examples from the literature.

Keywords: statistical efficiency, inclusion probability, non-linear programming