• Conference Proceeding

A paradox of multiple imputation

Citation

Kott, P. S. (1995). A paradox of multiple imputation. In Proceedings of the Survey Research Methods Section, American Statistical Association, pp. 380–383. .

Abstract

Repeated imputation as championed by Rubin (1987) provides a method of adjusting for survey nonresponse. When done correctly, inferences drawn from repeatedly imputed data sets are statistically valid under the right conditions. Some recent attempts to question the validity of repeated-imputation inferences (e.g. Fay 1992) have not fully understood what those conditions are.