|
Abstract:
|
Multiple imputation is a popular tool for handling item nonresponse in survey sampling. The current multiple imputation techniques with complex survey data are developed with the assumption that the sampling design is non-informative under the specified model. In this talk, we present a new multiple imputation procedure for parametric inference without requiring the sampling design to be non-informative. In the proposed approach, instead of using the sample data likelihood, we use the sampling distribution of the pseudo maximum likelihood estimator to derive the posterior distribution of the parameters. The asymptotic properties of the proposed method are investigated. A simulation study confirms that the new procedure provides unbiased point estimation and valid confidence intervals with correct coverage properties under informative sampling.
|
