Combining Probability and Nonprobability Samples by Using Multivariate Mass Imputation Approaches with Application to Biomedical Research

Nonprobability samples have been used frequently in practice including public health study, economics, education, and political polls. Naïve estimates based on nonprobability without any further adjustments may suffer from serious selection bias. Mass imputation has shown to be effective improve the representativeness of samples. It builds an model generates imputed values for all units probability In this paper, we compare two mass approaches latent joint multivariate normal (e.g., Generalized Efficient Regression-Based Imputation with Latent Processes (GERBIL)) fully conditional specification (FCS) procedures integrating multiple outcome variables simultaneously. The Monte Carlo simulation study shows benefits GERBIL FCS predictive mean matching terms balancing bias variance. We evaluate our proposed method by combining information Tribal Behavioral Risk Factor Surveillance System data files.