2020/11/10 by Olanrewaju Akande, Jerome P. Reiter, Akande, Olanrewaju +1
Mathematics · Social Sciences · #Advanced Causal Inference Techniques #FOS: Computer and information sciences #Methodology (stat.ME) #Statistical Methods and Bayesian Inference #Survey Methodology and Nonresponse
paper · pdf · doi:10.48550/arxiv.2011.05482
openalex publication_date 2020/11/10 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
We outline a framework for multiple imputation of nonignorable item\nnonresponse when the marginal distributions of some of the variables with\nmissing values are known. In particular, our framework ensures that (i) the\ncompleted datasets result in design-based estimates of totals that are\nplausible, given the margins, and (ii) the completed datasets maintain\nassociations across variables as posited in the imputation models. To do so, we\npropose an additive nonignorable model for nonresponse, coupled with a\nrejection sampling step. The rejection sampling step favors completed datasets\nthat result in design-based estimates that are plausible given the known\nmargins. We illustrate the framework using simulations with stratified\nsampling.\n