Inference for Multivariate Regression Model Based on Synthetic Data Generated Using Plug-in Sampling

In this article, the authors derive likelihood-based exact inference for singly and multiply imputed synthetic data in context of a multivariate regression model. The are generated via Plug-in Sampling method, where unknown parameters model set equal to observed values their point estimators based on original data, drawn from estimated version Simulation studies carried out order confirm theoretical results. provide test procedures, which case multiple datasets permissible, compared with asymptotic results Reiter. An application using 2000 U.S. Current Population Survey public use is discussed. Furthermore, properties proposed methodology evaluated scenarios some conditions that were used do not hold, namely nonnormal discrete distributed random variables, cases inferential procedures developed still show very good performances.