Imputation for factor analysis 1 Running head: IMPUTATION FOR FACTOR ANALYSIS A comparison of imputation methods for Bayesian factor analysis models

Imputation methods are popular for the handling of missing data in psychology. The methods generally consist of predicting missing data based on observed data, yielding a complete dataset that is amiable to standard statistical analyses. In the context of Bayesian factor analysis, this paper compares imputation under an unrestricted multivariate normal model (Multiple Imputation) to imputation under the statistical model of interest (Data Augmentation). The former method is popular in applied research, but the latter method is more straightforward from a Bayesian perspective. Simulations demonstrate that Data Augmentation yields less-biased parameter estimates for moderate sample sizes and high missingness proportions. Multiple Imputation, on the other hand, yields less-biased parameter estimates for large sample sizes with misspecified models. The incorporation of auxiliary variables in Data Augmentation is also addressed, and BUGS code is provided.