ABSTRACT
This chapter covers Bayesian perspectives on missing-data as they play out in psychometrics through the lens of Rubin's missing data framework. It reviews foundational concepts pertaining to missing data and introduce a running example. The chapter focuses on conducting inference when the missingness is ignorable. It also focuses on inference when the missingness is nonignorable. The chapter reviews multiple imputation, drawing connections to the fully Bayesian analyses of missing data. Multiple imputation tackles the problems posed by missing data for inference by imputing multiple values for each missing data point and guiding the subsequent analyses in a way that properly accounts for our uncertainty. The chapter discusses connections and departures among missing data, latent variables, and model parameters. Inference regarding model parameters can proceed using just the observed data when the missingness is ignorable. Missing data are prevalent in assessment.
