Handling Missing Data

This session will cover a basic background for missing data (e.g., reasons for missing data, missing data mechanisms (e.g., MCAF, MAR, NMAR), older conventional approaches and their limitations); introduce two most widely accepted approaches (e.g., full-information maximum likelihood (FIML) and multiple imputation (MI)) with more information on MI procedures, options, decisions and limitations; and provide a motivating example of MI in Stata. This session would be more appropriate for those who may have no or little experience using FIML or MI or who want to refresh their practical knowledge in handling missing data.