Multiple Imputation: An Investigation of the Missing Data Techniques Effectiveness

Abstract

Multiple Imputation (MI) is one of the most reliable techniques in addressing missing data due to partial or incomplete responses from a portion of the sample. MI has been particularly useful when handling missing data patterns such as Missing Completely at Random (MCAR) and Missing at Random (MAR). However, there have been some debates on its use when it comes to the Missing Not at Random (MNAR) pattern due to the bias it creates. This paper further examines the complexities of using MI to accurately complete missing data sets, exploring both its effectiveness and limitations.