Handling Missing values through Imputation techniques: A Quantitative Survey

Abstract

This study examines the various techniques used to address missing datain quantitative studies. It acknowledges that missing data poses challengesto data analysis and presents a range of solutions, from simple methodslike deletion and single imputation to advanced approaches involving mod-eling, multiple imputation, and machine learning. The survey takes both abottom-up and top-down approach in selecting an appropriate method forhandling missing data, where the analysis of the data helps narrow down theoptions for the selection process. The study explores the qualitative benefitsand drawbacks of each technique, considering factors such as the analyticalapproach, missing data handling strategy, sample size, and proportion ofmissing data to assess estimation precision. The researchers employed a the-oretical approach using fictitious data to evaluate the significance of missingdata treatment on estimate precision. The statistical analysis conducted inthis study provides practical guidelines for educational researchers, assistingthem in determining the optimal amount of data to collect, the type of anal-ysis to perform, and the appropriate strategy for addressing missing data toensure the reliability of their studies