Online Program

Dealing With Missing Data; From Small to Large Scale (302414)

 View Presentation

Keywords: Multiple Imputation, Non-Response, Analysis of Incomplete Data

Missing or incomplete responses are a common feature in observational studies, sample surveys, clinical trials & census data. These missing values not only mean less efficient estimates due to the reduction in sample size, but also that standard complete-data methods cannot be used to analyse the data. Furthermore, bias could become an issue in the analysis due to responses often being systematically different to nonresponses. Originally proposed by Rubin (1977b, 1978a), multiple imputation is the technique that replaces each missing value with two or more acceptable values representing a distribution of the possibilities. The paper will discuss the following multiple imputation techniques: • Propensity Score Based Multiple Imputation • Predictive Mean Matching Method • Mahalanobis Distance Matching Method • Predictive Model Based Multiple Imputation • Propensity Score/Predictive Mean Matching/Mahalanobis Distance Combination Method. SOLAS for Missing Data Analysis (developed with guidance from Donald B. Rubin) will be used to demonstrate each technique using examples from the various industries.