Working with missing data in large-scale assessments
Francis L. Huang, Brian T Keller
Abstract
Abstract Missing data are common with large scale assessments (LSAs). A typical approach to handling missing data with LSAs is the use of listwise deletion, despite decades of research showing that approach can be a suboptimal strategy resulting in biased estimates. In order to help researchers account for missing data, we provide a tutorial using R and the freely available Blimp program to impute and analyze multiply imputed datasets.
Topics & Concepts
Scale (ratio)Missing dataData scienceComputer scienceGeographyCartographyMachine learningStatistical Methods and Bayesian InferencePsychometric Methodologies and TestingStatistical Methods and Inference