Sophia Rabe-Hesketh and Anders Skondal: Standard and new methods for handling missing data

April 2, 2024

Tuesday, April 2, 2024

2:00 - 4:00 PM (PDT) at Berkeley Way West 1212 and via Zoom

Open to GSE faculty, students, and community.

Request a zoom link from


In linear/logistic (and similar) regression with or without latent variables/random effects, the default way of handling missing data is complete-case (CC) analysis, also known as listwise deletion. CC analysis discards all units (e.g., item-person combinations in IRT) with missing values of the response variable or of any of the explanatory variables. This simple-minded method actually makes relatively mild assumptions in comparison to methods often considered to be more valid, such as multiple imputation (MI) or joint modeling of response and explanatory variables by "full information" maximum likelihood (FIML) estimation of all available data. After briefly reviewing CC analysis, MI and FIML and their assumptions, I will show how some of the assumptions of MI/FIML can be relaxed by creating more missing data.

About the authors:

Sophia Rabe-Hesketh is a statistician whose research interests include multilevel/hierarchical modeling, item response theory, longitudinal data analysis, and missing data. She has over 130 peer-reviewed articles in over 70 different journals including Psychometrika, Journal of Econometrics, Biometrics, Journal of the Royal Statistical Society, Series A, with an h-index of 74 in Google Scholar. She has developed a modeling framework for a wide range of multilevel and latent variable models called GLLAMM (Generalized Linear Latent and Mixed Modeling) and written a publicly available software package called gllamm to estimate these models. The theory of these models is published in Generalized Latent Variable Modeling, co-authored with Anders Skrondal. Since 2002, her gllamm software has been used in over 1000 peer-reviewed papers in over 700 different journals by researchers in education, sociology, political science, economics, medicine, and statistics.

Anders Skrondal is a Senior Scientist at the Division of Epidemiology. He started his career as a Research Fellow in the Department of Economics and as a Researcher in the Department of Biostatistics, both at the University of Oslo. His Doctor Philos thesis in Statistics was awarded the 1997 Psychometric Society Dissertation Prize, judged on "the level of originality in the ideas and techniques, the possible applications and their treatment, and potential impact". He was subsequently awarded a personal Postdoctoral Fellowship from The Research Council of Norway which was spent at the University of London and the University of Manchester. After returning to Norway in 2001, Skrondal became a Senior Researcher and subsequently Head of the Biostatistics Group at the Norwegian Institute of Public Health. In 2005, he took up a position as Professor of Statistics in the Department of Statistics at the London School of Economics (LSE) and was subsequently appointed as Director of the LSE Methodology Institute. After spending almost four years at the LSE, Skrondal returned to the Norwegian Institute of Public Health where he is currently a Senior Scientist. He recently served as Adjunct Professor of Biostatistics in the Department of Mathematics, University of Oslo.