Keynote 3: Two Key Ideas in Missing Data – Missing at Random and Response Propensity
Speaker: Prof. Rod Little, University of Michigan
Moderator: Alfred H. Balch
I discuss the definition of two key concepts in the analysis of data with missing values, missing at random (MAR) and response propensity. Don Rubin defined MAR as a sufficient condition for ignoring the missingness mechanism for likelihood-based and Bayesian inference. Some simple examples of MAR are described, and related concepts, including missing completely at random, always missing at random, always missing completely at random, partially missing at random, informative missingness, informative censoring, and coarsening at random are discussed. I present a formal argument for weakening Rubin’s sufficient conditions for frequentist maximum likelihood inference with precision based on the observed information. The definition of response propensity is often misunderstood, particularly in the survey sampling literature, because of ambiguity about conditioning. A definition that resolves this ambiguity is suggested, by clarifying that dependence on unobserved variables is superfluous for the purpose of nonresponse adjustments.
Rod Little is Richard D. Remington Distinguished University Professor of Biostatistics at the University of Michigan, where he also holds appointments in the Department of Statistics and the Institute for Social Research. He chaired the Biostatistics Department at Michigan for 11 years. He has over 250 publications, notably on methods for the analysis of data with missing values and model-based survey inference, and the application of statistics to diverse scientific areas, He chaired an influential National Research Council study on the treatment of missing data in clinical trials. Little is an elected member of the International Statistical Institute, a Fellow of the American Statistical Association and the American Academy of Arts and Sciences, and a member of the National Academy of Medicine. In 2005, Little was awarded the American Statistical Association’s Wilks Medal for research contributions, and he gave the President’s Invited Address at the Joint Statistical Meetings. He was the COPSS Fisher Lecturer at the 2012 Joint Statistics Meetings.