Distinguishing “Missing at Random” and “Missing Completely at Random”

1996/08/01 by Daniel F. Heitjan, Srabashi Basu · 216 citations
Computer Science · Mathematics · #Advanced Causal Inference Techniques #Artificial intelligence #Bayesian Modeling and Causal Inference #Computer science #Econometrics #Inference #Mathematics #Missing data #Statistical Methods and Bayesian Inference #Statistics

paper · doi:10.1080/00031305.1996.10474381

published in The American Statistician 50(3), 207-213 (Taylor & Francis)

openalex publication_date 1996/08/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/03

Abstract

Missing at random (MAR) and missing completely at random (MCAR) are ignorability conditions—when they hold, they guarantee that certain kinds of inferences may be made without recourse to complicated missing-data modeling. In this article we review the definitions of MAR, MCAR, and their recent generalizations. We apply the definitions in three common incomplete-data examples, demonstrating by simulation the consequences of departures from ignorability. We argue that practitioners who face potentially non-ignorable incomplete data must consider both the mode of inference and the nature of the conditioning when deciding which ignorability condition to invoke.

Citations

• Inference and missing data

Cited by

• An Interdisciplinary and Cross-Task Review on Missing Data Imputation
• Best practices for addressing missing data through multiple imputation
• Imputation by power transformation
• Diagnostic Test for Realized Missingness in Mixed-type Data
• Learning representations for multivariate time series with missing data using Temporal Kernelized Autoencoders
• A Relative Ignorability Framework for Decision-Relevant Observability in Control Theory and Reinforcement Learning
• What Makes Mental Modeling Difficult? Normative Data for the Multidimensional Relational Reasoning Task. [europepmc]
• An Exponential-Cum-Sine-Type Hybrid Imputation Technique for Missing Data. [europepmc]
• Some Classes of Logarithmic-Type Imputation Techniques for Handling Missing Data. [europepmc]
• A novel machine learning-based imputation strategy for missing data in step-stress accelerated degradation test. [europepmc]
• An evaluation of synthetic data augmentation for mitigating covariate bias in health data. [europepmc]

Related

• A fresh look at ignorability for likelihood inference
• Three issues impeding communication of statistical methodology for incomplete data
• Statistical Inference with Different Missing-data Mechanisms
• Missing at Random or Not: A Semiparametric Testing Approach
• Assumptions and analysis planning in studies with missing data in multiple variables: moving beyond the MCAR/MAR/MNAR classification