A Latent Variable Framework for Multiple Imputation with Non-ignorable Missingness: Analyzing Perceptions of Social Justice in Europe

2025-09-25

AI Summary

In large-scale social surveys, nonignorable missingness (MNAR) biases conventional complete-case analysis. To address this, we propose a latent-variable-based multiple imputation framework tailored to analyzing European public perceptions of income distribution fairness. Our method innovatively embeds latent variables directly into the imputation mechanism to explicitly model the nonignorable missingness process—where missingness depends on unobserved latent constructs. We employ joint modeling of the full data distribution combined with frequentist imputation algorithms, enabling flexible handling of mixed variable types and complex dependency structures. We further establish asymptotically valid statistical inference theory for the resulting estimators. Empirical results demonstrate that our approach substantially improves estimation consistency and robustness compared to standard methods. In contrast, ignoring the MNAR mechanism—e.g., via complete-case analysis—leads to upward bias in the estimated association between personal income and fairness perceptions.

Technology Category

• Reasoning under Uncertainty: Causality
• Game Theory and Economic Paradigms: Imperfect Information
• Knowledge Representation and Reasoning: Preferences

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Social Networks and Social Media: Fairness and bias in social network and social media analysis
• Economics, Online Markets and Human Computation: Fairness, privacy, and diversity in economic environments

Abstract

This paper proposes a general multiple imputation approach for analyzing large-scale data with missing values. An imputation model is derived from a joint distribution induced by a latent variable model, which can flexibly capture associations among variables of mixed types. The model also allows for missingness which depends on the latent variables and is thus non-ignorable with respect to the observed data. We develop a frequentist multiple imputation method for this framework and provide asymptotic theory that establishes valid inference for a broad class of analysis models. Simulation studies confirm the method's theoretical properties and robust practical performance. The procedure is applied to a cross-national analysis of individuals' perceptions of justice and fairness of income distributions in their societies, using data from the European Social Survey which has substantial nonresponse. The analysis demonstrates that failing to account for non-ignorable missingness can yield biased conclusions; for instance, complete-case analysis is shown to exaggerate the correlation between personal income and perceived fairness of income distributions in society. Code implementing the proposed methodology is publicly available at https://anonymous.4open.science/r/non-ignorable-missing-data-imputation-E885.

Problem

• Handling non-ignorable missing data in large-scale surveys
• Developing flexible imputation for mixed-type variables using latent models
• Addressing bias from missingness in social justice perception analysis

Innovation

• Latent variable model for mixed-type data imputation
• Handles non-ignorable missingness using latent variables
• Frequentist multiple imputation with asymptotic validity guarantees