Inference for the Complier-Average Causal Effect From Longitudinal Data Subject to Noncompliance and Missing Data, With Application to a Job Training Assessment for the Unemployed

• Published in: Journal of the American Statistical Association Vol. 96; no. 456; pp. 1232 - 1244
• Main Authors: Yau, Linda H Y, Little, Roderick J
• Format: Journal Article
• Language: English
• Published: Alexandria, VA Taylor & Francis 01.12.2001; American Statistical Association
• Subjects: Algorithms; Applications; Applications and Case Studies; Attrition; Bayesian inferential method; Bayesian method; Causal inference; Causality; Control groups; Covariance matrices; Economic depressions; EM algorithm; Estimation; Estimators; Exact sciences and technology; Gibbs sampler; Inference; Instrumental variable; Insurance, economics, finance; Mathematics; Missing data; Modeling; Parametric inference; Parametric models; Probability and statistics; Questionnaires; Sampling theory, sample surveys; Sciences and techniques of general use; Statistical methods; Statistics; Sufficiency and information; Training; Unemployed; Unemployment; Causal inference; Instrumental variable; Job analysis; Attrition test; Statistical analysis; Econometric model; Statistical estimation; Missing data; Gibbs sampler; Randomized design; Follow up study; Repeated measurement; Maximum likelihood; Bayes methods; EM algorithm
• ISSN: 0162-1459; 1537-274X
• DOI: 10.1198/016214501753381887

Longitudinal studies involving human participants are often complicated by subjects who do not comply with their treatment assignment or do not provide complete data. A treatment effect of interest in the presence of noncompliance is the complier-average causal effect (CACE; Imbens and Rubin 1997a), which is the treatment effect for subjects who would comply regardless of the assigned treatment. Imbens and Rubin (1997a,b) proposed maximum likelihood and Bayesian inferential methods for CACE, which make explicit assumptions for causal inference in the presence of noncompliance and are more efficient than standard instrumental variable methods. A model for inference about the CACE based on this approach is developed which allows for the inclusion of baseline covariates and handles missing data in the repeated outcome measures. Our methods are applied to a randomized trial of a job training intervention for unemployed workers. Results suggest that the intervention trial significantly reduced depression for high-risk compliers up to six months postintervention but not for low-risk compliers.