Bayesian Estimation of Hierarchical Linear Models From Incomplete Data: Cluster‐Level Interaction Effects and Small Sample Sizes
Open Access
- 16 May 2025
- journal article
- Published by Wiley in Statistics in Medicine
- Vol. 44 (10-12) , e70051
Abstract
We consider Bayesian estimation of a hierarchical linear model (HLM) from partially observed data, assumed to be missing at random, and small sample sizes. A vector of continuous covariates includes cluster‐level partially observed covariates with interaction effects. Due to small sample sizes from 37 patient–physician encounters repeatedly measured at four time points, maximum‐likelihood estimation is suboptimal. Existing Gibbs samplers impute missing values of by a Metropolis algorithm using proposal densities that have constant variances while the target posterior distributions have nonconstant variances. Therefore, these samplers may not ensure compatibility with the HLM and, as a result, may not guarantee unbiased estimation of the HLM. We introduce a compatible Gibbs sampler that imputes parameters and missing values directly from the exact posterior distributions. We apply our Gibbs sampler to the longitudinal patient–physician encounter data and compare our estimators with those from existing methods by simulation.Keywords
All Related Versions
Funding Information
- National Institute of Diabetes and Digestive and Kidney Diseases (R01DK112009)
- National Institute of Nursing Research (R01NR020030)
- Institute of Education Sciences (R305D210022)
This publication has 30 references indexed in Scilit:
- A model-based imputation procedure for multilevel regression models with random coefficients, interaction effects, and nonlinear terms.Psychological Methods, 2020
- Multiple imputation with non-additively related variables: Joint-modeling and approximationsStatistical Methods in Medical Research, 2016
- Multilevel multiple imputation: A review and evaluation of joint modeling and chained equations imputation.Psychological Methods, 2016
- Dealing with missing covariates in epidemiologic studies: a comparison between multiple imputation and a full Bayesian approachStatistics in Medicine, 2016
- Evaluating model‐based imputation methods for missing covariates in regression models with interactionsStatistics in Medicine, 2015
- Multiple imputation of covariates by fully conditional specification: Accommodating the substantive modelStatistical Methods in Medical Research, 2014
- Multiple imputation of discrete and continuous data by fully conditional specificationStatistical Methods in Medical Research, 2007
- Just‐Identified Versus Overidentified Two‐Level Hierarchical Linear Models with Missing DataBiometrics, 2007
- Understanding and Using the Implicit Association Test: II. Method Variables and Construct ValidityPersonality and Social Psychology Bulletin, 2005
- Measuring individual differences in implicit cognition: The implicit association test.Journal of Personality and Social Psychology, 1998