Abstract:Summary Generalised linear mixed regression models are fundamental in statistics. Modelling random effects that are shared by individuals allows for correlation among those individuals. There are many methods and statistical packages available for analysing data using these models. Most require some form of numerical or analytic approximation because the likelihood function generally involves intractable integrals over the latents. The Bayesian approach avoids this issue by iteratively sampling the full conditional distributions for various blocks of parameters and latent random effects. Depending on the choice of the prior, some full conditionals are recognisable while others are not. In this paper we develop a novel normal approximation for the random effects full conditional, establish its asymptotic correctness and evaluate how well it performs. We make the case for hierarchical binomial and Poisson regression models with canonical link functions, for hierarchical gamma regression models with log link and for other cases. We also develop what we term a sufficient reduction (SR) approach to the Markov Chain Monte Carlo algorithm that allows for making inferences about all model parameters by replacing the full conditional for the latent variables with a considerably reduced dimensional function of the latents. We expect that this approximation could be quite useful in situations where there are a very large number of latent effects, which may be occurring in an increasingly 'Big Data' world. In the sequel, we compare our methods with INLA, which is a particularly popular method and which has been shown to be excellent in terms of speed and accuracy across a variety of settings. Our methods appear to be comparable to theirs in terms of accuracy, while INLA was faster, for the settings we considered. In addition, we note that our methods and those of others that involve Gibbs sampling trivially handle parameters that are functions of multiple parameters, while INLA approximations do not. Our primary illustration is for a three‐level hierarchical binomial regression model for data on health outcomes for patients who are clustered within physicians who are clustered within particular hospitals or hospital systems.

Statistical Modeling of Within-Laboratory Precision Using a Hierarchical Bayesian Approach

A model and method for analyzing the precision of binary measurement methods based on beta-binomial distributions, and related statistical tests

A simple Bayesian model to estimate proportions and ratios from count data with a hierarchical error structure with an application to droplet digital PCR experiments

A Hierarchical Bayes Unit-Level Small Area Estimation Model for Normal Mixture Populations

Hierarchical Stochastic Model in Bayesian Inference: Theoretical Implications and Efficient Approximation

Bayesian hierarchical modeling of drug stability data

Approximate inferences for Bayesian hierarchical generalised linear regression models

Hierarchical Bayesian Operational Modal Analysis: Theory and Computations

Bayesian hierarchical reconstruction of protein profiles including a digestion model

A Bayesian Measure of Model Accuracy

A Bayesian Approach to Modeling Variance of Intensive Longitudinal Biomarker Data as a Predictor of Health Outcomes

Bayesian Multisource Hierarchical Models with Applications to the Monthly Retail Trade Survey

A Bayesian Model for Forecasting Hierarchically Structured Time Series

A method for Bayesian regression modelling of composition data

A Bayesian Hierarchical Modeling Approach to Dietary Assessment via Food Frequency

Does the reliability of computational models truly improve with hierarchical modeling? Some recommendations and considerations for the assessment of model parameter reliability

A novel multi-stage precision reliability assessment method for mechanical system by Bayesian fusion

A Gentle Introduction to Bayesian Hierarchical Linear Regression Models

Non-Gaussian Normative Modelling With Hierarchical Bayesian Regression

A Bayesian approach for parameter estimation and prediction using a computationally intensive model

Improving Bayesian radiological profiling of waste drums using Dirichlet priors, Gaussian process priors, and hierarchical modeling