Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

A Bayesian hierarchical model for categorical data with nonignorable nonresponse.

Paul E Green1, Taesung Park

  • 1Department of Epidemiology, University of Michigan, Ann Arbor, Michigan 48105, USA. pgreen@umich.edu

Biometrics
|February 19, 2004
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Parametric hypothesis testing for pathway based hierarchical structural component models.

Genes & genomics·2026
Same author

Analysis of severity in COVID-19 patients by using longitudinal immune profiles.

iScience·2026
Same author

Enhancing polygenic risk prediction by modeling quantile-specific genetic effects.

Scientific reports·2026
Same author

Optoelectronic Synaptic Transistors Based on Colloidal CdSe Nanowires for Energy-Efficient Neuromorphic Computing.

ACS applied materials & interfaces·2026
Same author

Impact of ACEI/ARB use on COVID-19 mortality in patients with ischaemic heart disease: insights from South Korean National health insurance service data.

BMC infectious diseases·2025
Same author

DeepIMB: Imputation of non-biological zero counts in microbiome data.

Genes & genomics·2025

This study introduces a Bayesian hierarchical model to improve estimates for nonrespondents in contingency tables, addressing issues with existing log-linear models. The new approach uses stochastic search variable selection for more robust results in analyzing nonignorable nonresponse data.

Area of Science:

  • Statistics
  • Biostatistics
  • Data Analysis

Background:

  • Log-linear models smooth contingency tables with nonignorable nonresponse.
  • Existing maximum likelihood estimates can be unstable near boundaries.
  • Previous empirical Bayes models smoothed estimates but had limitations.

Purpose of the Study:

  • Develop a Bayesian hierarchical model for nonignorable nonresponse.
  • Incorporate log-linear models into prior specifications.
  • Enhance estimation stability and accuracy for nonrespondent data.

Main Methods:

  • Utilized a Bayesian hierarchical model with log-linear priors.
  • Integrated Stochastic Search Variable Selection (SSVS) for model uncertainty.
  • Employed Markov Chain Monte Carlo (MCMC) for estimation.

Related Experiment Videos

Main Results:

  • The proposed model provides smoothed estimates for nonrespondent cell frequencies.
  • SSVS integrates multiple log-linear models for averaged estimates.
  • Demonstrated with a renal transplant patient creatinine level dataset.

Conclusions:

  • The Bayesian hierarchical model with SSVS offers a robust approach to nonignorable nonresponse.
  • This method improves upon existing techniques by averaging across models.
  • The approach is effective for complex categorical data analysis.