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

[Modeling correlated data in epidemiology: mixed or marginal model?].

M Chavance1

  • 1INSERM U472, 16 Av. P. Vaillant-Couturier, 94807 Villejuif.

Revue D'Epidemiologie Et De Sante Publique
|February 16, 2000
PubMed
Summary

Epidemiological studies require modeling the variance-covariance matrix for accurate inferences, especially when analyzing correlated observations. This study explores mixed and marginal models for covariance, highlighting their impact on parameter interpretation in linear and logistic regression.

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

Sensitivity analysis of incomplete longitudinal data departing from the missing at random assumption: Methodology and application in a clinical trial with drop-outs.

Statistical methods in medical research·2013
Same author

[Case-cohort surveys].

Revue d'epidemiologie et de sante publique·2013
Same author

Misspecification of the covariance structure in generalized linear mixed models.

Statistical methods in medical research·2012
Same author

[The case series method].

Revue d'epidemiologie et de sante publique·2010
Same author

Relative-risk ratio was a useful measure of differential association in cohort and case-series studies.

Journal of clinical epidemiology·2007
Same author

Disability evolution in multiple sclerosis: how to deal with missing transition times in the Markov model?

Neuroepidemiology·2007

Area of Science:

  • Epidemiology
  • Biostatistics
  • Statistical Modeling

Context:

  • Correlated observations are prevalent in epidemiological studies, necessitating appropriate statistical methods.
  • Accurate inference on risk factors requires modeling the variance-covariance structure of the data.
  • Understanding and quantifying correlations or random effects is often a primary research objective.

Purpose:

  • To discuss the implications of two covariance modeling choices: mixed models and marginal models.
  • To examine these choices within the context of linear and logistic regression models.
  • To clarify the distinction and interconnection between mixed and marginal approaches, particularly for non-linear models.

Summary:

  • The study contrasts mixed models, which incorporate unobserved random effects to explain data similarity (e.g., subject-specific trajectories in longitudinal data), with marginal models, which separately analyze means and covariance structures.

Related Experiment Videos

  • It emphasizes the critical differences between these approaches for non-linear models, such as logistic regression.
  • The research highlights how the choice of model impacts the interpretation of parameters.

    • Impact:

    • Provides guidance on selecting appropriate covariance modeling strategies in epidemiological research.
    • Enhances understanding of how different modeling approaches affect the interpretation of results, particularly in complex data structures.
    • Contributes to more robust statistical inference in studies involving correlated observational data.