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 comparison of statistical methods for clustered data analysis with Gaussian error

Z Feng1, D McLerran, J Grizzle

  • 1Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA 98104, USA.

Statistics in Medicine
|August 30, 1996
PubMed
Summary

For small numbers of clusters in group randomized trials, bootstrap resampling is preferred for correlated data analysis. Maximum likelihood and four-stage methods are better when distributional assumptions are acceptable.

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

Prediction and parameter inversion of wavelength-tunable fiber laser based on machine learning.

Applied optics·2026
Same author

Evidence for the Collective Nature of Radial Flow in Pb+Pb Collisions with the ATLAS Detector.

Physical review letters·2026
Same author

[Surgical treatment outcomes of different stages of maxillary medication-related osteonecrosis of the jaw].

Beijing da xue xue bao. Yi xue ban = Journal of Peking University. Health sciences·2026
Same author

Evidence for the Dimuon Decay of the Higgs Boson in pp Collisions with the ATLAS Detector.

Physical review letters·2025
Same author

[Advances in minimally invasive surgical techniques for liver transplant recipients].

Zhonghua gan zang bing za zhi = Zhonghua ganzangbing zazhi = Chinese journal of hepatology·2025
Same author

[Prevalence of <i>Schistosoma japonicum</i> infections in wild rodents in key areas during the elimination phase].

Zhongguo xue xi chong bing fang zhi za zhi = Chinese journal of schistosomiasis control·2025

Area of Science:

  • Statistics
  • Biostatistics
  • Epidemiology

Background:

  • Correlated data present challenges in statistical analysis, particularly in group randomized trials.
  • Accurate estimation is crucial for drawing valid conclusions from clustered data.

Purpose of the Study:

  • To compare the performance of four estimation procedures for correlated data under a linear model.
  • To identify the most suitable method for group randomized trials with specific characteristics.

Main Methods:

  • Simulation study evaluating maximum likelihood (ML), generalized estimating equations (GEE), a four-stage method, and a bootstrap resampling method.
  • Focus on scenarios with small numbers of independent clusters, large cluster sizes, and weak within-cluster correlation.
  • Analysis of balanced and near-balanced data structures.

Related Experiment Videos

Main Results:

  • Bootstrap method demonstrates superiority when strong distributional assumptions are to be avoided, especially with <= 10 clusters.
  • Maximum likelihood and four-stage methods show slightly better performance when distributional assumptions are met.
  • All four methods exhibit good performance when the number of independent clusters reaches 50.

Conclusions:

  • The choice of estimation method depends on the number of clusters and the willingness to make distributional assumptions.
  • Bootstrap offers a robust alternative for correlated data analysis in group randomized trials with few clusters.
  • Effective estimation strategies exist for various scenarios in correlated data analysis.