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

Simple robust procedures for combining risk differences in sets of 2 x 2 tables

J D Emerson1, D C Hoaglin, F Mosteller

  • 1Department of Mathematics and Computer Science, Middlebury College, VT 05753, USA.

Statistics in Medicine
|July 30, 1996
PubMed
Summary

Trimmed meta-analysis methods can improve risk difference estimates by reducing the influence of outlier studies. A modified DerSimonian-Laird estimator, particularly a trimmed version, shows robust performance in meta-analyses with heterogeneity.

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

The distribution of sums of rounded percentages.

Demography·2011
Same author

Access to and utilization of neurologists by people with multiple sclerosis.

Neurology·2008
Same author

Access to health care for people with multiple sclerosis.

Multiple sclerosis (Houndmills, Basingstoke, England)·2007
Same author

The Sonya Slifka Longitudinal Multiple Sclerosis Study: methods and sample characteristics.

Multiple sclerosis (Houndmills, Basingstoke, England)·2006
Same author

Overview of the sampling design and statistical methods used in the National Immunization Survey.

American journal of preventive medicine·2001
Same author

Meta-analysis of multiple outcomes by regression with random effects.

Statistics in medicine·1998

Area of Science:

  • Biostatistics
  • Epidemiology
  • Medical Research Methodology

Background:

  • Meta-analyses commonly employ random-effects models to address study result heterogeneity.
  • Anomalous studies can disproportionately influence meta-analytic findings.
  • Risk difference is a key measure in meta-analyses, especially for binary outcomes.

Purpose of the Study:

  • To compare the performance of trimmed versus untrimmed meta-analytic estimators for the risk difference.
  • To evaluate the robustness of trimmed estimators against anomalous study results.
  • To identify the most effective trimmed meta-analytic procedure for handling heterogeneity.

Main Methods:

  • A simulation study was conducted to compare four meta-analytic procedures for risk differences.

Related Experiment Videos

  • Trimmed and untrimmed versions of these procedures were evaluated.
  • A modified DerSimonian-Laird estimator and its trimmed adaptation were specifically assessed.
  • Winsorization was used to adapt variance component estimation for robustness.

    • Main Results:

    • A modified DerSimonian-Laird estimator is effective when random-effects models capture study variability.
    • A 20% trimmed, weighted version of this estimator demonstrated resistance to highly anomalous study results.
    • Among the four trimmed procedures, the modified DerSimonian-Laird trimmed version performed best across various simulations.
    • No single method (trimmed or untrimmed) was universally superior.

    Conclusions:

    • Trimmed meta-analytic procedures, particularly a modified DerSimonian-Laird estimator, offer enhanced resistance to outlier studies in meta-analyses.
    • The choice of meta-analytic method depends on the specific characteristics of the data and the presence of heterogeneity.
    • Further research may be needed to establish optimal trimming strategies for different scenarios.