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

On logit confidence intervals for the odds ratio with small samples.

A Agresti1

  • 1Department of Statistics, University of Florida, Gainesville 32611-8545, USA. aa@stat.ufl.edu

Biometrics
|April 25, 2001
PubMed
Summary

Simple confidence intervals for odds ratios, like the Woolf and Gart intervals, perform well even in small samples. Adjusting the Gart interval by smoothing toward the independence model improves coverage probabilities.

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

Author Correction: Titanium-carbide MXenes for work function and interface engineering in perovskite solar cells.

Nature materials·2019
Same author

Titanium-carbide MXenes for work function and interface engineering in perovskite solar cells.

Nature materials·2019
Same author

Graphene-Induced Improvements of Perovskite Solar Cell Stability: Effects on Hot-Carriers.

Nano letters·2019
Same author

Nucleosome loss facilitates the chemotactic response of macrophages.

Journal of internal medicine·2014
Same author

Hepatitis B virus blood screening: impact of nucleic amplification technology testing implementation on identifying hepatitis B surface antigen non-reactive window period and chronic infections.

Vox sanguinis·2010
Same author

Tunneling as a stochastic process: a path-integral model for microwave experiments.

Physical review. E, Statistical, nonlinear, and soft matter physics·2005

Area of Science:

  • Statistics
  • Biostatistics
  • Epidemiology

Background:

  • Confidence intervals for odds ratios are crucial in statistical analysis.
  • Traditional methods like the Woolf logit interval have limitations in small samples.
  • The Gart interval offers an improvement by smoothing observed counts.

Purpose of the Study:

  • To evaluate the performance of simple large-sample confidence intervals for odds ratios in small samples.
  • To compare the effectiveness of the Woolf and Gart intervals.
  • To explore methods for improving interval coverage probabilities.

Main Methods:

  • Utilizing the delta method for calculating confidence intervals.
  • Implementing the Woolf logit interval.
  • Applying the Gart interval with adjustments for smoothing and zero cell counts.

Related Experiment Videos

Main Results:

  • Simple large-sample confidence intervals, including Woolf and Gart, demonstrate good performance even with small sample sizes.
  • The Gart interval's performance is enhanced by smoothing toward the independence model.
  • Adjusting the interval direction for zero cell counts further improves coverage probabilities.

Conclusions:

  • Standard confidence intervals for odds ratios are reliable for small samples unless the true association is weak.
  • The Gart interval, particularly when smoothed towards the independence model and adjusted for zero counts, offers superior coverage probability.