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

Identifying important results from multiple statistical tests.

R A Parker1, R B Rothenberg

  • 1Department of Preventive Medicine, Vanderbilt University School of Medicine, Nashville, TN 37232.

Statistics in Medicine
|October 1, 1988
PubMed
Summary

This study introduces a novel statistical method using mixture distributions to analyze multiple P-values. It offers a less conservative approach for identifying potential research findings when many statistical tests are performed.

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

Uptake of HIV pre-exposure prophylaxis and contraception in South African hair salons: a study protocol for a pilot cluster randomised controlled trial.

BMJ open·2025
Same author

Decentralised HIV care during the COVID-19 pandemic: Ensuring safe community-based services.

South African medical journal = Suid-Afrikaanse tydskrif vir geneeskunde·2022
Same author

NAA is a Marker of Disability in Secondary-Progressive MS: A Proton MR Spectroscopic Imaging Study.

AJNR. American journal of neuroradiology·2020
Same author

Chondromalacia of the cranial medial femoral condyle; its occurrence and association with clinical outcome in a population of adult horses with stifle lameness.

Equine veterinary journal·2019
Same author

The effect of age on fitness among female firefighters.

Occupational medicine (Oxford, England)·2017
Same author

The cascade of HIV care among refugees and nationals in Nakivale Refugee Settlement in Uganda.

HIV medicine·2017

Area of Science:

  • Statistics
  • Biostatistics
  • Data Analysis

Background:

  • Performing numerous statistical tests simultaneously inflates the overall risk of a type I error (false positive).
  • Existing methods to control experimentwise error rates often employ overly stringent significance criteria, potentially masking important findings.
  • Conservative statistical adjustments may be inappropriate for exploratory analyses aiming to identify results for further research.

Purpose of the Study:

  • To present a new statistical approach for analyzing multiple P-values or test statistics.
  • To develop a method that models P-values using mixture distributions, distinguishing between null and non-null results.
  • To provide a less conservative alternative for identifying potentially significant findings in exploratory data analysis.

Main Methods:

Related Experiment Videos

  • The proposed method utilizes a mixture of distributions to model the set of P-values or test statistics.
  • One component of the mixture models results consistent with the null hypothesis.
  • Other components model results inconsistent with the null hypothesis, including those not conventionally significant.

Main Results:

  • The mixture model approach provides a framework for analyzing multiple test results.
  • It allows for the identification of results that may warrant further investigation, even if not strictly statistically significant by traditional standards.
  • The method was illustrated using national mortality data and previously analyzed datasets.

Conclusions:

  • The mixture distribution approach offers a flexible and less conservative alternative for analyzing multiple statistical tests.
  • This method can be valuable in exploratory research settings for identifying potentially meaningful results.
  • The approach facilitates a nuanced interpretation of P-values when numerous tests are conducted.