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

Sample size calculations for paired or matched ordinal data

S A Julious1, M J Campbell

  • 1University of Southampton, Southampton General Hospital, U.K.

Statistics in Medicine
|August 12, 1998
PubMed
Summary
This summary is machine-generated.

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

Structural intracranial abnormalities and brain volumes in foetuses with congenital heart disease.

Clinical radiology·2026
Same author

Analysis of apheresis outcomes in a cohort of Chilean patients treated with autologous stem cell transplantation: A single center real-world experience.

Transfusion and apheresis science : official journal of the World Apheresis Association : official journal of the European Society for Haemapheresis·2024
Same author

High-dimensional immune cell profiling of cerebrospinal fluid from patients with metastatic breast cancer and leptomeningeal disease.

NPJ breast cancer·2023
Same author

Do studies published in two leading reproduction journals between 2011 and 2020 demonstrate that they followed WHO5 recommendations for basic semen analysis?

Human reproduction (Oxford, England)·2022
Same author

Cardiac arrhythmias and electrophysiologic responses during spontaneous hyperglycaemia in adults with type 1 diabetes mellitus.

Diabetes & metabolism·2021
Same author

Myeloperoxidase inhibitor AZD5904 enhances human sperm function in vitro.

Human reproduction (Oxford, England)·2021

Calculating sample size for ordinal outcomes in matched studies is complex. This study proposes an approximation for discordant pairs, improving accuracy with more categories and offering a rule of thumb for sample size estimation.

Area of Science:

  • Biostatistics
  • Clinical Trials
  • Statistical Methods

Background:

  • Determining adequate sample size is crucial for the statistical power of paired or matched studies.
  • Ordinal outcome variables present unique challenges in sample size calculations compared to continuous or binary outcomes.
  • Existing methods often simplify ordinal data into binary categories, potentially affecting accuracy.

Purpose of the Study:

  • To address the challenge of calculating sample sizes for paired/matched studies with ordinal outcome variables.
  • To propose and evaluate an approximate method for estimating the number of discordant pairs for ordinal data.
  • To compare the proposed method with simulation-based estimates and assess its performance across different numbers of categories.

Main Methods:

Related Experiment Videos

  • Development of an approximate formula for the number of discordant pairs in ordinal matched studies.
  • Comparison of sample sizes estimated by the approximation with those obtained through simulation.
  • Analysis of sample size requirements as the number of categories in the ordinal variable increases.

    • Main Results:

    • The proposed approximation underestimates sample sizes for two-category variables but improves with more categories.
    • Collapsing an ordinal variable into two categories may not significantly reduce statistical power but can increase the total sample size needed.
    • The expected number of discordant pairs generally decreases as the number of categories increases.

    Conclusions:

    • The approximation for discordant pairs provides a viable method for sample size calculation in ordinal matched studies.
    • A rule of thumb is suggested: use the two-category discordant pair sample size as an approximation for studies with more than two categories.
    • Researchers should consider the trade-offs between category reduction and total sample size requirements in study design.