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

Low power, type II errors, and other statistical problems in recent cardiovascular research

J L Williams1, C A Hathaway, K L Kloster

  • 1Department of Physiology and Pharmacology, School of Medicine, University of South Dakota, Vermillion 57069, USA.

The American Journal of Physiology
|July 1, 1997
PubMed
Summary

Many biomedical studies incorrectly conclude no difference due to low statistical power or large variability, leading to Type II errors (false negatives). This research highlights common statistical errors in physiology journals, emphasizing the need for improved statistical practices.

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 role of area deprivation index in health care disruptions among cancer survivors during the SARS-CoV-2 pandemic.

Public health·2024
Same author

Implications of miRNA expression pattern in bovine oocytes and follicular fluids for developmental competence.

Theriogenology·2020
Same author

Understanding the effects of the bovine POLLED variants.

Animal genetics·2020
Same author

The genetic background of clinical mastitis in Holstein-Friesian cattle.

Animal : an international journal of animal bioscience·2019
Same author

Landscape of copy number variations in Bos taurus: individual - and inter-breed variability.

BMC genomics·2018
Same author

Evaluation of responses to vaccination of Angus cattle for four viruses that contribute to bovine respiratory disease complex.

Journal of animal science·2018

Area of Science:

  • Biostatistics
  • Cardiovascular Physiology Research

Background:

  • Biomedical research often relies on P values < 0.05 to determine statistical significance.
  • A non-significant P value may be misleading due to insufficient sample size or high data variability, potentially indicating a Type II error (false negative).

Purpose of the Study:

  • To evaluate the statistical power and Type II error rates of unpaired t-tests in the American Journal of Physiology: Heart and Circulatory Physiology.
  • To identify other statistical errors in published biomedical research.

Main Methods:

  • Analysis of unpaired t-tests from Volumes 246 and 266 of the American Journal of Physiology: Heart and Circulatory Physiology.
  • Calculation of statistical power to detect specified effect sizes (20% and 50% change).
  • Examination of all articles for statistical reporting accuracy and adherence to assumptions.

Related Experiment Videos

Main Results:

  • Median power was insufficient to detect smaller effects (approx. 0.55 for 20% change).
  • Approximately 80% of studies with nonsignificant t-tests had a high probability (> 0.30) of Type II errors.
  • Frequent issues included vague statistical reporting, misuse of t-tests, and lack of examination of statistical assumptions.

Conclusions:

  • Low statistical power and a high incidence of Type II errors are prevalent in the analyzed journal volumes.
  • Statistical reporting and application in biomedical literature require significant improvement to ensure accurate interpretation of results.