Related Experiment Videos
Why we need to report more than 'Data were Analyzed by t-tests or ANOVA'
Tracey L Weissgerber1,2, Oscar Garcia-Valencia1, Vesna D Garovic1
1Division of Nephrology and Hypertension, Mayo Clinic, Rochester, United States.
Elife
|December 22, 2018
Summary
Reporting of statistical methods like t-tests and ANOVA in physiology studies is often incomplete. Many papers omit crucial details needed to evaluate these common statistical tests, hindering scientific reproducibility.
Area of Science:
- Biomedical Sciences
- Physiology Research
Background:
- Transparent reporting is critical for evaluating scientific studies.
- Reporting of statistical methods in biomedical research is frequently insufficient.
Purpose of the Study:
- To systematically review the reporting quality of t-tests and Analysis of Variance (ANOVA) in physiology journals.
- To identify deficiencies in the reporting of these statistical methods.
Main Methods:
- Systematic review of original research articles published in selected physiology journals in June 2017.
- Analysis of the reporting completeness for t-tests and ANOVA in 328 articles.
- Examination of essential information required for test verification and type determination.
Main Results:
- 84.5% of articles used t-tests or ANOVA.
- 95% of papers using ANOVA lacked information on the specific type of ANOVA employed.
- 26.7% of papers did not specify post-hoc tests for ANOVA.
- Essential details for verifying ANOVA and t-test results were frequently missing.
Conclusions:
- Significant gaps exist in the reporting of statistical methods, particularly for t-tests and ANOVA, in physiology research.
- Incomplete reporting compromises the critical evaluation and reproducibility of study findings.
- Measures are needed to enhance the quality of statistical reporting in scientific publications.