Statistics in Medical Educational Studies
Author Information
Author(s): Desbiens Norman A
Primary Institution: University of Tennessee College of Medicine
Hypothesis
How often are statistical tests reported in survey studies that query an entire population in medical education literature?
Conclusion
Educational enumeration surveys frequently report statistical tests, but researchers should avoid performing them when studying an entire population.
Supporting Evidence
- 45% of the articles reported or stated that they calculated statistics.
- 73% of articles in the Journal of General Internal Medicine reported statistics.
- 86% of the articles that reported statistics also reported p-values.
Takeaway
This study looked at how often medical education surveys use statistics. It found that many studies report statistics when they shouldn't, especially when asking everyone in a group.
Methodology
The study identified survey studies published in two journals and tallied the use of inferential statistics and p-values.
Potential Biases
Potential bias due to the non-random sampling of subjects from specific institutions.
Limitations
The study only included articles from two journals and may not represent the entire medical literature.
Participant Demographics
The studies surveyed responders at all U.S. or Canadian medical schools, program directors, or clerkship directors.
Statistical Information
P-Value
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website