Analytic Perspective (Errors in statistical tests)
2008

Critique of Statistical Errors in Biomedical Research

Sample size: 371 publication Evidence: high

Author Information

Author(s): Carl V. Phillips, Richard F. MacLehose, Jay S. Kaufman

Primary Institution: University of Alberta School of Public Health

Hypothesis

Are the terminal digits of p-values in biomedical research uniformly distributed?

Conclusion

The study found that the terminal digits of p-values diverge from a uniform distribution, suggesting potential biases in reporting.

Supporting Evidence

  • The analysis included 190 p-values from Nature articles, revealing significant deviations from uniformity.
  • Combined data from previous critiques confirmed a strong digit preference in reported p-values.
  • Statistical significance was found to be misleading in interpreting the results of the studies analyzed.

Takeaway

This study looked at the last digits of p-values in research papers and found they don't follow the expected pattern, which might mean researchers are not reporting their results accurately.

Methodology

The authors analyzed p-values from articles in the journal Nature using chi-square tests to assess the distribution of terminal digits.

Potential Biases

Publication bias may lead to an overrepresentation of statistically significant results, affecting the distribution of reported p-values.

Limitations

The study may not account for all factors influencing digit distribution, such as varying reporting practices across different journals.

Statistical Information

P-Value

p<0.001

Statistical Significance

p<0.001

Digital Object Identifier (DOI)

10.1186/1742-7622-5-9

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication