Improved reporting of statistical design and analysis: guidelines, education, and editorial policies

Methods Mol Biol. 2010:620:563-98. doi: 10.1007/978-1-60761-580-4_22.

Abstract

A majority of original articles published in biomedical journals include some form of statistical analysis. Unfortunately, many of the articles contain errors in statistical design and/or analysis. These errors are worrisome, as the misuse of statistics jeopardizes the process of scientific discovery and the accumulation of scientific knowledge. To help avoid these errors and improve statistical reporting, four approaches are suggested: (1) development of guidelines for statistical reporting that could be adopted by all journals, (2) improvement in statistics curricula in biomedical research programs with an emphasis on hands-on teaching by biostatisticians, (3) expansion and enhancement of biomedical science curricula in statistics programs, and (4) increased participation of biostatisticians in the peer review process along with the adoption of more rigorous journal editorial policies regarding statistics. In this chapter, we provide an overview of these issues with emphasis to the field of molecular biology and highlight the need for continuing efforts on all fronts.

Publication types

  • Research Support, N.I.H., Extramural
  • Research Support, U.S. Gov't, P.H.S.

MeSH terms

  • Animals
  • Biology
  • Biomarkers, Tumor / metabolism
  • Biomedical Research
  • Editorial Policies*
  • Genome-Wide Association Study
  • Guidelines as Topic*
  • Humans
  • Meta-Analysis as Topic
  • Models, Animal
  • Neoplasms / diagnosis
  • Neoplasms / metabolism
  • Oligonucleotide Array Sequence Analysis
  • Prognosis
  • Proteomics
  • Research Design
  • Statistics as Topic / education*
  • Statistics as Topic / methods*

Substances

  • Biomarkers, Tumor