From Bayes through marginal utility to effect sizes: a guide to understanding the clinical and statistical significance of the results of autism research findings

J Autism Dev Disord. 2011 Feb;41(2):168-74. doi: 10.1007/s10803-010-1035-6.

Abstract

The objectives of this report are: (a) to trace the theoretical roots of the concept clinical significance that derives from Bayesian thinking, Marginal Utility/Diminishing Returns in Economics, and the "just noticeable difference", in Psychophysics. These concepts then translated into: Effect Size (ES), strength of agreement, clinical significance, and related concepts, and made possible the development of Power Analysis; (b) to differentiate clinical significance from statistical significance; and (c) to demonstrate the utility of measures of ES and related concepts for enhancing the meaning of Autism research findings. These objectives are accomplished by applying criteria for estimating clinical significance, and related concepts, to a number of areas of autism research.

MeSH terms

  • Analysis of Variance
  • Autistic Disorder*
  • Bayes Theorem
  • Biomedical Research / standards
  • Biomedical Research / statistics & numerical data*
  • Data Interpretation, Statistical*
  • Humans
  • Reproducibility of Results
  • Sample Size