Significance Testing in Psychological Research: Some Persisting Issues

Publisher Website
Google Scholar
Abstract
Empirical surveys show that reports of significance tests appear in the vast majority of articles in psychological research journals and are relied on by both investigators and journal reviewers when making decisions about replication of experiments and submission and acceptance of research reports Evaluation of claims by critics that significance testing is, nonetheless, counterproductive yields suggestions for improvements of prevailing practices, but also the recommendation that decisions about significance testing versus alternative procedures should remain the province of the individual investigator, unconstrained by mandates or prohibitions by publishers or officials of scientific associations The prime focus for efforts toward improvement of research design and data analysis in psychology and other behavioral sciences should be upgrading the mathematical and general scientific preparation of students preparing for research careers in these fields