WHAT STATISTICAL SIGNIFICANCE MEANS

Sohn (1998) presents a good argument that neither statistical signific ance nor effect size is indicative of the replicability of research re sults. His objection to the Bayesian argument is also succinct. Howeve r, his solution of the 'replicability belief' issue is problematic, an d his verdict that significance tests have no role to play in empirica l research is debatable. The strengths and weaknesses of Sohn's argume nt may be seen by explicating some of his assertions.