Statistical power of tests and publication date: A spurious correlation?

Affiliation auteursAffiliation ok
TitreStatistical power of tests and publication date: A spurious correlation?
Type de publicationJournal Article
Year of Publication2018
AuteursBonache ABernard
JournalCOMPTABILITE CONTROLE AUDIT
Volume24
Pagination13-41
Date PublishedAPR
Type of ArticleArticle
ISSN1262-2788
Mots-clésPower, PUBLICATION BIAS, SAMPLE SIZE, SPURIOUS RELATION
Résumé

This article synthesizes a study of the statistical power of tests of the studies published in Comptabilite - Controle - Audit during the past twenty years and its determinants. To this end, we performed a power analysis, and bivariate and multivariate analyses. Similar to previous studies, the power and bivariate analyses demonstrated a statistically significant increase in the statistical power of tests of the articles studied. However, the multivariate analysis uncovered the spuriousness of the correlation between the publication date and the power of tests of these articles. This spurious relation could stem from a publication bias due to the use of statistical significance as a criterion for publication.

DOI10.3917/cca.241.0013