Published correlational effect sizes in social and developmental psychology release_axtqkurnlvdjxnqeghfndrtawq

by Josefína Weinerová, Denes Szucs, John Ioannidis, Apollo-University Of Cambridge Repository

Published by Apollo - University of Cambridge Repository.

2022  

Abstract

The distribution of effect sizes may offer insights about the research done and reported in a scientific field. We have evaluated 12 412 manually collected correlation effect sizes (Sample 1) and 31 157 computer-extracted correlation effect sizes (Sample 2) published in journals focused on social or developmental psychology. Sample 1 consisted of 243 studies from six journals published in 2010 and 2019. Sample 2 consisted of 5012 papers published in 10 journals between 2010 and 2019. The 25th, 50th and 75th effect size percentiles were 0.08, 0.17 and 0.33, and 0.17, 0.31 and 0.52 in Samples 1 and 2, respectively. Sample 2 percentiles were probably larger because Sample 2 only included effect sizes from the text but not from tables. In text authors may have emphasized larger correlations. Large sample sizes were associated with smaller reported correlations. In Sample 1 about 70% of studies specified a directional hypothesis. In 2010 no papers had power calculations, while in 2019 14% of papers had power calculations. These data offer empirical insights into the distribution of reported correlations and may inform the interpretation of effect sizes. They also demonstrate the importance of computation of statistical power and highlight potential reporting bias.
In text/plain format

Archived Files and Locations

application/pdf  1.2 MB
file_kuj6leo7nzfvpk4zcqh3csmqd4
www.repository.cam.ac.uk (publisher)
web.archive.org (webarchive)
Read Archived PDF
Preserved and Accessible
Type  article-journal
Stage   published
Date   2022-12-21
Work Entity
access all versions, variants, and formats of this works (eg, pre-prints)
Catalog Record
Revision: 578d8dd6-6789-4025-81c3-2d95c4e6d5b1
API URL: JSON