Effect sizes and confidence intervals
chapterposted on 12.03.2021, 06:12 authored by Geoffrey CummingGeoffrey Cumming, Fiona Fidler
This is an Accepted Manuscript of a book chapter published by Routledge in "The Reviewer's Guide to Quantitative Methods in the Social Sciences", 2nd edition, on 19 November 2018, available online: https://www.routledge.com/The-Reviewers-Guide-to-Quantitative-Methods-in-the-Social-Sciences/Hancock-Stapleton-Mueller/p/book/9781138800137
Chapter abstract: An effect size is simply an amount of something of interest. It can be as simple as a mean, a percentage increase, or a correlation; or it may be a regression weight, a standardized measure of a difference, or the percentage of variance accounted for. Most research questions in the social sciences are best answered by finding estimated effect sizes, meaning point estimates of the true effect sizes in the population. Grissom and Kim (2012) provided a comprehensive discussion of effect sizes, and ways to calculate effect size estimates. Our discussion mainly focuses on experimental designs, but much of the discussion is relevant also to other types of research.