Bulletproof Bias? Considering the Type of Data in Common Proportion of Variance Effect Sizes

Erin Michelle Buchanan and John E. Scofield
Additional contact information
Erin Michelle Buchanan: Harrisburg University of Science and Technology

No cs4vy, OSF Preprints from Center for Open Science

Abstract: As effect sizes gain ground as important indicators of practical significance and as a meta-analytic tool, we must critically understand their limitations and biases. This project expands on research by @Okada2013, which highlighted the positive bias of eta squared and suggested the use of omega squared or epsilon for their lack of bias. These variance overlap measures were examined for potential bias in different data scenarios (i.e. truncated and Likert type data) to elucidate differences in bias from previous research. We found that data precision and truncation affected effect size bias, often lowering the bias in eta squared. This work expands our understanding of bias on variance overlap measures and allows researchers to make an informed choice about the type of effect to report given their research study. Implications for sample size planning and power are also discussed.

Date: 2017-08-24
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://osf.io/download/599f3a566c613b0251368d6a/

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:osf:osfxxx:cs4vy

DOI: 10.31219/osf.io/cs4vy

Access Statistics for this paper

More papers in OSF Preprints from Center for Open Science
Bibliographic data for series maintained by OSF ().