2023
DOI: 10.31234/osf.io/rv3kw
|View full text |Cite
Preprint
|
Sign up to set email alerts
|

Power to Detect What? Considerations for Planning and Evaluating Sample Size

Abstract: In the wake of the replication crisis, social and personality psychologists have increased attention to power analysis and the adequacy of sample sizes. In this paper, we analyze current controversies in this area, including choosing effect size, why and whether power analyses should be conducted on already-collected data, how to mitigate negative effects of sample size criteria on specific kinds of research, and which power criterion to use. Because our conventions about effect sizes are less clear than some … Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1

Citation Types

0
24
0

Year Published

2023
2023
2024
2024

Publication Types

Select...
8
1

Relationship

0
9

Authors

Journals

citations
Cited by 24 publications
(24 citation statements)
references
References 141 publications
0
24
0
Order By: Relevance
“…Past research on masculinity threats has observed medium to large effects (Bosson et al, 2009; Netchaeva et al, 2015; Vandello et al, 2008; Weaver et al, 2013). Thus, to remain conservative in our sample size determination, our power analysis was based on an effect of d = 0.50, and we opted for a power of 95% (e.g., Giner-Sorolla et al, 2020). A power analysis, conducted with G*Power (Faul et al, 2007), for a design examining the difference between two independent means with a t test (two-tailed) suggested a sample size of 210.…”
Section: Methodsmentioning
confidence: 99%
“…Past research on masculinity threats has observed medium to large effects (Bosson et al, 2009; Netchaeva et al, 2015; Vandello et al, 2008; Weaver et al, 2013). Thus, to remain conservative in our sample size determination, our power analysis was based on an effect of d = 0.50, and we opted for a power of 95% (e.g., Giner-Sorolla et al, 2020). A power analysis, conducted with G*Power (Faul et al, 2007), for a design examining the difference between two independent means with a t test (two-tailed) suggested a sample size of 210.…”
Section: Methodsmentioning
confidence: 99%
“…The effect‐size sensitivity analysis (as suggested by Giner‐Sorolla et al, 2019) using GPower (version 3.1.9.7, Faul et al, 2007) showed that the final analysis had 80% power to detect an interaction effect of η 2 = 0.028 between the experimental trustworthiness condition factor and the CM covariate when treating the conspiracy mentality covariate conservatively as a two level between‐subjects factor.…”
Section: Methodsmentioning
confidence: 99%
“…All findings should be taken as preliminary and offering inspiration for further, focused studies. Statistical power is a serious limitation for studies in this context due to high variances and low likely effect sizes [45]. As a power sensitivity analysis, based on the variances in total points we observed, for example, a study of this size (~40 pairs in each condition) would be powered to detect a difference by condition of approximately 1230 points (this is the minimum detectable effect).…”
Section: Limitations and Future Directionsmentioning
confidence: 97%