Hello Kaleb,

I'm not clear on how you arrived at a recommendation of 47 cases using g*power. Partial eta-squared of .256 converts to an approximate Cohen's f of 0.587. With that as the target ES, and your settings, the resultant N varies from 16 (using inter-measure correlation of 0.1) to 4 (using r = 0.9). Why so few? Because your target ES is pretty darn large.

As to why g*power and morepower differ, be sure that you are in fact trying to elicit an apples to apples comparison. The original (2012) article in which the authors of morepower outline their methods would be a good place to start, vs. the documentation for g*power (2007) (links are given below).

Good luck with your work.

Campbell, J. I. D., & Thompson, V. A. (2012). MorePower 6.0 for ANOVA with relational confidence intervals and Bayesian analysis. Behavior Research Methods, 44(4), 1255-1265. doi:10.3758/s13428-012-0186-0

https://www.psychologie.hhu.de/fileadmin/redaktion/Fakultaeten/Mathematisch-Naturwissenschaftliche_Fakultaet/Psychologie/AAP/gpower/GPower3-BRM-Paper.pdf

Hi David,

Thanks for your response.

I calculated 47 participants by selecting the “as in Cohen (1988)” effect size specification under the options tab. G*Power lists that as the recommended option. I believe you used the default “as in GPower 3.0” option.

It appears that the “as in SPSS” option is consistent with the MorePower sample size estimate. I found this paper (Lakens, 2013) describing the differences between these approaches in G*Power.

Lakens, D. (2013). Calculating and reporting effect sizes to facilitate cumulative science: a practical primer for t-tests and ANOVAs. Frontiers in psychology, 4, 863. doi:10.3389/fpsyg.2013.00863

Hello again Kaleb,

Again, apples to apples! Glad you sorted out the apparent discrepancy.