Dear researchers,

I am calculating sample sizes for two repeated measures ANOVAs using G*Power and have encountered a discrepancy in the results that I would appreciate your insights on.

For both designs, I used the following parameters in G*Power:

• Alpha = .05
• Power = .80
• Number of groups = 1 (since both are fully within-subject designs)
• Effect size f(U) = .52, based on a previous effect size of η² = .212 (options: as in SPSS).

The resulting sample sizes are:

• N = 21 for the 2 x 3 design, and
• N = 14 for the 3 x 3 design.

I am confused as to how the 2 x 3 ANOVA requires a larger sample size than the 3 x 3 ANOVA, as I would expect the more complex design (3 x 3) to require more participants. Could you explain why this might be the case? Shall I opt for an alternative power calculation program?

Thank you very much in advance!

Regards,

Tatiana