As researchers, we’ve all been there – stuck in the midst of a statistical conundrum, wondering if we’re doing it right. I recently found myself in a similar situation while calculating a priori sample size in GPower for an F-test. My study involves a 3 × 3 × 2 mixed design, and I was unsure about the sample size allocation in GPower.
To give you some context, I initially attempted an R simulation, but my supervisors thought it was too complex. So, I opted for a mixed ANOVA in GPower instead, following the advice of our university statistician. I entered all the required values – alpha, effect size, power – and specified 6 groups to reflect the Group × Order cells.
Here’s the question that’s been bugging me: does the total sample size returned by GPower assume equal allocation of participants across the 6 groups? From what I understand, in GPower’s repeated-measures ANOVA modules, you can’t enter unequal cell sizes. This led me to believe that the reported total N should correspond to equal n per group. But, I’m not entirely sure.
After digging through GPower’s documentation, I couldn’t find an explicit source confirming this. That’s why I turned to the Reddit community for help. If you’ve encountered a similar situation or know of a reliable source that clarifies this, please share your insights.
**The Importance of Sample Size Allocation**
Getting the sample size right is crucial in ensuring the validity of our results. In this case, understanding how GPower allocates the sample size can significantly impact our study’s outcome. If we assume equal allocation, but the actual sample size is unequal, it could lead to incorrect conclusions.
**Seeking Clarity**
If you’re familiar with GPower or have experience with similar studies, I’d love to hear your thoughts on this matter. Is there a reliable source that confirms GPower’s allocation method? Have you encountered a similar situation in your research? Share your experiences and let’s unravel this mystery together.
—
*Further reading: GPower Documentation*