I really appreciate the opportunity to ask this question here and hope that it's comprehensible.

For an upcoming evaluation with 2 experimental groups and 1 control group, we would like to estimate the necessary sample size using G*Power. In addition, we would like to consider the baseline symptom severity (in this case the baseline depression score), which will be measured in three categories: low, moderate and high. We plan to use stratified randomization (depression severity as 1 stratum). This way we hope to create a balance for this clinical factor, because our trial might not have valid results if these three groups are not balanced.

So my question now is:

Do I have to consider this stratum with 3 categories in our testpower analysis, thus calculating not with 3 groups (EG1, EG2 and CG) but with 9? Originally I used the repeated measure "ANOVA within-between interaction" setting in G*Power.

I hope I made myself clear, help is much appreciated and thank you in advance.

Best regards,

Angelika