JMP vs Minitab Sample size for ANOVA/Fit Models

[mchenj1]

Hello,

I do a lot of tests with multiple groups using the fit models for comparison. To date, I use Minitab to calculate a one way anova sample size. This is simple and allows me to input number of groups and our difference to detect.

I would like to use JMP to do this calculation and I know i should use k sample means. But i dont want to put in guess means. i want to say i want a difference to detect of 100. If i enter in dummy means of 0 and 100, the number of 0's vs 100's in the 4 fields drastically changes my sample size. 

Is there a rule of thumb on how to pick means to be as close to minitabs one way anova sample size calculation? If not, can someone please explain why these two platforms do it differently and what that means for the power of my study?

Thanks

[PaigeMiller]

You might want to post this in the JMP forum.