This question concerns the "power repeated" command in Stata 13, which allows one to determine sample size, or power, for repeated measures designs.
An example provided in the help documentation is as follows:
. power repeated 145 135 130\145 130 120, covmat(225 158 158\158 225 158\158 158 225)
I'm assuming that the six numbers following "repeated" represent the cell means at three time points, in two separate groups. That is,
145 = mean for group 1, time 1
135 = mean for group 1, time 2
130 = mean for group 1, time 3
145 = mean for group 2, time 1
130 = mean for group 2, time 2
120 = mean for group 2, time 3
My questions:
[1] Am I right so far?
[2] What, specifically, do the nine repsective numbers in the covariance-matrix, or "covmat" option, represent?
[3] If I have some advance idea of what the mean and standard deviation in each design cell will be, is there a quick and easy way to pre-estimate the numbers for the "covmat" option (whatever they mean)?
An example provided in the help documentation is as follows:
. power repeated 145 135 130\145 130 120, covmat(225 158 158\158 225 158\158 158 225)
I'm assuming that the six numbers following "repeated" represent the cell means at three time points, in two separate groups. That is,
145 = mean for group 1, time 1
135 = mean for group 1, time 2
130 = mean for group 1, time 3
145 = mean for group 2, time 1
130 = mean for group 2, time 2
120 = mean for group 2, time 3
My questions:
[1] Am I right so far?
[2] What, specifically, do the nine repsective numbers in the covariance-matrix, or "covmat" option, represent?
[3] If I have some advance idea of what the mean and standard deviation in each design cell will be, is there a quick and easy way to pre-estimate the numbers for the "covmat" option (whatever they mean)?