how marginal means (also called least square means) and standard deviation are calculated from repeated measure model? I do not have any "groups" in my data, I only have 76 samples each measured in 6 locations (1 location missing in one sample)

I would like to calculate marginal means and std from repeated measure model or directly from raw data. I do not have any group in my data. I have 76 samples and each measured in six different location in the same sample. I would like to see effect of locations and I used repeated measure but I would like to calculate marginal mean and standard deviation of the data from the repeated measure model.

Answer 1

[Jeff Miller]

With this model I believe the marginal means and standard deviations for the different locations are simply the raw means and standard deviations of the scores at each of the locations. The model does not seem to include anything other than location that could be adjusted for (e.g., covariates).

[Mohammed Kamruzzaman]

There is one function to compute marginal mean tbl = margmean(rm,vars) but I dont have variables for which to compute the marginal means. Is there anyway to do this? @Jeff Miller

[Mohammed Kamruzzaman]

Is there any way to use repeated measure model in multcompare function as below? Again I dont have any grouping in the data. T = multcompare(rm,'Group')

[Jeff Miller]

I'm not sure what command you used to fit the repeated measures model in the first place. If it was fitrm, then that command will assign the name 'Time' to you location factor as a default. Then you can use this factor name in multcompare. So, I guess you want something like this:

rm=fitrm(t,'Location1-Location6~1')
multcompare(rm,'Time')

[Mohammed Kamruzzaman]

Thank you. I used the following command for rm model. where Meas = table([1 2 3 4 5 6]','VariableNames',{'Measurements'});

rm = fitrm(data,'L1-L6 ~ 1','WithinDesign',Meas);

[Jeff Miller]

Look at the value of rm.WithinFactorNames. This is the name of the variable that you shouldo pass to multcompare. Or you could just use this:

multcompare(rm,rm.WithinFactorNames)

[Mohammed Kamruzzaman]

Thanks a million. I got exactly the same results when I used the function you mentioned. Because WithinFactorNames in rm is Time. Thanks again.