I've been trying to use this function to replace the ticks with letter strings (e.g., 'frontal', 'temporal', etc.) that are stored in cell array. I get an error message saying that the function does not accept cell as input. How can I solve the problem?
I found the solution, sorry for my question, if somenone has the same issue, a slight change in the function teg_report.m fixed the problem.
At the line 115 of teg_report.m, change .05 by 1. :
"[...] if p < .05 [...]"
"[...] if p < 1. [...]"
It should make the job !
Still thanks your work Thomas !
Oh, sorry: the missing post said that the code does correct for sphericity, using this calculation for epsilon, from line 103 in teg_RMF_ANOVA:
vars = var(transformed0);
eps_num = sum(vars) .^ 2;
eps_den = length(vars) * sum(vars .^ 2);
eps0 = eps_num / eps_den;
The matrix transformed0 contains the difference-score columns. eps0 is 1 if their variances are identical, and has a lower bound of 1 / (nVars - 1) as the variances differ.
The DFs are corrected a la Greenhouse-Geisser (i.e., just multiplied by epsilon), but the adjusted DFs aren't printed.
The problem is that the tutorials I found (and on a quick look also the one you linked) either don't define how they calculate epsilon or only define it for main effects. I took this from an ancient '70s article, if I recall correctly.