Friedman
Defining my study.
After checking normality I obtained a value less than 0.001 in the Kolgomorov test and I opted for Friedman.
I have a total of 72 measurements, from a single subject, using 3 different formulas to calculate power in watts, and 6 different postures were carried out, in two different conditions (in and out) and each was repeated 6 times. 6x2x6=72.
I understand that these are repeated measurements. I select repeated measures anova, in repeated measures factors I define "POWER" as a factor, and in levels I put A, B and C. Just below in repeated measures cells I put each variable corresponding to each method. In Factors between elements I select Condition (in and out) and Settings (the 6 positions). After this, in the non-parametric tab in MR Factor I have POWER, and I select post-hoc conover contrasts. However, with this I can only see the differences between the 3 power calculation models. I can't see anything about the differences between condition (in and out) or between settings (6 positions).
Is there something I could be doing wrong?
Comments
Hi @gvt ,
The friedman test is only appropriate for univariate designs, not for designs that have more than 1 predictor variables, as seems to be the case for you. I understand that makes the Friedman test fairly limited, but for now that is beyond our control.
Kind regards
johnny
Ok.
I decide to do Conover post hoc but the results are different from JAMOVI.
Which is the reason?
Could you share the JASP file you used? You can send it to j.b.vandoorn<at>uva.nl
Without it I do not have a way to analyze your situation.
In the meantime I doublechecked the results of JASP with the PMCMRplus package in R, and get the same results for the Conover tests - they do differ from jamovi's, since it seems they use an outdated package (PMCPR instead of PMCMRplus).
Kind regards
Johnny
Ok I understand!