Bayesian on a repeated-measures ANOVA with 5 factors
Hello,
I have to run a bayesian repeated-measures ANOVA with 2 within-subject factors and 3 between-subject factors, so 5 factors in total. While running a ANOVA with 4 factors work works perfectly, running it with 5 factors does not work. The computer (a very powerful one used to process MRI data) runs for several hours and then the page on JAPS becomes entirely blank (see attached file). The JASP version is ont of the latest and we have already reinstalled it. We have also tried on other computers, with different JASP version, same issue.
Is there a maximum in JASP regarding its possibility to compute BF?
Thank you,
Emilie
Comments
Hi Emilie,
There shouldn't be any limit to the number of factors. Would it be possible to share the jasp file so I can take a look at what goes wrong? If yes, you can email it to me confidentially at d.vandenbergh<at>jasp-stats.org.
Cheers,
Don
Hi Emilie,
the issue is caused because Jasp considers all models and by default also all possible interactions among the variables. This simply blows up the model space to the extent where the Jasp output presenter crashes. In your case, it's trying to display a table with 7580 rows and fails to do this. I will open an internal issue for that.
You can avoid this by removing the interactions among the variables. Using the same structure as the example you send me via email, first enter all but one Repeated Measures Cell:
Next, under Model highlight all interactions and remove these.
You can select multiple terms by holding shift or Ctrl, similar to selecting multiple files in windows explorer. Clicking the left arrow highlighted in the screenshot removes all selected effects. Afterward. add the final Repeated Measures Cells and the analysis should start. If there are any particular interaction effects that you are interested in then you should leave those in of course.
Let me know if that solves the problem!
Cheers,
Don