Bayesian Repeated Measures ANOVA: same data, different output
Hello everyone!
I'm writing here because I analyzed soma data and I encountered a problem.
I did "regular" a Repeated Measures ANOVA using both JASP and R, and I obtained the same output.
Then, I tried to do a Bayesian Repeated Measures ANOVA, using the same data, and everything seemed smooth. But, to be sure, I made again the same analysis (same data, same procedure) and I obtained slightly different outputs. Therefore, I did it again many times: the outputs are very similar, but every time they differ a little bit.
How can it be possible? I have no idea.
Thank you very much!
Comments
Hi Gaffiere,
This happens because RM ANOVA is based on a numerical procedure. You can up the accuracy by increasing the samples under "Additional Options" -> "Numerical Accuracy". If you find it annoying to get slightly different results every time you can also set a seed (under "Repeatability"). See also https://psyarxiv.com/spreb where the meaning of the error% is discussed.
Cheers,
E.J.