search

richarddmorey / BayesFactor

BayesFactor R package for Bayesian data analysis with common statistical models.
https://richarddmorey.github.io/BayesFactor/
131 stars 49 forks source link

Inconsistent Output for Top-Down Analysis with Fine-Tuned Prior Scale #119

Open mattstern2 opened 6 years ago

mattstern2 commented 6 years ago

Basic issue: the Bayes Factor seems to depend on the arbitrary labeling of ANOVA factors.

The attached code creates a data set with two factors. It then flips the labeling of the factors so that the first factor is the second factor and the second factor is the first. Finally, it computes BFs for the original data and the "flipped" data using the top-down approach. The BFs do not tie across the two outputs.

One potential issue might be how I specify the rscaleEffects because the problem doesn't seem to exist when I use the defaults. Perhaps I am missing something, but this appears to be a bug.

BayesFactor Issue.txt

richarddmorey commented 6 years ago

I'll check on this; could be an issue with the rscaleEffects code.