Bayes Factor 10 = 66.82 while frequentist ANOVA gives p =.159

tea1996

Hi I'm doing an experiment with repeated-measures 3x3 design.

I ran Bayesian Repeated Measures ANOVA in JASP and it gives me a BF10 for one independent variable (TMS regions) of 66 - indicating strong evidence for the alternative hypothesis.

However when I ran a frequentist repeated-measures ANOVA it gives a p-value of 0.159 of the main effect of TMS region on reaction time.

Does anyone knows why they disagree so much?

Thanks

tea1996

https://s31.postimg.org/8pg3gqz2j/ANOVA_2.png

Please see the link above for the Bayesian repeated-measures ANOVA results

https://s31.postimg.org/z7yqpgxsb/ANOVA_1.png

And the link above for frequentist ANOVA

MSB

Looking at your results, it would seem that this is due to a lack of effect for congruency - adding it to you ANOVA seems to hurt the overall explained variance, as can be seen in how the BF for a model containing both main effects is much weaker than a model with only TMS Region.