bayesian two way anova interaction

narcilili

hello jaspers,

I analyzed the andy field’s beer goggles data set which is the effect of amount of alcohol consumed and sex on attractiveness of partner using Bayesian two-way anova. The model comparisons table below shows there is very large evidence for the interaction between alcohol and sex on attractiveness of partner.

Then, I tried to determine the simple effects of sex within each level of alcohol on attractiveness of partner using Bayesian independent t test. The table below shows that there is very large evidence for the alternative hypothesis at four beers. 

I would like to know if there is a need to correct for multiplicity in the follow-up Bayesian independent samples t test. And if there is a need, how to correct for multiplicity.

Thanks and cheers

narcilili

EJ

Hi Narcilili,

The Bayesian ANOVA includes a "Post-hoc Tests" section that reports the t-tests, corrected using the Westfall method (see also https://psyarxiv.com/s56mk). However, I cannot reproduce your numbers -- you have 243 whereas I get 35. Are you using the data set from the JASP data library?

Cheers,

E.J.