@cluhmann I am also confused about this application. I think what @dv_mepi is saying is: shouldn’t we compare the posterior mean of the difference of group2 posteriors distribution and group1s posterior distribution. I don’t see how the current deterministic “diff of means” is connected to the data. I do recognize that the priors “pre_mean” and “post_mean” are informed by “pre_mu”, “pre_sd”, “post_mu”, and “post_sd”, but it is not clear that these estimates are informed by the observations in the data. Does the code in the with block not run sequentially top to bottom?
1 Like