Phew—my weekend project is complete! Check out this guide on everything you ever wanted to know about (zero|one|zero-one)-inflated regression! Use #rstats and brms to correctly model proportion data + learn all about the beta distribution along the way. https://www.andrewheiss.com/blog/2021/11/08/beta-regression-guide/

The combination of {brms}, {tidybayes}, and {emmeans} makes it incredibly easy to calculate complex average marginal effects for these model parameters and it's so neat to do. Like, this is all you have to do! #rstats

Pinging @IsabellaGhement, @mjskay (whose incredible {tidybayes} and {ggdist} package make this possible) and @VincentAB (whose new {marginaleffects} package is really useful for beta models too!)

And shoutout to @stevenvmiller whose post here helped me *finally* understand how to interpret logistic regression coefs—my whole stats-knowing life I've just been using odds ratios, but it turns out you can work directly with probabilities instead! http://post8000.svmiller.com/lab-scripts/logistic-regression-lab.html

(apologies tho)