New blog post: using Stan/brms to get a covariate-adjusted average treatment effect with uncertainty

Hi all! FYI: we (Gertjan Verhoeven & Misja Mikkers) wrote up a straightforward application / tutorial of Poisson regression using brms on a simulated dataset. The dataset is generated mimicking a perfect Randomized Controlled Trial (RCT). The post explains how to use the brms::posterior_predict() function with new data to calculate the posterior distribution of model derived quantities such as an average treatment effect (ATE).

The post is here:

Feedback welcome!


Thanks for sharing it with everyone on the forums!