📚 Finished reading Statistical Rethinking by Richard McElreath.

This textbook that teaches the use of Bayesian techniques and casual inference approaches (directed acyclic graphs and the like) as a way to test and compare scientific, as opposed to purely statistical, theories. There’s chapters on important topics that other approaches often ignore such as measurement error.

Honestly it felt like hard going at times, particularly as I was reading it as context for an associated course that the author was kindly providing, meaning that there were deadlines. But it did open up an interesting range of new possibilities that I’m looking forward to integrating in my future work, as well as new ways of thinking about modelling, so hopefully very worthwhile!

The code examples in the book mostly use an accompanying R library, rethinking, along with base R techniques. Although people have rewritten the examples to use e.g. tidyverse and brm, or Python, Julia and possibly others - links to those can be seen on the author’s book page.