Shravan points us to these materials:
Hierarchical models are bread and butter stuff for psycholinguists, so we are trying hard to make Stan/brms mainstream through various means. Teaching this stuff feels like the most important work I am doing right now, more important even than the scientific side of things.
We have chapters on hierarchical modeling in our book (to be published soon with CRC Press), we use both brms and Stan:
https://vasishth.github.io/bayescogsci/book/ [edit: made it a live link]
The online version will remain available for free. Comments/corrections are welcome; one can open issues: https://github.com/vasishth/bayescogsci/issues
This summer, I [Shravan] am teaching an intro to Bayes using brms/Stan, with a focus on hierarchical modeling, especially directed at researchers in linguistics who do experimental work:
https://www.mils.ugent.be/courses/module-9-bayesian-data-analysis/
Plus, at Potsdam, for the last seven years I have been running an annual summer school on stats for linguistics and psych, where our focus is on hierarchical modeling using Stan/brms:
https://vasishth.github.io/smlp2024/
Here, we teach both frequentist and Bayesian approaches to hierarchical modeling.
Cool! Good to have these resources out there.
Cool! Thanks to Daniel, Bruno and Shravan (and CRC!) for making this open access. I had the pleasure of reviewing the hierarchical modeling chapters for CRC (for which they sent me $400 in books, so I now have the Handbook of X for a lot of X!). This is a really nice intro int erms of learning how to model and workflow.
Thanks for reviewing those chapters, Bob; your review was very helpful.
The book is now almost done and will be sent to production in April or so. I wish CRC Press would
bring out soft covers right away with the first printing. It’ll be some 900 pages and a
hard cover will make it even heavier than it needs to be.
Congrats Shravan, Bruno, and Daniel! It looks great based on a quick browsing.
Had an idea for a homework problem…. Could have readers try to fit a model for the Stroop effect:
rt ~ cnd + (1+cnd|subject) + (1+cnd|word) + (1+cnd|color)
I could not get this to work cleanly a few years ago, maybe things have changed?
p.s., In the references, maybe there is a missing author list for the bridgesampling paper (after Gronau et al.)?
Thanks Doug. I will test out that model and maybe add it as an exercise. How was cnd contrast coded?
I believe it was 1 0 -1 for facilitation, baseline, and interference, respectively.
Shravan, Bruno, and Daniel – great book thank you. I like the log normal race model. Coding the predictors as a matrix and including an interaction term would be even more useful for that model but it is still great as it is.
Thanks, Jason. I didn’t understand what you wanted exactly, but I will talk to Bruno about it.
If you read parts of the book and notice errors, please open an issue on the github page; thanks :)