Model fitting

Don't introduce the file= argument of brm() until much later. It causes more harm than good. (A lot of student follow good practices of building the model incrementally, e.g., fitting intercept-only first to make sure things run at all, but then that's the model that gets saved to file and read in, causing confusion later when they expand the model).
- Maybe include it toward the end of the course as a helpful tip for when people are fitting more sophisticated models that can take a long time to run?

Data summary + visualisation

Motivate why certain measures are suitable, and others are not. Would be good to include the reasoning on the slides, rather than just "do" and "don't".
use conditional_effects()

We should repeatedly drive home the difference in interpretation between model coefficient estimates and the conditional posteriors, and why they are different.

The class gives a lot of tools, but doesn't really talk about how to apply them (based on my experience and also some feedback from first-semester and current students). Is there a way to talk about the model selection / decision-making process?

e2_1.csv, change "repose" to "Response"
fix wording in exercise 2.6 to make it clearer that we're looking for an interaction between condition * response

In Piazza, include folders for each of the assessments in addition to the folders for each week (for easier filtering, and since people work on them over many weeks)

Starting with categorical predictors (getting people comfortable with 0s and 1s), and then complexifying with continuous ones. This worked SUPER well and I love that we did it this way—I could say "a lot of this is similar to how it was before, but now with a little tweak! and this tweak actually gives us the more general case! (a unit change)". Super good.