stepwise selection using p-values, AIC, Adjusted R2R2R^2).
This makes me wonder if there isn't a missing chapter that hits p-values, AIC, and BIC before hopping into linear regression (which, great intro to p-values, but, maybe portioning off modes of inference into one section or chapter might be useful)
the use of the bootMer function to calculate uncertainty in our subject-specific predictions, below:
Have you looked at predictInterval in the merTools package at all? One of the vignettes talks about comparison to bootMer and finds that it's a better methodology to use. That, and for the simulation geeks out there, you can get raw simulations and play with visualizing those.
Comparisons using AIC
Hrm. This reminds me of ye olde discussions centering on Vaida and Blanchard and CAIC versus Marginal AIC. My take away from that was use CAIC to compare RE structure, and never use Marginal AIC. CAIC is implemented in R as well in the cAIC4 package. It's been years since I was deeply steeped in that literature, though, so I don't know how much things have changed.
(type III sums of squares)18.
As Anova() uses Type II by default, it might be worth including Type II here and talking about type III in the footnote? And/or a brief sentence or table about the differences between them (I have one floating around if ya need it!)
I(Chirps - mean(Chirps))
I wonder if it might be easier to create a chirps_centered variable in CricketChirps intead of using the I() notation here. I find students have a hard time with it.
per unit
Maybe "per one unit" to be clearer?