use gganimate or at least add another level of transparency to clean up the spaghetti plots

In keeping with glmmTMB and brms notation, and following Bolker's demo. Probably best for appendix.

The initial part where using of merTools came off as confusing to most.

Hi Clark,

first of all thank you for this precious guides. I just want to ask you if there is the possibility to publish the related R snippets code (eg all the plots codes). Thanks in advance

If possible try to provide format that might be more print friendly. Currently epub will break with html based images (won't print any subsequent text), but perhaps it will be more viable in the future. PDF at present seems incompatible with the current approach.

A current solution is to minimize the menu, which is more print friendly than retaining.

In Issues/Convergence, add link to post

Your deck is truly an excellent introduction to mixed models. I wish it had existed 3 years ago!
One thing I think is missing is a note on how to do hypothesis testing using LME.
I ended up using something like this

    #adding groups fixed effect
      model.test <-lmer(metric~experiment_treat+
                           (1|user_id)+(1|days_in),data=experiment_session_df,REML=F)
    # create a test
    anova(model.null,model.test) %>% 
      data.frame() %>%
      display(.)

    #display the model
    model.test %>%
      display(.) 

Where experiment_treat is the independent variable I really care about, but I want to extract the effect of user and time, while seeing the effect of time (days_in).
It would be good to have something about doing these kinds of hypothesis tests in your guide.
Cheers,
James

https://m-clark.github.io/sem/latent-growth-curves.html
in issues.Rmd is not working

in first nested example use the default ward identifier (i.e. 'ward'), making for more comparison to nesting vs. crossed section later.

In the issues section, add link for mgcv, lme4, glmmTMB comparisons

currently is a folder within the older version of the folder

Great book! Just finished it up. In the appendix you mentioned that Python and Julia offer a subset of modeling options that R does. You also mentioned that you've been able to build some models in R with over a million records. Do you know if Julia (or Python) can process larger amounts of data than R for running these types of models or any run time advantages? I would like to think Julia could but I'm not sure what the bottleneck is in the training of these models or to what extent there are multi-threading opportunities... Again, great book!

Discuss issues/difficulty, demo no more than two ways.

Perhaps make a separate chapter given the utility/importance

• Intro
• Random Intercepts
• Random slopes
• Extensions
• Issues
• Bayesian
• Supp/Appendix

Thank you for such a nice Package.

1. Can you add some special covariance structures? For example, FA0(2) or CSH(Heterogeneous Compound Symmetry).
2. This package does not seem to support ANOVA.