lsmeans -> emmeans

[dereksonderegger]

Russ Length has been expanding the lsmeans package to be more general and the new package name is emmeans (expected marginal means).

The lsmeans package has been deprecated and we should switch everything to emmeans. Annoyingly.

[dereksonderegger]

Oh cool. the predict() function has an at= argument and you can give it a data.frame.

[dereksonderegger]

I think that we actually ought to do contrasts SOONER. After the inference chapter, just dive into the contrasts and do the regression and 1-way ANOVA contrasts using emmeans.

Skip all of the glht() crap and just pass in the contrast vector in lsmeans() and then show how to use the package nicely for it.

Then the next two chapters ought to be ANCOVA and 2-way ANOVA with all the contrast stuff in there.

After the ANCOVA and 2-way ANOVA chapters we'll move onto the transformations chapter. At the end of that chapter, we should walk through the University and Toads questions.