Generalized linear regression

[rosemm]

New module!

[rosemm]

Okay, @drelliche it's ready for you again! I implemented several (hopefully!) improvements:

• Moved around some existing content to improve the pedagogical flow, as we discussed, especially breaking up the sepsis example so the problematic attempt with the plain linear model occurs early on, and the better attempt with the logistic regression is toward the end
• Smoothed out the new transitions, added a little more scaffolding around, in particular, the plot of the bad linear model
• Substantially reworked the "probability and odds" bit, which is now under the heading "talking about chance" -- lemme know what you think
• Added a whole new quiz to try to more effectively hammer home the question of what is (and isn't) a continuous unbounded outcome, to better support learning objective "identify when a logistic regression model is appropriate"
• Added a little more explanation of plain linear models at the very beginning. I'm happy to take that out again, though, if you think it's overkill given that OLS is listed as a prereq.

[drelliche]

@rosemm This looks great! I'll go over it for all of the checks/typos, etc in the next few days.

[franzenr]

Looking great! Just a few small suggestions, and a point or two of clarification :)

I've always wanted to know more about logistic regression in particular, but never gotten around to reading up on it -- this was such a wonderful intro!! 😃

[rosemm]

Thanks for the great edits, @franzenr! I think it's ready for you again -- lemme know if you want further clarification or anything :)

[franzenr]

@rosemm I've got two replies for you to mull over, but that's it!

[rosemm]

Thanks, @franzenr ! Back to you :) 🎾

[franzenr]

one last(?) lob @rosemm 🎾

[franzenr]

Great work Rose! This is an awesome addition to our module collection :)