Skip to content

Sorry, no ARM solutions

Daniel Gerlanc asks:

I’ve been reading your Regression and Multilevel Modeling book. Do you have a set of example solutions for the problems in the book?

Henning Piezunka, Adam Lynton, and others have asked the same question.

My universal response:

I’m glad you like our book. Unfortunately, we have no solution sets. I made a bunch of solutions for my earlier book but it was so much work that I decided not to do it a second time!


  1. zbicyclist says:

    If neither you nor a lackey worked out the problems, are you sure there are solutions?

  2. John says:

    Someone, somewhere, is working out the solutions. Probably multiple people are.

    What about crowdsourcing the solutions: creating an official site for the book at which people could post their own answers to the problems? In principle, anyone could create such a site. But it would attract more attention (and solutions) if it came from you.

  3. Who you really need is Chengxiang Zhai (now of UI Computer Science). He took my class in natural language semantics at Carnegie Mellon and submitted homework based on questions in my book (to which I hadn’t developed answers). Chengxiang’s ansers were formatted in LaTeX and were better than I could’ve produced myself!

    @zbicyclist You always wind up with the problem of difficult-to-answer questions, or at least questions whose answers are much more complicated than you thought they’d be when you wrote them. But providing the answers you intended doesn’t solve the problems of people interpreting a vague question the wrong way, which is a large part of the problem.

    I think there’s too much focus on getting the “right” answers to rote exercises and not enough focus on hard exercises that make you think. Original readers complained so much that I wound up adding a bunch of rote exercises to my semantics book. But I find a question without a clear right answer often teaches you more than a pure “exercise”. But students and lazy faculty don’t like it in my experience — I fielded no end of questions about the exercises in my semantics book, many of which were thought exercises I didn’t have an answer for.

    Seriously, though, vetting a crowdsourced official answers site (as in being the Wikipedia owner of a page) would probably be as much work as doing it yourself.