Teaching Statistics

29 Dec 2015 12:26

Doing this is now, officially, what I am paid for. I am basically unembarrassed about doing this while never having taken a statistics class --- after all, I do statistical research, so it's not exactly like a celibate man offering advice on marriage --- but I do want to do it better.

One thing which particularly concerns me is that almost all the introductory textbooks I run across seem like either cookbooks, or lower and distorted forms of Cramér's Mathematical Methods of Statistics. Cramér's book is great, but giving a debased version of it to engineers or social scientists doesn't seem all that effective. So I'm interested in good approaches to teaching statistics as a way of learning about the world from data, not a set of rituals or a calculational exercise in basic probability theory. If they do a good job of teaching about computer-intensive methods and applied probability, so much the better.

In fact, what I'd really like is for somebody to write a popular book on "better living through data analysis". I wish I could say that Freakonomics was that book, but it isn't.