The new book titled: Past, Present, and Future of Statistical Science is now available for download.
The official description makes the book sound pretty stuffy:
Past, Present, and Future of Statistical Science, commissioned by the Committee of Presidents of Statistical Societies (COPSS) to celebrate its 50th anniversary and the International Year of Statistics, will be published in April by Taylor & Francis/CRC Press. Through the contributions of a distinguished group of 50 statisticians, the book showcases the breadth and vibrancy of statistics, describes current challenges and new opportunities, highlights the exciting future of statistical science, and provides guidance for future statisticians. Contributors are past COPSS award honorees.
But it actually has lots of good stuff, including the chapter by Tibshirani which I discussed last year (in the context of the “bet on sparsity principle”), and chapters by XL and other fun people. Also my own chapter, How do we choose our default methods?
Nice! I particularly enjoyed the chapter by Jim Berger on conditioning.
E.J.
I guess my only comment is that there doesn’t seem to be any discussion of the Mayo/Spanos “error’ approach to statistics in the volume (it may be discussed in some of the chapters, though your chapter does not mention this, which one would think this approach pertains to model selection). It is convenient to have a collection like this in one place–seems a bit of a time warp to read Anderson’s chapter though (haven’t seem him in 35 years–glad to know he’s still alive).
Numeric:
Mayo’s approach is important as philosophy but I don’t see it having any direct impact on statistical practice so I’m not surprised that this work was not mentioned in a book written by a bunch of statisticians. If any of the authors would’ve mentioned this stuff, it would’ve been me, and indeed I do discuss model checking in my digression near the top of p. 293. I don’t discuss Mayo’s ideas there but I mention my paper with Shalizi which does discuss that work.
I’m thinking of such concepts as severity and model checking through residual checking–your comment on no direct impact on statistical practice gives me an answer, though not particularly one that I find palatable. As I glance more at the on-line book, I don’t see anything about counter-factuals or casual modeling. Econometrics is ignored also (deservedly so, in my opinion :-)). What is truly bizarre about this edited collection is that half of it is advice on how to function as a statistically professional and half is articles on fields of statistics. This is mash-up gone wild–the two types of chapters should have been separated and two books prepared. Part IV should be the separate volume. As it is, it reminds me of one of those all-star groups thrown together for one night that doesn’t work that well–like the Rainbow Concert.
I like the chapters that are about statistics, like my chapter and Tibshirani’s. Career advice, I could care less about. But, then again, that’s just me. On the plus side, the book is free and you only have to read the chapters that you like.
Mayo gets a brief mention in Jim Berger’s contribution – he notes that there’s more than one way to define frequentism.
So it turns out that Herman Chernoff didn’t prove Chernoff’s bound — the proof is due to Herman Rubin.
I put that fact in Wikipediadirectly.
Herman Rubin apparently was very unappreciated especially in regard to his mathematical/technical leadership in statistics.
And one of the pioneer bloggers on sci.stat.theory or whatever it was many years ago.
Maybe sci.stat.consult. I remember Rubin as being helpful there, and solid. I’ve often thought about what a vast improvement CrossValidated is over those sci.stat boards, with regard to civil discourse. http://stats.stackexchange.com/
So, I finally finished reading the whole thing after seeing it here! My review: https://www.goodreads.com/review/show/959222084
Pingback: 2 – Gwern’s book review on Cat Sense | Traffic.Ventures Social