A Wonderful Panorama of Statistics
Bob and I have been off-air for a while, but we haven’t gone away. I’ve been meaning for ages to blog about a wonderful book. Here it is:
Sowey, E., & Petocz, P. (2017). A panorama of statistics: Perspectives, puzzles and paradoxes in statistics. Wiley.
And the flyer with succinct information about the book is here. (Enjoy the full spread of the fine artwork that wraps around the book’s cover. The original painting, by Jeffrey Smart, is an enormous and wonderful sight in the foyer of one of Melbourne’s prominent theatres. Worth visiting!)
Panorama has been my beside-the-bed book for a while now. You could read it straight through, but I’ve preferred to dip in haphazardly, just about always finding something intriguing. It’s a cornucopia of statistical ideas, examples, oddities, paradoxes, historical tales, and more.
The back story: The journal Teaching Statistics, from 2003 to 2015 published the Statistical Diversions column by Peter Petocz and Eric Sowey. Peter and Eric are distinguished statisticians–and statistics teachers–based in Sydney.
Whenever a new issue of Teaching Statistics arrived I would first turn to their column to check out the new goodies, and the commentaries they gave on the questions they’d posed in the previous issue.
Eric and Peter now present the content of those columns, and more, assembled into coherent chapters as their Panorama book. It’s a great resource for any teacher looking for ways to engage or extend their students, or for anyone simply interested to explore–and be fascinated by–the discipline of statistics.
Here are a few tastes:
Over about 20 years I built ESCI and wrote two books (the second with Bob, of course). For all that time I played with simulations of randomness, notably ESCI’s dances–of CIs in particular. I concluded early on that randomness is endlessly surprising and fascinating. It’s amazingly lumpy in the short term, while in the very long term fits exactly with what theory says we should expect. Even with this long experience, I found Chapters 11 (Some paradoxes of randomness) and 12 (Hidden risks for gamblers) especially interesting.
My brief version of Q11.5 (p. 89): Two people, Alice and Bert, toss a fair coin numerous times. Alice scores a point when a Head turns up, Bert a point for a Tail. How often is the lead likely to change? See pp.243-244 for the authors’ discussion–which may help us avoid unwarranted conclusions about what the movements of stock prices mean.
Getting the answer you want
You are teaching about questionnaires and wish to explain how a sequence of slanted questions can steer respondents in any direction you choose. A short and sharp satirical example is from the classic British Yes Minister program. See pp. 64 and 228 in the book, and the video here. (There are numerous links in the book. A list of all those links, in clickable form, is here.)
Eponymy and Stigler’s Law
We’re all familiar with many statistical eponyms (the Fisher exact test…). What is the relevance of Stigler’s law? Is that law true or false? If true it must be misleadingly named? See pp. 178-181 for an intriguing discussion, and pp. 292-295 for discussion of the questions posed in the earlier pages.
P.S. On a totally different topic, one of the reasons I’ve been off-air is that Lindy and I joined a two-week tour of Greenland. It was fascinating. For example we visited the Ilulissat Glacier, which drains about 7% of the huge Greenland icecap, and which may have been the source of the iceberg that sank the Titanic. The most scary statistics I’ve seen for some time describe how that giant glacier–and others in Greenland–have greatly increased their rate of retreat in the last 10-20 years. In the case of the Ilulissat Glacier the calving is now no longer from a vast floating ice tongue, but from the much-retreated glacier front sitting on solid rock. So now the massive new icebergs all contribute to rising sea levels. That’s accelerating climate change in action, which is truly scary.