- 3rd Anglican Church, Sandwich, Ont.
- Point of Grace Church, Waukee, Iowa
- 1st Unitarian Church of Portland, Ore.
- 4th Presbyterian Church, Chicago, Ill.
- 1st Church of the Nazarene, Nashville, Tenn.
- 5th Reformed Church, Grand Rapids, Mich.
- 9th Church of Christ, Scientist, New York, N.Y.
- 2nd Congregational Church, Attleboro, Mass.
- 6th Mount Zion Baptist Temple, Hampton, Va.
- 5th Avenue Christian Church, Harve, Mont.
I composed this list weeks ago, and a good thing too, because it was a big pain in my tuchus. I don't think it's any secret that most of my lists are not really random; they're almost always consciously ordered in one way or another. (This one is less random than most, obviously, though as mrghoul pointed out last year, it is irrational.) Also, I have a tendency to try to make the lists look more random than they would be if they really were random. In this case, I went to great lengths to make sure that no state and no denomination was represented more than once, whereas in a truly random list there would have been two Reformed churches, three Presbyterian, and three Christian Science. (Which makes sense, sort of, insofar as the Reformed tradition is somewhat more prone to schism than many other denominations.)
This tendency to over-randomize is a very common phenomenon; as Natalie Angier wrote in The Canon: A Whirligig Tour of the Beautiful Basics of Science, which I just finished reading last week, real randomness often makes people uncomfortable by not looking random enough. She relates an experiment performed at the beginning of each semester at UC Berkeley by a statistics professor named Deborah Nolan. Nolan divides her class in two, telling one half to flip a coin 100 times and record the results, and the other to pretend to flip a coin 100 times and record what they think the results would be. She further instructs them to mark their work with a symbol known only to themselves, and place their work on her desk when finished. And then she leaves the room. When she returns, she looks at each spreadsheet and judges them either genuine or fake, and gets it right almost every time. Nolan says:
When I look at the fabricated coin tosses, the length of the longest run of heads or tails is way too short. And overall, the number of switchbacks between heads and tails is way too high. People want to apply the fifty-fifty rule over a very short period of time. They have a skewed sense of probabilities, and they think the odds of getting multiple heads or tails in a row are much smaller than they are. In fact, the probability of getting four heads in a row is one in eight, so there's a pretty high chance of it happening.
So anyway, Happy Pi Day. I hope it was an irrational one.