Lies, Damn Lies and Statistics

Numeri (Numbers) Plate, Laura de Santillana, 1977

Neither snow nor cold nor gloom of night shall stay this blogger from the swift completion of his appointed rounds. Braving the polar vortex, Anne and I first did our Gyrotonic workout and then continued on to Science on Tap. Liberty Vittert, a visiting professor from Glasgow in Mathematics and Statistics was the night’s speaker. Her talk was entitled, How to Win the Lottery and Get Away with Murder. Here she is via YouTube giving the TEDx version of this talk.

With our previous exercise, we arrived later than normal, but Joanie had our table ready and Pat just beat us there. The usual hosts were absent due to a family emergency and without their tutelage the venue’s flakey AV systems reared their awful heads and plagued Dr. Vittert’s talk. She straight off wrecked my planned question, “Which high school did you go to?” Growing up in Saint Louis, she knew that such a question would be code for who are you. Burroughs indicates that she was a local high flyer. 

The gist of her talk was that we can defend ourselves from the chronic misuse of statistics through common sense. The get away with murder portion of the title comes from the OJ Simpson trial. In it, one of the defense lawyers argued that only one in 1/2500 of women who are abused, were murdered by their abuser and you can’t convict on a 1/2500 chance. A more correct way of viewing it is that nine out of ten women who are murdered by a spouse had been abused. The police have always known this and that is why they always suspect the husband.

The how to win the lottery part of the title presents a strategy for dealing with incomprehensibly big numbers, by characterizing them using a real world situation. Imagine a bathtub, the biggest that you’ve ever seen and then imagine that that tub is filled to almost overflowing with kernels of dried rice. Take one of the kernels, paint it gold and then plunge it down into the rice. Then standing at the bathroom door, charge people two bucks to blindfold themselves and pick one grain of rice. This example illustrates both the futility of buying lottery tickets and why government loves them. As the saying goes, it’s a tax on people who can’t do math or more correctly can’t visualize the math.

7 thoughts on “Lies, Damn Lies and Statistics

    • Another of her examples was the old Colgate toothpaste advert claim that “8 out of 10 dentists recommend Colgate.” Crest objected and in discovery it was found that the survey was a list of 10 name brands and dentists were asked to cross out those brands that they would not recommend. Their logic dictated that if they did not “not recommend” Colgate, then they must have recommended it. Crest further protested and examining the raw data found that using the same criteria that “9 out of 10 dentists recommend Crest.”

Leave a Reply to ksbethCancel reply