Passion and Data – Part II
At the New Yorker, a small article hits the point I was trying to make in my previous post pretty squarely. Here is the last paragraph from Risk Management by Lauren Collins:
He acknowledged that the paragraph had come from a forwarded e-mail, but said that, before pasting it into his pitch, he had done “some simple calculations” that supported its conclusions. “In what I’ve seen in dealing with the war and the misperceptions of it,” he said, “it seemed to me like those would be the right numbers.” He went on, “I work in D.C. on a daily basis, and I’m afraid to get out of my car in a lot of places. I hear about police officers being murdered every day in D.C. and Baltimore. And I’ve had thousands of friends and colleagues go to Iraq and come back safely.”
Even without the explanation of the particulars of the question (murder rate in Washington DC vs soldier death rate in Iraq) this quote illustrates two habits of humans: our feelings about our personal experiences count for more than data (anecdotal evidence weighs a lot, passion over data!) and we are profoundly unskilled at probability and statistics. Policies made on this kind of thinking may have good or bad results. But they will be consistent only with the bias, not the way the world works.