Friday, May 10, 2013

Thinking Slow

In "Thinking Fast and Slow", Daniel Kahneman discusses research into our thinking processes and why they sometimes seem unreasonable. One factor is his belief that we read statistics intuitively. Is a shy, tidy, helpful man more likely a librarian or a farmer? Most say librarian although there are 20 times more farmers. Whether this is truly a statistical error or a logic error is open to debate but see what you think of these numbers:

Mexico found 29,000 guns at crime scenes in 2007-2008 and submitted 11,000 of them to the Bureau of Alcohol, Tobacco, Firearms and Explosives for tracing. Close to 6,000 were successfully traced, and of those, 90% — 5,114 according to testimony in Congress by William Hoover, assistant director for field operations of the ATF — were found to have come from the U.S.. That 90% is a number that is repeated over and over as a reason to control American gun sales. And it is impressive--except for the fallacy. 68% could not be traced or were clearly not American. That means 83% of the guns found at crime scenes in Mexico  were not traced to America. Only 17% were.

This may not be a failure to understand statistics because of some instinctive bias, however. There is another bias, far more prevalent in our weary world. When you are doing any research, as soon as you have a cause, you have a conflict of interest.

No comments: