Alan Blinder left his position as a Princeton economist to serve as vice chairman of the Fed. After he went back to academia he wrote a great short book that I see as a crypto-memoir, Central Banking in Theory and Practice (GMU students: it’s free here).  One big lesson from the book: Central bankers are suckers for the law of small numbers, for the unrepresentative anecdote, for the hasty generalization. In his words, central bankers do a lot of uncle-asking (PDF):
You can get your information about the economy from admittedly fallible statistical relationships, or you can ask your uncle. I, for one, have never hesitated over this choice. But I fear there may be altogether too much uncle-asking in government circles in general, and in central banking circles in particular. 

I’m glad to draw attention to the individual problems of economic statistics, and I’m glad to judge a country’s economy partly by walking around and looking, especially in countries with weak economic statistics divisions.  But Blinder draws attention to something I’ve seen in Washington quite a lot over the years (yes, my own embrace of the law of small numbers): A willful embrace of the intimate and individual over the abstract and averaged.  

“I knew a guy” goes a long way in DC, a lot longer than it should.  If central bankers are doing it, politicians are doing it even more: So if you’re wondering, regardless of which party is in power, “On those occasions, however rare, when policymakers genuinely try to get their facts right, how do they try to get those facts?”
According to Blinder, too many of those facts come from uncle-asking.  If central banking is government near its best, as Blinder has written elsewhere, then one can only imagine where the supposed facts come from in other branches of the bureaucracy.