“Businesses conduct hundreds of thousands of randomized trials each year. Pharmaceutical companies conduct thousands more. But government? Hardly any.”
–David Brooks, The New York Times, 4/26/12 editorial seen here
For those of us in the know about statistical tools this statement provides light at the end of a long tunnel. However, this columnist gets a bit carried away by the idea that an FDA-like agency inject controlled experiments throughout government.
Although it’s great to see such enthusiasm for proactive studies based on sound statistical principles, I prefer the lower-profile approaches documented by Boston Globe Op-Ed writer Gareth Cook in this May 2011 column. He cites a number of examples where rigorous experiments solved social problems, albeit by baby steps. Included in his cases are “aggressively particular” successes by a group of MIT economists who are known as the “randomistas”—a play on their application of randomized controlled trials.
Evidently the obvious success of Google (12,000 randomized experiments in 2009, according to Brooks) and others reaching out over the internet has caught the attention of the mass media. Provided they don’t promote randomistas running wild, some good will come of this, I feel sure.