(On The Media/WNYC)
Our friends at On The Media just posted a new episode all about data, featuring Radiolab’s story on the decline effect and other stories about the onslaught of information: personal, political, seductive, disastrous. Listen here!
I loved this bit, but I can't help but wonder if you (and perhaps a lot of other people) are overlooking some subtle statistical something. Maybe your friend Steve Strogatz should have been consulted.
Possibly the explanation lies in a purely numeric problem that all these kind of studies will have. Remember Benford's law? Maybe Darrell Dorrell can help.
I heard your segment on the peculiar outcome of the experiment with the bank robbery. I think a fairly obvious explanation for the regression to the norm phenomenon was overlooked. If the initial outcome had been that the subjects who wrote down characteristics of the bank robber were later slightly better at identifying him in a line-up, it unlikely the experiment would have ever been repeated. The same would be true of any experiment. If the result is counterintuitive it will invite replication, and if the initial outcome is a statistical anomaly it will be followed by regression to the norm. No need for invoking the spirit world here.
Email addresses are required but never displayed.
Radiolab is supported, in part, by the National Science Foundation
and the Alfred P. Sloan Foundation, enhancing public
understanding of science and technology in the modern world.
Radiolab is distributed by