Economists and Data Mining

Economists are some of the strongest critics of data mining. They often (perhaps partially correctly) argue that the theory must come before data; data driven theory is "bad science," data mining is data fishing and data dredging etc. So I was quite surprised when I read the the following article by Ted Goertzel. As it turns out most econometric studies which use regression never use it for prediction but just for description. In other words many of them don't understand the concept of "generalization" and tend to use overfitted models! Unbelievable.