By J. Scott Armstrong, forthcoming in the International Journal of Forecasting.
It's an excellent summary of some important but underappreciated weaknesses of regression analysis as it is often practiced. (Note, though, Professor Armstrong's warning: "if you are a user or consumer of regression analysis or 'econometrics' or the editor of a journal, you will likely find this material unsettling.")
Here's one fine paragraph:
This illusion [that correlation implies causality] has led people to make poor decisions about such things as what to eat (e.g., coffee, once bad,is now good for health), what medical procedures to use (e.g., the frequently recommended PSA test for prostate cancer has now been shown to be harmful), and what economic policies the government should adopt in recessions(e.g., trusting the government to be more efficient than the market).
Do not use regression to search for causal relationships. And do not try to predict by using variables that were not specified in the a priori analysis. Thus, avoid data mining, stepwise regression, and related methods.