One of the problems with standard economics training is that the econometrics focuses on linear regression and tends to skip over non-linear functions. Most economists would never consider a curve may be, say, a Poisson distribution. Many sem to think adding a squared or cubed term to their model is the height of sophistication!
Part of the problem is that few economists seem to eyeball the data before crunching it. Now there's a tool to help. In a new post on the Statistical Modeling, Causal Inference, and Social Science blog, Aleks Jakulin at Columbia University points us to a great online tool, ZunZun. It lets you use 2 and 3 dimensional 'Function Finders' to "help determine the best curve fit for your data".
I've tried it and it works. Of course, one needs to avoid over-fitting the data. But the developer of the software, James R. Phillips, has promised to add the AIC sooon; this will help select the more parsimonius function.