Quote Originally Posted by Ken White View Post
human interaction will always show patterns -- and different modelers will draw different patterns from the same data. You cannot put people in boxes IMO; you have to deal with the person or group as they are and as they constantly shift and change.
Modelers will almost certainly not draw different patterns from the same data. To arrive at a different distribution, you'll need to infer that the domain (in this case the sample size) is too small to rule out piecewise or differential behavior, or that the data set is inconsequential to your object of study. Either way, the fact remains to a degree of accuracy clusters of thinking human beings can be modeled successfully and have been for decades.

Now the data itself--particularly the chosen input streams--can definitely be challenged. Though it's unlikely that three independent studies happened upon k-power polynomial relationships between different sets of variables on their own, they can differ wildly in their constants.

Well, you can put 'em in boxes and rely on trends, I suppose. Seen a lot of folks do some fascinating variations on that. None successfully, as I recall...
If that's the case, we're wasting a lot of money on pshrinks, term insurance, and advertising with absolutely zero discernible benefit.

All things considered, though, I don't guess a Physicist playing around with the People thing is any worse than Economists trying to do that...
A physicist is generally a better working mathematician and statistician than an economist. We've had too few of those in the social sciences in recent decades.