Is there a fancy word for, better yet papers on "overkill modeling"? I need a term to get the google scholar search going.
Am reading some sociological papers and they use very little unrepresentative data, then from this data they build the weirdest and most complex structural equations. It all sure looks real sciencey, but I think it's crap à la: Likelihood that consumer A buys from brand B = consumer age * 0.377 y + education / 0.14x + (number of pans in the kitchen)^2 - 0.000012 * 1/sqrt(ln(e)) * sqrt(price).
I mean, seriously people?: