Suffice to say that I\’m reminded of a newspaper cartoon that circulated a few years ago. It featured a mum and a dad and their superbaby, posing together. “Thank God we found the sperm of that astronaut,” the mother said. “Thank God we found the eggs of that supermodel,” the dad said. The superbaby, meanwhile, was looking up at its doting creators. Above its head was a thought bubble: “Who the hell are these stupid, ugly people?”
I wonder if any of these people questing for a ‘superbaby’ have ever heard of ‘regression to the mean’.
Actually, David, I’d suspect that they have–as a result of reading to familiarize themselves with their project. And, of course, selecting from examples exhibiting more of the desired traits is indicative of a higher mean (with respect to such traits) than not selecting.
I have a particular interest in “regression to the mean.” When I first encountered the concept, quite some years ago, I said to myself, “That’s just the Law of Large Numbers expressed in statistical terms.” But I have no education in either statisitics or mathematics. I’ve asked a fair number of people presumably knowledgeable in those fields, all telling me that I’m wrong but without any explanation (and that includes Godless Capitalist and other statistics types over at gnxp.com; LaGriffe DuLion didn’t bother to answer). Actually, I’m more convinced now that I’m correct because I wrote out what I feel is a “proof” of the matter and it seemed air-tight to me. Any thoughts on the matter?