Hi Tom. Thanks for the response. I totally get your point and read the Pew article. It is, of course, no surprise that the entire polling industry has come to the collective view that despite these complete failures to provide the outcome that users want - i.e. the correct prediction - the tool that is the complete basis of its business models is just fine. Of course, the pollsters would care that their models were actually 'close' from their internal criteria. But remember, there wouldn't be a polling industry if pollsters didn't provide an outcome valuable to the readers. It is no comfort to users to have pollsters say, after the fact - "No, no you don't understand. We were really, really close." They all said Hilary would win.

As you know, Thomas Kuhn wrote eloquently about the phenomenon we are discussing. You are on the side of normal science - attempting to justify glaring anomolies. And the justifications make oodles of sense to normal scientists. But as the anomolies mount - and they will - people will start to open their eyes to the possibility that there is a flaw in the prevailing model (that being that people will continue to acquiese to the use against them of a tool that requires their cooperation). That takes years or decades. So I don't expect my suggestion to be accepted. No, it will be fought - for years or decades. But my friend, that doesn't mean it is 'so wrong.' It just means that it isn't acceptable to say yet. Best, Roger