>Didn't you once say that the Statisticians have been testing these
>web polls and, because of the huge numbers, they aren't actually
>terribly bad or much different from the accuracy of typical
>1000-1500 N surveys of random individuals?
What I remember the AAPOR-ites saying was that web polls were turning out to be more accurate than they should be, but no one could figure out why, since they're totally unscientific. That was a couple of years ago, though - I haven't heard anything lately about that.
Doug