Not just the Democrats got a shellacking Tuesday. The efficacy of the polls took another hit.
Virginia seems especially problematic. The polls missed Eric Cantor’s fall from grace in the primary, and then missed the closeness of the Senate race was November 4.
The problem is sampling. Since polls indicated Dewey beating Truman in 1948, we’ve recognized the difficulty of obtaining a representative sample. Pollsters then accidentally ignored folks too poor to have telephones.
The problem today may be more intentional. Some people avoid being polled. We, for example, never respond to any unsolicited call on the phone. Yes, some of them say they are polls, but we’ve had such calls turn into sales pitches.
Ask yourself, what kind of person is more likely not to respond to a poll? Perhaps people who like their privacy skew the data. Would people protective of their privacy be broadly spread over the electorate, or bunched?
Of course, the GOP can’t count on “hidden voters.” Remember how Karl Rove kept saying Romney was winning in 2012?
Both parties need to spend less time assassinating each other and more on selling their program (assuming they have a program, which some of us voters doubt. Which won’t show in polls either.)
So, the pollsters are missing a significant segment of the population, and don’t know it. They seem blind to what is and is not representative of the whole.
The next question is, do they care?