Joe's Road Trip 2012

2012 Road Trip Day 15: How Pollsters Work

  • Share
  • Read Later

Columbus, Ohio

We’re heading to Michigan tomorrow and this is a very smart analysis of the skewed results of recent polling there and the problems with polling in general. Note the response rate to telephone polling–4%. Hmmm. My own rule is to believe the results when they’re weighted overwhelmingly one way or the other, but to consider anything within 10 points a tossup. It irks me mightily when you see one of these cable TV dunderheads breathlessly announcing that Governor Romney (or President Obama) has surged into a 2 point lead in the state of West Neuralgia. We’ve got a long way to go before humans decide this thing. And even at the end, the response rate probably won’t be more than 10%, and in the better polls that will be reassigned according to voter participation models that are theories the pollsters have about who will actually vote. Some of these theories will be more accurate than others–but we, all of us, including me, will suck them like crack cocaine. If you notice me doing any such thing come September, please spank me with comments.