Polling Isn’t Broken

Lynn Vavreck: “The science of polling is sound, but if you ask the wrong group of people your poll questions, you can get the wrong answers. Think of it this way: An arrow shot by an expert marksman has some chance of hitting the target depending on the wind, the distance and any number of other things, but if the marksman aims at the wrong target, those other things have nothing to do with why the arrow misses. Saying that 1 out of 20 polls is “wrong,” and therefore polls cannot predict elections conflates the chance that comes from drawing a sample of voters with what happened in Mr. Cantor’s poll — drawing the sample from the wrong set of people.”