We were reminded of the huge limitations of voter polls this past week when every major news source who made a prediction in the Michigan primary proved to be wrong by a margin of 25% or more. The simple explanation seems to be that traditional polling methods do not reach a large portion of the population being sampled and that those not reached are significantly different than the sample population. We are beginning to understand that younger voters, for example, do not use their telephones to respond to pollsters in the same way as older voters.
NPR coverage this week went into even more depth to show the weaknesses in polling methods and the inappropriate use that news media make of the errant information.
All of this caused me to questions the most recent national poll by Wall Street Journal/NBC News that showed that two-thirds of registered voters say they couldn’t see themselves supporting Mr. Trump and 56% say they couldn’t vote for Clinton.
I had two reactions:
- I’m personally glad to find myself among the majority of responders in each of these two poll questions. This help abate my recent feelings of political disenfranchisement. Maybe my fellow countrymen aren’t as crazy as they are portrayed yo be in the other news stories.
- If we apply the same line of thinking that caused the Michigan polls to be skewed then this would seem to increase the “two thirds” and “56%” to even larger numbers. This is a problem for the political system; i.e. #worstelectionever. This poll result, if taken literally, makes either party’s candidate unelectable.
The first thing to recognize is that American voters have short memories and weak convictions. Both candidates will move toward the center after their parties’ national convention and voters will believe the new rhetoric as the basis for their voting decision.
The second reality is that we do not have any reasonable basis to predict the voter outcome and anything can happen in this year’s national election.