Why The Polls May Be Wrong, in One Chart

The handful of “outlier” polls, like the Democracy Institute poll out over the weekend that finds Trump tied or ahead in the popular vote, all have one thing in common: they don’t rely on traditional polling sample methods such as telephone calls, email or other internet contact. They are coming up with estimates derived from a number of techniques, mostly proprietary, that may presage the future of opinion surveys if the election results tomorrow diverge sharply from the final estimates of the “mainstream” polls. We will find out tomorrow if their “secret sauce” works.

Why might the standard polls be underestimating the Trump vote, despite the efforts of pollsters to give greater weight and coverage of white working class voters they undercounted last time? Roll back the tape to a Cato Institute/YouGov survey last summer that found 62% of Americans say they are now afraid to share their political views. This in particular jumps out: “Majorities of Democrats (52%), independents (59%) and Republicans (77%) all agree they have political opinions they are afraid to share.” Majority of Democrats too? Given Trump’s crossover appeal, polls might not be missing only Republican voters.

Here’s the chart that shows how that 62% overall number understates how polls might not be capturing the voting choices of many people:

Given these startling figures—the result of the unrelenting censoriousness of the media, Hollywood, academia, etc—ask yourself whether you’d tell a stranger on the phone or in an email how you’re going to vote.

There are other reasons, but I’ll break off and await the vote tomorrow with this sendoff:

Responses