Pardon my popping in from vacation to notice this morning’s New York Times column from Thomas Edsall contending that the exit polls from the 2016 election were inaccurate, and overestimated the percentage of the electorate with college degrees by a significant margin:
Pew found that whites with college degrees made up 30 percent of the total electorate, not the 37 percent reported in the exit polls. In other words, Pew found that white working-class voters outnumbered white college voters among all voters, while the exit polls reported just the opposite.
These numbers have powerful ramifications for both Democrats and Republicans preparing for the 2018 and 2020 elections.
By showing that the white working class makes up a larger proportion of the electorate than previously reported, the Pew report — taken together with similar results in a study sponsored in November 2017 by the liberal Center for American Progress — strengthens the case made by Democratic strategists calling for a greater emphasis on policies appealing to working class voters and a de-emphasis on so-called identity issues.
It reminded me of the conclusions after the 2004 election that the exit polls overestimated the vote for John Kerry in 26 states and the vote for George W. Bush in four. And how the 2006 exit polls in several Senate races showed significantly bigger margins of victory than the actual results. Or the 2008 exit polls that showed Obama’s margin three to four points larger than the actual voting margin in seven swing states. Or the 2012 exit poll in Wisconsin that implied a 90 percent turnout rate among African-Americans.
Let me offer a simple theory about why working-class whites would be so under-represented in exit polls. Exit pollsters conduct phone interviews before Election Day to get a sample of those voting absentee and early. On Election Day, exit pollsters send staff to a hopefully representative selection of polling places across the state, and ask people to fill out questionnaires about their demographic information and their ballot choices. About 50 to 60 percent refuse to participate; when someone refuses, the pollster notes the person’s rough age, race, and gender. (Notice you can’t measure someone’s education level by looking at them.) They then weight their data to match the population that voted at that location.
If the largest group of persons conducting the exit polls are college-educated whites, then it’s likely that despite their best efforts to get responses from a representative selection of voters at the polling place, they are more likely to get voluntary responses from other college-educated whites.
Certain groups may be particularly wary of revealing their demographic information and voting choices to a professional-looking person with a clipboard. In an angry, partisan, judgmental political culture, why would anyone want to tell a stranger who they voted for? Particularly when they know that in the eyes of the elites, their candidate preference is the “wrong” one?