Let the Great Unskewing Begin

Looking for the “hidden Trump vote” in online polls.

Some of us are old enough to remember that back during the 2016 Republican primary, Donald Trump could hardly talk about anything other than how well he was doing in the polls. Of course, now that his opponent is leading in all of them, the great unskewing has begun.

Most often those attempts focus on the idea that pollsters are oversampling Democrats and/or Democratic-leaning groups. But over the weekend, word started spreading about a “hidden Trump vote.”

Longtime Republican pollster and Trump advisor Kellyanne Conway explained on the Today show that the hidden vote is seen when anonymous online polls are compared to phone call interview polls where people have to be more public in stating who they support.

For example, the new Reuters poll had Trump and Hillary Clinton just three points apart, where others that are phone survey based out Sunday showed a greater gap.

“The Reuters poll, which is an online poll, where Donald Trump is three points behind Hillary Clinton nationally, and I think that the important point to note there is that when you have online polls as opposed to telephone polls, Mr. Trump tends to do better, and that’s because the online polls approximate the ballot box, where you’re issuing your vote privately,” she said.

“We think there’s a big hidden Trump vote in this country,” said Conway, who added that Trump internal polls project “tighter” results in battleground states.

That’s an interesting take. It is the mirror image of what has in the past been postulated as the “Bradley effect.”

The theory proposes that some voters who intend to vote for the white candidate would nonetheless tell pollsters that they are undecided or likely to vote for the non-white candidate. It was named after Los Angeles Mayor Tom Bradley, an African-American who lost the 1982 California governor’s race despite being ahead in voter polls going into the elections.

The Bradley effect posits that the inaccurate polls were skewed by the phenomenon of social desirability bias.

What Conway is suggesting, without really saying so, is that there is a “social undesirability bias” against Trump that surfaces in telephone polls but not online polls. Who am I to argue against that? The fact that her candidate might suffer from that kind of bias is a statement in and of itself.

But with all the tools we have at our disposal these days, it is a theory that is easily tested. Rather than rely on the one Reuters poll, we can take a look at the averages of online only polls and live telephone polls at Huffington Post Pollster. Here is how that looks today:

Online: Clinton 43.9 / Trump 39.5

Telephone: Clinton 47.1 / Trump 41.6

While it’s true that the margin of Clinton’s lead is less in the online polls (from 5.5 to 4.4), Trump’s support actually drops – just less than Clinton’s. So much for the “hidden Trump vote,” I guess.

There are a lot of new challenges and developments in political polling that are affecting the industry right now, including the prevalence of cell phones and the inclusion of internet polling. For a more thorough look at that, I suggest reading this summary by Cliff Zukin, a professor of public policy and political science at Rutgers University and a past president of the American Association for Public Opinion Research. Here is what he says about online polls:

The new economics have driven many election pollsters to the Internet, where expenses are a fraction of what it costs to do a good telephone sample. However, there are major problems with Internet polls. First is what pollsters call “coverage error.” Not everybody is reachable online; Pew estimates that 87 percent of American adults are Internet users.

But Internet use correlates inversely with age and voting habits, making this a more severe problem in predicting elections…

A much bigger issue is that we simply have not yet figured out how to draw a representative sample of Internet users…

Almost all online election polling is done with nonprobability samples. These are largely unproven methodologically, and as a task force of the American Association for Public Opinion Research has pointed out, it is impossible to calculate a margin of error on such surveys.

In other words, this is new territory for pollsters and there is a lot of work left to do in figuring out which method (phone/internet) produces better results. He simply points out that pollsters relied heavily on internet polls and did a terrible job of predicting the 2014 midterm elections as well as those in Israel and Britain. It could be that online polls are the ones that are skewed.

Nancy LeTourneau

Nancy LeTourneau is a contributing writer for the Washington Monthly. Follow her on Twitter @Smartypants60 .