Here are the results of the recent Wikileaks online poll which asked the simple question: “Who will you vote to become president?”
At first glance, many would dismiss this result as unsurprising. After all, Wikileaks founder, Julian Assange, makes no secret of his contempt for Hillary Clinton. Recently, his organization has been criticized for releasing leaks from the DNC on the eve of the Democratic convention. Certainly, we would expect more anti-Clinton voters to be visiting the Wikileaks’ Twitter page than pro-Clinton supporters. On the other hand, many of the Wikileaks visitors are ex-Bernie Sanders supporters who are not necessarily pro-Trump. This could explain why third party candidates did so well in these polls.
There are other problems with the poll. Voters didn’t need to be from the U.S. to vote. And, like many online polls, there were probably ways to cheat if you wanted to take the time to do so. A voter could use a VPN, for example, to hide their IP address, if that was what was being tracked to prevent multiple votes. However, to vote, the voter must, at least, have had a Twitter account. This would make it difficult for bots to vote. Then again, like most popular sites, Wikileaks has its share of fake followers. The Twitter Audit site found that 30% of Wikileaks’ followers were fake, which is actually quite low. Nonetheless, even given the worst case scenario (most bots, cheats, and fake followers voting for Trump), it cannot fully explain Trump’s margin of victory. This is because Trump has repeatedly beaten Clinton by landslide margins in nearly every online poll. A recent NBC online poll had Trump winning over Clinton by 50% to 19%, while last week’s ABC poll had Trump at a stunning 70% to Clinton’s 5%. (Jill Stein had 17%/ total voters – 63,536). Neither NBC nor ABC would be considered as pro-Trump strongholds. Something else is going on here.
To find out what might be happening, it is necessary to go back to December, 2015, when Trump was inexplicably (at least to many) getting more votes than anyone thought he should be getting and getting more votes than the polls had been predicting. He was also doing surprisingly well in online polls. This phenomenon stimulated an in-depth study on Trump’s online support among Republicans and found that “Trump’s advantage in online polls compared with live telephone polling is eight or nine percentage points among likely voters.” They also found that “a key implication of the study is that many national polls may be underestimating Trump’s support levels.” The reason for the difference between the online and interview polls was connected to “a social desirability bias in which respondents answer questions in a manner they believe will be viewed favorably by others.” This was closely correlated with education level. In other words, well-educated voters did not want to admit, at least to interviewers, to being Trump supporters for fear that it would be looked upon negatively. And remember, this study was only among Republicans. Among Democrats or minority groups, where the stigma of being a Trump supporter would be much stronger, one would expect even fewer people willing to admit they may vote for him. Some anecdotal evidence seems to support this.
The Pew Research Center came to similar conclusions when they compared phone interviews with online polls.
“The social interaction inherent in a telephone or in-person interview may also exert subtle pressures on respondents that affect how they answer questions. Respondents may feel a need to present themselves in a more positive light to an interviewer, leading to an overstatement of socially desirable behaviors and attitudes and an understatement of opinions and behaviors they fear would elicit disapproval from another person.”
The following shows some comparisons in the two types of polling that they conducted in 2014.
The only question that remains is whether being a Trump supporter carries such a negative perception. Accusations of racism, sexism, and homophobia associated with Trump by the media may make Trump supporters hesitant to voice their true opinion of him. Numerous stories in the news of cars with Trump bumper stickers being vandalized and a 62-year-old man being beaten with a crowbar for wearing a Trump t-shirt may have wider implications than just being interesting anecdotes. There really are nervous Trump supporters out there. Actually, there’s very little argument about this as both parties agree that Trump has voters who do not want to openly admit it. The real question is just how many.
Back in April, when the vast majority of pundits were predicting a contested Republican convention, I did an extensive analysis of the social media campaigns of all the candidates and predicted “Trump to win the Republican nomination hands down”. I further predicted that “Sanders will make a close race of it, even if he ends up losing the nomination”. Both of the above predictions came to pass. I further predicted that if the social media statistics stayed the same, Trump would win an election over Hillary Clinton. Since then, Clinton has stepped up her social media presence, especially on Twitter. But she is still 3 million followers behind Trump.
In an analysis of Google Trends, we can still see Trump leading the way by a large margin (red line = Trump; blue line = Clinton).
In fact, Trump has about a 63% to 37% lead in this statistic, which is surprisingly close to what one would find in online polling.
Trump often criticizes the media for working against him. That may be, but it does not seem like something he needs to worry about. The more the media try to demonize him, the more inaccurate interview polls will be. The media may believe their own polls and believe they are making more of an impact on the election than they truly are. By extension, the Clinton campaign may begin to relax, believing the polls and believing the election is in the bag. I would caution them against such behaviors and point to the conclusion reached by the Washington Post when they analyzed the inaccuracy of the Brexit polls before the vote. It should be remembered that the majority of polls showed, by a wide margin, that Brits would vote to remain in the EU. After analyzing the polling flaws of various polling companies, the Post concluded, “two of these companies conducted their polling online, and generally online polls — which had provided lower figures for Remain throughout the campaign — were more accurate.” In other words, online polls, like the Wikileaks poll, should, perhaps, be treated far more seriously than they currently are.