VOLUME 34, ISSUE 4
Distributed quarterly by mail and email, the Conservative Caucus of Delaware's newsletter contains relevant information and insights from noted leaders, authoritative stakeholders and like-minded members who demonstrate their passion for the truths we hold dear by putting pen to paper!
How the Polls Work and Why Trump Can Win
By Sam Friedman
As the weeks turn into days before arguably the most important election since 1980, if not in our lifetimes,
more and more people are increasingly following this year’s presidential polls in anticipation of who’s in the lead and how the third-party candidacies of Governor Gary Johnson (Libertarian) and Dr. Jill Stein (Green) are affecting the race between Donald Trump and Hillary Clinton.
The problem is, more than in
any presidential election since polling began, no one really has a solid idea of who’s winning. Polls over the
years have sometimes missed major
elections, such as the 1948 presidential
election when Truman surprised
Dewey because he made a late push for voters that pollsters didn’t anticipate and the “Doug Wilder” effect where Black candidates perform better in the polls than on election day.
I strongly suspect this year’s polling is no different. Anyone admitting to
Voting for Trump is likely to be tagged as ‘racist,’ ‘misogynist,’ ‘homophobic,’ or ‘crazy,’ whether they actually are or not. Because of this, I suspect many Trump supporters are either not answering pollsters’ questions or are outright refusing to take part in surveys. This does not mean, however, Trump is going to take the election easily, or at all.
So how do polls work? I had the fortune of spending four semesters working for the Quinnipiac University (QU) Poll while a student at QU. The poll is funded by the University but is run separately from them, so there’s no conflict of interest and the poll is free to run as it sees fit (though one of my supervisors did have a Cheney-Satan 2008 bumper sticker).
My work there gave me some
insights into how polls are conducted and who exactly answers them.
QU’s poll is done by live-calling and is based on asking the first person to pick up the phone which adult in the home has the next birthday. I was told by my supervisors this is done to reduce oversampling women, since
they are more likely to be home in the early evenings than men when the poll begins its calls (typically calls begin around 5:30 and end around 9:00 pm). If the person with the next birthday was not available, we scheduled callbacks. However, there is no way to verify the information and I am sure on multiple occasions the respondent, usually a middle-aged woman, was probably lying about the next birthday, but we had to take the respondent’s word and proceed.
I remember I spoke to more moderates and liberals than conservatives, and I can hazard a guess why. Back in 2008, I spoke to a Cuban-American living in Florida who was a registered Republican got angry at me because she said something outlandish about President Obama (it was very far out there) and I chuckled, which she took as proof that I was part of the liberal establishment trying to embarrass
conservatives like her. She then hung up on me before finishing the survey (QU trashes unfinished surveys). Anyone who knows me knows I am not part of the “liberal elite,” or any elite for that matter. But I remember it took me almost four minutes to persuade her to take the survey, because she was on guard the whole time against bias.
We occasionally did national polling (which meant I worked a shift from 9:00 pm-midnight on Tuesdays) to get the entire country, though we got many more on the West Coast than in rural conservative states like Wyoming or Montana. Dialed numbers are
auto-generated via a software program. This meant I could call a house, a business, a hospital, or one of the CIA hotlines (this really happened) and we never knew in advance who we called.
When cell phones were added later to account for those who did not own landlines (we had to manually dial cell phones per U.S. law), I remember fewer completions with a crowd that was mostly younger and diverse.
Overall, the QU poll has a reasonable track record of being with-in the margin of error for accuracy. But this year could be different and I think their frustration is the same every pollster has this year, especially since I think a lot of Trump supporters are keeping their true feelings unknown. Yes, I believe there are people attacking Trump on social media and refusing to support him in the polls who intend to vote for him on Election Day!
The only thing guaranteed is that no one knows exactly how the election will go, but my best guess as of now is: Trump 49 percent, Clinton 47 percent, Johnson 2.5 percent, Stein 1 percent, everyone else 0.5 percent in a low-turnout election.
My electoral map has Trump 295, Clinton 243, and Johnson running big in Maine’s 2nd district and in New Mexico and Colorado. This of course does not factor in the possibilities of Trump saying something severely damaging to his campaign, Hillary falling very ill, or a major terrorist attack (I mean sporting-event sized, not one Jihadi). ■