Could a team of George Mason University students accurately predict the winners of Super Tuesday’s voting before the polls closed?
That was the task for the 36 students in George Mason’s GOV 311 Public Opinion and Electoral Behavior class. Jeremy Mayer, a professor in Mason’s School of Policy, Government, and International Affairs, coached the class through the intricacies of political polling and turned them loose on March 1 at three Fairfax County precincts.
The precincts were chosen to reflect demographic characteristics leaning toward a majority conservative or liberal bent. A third district considered “bellwether” was also included, said Katie Garay, a junior government and international politics major.
Armed with Dell tablets loaded with a special data collection app, three teams of exit pollsters fed information to another team at the project’s headquarters at the Hub on Mason’s Fairfax Campus. Students worked in three- to four-hour shifts until the final numbers were in after the polls closed at 7 p.m. (As a practice, exit poll predictions are not made public by media outlets as they may influence the outcome.)
Project leader Julian Ausan, a junior government and international politics major and member of Mason’s ROTC, said the idea was not so much to project the outcome of the voting but to explore the research aspects.
“We’re trying to figure out what’s driving people to vote the way they do,” he said.
Students requested information not only about voters’ first choice for candidate, but also possible alternate candidates in both parties, said Alexa Rogers, a junior communication major and editor-in-chief of Mason’s Fourth Estate. Voters were also asked what motivated them the most to vote in that day’s election and whether they had voted in a primary before, among other questions.
Lauren Ernst, a sophomore government and international politics major, learned that the phrasing of a question is key to eliminating bias. “There are a lot of little things I never would have thought of as ‘wrong’ until Professor Mayer explained it to us,” she said. For example, a question about whether the government should provide health insurance to those who can’t afford it should not say the reason is “to prevent bankruptcies and unnecessary deaths,” because this wording gives the respondent reasons to say yes, not no.
Sophomore public administration major Colin Sapko was intrigued by the GOP’s John Kasich voters, 42.3 percent of whom considered themselves to be Democrats and 32.1 percent of whom said they had no second choice.
The students also tracked how many voters refused to respond to the exit poll, said Caleb Kitchen, a junior government and international politics major. He said about 25 percent of those asked to participate declined.
“That’s not as bad as I thought it could be,” he said, surmising that because this was an academic project, people were more likely to participate than if it were spearheaded by a professional political organization.
What was the final outcome? Mason’s poll predictions were accurate. Bonus: Mayer’s prediction that Donald Trump would not win Fairfax County was accurate; that went to Marco Rubio, even though Trump won Virginia.