See an interactive visualization of the poll results here.
Hillary Clinton led Donald Trump by nearly 70 points in a representative poll of Duke undergraduates.
The poll, which had a 41.2 percent response rate, showed 75.7 percent of students intend to vote for Clinton, compared to 6.4 percent who say they intend to vote for the Republican nominee.
Among fraternity members and students who identified as Republicans, Clinton's support was notably lower than in the general student body. Approximately 62.1 percent of fraternity members said they would vote for Clinton. Eleven percent said they would vote for Trump, and 12.6 percent said they would opt for Johnson. In 2012, Romney earned 37.5 percent of fraternity members’ votes.
Among Republicans, Trump edged Clinton out with 32.6 percent of the vote compared to Clinton’s 31.2 percent.
The Interfraternity Council and junior Colin Duffy, the president of Duke College Republicans, could not be reached for comment.
The representative poll, conducted with assistance from the Office of Institutional Research, was sent to one-third of undergraduates, or 2,230 students, via email and was open Oct. 5 - Oct. 12
“A one-third random sample is gigantic and to have a 40 percent response rate I think gives you a lot of confidence that you're getting a reasonably good reflection of Duke students,” said David Rohde, Ernestine Friedl professor of political science.
Clinton’s percentage is the highest among any presidential candidate since The Chronicle began polling about the presidential election in 2008. Polling in those years showed President Barack Obama winning 74.8 percent of the student vote in 2008 and 65.6 percent of students supporting him in 2012.
“I would bet that the people who favored Obama in previous polls were more strongly positive about him than people who favor Clinton are about her," Rohde said. "It's just that it's measured against the alternatives."
Support for Trump is much lower than for previous Republican nominees. Among Duke students, John McCain and Mitt Romney polled at 23.3 percent and 26.5 percent, respectively.
“[Trump’s support rate] is even smaller than I thought it would be,” Rohde said. “Duke students aren't all automatically liberal, but it would take a different kind of Republican to appeal to them. Neither McCain nor Romney were candidates who alienated voters the way that Trump has.”
Libertarian candidate Gary Johnson received 3.7 percent of the overall vote, while 0.9 percent of students said they intended to vote for someone other than Trump, Clinton or Johnson, and 6.7 percent of voters said they were undecided.
The remaining 6.7 percent of respondents said they did not intend to vote for any candidate. Since the poll sampled all undergraduates, this may include international students.
“Especially from a Democratic point of view, we are the party who really just wants higher voter turnout,” said junior Amy Wang, president of Duke for Hillary and a vice president of Duke Democrats. “It would make me incredibly happy to hear that 94 percent of eligible voters vote, but I wouldn't be surprised if it was a lot lower of that.”
Towerview also asked students which words they would use to describe each candidate. Reflecting on-campus support for Clinton, the most common words students used to describe Clinton were primarily positive, while Trump’s were primarily negative.
The top three words used to describe Clinton were "experienced," "qualified" and "strong," but they were followed by unfavorable descriptors—"liar" and "corrupt."
“It's very easy for the Republican Party to then spin [her political actions] into whatever they feel like or for Donald Trump to come by and just call her a liar when people are already kind of tired to hear that to begin with,” Wang said.
All five most commonly used words for Trump were negative: "bigot," "racist," "idiot," "crazy" and "asshole."
Although Clinton leads Trump by 69.5 percentage points on campus, the race is much closer statewide, where current polls from FiveThirtyEight show Clinton up 48.5 percent to Trump's 45.8 percent.
The growing support for the Democratic candidate may also reflect an increase in liberal-leaning students on campus. Of students who disclosed their party affiliation, 58.8 percent said they aligned closest to the Democratic Party. Comparatively, 45.3 percent of students said they were registered Democrats in 2012.
In contrast, although 18.5 percent of respondents in 2012 said they were registered Republicans, students that aligned with the Republican Party made up 15.9 percent of the respondents in the 2016 poll.
“I wouldn't be surprised if some Republicans did say that this is such a liberal campus that they would feel worried about being Republican,” Wang said. “There is a dialogue that needs to be established because ultimately, we're not looking for people to support us, we're looking for people to genuinely feel strongly about politics, period.”
Democrats were substantially more united behind their nominee, with 94 percent of students who identified as Democrats saying they would vote for Clinton. No Democrats intended to vote for Trump.
But even most who don’t intend to vote for Clinton expect her to win—94.3 percent said they thought she would be our next president. Just 5.5 percent students expect a Trump victory, and the remaining 0.2 percent—two respondents—predicted an upset victory for Johnson.
Still, the poll data should not be taken at face value, warned David Banks, professor of the practice of statistics.
“The people who decided to respond are not representative of the people who did not respond,” Banks said. “60 percent of the people did not return their forms. Why not? They may be systematically different from the people who did return their forms.”
The demographics of poll respondents did, however, roughly correspond to the demographics of Duke as a whole.
Banks noted that while the response rate was higher than most government surveys, non-respondent voting patterns may be different.
“If people who responded first may have been, for example, pro-Clinton, and people who responded late are more pro-Trump, then you can extrapolate that the people who did not respond might be more pro-Trump than you'd expect,” he said.