As Donald Trump has inched closer and closer toward the nomination, the possibility that he could take the whole damn Republican Party down with him in November has become even more likely, so much so that archconservatives like Erick Erickson are plotting to drum up support for an “independent conservative” to keep Trump from winning, which would effectively hand the election to the Democratic nominee.

But according to a new poll on North Carolina’s elections, a Trump nomination could have trickle-do effects as well: it seems Senator Richard Burr isn’t in a great position to keep his job if Trump drives the GOP straight over a cliff.

A new poll released by Public Policy Polling indicates that the gap between Burr and his Democratic challenger, former state Representative Deborah Ross, has closed since both handily won their respective primaries on March 15. Burr now has just a five-point lead over Ross, which PPP notes is the same spot Kay Hagan started out at before she defeated Elizabeth Dole in 2008.

Even worse news for Burr: 48 percent of voters say that his support of Trump in November would make them less likely to support him in the general. Burr, you might remember, backtracked on his comment that he would vote for Bernie Sanders over Ted Cruz by throwing his hat in with whomever the GOP decides to nominate in November.

The PPP poll found that Clinton narrowly leads Trump in a head-to-head matchup in North Carolina, putting what’s been a relatively reliable win for Republican nominees into serious play. And given that the last year a Democrat presidential nominee won North Carolina was also, conveniently, the last time a Republican senator here lost a seat, Burr could be in for a very tough night on November 8.

On a somewhat related note, over the weekend, Catawba College political science professor Michael Bitzer tried to gauge the variables, across one hundred counties, that powered Trump’s win in North Carolina. “I tried to answer the question, ‘What is the main attraction support for Donald Trump?’” Bitzer tells the INDY. “There’s a lot of variables out there trying to answer where his support is coming from.” (Trump won seventy-eight North Carolina counties.)

The most widely assumed factors in Trump’s success—racial resentment and generational gaps—didn’t produce much of a statistical relationship to Trump votes. Even January unemployment numbers only produced a moderate relationship. But when Bitzer looked at Trump’s support among voters with a bachelor’s degree or higher, he found a strong negative correlation. In other words: “the higher the education in a county, the lower the Trump vote was,” Bitzer says. “In political-science terms, I can explain almost half of the Trump vote based on that one factor.”

Trump, you’ll recall, loves the poorly educated.