The poll also shows Clinton has a significant lead among black voters, with 98 percent indicating they plan on voting for her in a two-party vote.
However, Clinton's lead could change in the swing state, Husser said.
He said although polls tend to be very accurate, the public should interpret polls as estimates.
“People should take the margin of error into account and then realize that there are probably a few extra points of error here and that they are just due to the realities of surveys," he said.
The target sample of the Elon University Poll was registered North Carolina voters. Husser said polling registered voters allows the results to be more representative.
“When (registered voters) tell us about their likelihood to vote we will get a more credible response than just using random digit dial which we can’t verify if the person is actually registered to vote," he said.
Using registered voters' phone numbers to conduct polls has its pros and cons, said Thomas Carsey, political science professor at UNC.
“What can happen sometimes by interviewing registered voters, is that all registered voters don’t in fact show up to vote on Election Day," he said.
Carsey said survey-based polling can provide an accurate representation of the North Carolina voting population.
"People who actually answer the survey are folks that are more interested in politics and are thus pretty likely to vote," he said.
N.C. political analyst John Davis said the most important thing about polling is to follow the trends.
“The momentum in North Carolina is not fixed," he said. "One week Trump is up, one week Clinton is up — next week you have no idea who is going to have the momentum.”
Husser said polls can provide a snapshot of a particular moment in time, but with this election’s unpredictability it is difficult to assess who will win at the end.
“Things could change between now and the election," he said. "We think our results are accurate on how things were between Sept. 27th and Sept. 30th, but that’s a long way from November.”