In terms of the 2.8 million North Carolinians who cast accepted in-person absentee ballots and were able to match their 2012 registration status (whether they were registered before or after 2012) and, if they were registered, did they vote and what voting method (mail-in, in-person early absentee, on election day) did they use.
For the overall state-wide numbers:
So, among the 2016 in-person early voters in NC, 57 percent of them used the same method, with another 2 percent voting early through mail-in ballots.
There was 19 percent of this year's in-person early voters who changed their voting method from waiting until election day 2012 and decided to cast an early in-person ballot this year, with another 22 percent having not voted in 2012 (but were registered) combined with those voters who registered after 2012's election.
In response to a reporter's query, I broke these voters into their party registration (registered Democrat, registered Republican, and registered unaffiliated (and yes, there are Libertarians, but their numbers are small)) and ran the same analysis:
Among the partisans, we see that majorities of this year's registered Democrats and Republicans who voted early in-person used the same method in 2012. However, there is one noticeable difference: registered Republicans had 21 percent shift from voting on election day four years ago to casting in-person absentee ballots this year, while 17 percent of registered Democrats did the same thing (cast ballots on election day in 2012 but cast early in-person this year).
Among this year's early unaffiliated voters, 19 percent (matching the state-wide percentage) voted on election day in 2012. So Republicans have drawn more of their voters into the early voting rather than voting on election day than did unaffiliated or Democrats this year.
But another interesting dynamic is present among registered unaffiliated voters, who I consider the wild card in this year's election: nearly one-third (32 percent) of registered unaffiliated voters who have cast in-person absentee ballots this year were either registered in 2012 but didn't vote (6 percent) or registered after the 2012 election year (26 percent). These 'new voters' are higher than the same kind of partisan registered voters (registered Democrats at 18 percent combined while 20 percent of registered Republicans combined).
So, among the 2.8 million early voters who cast accepted in-person absentee ballots, 238,430 registered unaffiliated voters, 224,445 registered Democrats, and 174,779 registered Republicans could be considered 'new' voters in this year's general election.