The latest Internet-based poll from YouGov shows Thom Tillis leading Kay Hagan by a point in the U.S. Senate race. It confirms what we already know: the race is extremely tight and in all likelihood we will not know the winner until very late on Election Night, possibly not until the early hours of November 7th. But if anyone has an advantage, it’s Tillis: with this newest poll, he now leads by 0.8% in the RCP average. But if the Democratic ground game here is stronger than what polls are picking up, then it is easy to see Hagan pulling off a win.

General election
Tillis – 43%
Hagan – 42%
Haugh – 5%

Tillis leads by 1, even with Haugh included. Haugh is taking 5% of the vote. This still exceeds what Libertarians usually get on Election Day in North Carolina, which is around 3%. At this point in the 2010 Senate race, polls showed the Libertarian candidate, Michael Beitler, with 6% of the vote. In a race this close, the presence of a Libertarian can change everything. It is not necessary for him to outperform Libertarian candidates of the past, though the smaller the Haugh vote, the better for Tillis (from this poll, however, it appears that many conservatives who in the past voiced support for Haugh have either dropped out of the electorate or decided to suck it up and vote for Mr. Tillis, making the remaining Haugh supporters ideologically more typical of third party supporters in the past and less disproportionately Republican.)

Now for a look at the crosstabs.

Men
52% Tillis
37% Hagan
6% Haugh

Tillis leads by 15 with men. In this poll, they are 50% of respondents. The gender gap cuts both ways. Some have asked whether Tillis has a problem with women, because he referred to Ms. Hagan by her first name. But perhaps it is time to ask whether Hagan has a “man problem.” If Tillis wins in November, then he will have won enough women. But looking at it another way, Hagan would have lost too many men.

Women
47% Hagan
34% Tillis
4% Haugh

Hagan leads by 13 with women. It is a shame that there is no further breakdown between single and married women. Single women are likely overwhelmingly Hagan. Married women are probably more competitive and favor Tillis (married white women are probably as pro-Tillis as men overall).

Independents
47% Tillis
32% Hagan
10% Haugh

There is almost no crossover support for either candidate with voters identifying with either the Democratic or Republican Party. But with independents, Tillis leads by 15. This is crucial when independents make up 34% of likely voters. Democrats make up 35%. Republicans, less than 30%. (It seems there are a lot of ‘independents’ who are really Republicans, are at least vote like Republicans, but don’t like the label.)

Liberals
86% Hagan
4% Tillis
4% Haugh

Liberals are overwhelmingly pro-Hagan, though 4% are still not sure. If they are not sure by this point, they might not vote at all. But in all likelihood, these voters will come around for Hagan. The 4% of liberals voting for Tillis is evidence of the need for greater access to mental health care services in this state.

Conservatives
76% Tillis
10% Hagan
4% Haugh

The 10% of conservatives going for Hagan is more easily explained. There are probably some African Americans who identify as conservative but really vote for liberal candidates up and down the ballot. That is my theory, anyway.

Aged 18 to 29
48% Hagan
18% Tillis
15% Haugh

Young people are overwhelmingly pro-Kay. This can be accounted for by the large non-white share of those in this age cohort, who vote for Democrats. Young whites, however, are almost certainly more liberal and Democratic than their elders, but not by much as one might think based off the above numbers.

It is very important to Kay Hagan’s campaign that young people turn out. This necessitates organizing the college campuses. But it is very, very difficult to energize young voters. Barack Obama is the rare candidate who could. In a Senate race, with someone like Hagan who is bland in comparison, that is a much more difficult task. Right now, this poll has young people making up 12% of the electorate. The higher, the better, for Hagan.

Aged 30 to 44
51% Hagan
39% Tillis
4% Haugh

Aged 45 to 64
42% Tillis
41% Hagan
5% Haugh

Aged 65 and older
60% Tillis
33% Hagan
2% Haugh

As age increases, so too does support for Tillis, according to this poll. Tillis’s lead, in fact, is based almost entirely upon his overwhelming lead with the oldest voters, the ones most likely to turn out. They are 27% of likely voters in this poll.

Whites
54% Tillis
32% Hagan
6% Haugh

71% of the electorate in 2012, whites are projected to make up 75% of the electorate this year. They go pretty strongly for Tillis, but it is a far cry from the 68-32 advantage with whites that Mitt Romney enjoyed in 2012. 8% are undecided.

Blacks
78% Hagan
4% Tillis
1% Haugh

23% of the state electorate in 2012, 20% this year. If Hagan can get African Americans to turn out at near-presidential levels, then she will be in a very good position. There will be a drop in turnout from this group; of that it is certain. The Democrats’ goal is to minimize the dropoff. 17% are undecided. That is potential good news for Hagan – if, if these voters turn out. They probably will. In the end, Hagan will almost certainly end up with 90% or more of the black vote, probably 95% or more.

Hispanics
About 1.5% of the likely electorate. The margin of error so large on the preferences of this group, no use in reporting it.

Other
40% Hagan
28% Tillis
13% Haugh

About 3.5% of respondents. Other – Asians, Native Americans, those who don’t feel they fall neatly into any one group. Hagan leads by 12. Large margin of error on the subsample.

Who is undecided?
75% Women
25% Men

46% Independent
29% Democrat
25% Republican

44% Moderate
32% Conservative
24% Liberal

46% aged 45 to 64
27% aged 65+
24% aged 18 to 29
16% aged 30 to 44

59% White
34% Black
7% Other

If we could construct a profile based on the average undecided voter in this race, we would come up with someone female, independent, moderate, middle-aged, and white. Keep that in mind as we get flooded with ads from both candidates and the third party groups supporting them.