I’m now featured on Politics NC, the website of political consultant and campaign strategist Thomas Mills. This is a great privilege and I’m excited to be a part of it. My posts can be found under the “Carolina Strategic Analysis” section. Check there for new stuff.
The Civitas Institute is out with a new poll, conducted by SurveyUSA, which also did a poll for WRAL last week. The results should be pretty similar. The WRAL poll was conducted from the 17th through the 19th of March, while the Civitas poll was conducted from the 19th to the 23rd. Civitas looked at the Supreme Court race, which we won’t cover here, and also the Republican primary, which we’ll take a look at now. It’s an interesting poll and people should take a look at it; there are some interesting findings. The link is here.
(The numbers in parentheses are the differences from the WRAL poll.)
Undecided – 38% (+15)
Tillis – 27% (-1)
Brannon – 13% (-2)
Harris – 9% (+3)
Grant – 5% (-6)
Snyder – 5% (+1)
Bradshaw – 2% (-2)
Alexander – 1% (-6)
Kryn – 0% (-3)
Notice the huge jump in undecided voters. The Civitas poll probably employed a different likely voter screen. Harris also does better here, but this is probably margin of error change. There are big drops in support for Grant and Alexander.
Tillis is taking 27% of the primary vote here. 62% of voters are decided. 27/62 = 43.5, so if undecided voters break down the same way as their decided counterparts, then Tillis will avoid a runoff. But it’s very close, too close for comfort for establishment Republicans. In the event of a runoff, Brannon looks like the most likely foe.
Notice that it appears that the Civitas poll had the names read in the order they will be on the ballot, with Tillis at the top. This did not cause any significant change, but polls who list the names in random order should probably be viewed more skeptically.
Favorable/Unfavorable (Republican primary voters
(First number is net favorability. Second number is difference in net favorability from the WRAL poll.)
Tillis 37/15 (+22) -1
Brannon 24/4 (+20) +4
Harris 18/4 (+14) No Change
Snyder 10/2 (+8) No Change
Grant 10/3 (+7) -4
Alexander 8/4 (+4) -4
Brannon’s support is 4 points higher, Grant’s and Alexander’s 4 points lower. Again, this is probably statistical noise. (Kryn’s favorability rating was not included, apparently because his name recognition was too low).
One thing I noticed – in the Civitas poll, females are 59% of the Republican primary electorate. That seems way too high.
Ideology: 48% of respondents reported themselves as ‘very conservative’ with 32% ‘somewhat conservative’ and only 16% ‘moderate’. That’s to be expected in the present day Republican Party. Interestingly, Tillis has the smallest lead with moderate voters, but again there’s a small sample size.
58% support the Tea Party. 28% do not, and 14% are not sure. 88% have lived in the state for more than 10 years. This could mean that recent migrants to the state skew Democratic, or are younger and less likely to vote in low turnout primaries.
72% say they are pro-life. 25% are pro-choice. That’s an amazing statistic, and why the GOP can’t get away with running a Todd Akin here. When it comes down to it, even a significant minority of Republicans are uncomfortable with prohibiting abortion. I’m surprised that more did not report themselves as undecided on this issue.
There is a consistent pattern emerging: Tillis is short of the 40% threshold, and it’s a 50/50 proposition on whether or not he can get there on May 6th. Should he fail to clear that 40% threshold, he will enter into a runoff with Greg Brannon. Mark Harris, however, is close on his heels, and at 9% does better here than in any recent poll. Harris has a chance to seize second place with a spirited television campaign.
SurveyUSA released their first poll of the North Carolina Senate race, conducted on behalf of WRAL. Unfortunately, they didn’t poll the general election, but they did poll the primary and the results are very interesting. Let’s take a look:
Tillis – 28%
Undecided – 23%
Brannon – 15%
Grant – 11%
Alexander – 7%
Harris – 6%
Bradshaw – 4%
Snyder – 4%
Kryn – 3%
This poll shows Brannon in second place, but unlike PPP, there is a big gap between the candidates. Tillis leads Brannon by 13 points, with Heather Grant in third. The poll is pretty terrible for Mark Harris. Many observers thought he would be Tillis’s strongest opponent, but here he’s in fifth place, behind Ted Alexander and just barely ahead of a bunch of non-serious candidates.
The path to a Tillis victory is clear: sweep the undecided voters through an onslaught of television ads. Assuming Tillis keeps his 28% support, he needs to win a little more than half of the undecided vote. Of course, if the undecideds break down the way that decided voters are breaking down, then Tillis will only get 36% of the vote, which means a runoff.
Favorable/Unfavorable (primary election voters)
Numbers in italics indicate net favorability)
Tillis 34/11 (+23)
Brannon 27/11 (+16)
Harris 23/9 (+14)
Grant 20/9 (+11)
Alexander 17/9 (+8)
Snyder 16/8 (+8)
Bradshaw 15/8 (+7)
Kryn 13/9 (+4)
SurveyUSA has a different methodology than PPP, resulting in higher favorability ratings. The numbers here roughly correspond to the candidates’ position in the primary. Republicans like everyone, but they like Tillis the most, and he’s also the most well-known of the candidates. This is in clear contrast to the PPP poll, which shows Republicans not liking anyone.
Burr 34/41 (-7)
Obama 39/54 (-15)
McCrory 34/51 (-17)
Hagan 34/54 (-20)
General Assembly 26/57 (-31)
The Burr and Obama numbers look accurate. The poll is a bad one for McCrory, at -17 this is worse than every other poll for him lately. The numbers for Kay Hagan are brutal. If accurate, then just about any Republican will be able to put up a good fight against her. The General Assembly numbers are also pretty bad, but most voters don’t like most legislatures.
Winners in this poll:
-Tillis. He has a 13-point lead on Brannon and is much closer to avoiding a runoff than PPP indicates.
-Brannon. In contrast to the Harris internal poll, he’s in second, and the most likely scenario at this point in time is a runoff between him and Tillis.
-Grant. Third place? And almost no money? It’s unlikely Grant will take off, but this is a good poll for her.
-Harris. Fifth place? That’s a terrible result. Harris has been much more visible than both Grant and Alexander, and he’s lagging behind both of them. Interestingly, he sports better favorable numbers than both of them. Could it be that voters like Harris personally, but just don’t want him as the nominee?
-Obama & Hagan. Senators with a -20 net approval don’t get reelected. Hagan better hope this poll is wrong.
-McCrory. These are his worst numbers since last fall, after the legislative session. It’s possible that his handling of the coal ash spill is hurting his popularity.
Averaging The Polls
There are now three polling organizations who have surveyed the Republican U.S. Senate primary – PPP, American Insight, and now SurveyUSA. Since the American Insight poll is over a month old, let’s leave that aside for now and average the PPP and SurveyUSA results. Doing so yields the following:
Tillis – 21%
Brannon – 15%
Grant – 11%
Alexander – 7%
Harris – 7%
Bradshaw – 5%
Snyder – 4%
Kryn – 2%
What’s incredible is how similar the results are in both polls. The numbers for the minor candidates are almost identical. What’s different is Tillis and Brannon’s performance relative to each other. PPP says we have a tied race, with both candidates at 14%, and Tillis nowhere near hitting the 40% threshold. SurveyUSA says Tillis is ahead by 13 and within striking distance to avoiding a runoff, assuming that the undecided candidates break down in a similar fashion. But both polls show him having more work to do. My guess is that the SurveyUSA poll is closer to reality, but regardless, it’s glad to see other organizations taking a look at the increasingly heated primary here.
It’s been quite a week for polls in NC politics. The first one is from American Insights, which purportedly wants to become the conservative alternative to PPP. They didn’t poll the Senate race, but they show President Obama with a 43/49 approval here, which is better than what other polling organizations have found for him.
They also show Republicans with a 44-38 lead on the generic congressional ballot, but redistricting makes that result pretty much irrelevant. But it does support the idea that 2014 is shaping up to be a GOP year, and might give comfort to Rep. Renee Ellmers in her bid against Clay Aiken. The poll also finds that voters do not want to see President Obama’s agenda “implemented on the state level in North Carolina” (38%-48%). Vulnerable GOP legislators in NC might seek to defend themselves by nationalizing their races.
The next two polls are internal polls. The first is from Clay Aiken’s campaign. It apparently shows Aiken with a 22-point lead, with Toni Morris in second and Keith Crisco lagging behind them both in third place. Nobody has a link to these results, and it’s possible that the poll wasn’t even conducted. Regardless, Crisco being in third place is questionable, even though Aiken could very well win the primary by 22 points. At this point, no one really knows what’s going on in this race. Is Aiken’s campaign ‘releasing’ an internal poll a sign of strength or weakness? We don’t know.
Finally, the poll that’s gotten the most coverage is the one from Mark Harris. It shows the following results for the U.S. Senate Republican primary:
This poll, if accurate, is good news for Tillis. It puts him closer to the 40% threshold than any other survey, the percentage of the vote required to avoid a runoff. However, there are a number of concerns with this poll; it draws from a sample of “very likely” voters and some have said it’s a push poll, intending to push voters into the Harris camp.
The GOP Senate primary is heating up. Both Greg Brannon and Mark Harris are attacking Tillis on ‘character’ issues. Republicans desperately need to avoid an ugly runoff here, but it looks like an ugly runoff is what they’re going to get. Yesterday, the Cook Political Report moved NC-Sen from “Leans Democrat” to “Toss-Up”. Given the developments over the past two weeks, that’s a curious decision.
It’s been 16 months since the November 6, 2012 general election. North Carolina is one of those states that has public voter registration information, which is ongoing as new voters register and other voters either die or move away. Looking at voter registration trends, one can get a glimpse of some of the small changes happening beneath the surface. Over time, they add up to massive change.
Since 2012, Democrats have lost 108,186 registered voters, while Republicans have lost 54,368. At first, this looks like good news for Republicans, as Democrats are losing voters at almost twice the rate. But remember, a lot of these registered Democrats are likely to be older, conservative Democrats, many of whom support Republican candidates, especially for federal office.
Unaffiliated voters, on the other hand, have gained. There are 22,515 more unaffiliated voters than there were on Election Day 2012. When it comes to unaffiliated voters, the most important thing to determine their ‘actual’ partisan affiliation is to look at where they’re registering. Unaffiliateds in Brunswick County, for instance, are far more likely to support Republican candidates. An unaffiliated voter in Durham County, on the other hand, is very likely to be Democrat-leaning. And of course, African American voters who are unaffiliated lean very heavily to Democrats.
In total, North Carolina has 136,000 fewer registered voters than in 2012. Does this mean the state is losing population? Of course not. After an election, certain inactive voters are taken from the voting rolls. And there are a large number of people who relocate to the state who do not register until the year of the next presidential election. This is especially true for Democratic-leaning voters. Overall, those who register as soon as they relocate are more likely to be Republicans. So, don’t despair: these kind of numbers are normal.
Overall, 90 out of 100 counties suffered a net loss in voters. This is, again, unsurprising. But what about the 10 that gained? This is a strong indication that these counties are, in fact, growing enough to offset the typical loss in registered voters. Take a look at the map below. The counties in green are the ten that have seen a net gain in total registration. The ten counties in red have seen the greatest falloff in registered voters.
Democrats should note, with dismay, that the ten counties with a net gain in voter registration tend to be strongly Republican suburban areas. In fact, of the ten, only Hoke County supported President Obama in 2012. The ten bottom, on the other hand, tend to be urban Democratic counties, with Obama winning 7 of the 10 in the last election.
What does this mean? Well, it does suggest that the ten counties in green are growing. But the absence of several urban counties on that list shouldn’t worry Democrats. Remember – Democrats wait until election year to register. And just because most of the green counties supported Romney, does not mean that the new voters are Republicans. They could be hardcore liberals for all we know. Notice, also, the absence of Wake on the “biggest losses” list. With Wake having the most voters, we would expect Wake to be at the bottom. The fact that it isn’t strongly suggests that Wake County is still growing, and faster than other urban counties like Mecklenburg and Durham.
Because looking at raw numbers can only tell us so much, we need to look at these changes in percentage terms:
Counties is dark green saw a net percentage growth in registered voters. Counties in light green saw a percentage decline, but did better than the state average. And counties in red saw a greater percentage loss in registered voters than the state average.
Conclusion: Overall gains in registration don’t tell us much at all, though it does serve to indicate which counties are growing. It doesn’t, however, tend to give us an accurate picture of the reverse. Nor can we tell from these numbers alone what the political implications will be going forward. Next post, we’ll look at the partisan numbers and see how they’ve changed, and what they can reveal about the future of NC politics, if anything.
Let’s be frank: PPP is a Democratic-leaning pollster, and they don’t like Thom Tillis. So take their results with a grain of salt. Personally, I’d like to see another pollster conduct a survey on the Republican primary, because I don’t buy these results for a second. Nonetheless, let’s take a step back and just accept these results as true.
Since last month, Hagan’s situation has improved slightly, but it’s still a very close race against all her opponents. The big news, however, is in the Republican primary, where Thom Tillis is now locked in a tie with Greg Brannon. PPP says Tillis’s comments about Obamacare and his gaffe on the minimum wage hurt him, and they use that to justify his drop in support. The problem, of course, is that nobody except the most energized conservative activists heard of Tillis’s Obamacare statement, and the minimum wage issue is more likely to harm him in the general than in the primary.
(Numbers in parentheses indicate change in margin from last poll
45% Alexander, 43% Hagan (-5)
43% Snyder, 42% Hagan (N/A)
43% Bradshaw, 43% Hagan (N/A)
43% Brannon, 43% Hagan (-3)
43% Hagan, 43% Harris (-2)
43% Hagan, 42% Grant (-3)
43% Hagan, 41% Kryn (-2)
45% Hagan, 43% Tillis (-2)
All of Hagan’s Republican challengers lost ground from last poll. Are North Carolinians slowly forgetting their anger about the ACA? Perhaps.
There’s nothing wrong with these results. But take a look at who does worst against Hagan – Edward Kryn, who is not a serious candidate, and Thom Tillis, who trails Hagan by 2. Now look at Greg Brannon. He ties her. Does this mean Brannon is more ‘electable’ than Tillis? Of course not. If voters know anything about Brannon, it’s that he was found liable in a civil trial a month ago. That said, I don’t think Brannon actually polls better than Tillis in the general election at this point in time. Maybe last month, but not now. Of course, I have no evidence for this, other than a Rasmussen poll that had Brannon doing worse than Tillis, but it’s intuition, and usually my intuition is good (except for that post I made about Brannon’s campaign being over).
Numbers in parentheses indicate change in support from last poll
Undecided – 36% (+2)
Tillis – 14% (-6)
Brannon – 14% (+1)
Grant – 11% (-2)
Alexander – 7% (-3)
Harris – 7% (-1)
Bradshaw – 6% (N/A)
Snyder – 4% (N/A)
Kryn – 1% (-1)
Tillis plummets in support, though instability is expected at this point in the primary. The number of undecideds is actually greater from last month. Every single candidate fell in support. Except, that is, for the guy who was found liable in a trial a couple of weeks ago. I’m not saying the results are inaccurate, but they’re definitely suspicious.
Favorable/Unfavorable (general election voters)
First number in parentheses – net favorability. Second number – change in support from last month)
Harris 11/16 (-5) +3
Grant 10/17 (-7) +4
Brannon 10/19 (-9) -2
Hagan 41/50 (-9) No change
Snyder 8/17 (-9) N/A
Alexander 8/19 (-11) No change
Kryn 6/17 (-11) +2
Bradshaw 7/19 (-12) N/A
Tillis 18/37 (-19) -5
This is a poll of general election voters. North Carolinians don’t like either of their options for Senate, even though they have quite a few of them. They dislike Mark Harris the least, who has a -5 favorability rating. Hagan is underwater at 41/50. Tillis is at -19. His favorability rating was already pretty bad, but it dropped 5 points from last month. I can buy that. There was the minimum wage snafu. What I don’t buy is Brannon’s numbers. Being found liable in a much-publicized civil trial should have cost him much more than 2 points.
Favorable/Unfavorable (primary election voters)
(First number: net favorability. Second number: change from last month.)
Harris 12/15 (-3) +5
Brannon 13/17 (-4) -2
Tillis 24/28 (-4) -2
Bradshaw 7/14 (-7) N/A
Snyder 7/15 (-8) N/A
Grant 8/18 (-10) No change
Alexander 6/17 (-11) No change
Kryn 4/17 (-13) +1
Nothing wrong with these results. Best-liked is Mark Harris, followed by Brannon and Tillis. PPP says Tillis’s Obamacare comments hurt him in the primary, but they didn’t hurt him much here. Why? Because nobody heard about Tillis’s Obamacare comments, except for the most engaged political junkies.
Conclusion: It’s hard to discern anything from a poll released by a questionable firm with a history of supporting Democratic candidates, who have many clients in North Carolina. It’s pretty clear that PPP wants Tillis to have a difficult primary. Conservatives should ask themselves why that is.
At this point, the most likely scenario is for a runoff between Tillis and Brannon. Obviously, there are a lot of people invested in making sure that doesn’t happen. This primary is going to be interesting.
In elections, candidates go where the votes are. In a statewide election, there are clearly areas that favor one party over the other. Where are those areas? And where should a candidate hoping to win in North Carolina try to max out their performance? The following is a list of counties ranked by net number of votes for Republican Mitt Romney. At the top of the list are blood-red counties where Republicans need to do well if they want to win. These are counties like Randolph, Davidson, Union – exurban areas with a lot of people.
At the bottom, there’s urban areas. Republicans mostly can’t win here anymore; their only hope is to cut down some of the huge margins that Democrats have been building up there recently. The most dangerous county for Republicans is Mecklenburg. Obama netted over 100,000 votes out of Mecklenburg County in 2012. You think it might have been a good idea for the Democrats to have their convention there? Combined, even Romney’s vote margin in his three best counties in the state couldn’t match Mecklenburg’s strength. In Randolph, Davidson, Union, and Iredell counties, Romney just barely exceed Obama’s vote margin in Mecklenburg. And even after Mecklenburg, there were still Democratic strongholds remaining such as Wake, Durham, and Guilford. Fortunately for Republicans, there were enough small, red counties that did the job for them, and Romney just narrowly eked out a victory here.
Consider also how few really big Republican counties there are. Romney’s highest margin was in Randolph County, where he netted 30,387 votes. But in five counties, Obama netted more votes than Romney did out of Randolph – each. Those were Orange (32,362 votes), Guilford (41,576 votes), Wake (55,666 votes), Durham (77,455 votes) and Mecklenburg (100,594 votes).
This illustrates the problem for Republican candidates going forward: Democrats are building up huge margins in urban areas, while red counties just aren’t growing as much. It’s possible that a decade from now, Republicans will sweep the vast majority of counties, but will lose because of Democratic strength in the I-85 corridor. There are only two solutions to this problem: (1) hope that the redder counties get bigger and stronger, or (2) become more competitive in urban areas. Republicans don’t have to win big urban counties like Wake, but they can’t afford to get blown out, especially as these areas continue to grow in size and importance.
Notice also the lack of importance of the heavily African American counties in the Northeastern part of the state. On a map, they’re very blue, but their population is relatively tiny compared to the huge urban counties. However, Democrats still need to do well here, because statewide wins, especially in federal races, leave little room for error.