2026 North Carolina General Election · Model

What North Carolina’s early vote says about Nov 3

Who has voted so far, who is still likely to, and what that means for the US Senate and Supreme Court races and the General Assembly. A model built from the State Board of Elections’ voter file, voter history, absentee files and precinct results, updated each morning.

Ballots through Oct 27, 7 days before Election Day.

The short version

• The early vote shows who is voting, not how they vote. Republicans are 25% of early ballots, against 31% at this point in 2022.
• The big question: are Republicans voting later, or voting less? Their most reliable voters are on pace; their occasional voters aren’t. In the two past pairs of elections I could check, gaps like this matched real turnout differences, so I lean toward “voting less,” two to one.
• The General Assembly is harder to flip than the statewide vote. On the 2023 maps, Republicans keep both majorities in 67 of 100 simulations, but three-fifths of both chambers, enough to override a veto, in only 15. The General Assembly
• The headline numbers lean on the polls. The early vote adds what polls can’t: the range of possible electorates and a benchmark for Election Night.

The forecast

My average of three poll averages has Cooper up about 10.7. In nine statewide races since 2016, Republicans beat their final North Carolina polling average by 2.4 points on average, so I shade the polls 2.4 points toward Whatley and allow for polling error with a standard deviation of 5 points. For the Supreme Court I start from the average of the last five public polls (fielded Sept 4 to Oct 20), Earls +7.5, with the same 2.4-point shading toward Stevens and the same polling error, a standard deviation of 5 points. Then part of the early vote’s signal passes through. The court race is closer, and probably less certain than the model shows: it has less polling, more undecided voters, and neither candidate is as well known as Cooper.

Why I shade the polls, and how much it matters

Across nine statewide races since 2016, Republicans beat their final North Carolina polling average by 2.4 points on average.

Race	Final poll average	Result	Republican beat polls by
2016 president	Trump +0.8 to +1.0	Trump +3.7	+2.7 to +2.9
2016 Senate	Burr +2.0	Burr +5.7	+3.7
2016 governor	Cooper +2.2	Cooper +0.2	+2.0
2020 president	Biden +0.6	Trump +1.3	+1.9
2020 Senate	Cunningham +2.7	Tillis +1.8	+4.5
2020 governor	Cooper +11.0	Cooper +4.5	+6.5
2022 Senate	Budd +5.1	Budd +3.2	−1.9
2024 president	Trump +1.2	Trump +3.2	+2.0
2024 governor	Stein +14.7	Stein +14.8	−0.1

Final averages from RealClearPolitics, 270toWin, FiveThirtyEight and other aggregators; for 2016 president and Senate and 2020 governor, RealClearPolitics’ final average (2016 president: head to head, and with Gary Johnson).

The polling misses cluster in presidential years with Trump on the ballot. The one midterm, 2022, went the other way, and Stein’s 2024 race, won by a lopsided margin, was polled almost exactly right. But Cooper’s own 2020 race, polled at about +11 like this one, had the biggest miss. So I shade the polls by 2.4 points, about the average miss. How much each choice matters:

Choice	Senate median	Court median
This forecast: 2.4-pt shading, “voting less” two to one	Cooper +9	Earls +6
1-pt shading	Cooper +10	Earls +7
No shading	Cooper +11	Earls +8
Both turnout readings weighted equally	Cooper +8	Earls +5

The General Assembly

The NC House has 71 Republicans, 47 Democrats and 2 unaffiliated members; the NC Senate, 30 Republicans and 20 Democrats. All 170 seats are on the ballot, on the maps the General Assembly drew in 2023. To see what my forecast means for them, I read the statewide party vote from the Supreme Court race, a down-ballot race decided more by party than by candidate: about D +5.7, with an 80% range of R +2.2 to D +13.4.

The maps hold up Republican majorities well past a tie. Republicans’ expected seats fall below a majority only when the statewide party margin passes about D +8.4 in the House and D +10.0 in the Senate, and Democrats don’t expect a majority of their own until about D +9.7 and D +13.9. A 25–25 Senate would leave tie votes to the lieutenant governor, a Democrat.

The veto is the long shot. Gov. Josh Stein’s vetoes stand unless three-fifths of those present and voting in each chamber override them. Republicans keep three-fifths of the Senate in 28 of 100 simulations and win it in the House in 15; both, enough to override on their own, in 15.

Seats at each statewide margin

Statewide party margin	Republican House seats, of 120 (80% range)	Republican Senate seats, of 50 (80% range)
R +4	73 72 to 75	31 30 to 33
Even	71 68 to 73	30 29 to 31
D +4	67 64 to 70	29 27 to 30
D +6	65 61 to 68	28 26 to 29
D +8	62 58 to 65	27 25 to 28
D +10	58 54 to 62	26 24 to 27
D +12	55 51 to 59	25 23 to 27
D +14	52 48 to 56	24 22 to 26

The closest seats

The contested seats closest to even in the forecast. Lean: how much more Democratic or Republican the district voted than the state in 2022’s three statewide races. Incumbent: the 2024 winner, running again; an appointed incumbent, named since 2024, gets no incumbent’s edge in the model.

District	Lean	Wins, of 100 simulations
House 105McCool (D) vs Cotham (R) · won by a Republican in 2024 (incumbent)	D +0.2	72 D · 28 R
House 35Hopkins (D) vs Schietzelt (R) · won by a Republican in 2024 (incumbent)	R +0.4	70 D · 30 R
House 25Wilkins (D) vs Chesser (R) · won by a Republican in 2024 (incumbent)	R +0.9	67 D · 33 R
House 37Decker (D) vs Pare (R) · won by a Republican in 2024 (incumbent)	R +1.1	66 D · 34 R
House 5Davis (D) vs Ward (R) · won by a Republican in 2024 (incumbent)	R +1.9	62 D · 38 R
House 20Merrick (D) vs Scalise (R) · won by a Republican in 2024	R +6.8	42 D · 58 R
House 74North (D) vs Zenger (R) · won by a Republican in 2024 (incumbent)	R +6.5	38 D · 62 R
House 47Hooper (D) vs Lowery (R) · won by a Republican in 2024 (appointed incumbent)	R +7.6	37 D · 63 R
House 63Baltutis (D) vs Moffitt (R) · won by a Republican in 2024	R +7.6	37 D · 63 R
House 7Speed (D) vs Thornton (R) · won by a Republican in 2024	R +8.5	33 D · 67 R
House 62Headen (D) vs Blust (R) · won by a Republican in 2024 (incumbent)	R +7.6	32 D · 68 R
House 119Burrows (D) vs Ferguson (R) · won by a Republican in 2024 (appointed incumbent)	R +9.6	27 D · 73 R

District	Lean	Wins, of 100 simulations
Senate 18Fatmi (D) vs Stock (R) · won by a Democrat in 2024 (appointed incumbent)	D +1.8	81 D · 19 R
Senate 11Gailliard (D) vs Barnes (R) · won by a Republican in 2024 (incumbent)	R +2.0	63 D · 37 R
Senate 7Bichler (D) vs Lee (R) · won by a Republican in 2024 (incumbent)	R +2.5	61 D · 39 R
Senate 24Hamm (D) vs Britt (R) · won by a Republican in 2024 (incumbent)	R +4.9	47 D · 53 R
Senate 4Rivera (D) vs Newton (R) · won by a Republican in 2024 (incumbent)	R +8.8	28 D · 72 R
Senate 34Cook (D) vs Crutchfield (R) · won by a Republican in 2024	R +10.3	21 D · 79 R
Senate 1Zehner (D) vs Tillett (R) · won by a Republican in 2024 (appointed incumbent)	R +10.8	19 D · 81 R
Senate 2Surrett (D) vs Sanderson (R) · won by a Republican in 2024 (incumbent)	R +14.7	8 D · 92 R

How I built it

Each district’s lean is the median, over 2022’s three statewide races (US Senate and two Supreme Court seats), of its margin less the state’s, from the General Assembly’s statistical reports on the 2023 maps. Tested on the 2024 legislative races, on the same maps, against 2024’s median statewide race, those leans missed the contested seats by about 3.6 points in the House and 3.3 in the Senate; leans from the 2020 races too missed by more. In 2024 the legislative races ran about half a point (House) and a point (Senate) more Republican than the median statewide race, and an incumbent ran about a point ahead in the House, with no clear edge in the Senate. I use those, and widen each seat’s own error for the four years since 2022.

Each simulation draws a statewide party margin from the Supreme Court forecast, with a typical error of 2 points (one standard deviation) for the gap between that race and the parties’ vote, one shift for the legislative races, and each seat’s own error. The chance of each outcome is its share of 100,000 simulations.

Election Night

North Carolina counties post mail ballots first when polls close at 7:30 p.m., then in-person early votes, usually within an hour or so, ahead of most Election Day precincts. Election Day voters lean more Republican, so Cooper’s margin should shrink as the night goes on. In every scenario, the final Senate margin comes out about 11 to 16 points more Republican than the count of all early and mail ballots, once every county has posted them (not the first batch posted, which is mostly mail ballots):

In the last midterm, all early and mail ballots counted showed Beasley ahead by 9, and Budd won by 3, a drop of about 12 points.

When does the Republican wall fall? The US House map

Each district’s lean is how much more Republican or Democratic it votes than the state as a whole: the median over 14 statewide races on the new lines, the 2022 Senate race and two Supreme Court races and the 2024 races for president, Supreme Court and the rest of the Council of State. I leave out the 2024 governor’s race, whose Republican candidate ran far behind the rest of his party. A district falls when the statewide Democratic margin passes its lean, if every district moves the same way. I grade each Republican-leaning district by how big a wave it takes: first to fall below D +10, needs a wave from D +10 to D +14, and only in a landslide beyond that.

In a tied statewide vote, the map gives Republicans 11 of the 14 seats. The first falls at D +5.9, and Democrats need about D +12 statewide to win a majority of the delegation. My forecast has Cooper up 8.9; a Democratic margin that size across the House races would take 5 seats. But House Democrats, mostly challengers, are unlikely to run as strong as Cooper, and incumbents usually run a few points ahead of their party, so the real wall is likely higher than this, except in the 1st, where the incumbent is a Democrat.

District	Grade	Falls at	Falls at, range over the races
US House 12Adams (D) vs Codiga (R) · held by a Democrat	Safely Democratic	—	Leans D in every race
US House 4Foushee (D) vs Ganorkar (R) · held by a Democrat	Safely Democratic	—	Leans D in every race
US House 2Ross (D) vs Douglass (R) · held by a Democrat	Safely Democratic	—	Leans D in every race
US House 11Ager (D) vs Balkcom (R) · held by a Republican	First to fall	D +5.9	D +4 to +9
US House 1Davis (D) vs Buckhout (R) · held by a Democrat, redrawn in 2025 to lean Republican	First to fall	D +8.1	D +6 to +10
US House 3Smith (D) vs Murphy (R) · held by a Republican	First to fall	D +9.0	D +8 to +12
US House 7Hardy (D) vs Rouzer (R) · held by a Republican	Needs a wave	D +10.4	D +8 to +12
US House 14Womack (D) vs Moore (R) · held by a Republican	Needs a wave	D +11.9	D +9 to +14
US House 9Ojeda (D) vs Hudson (R) · held by a Republican	Needs a wave	D +12.8	D +10 to +15
US House 13Barringer (D) vs Knott (R) · held by a Republican	Needs a wave	D +13.0	D +10 to +15
US House 6Jefferson (D) vs McDowell (R) · held by a Republican	Only in a landslide	D +14.2	D +12 to +17
US House 5Hubbard (D) vs Foxx (R) · held by a Republican	Only in a landslide	D +14.4	D +12 to +16
US House 10Bell (D) vs Harrigan (R) · held by a Republican	Only in a landslide	D +15.2	D +14 to +17
US House 8Watson (D) vs Harris (R) · held by a Republican	Only in a landslide	D +16.4	D +15 to +18

Statewide Democratic margin	Democratic seats, of 14
Even	3
D +4	3
D +6	4
D +8	4
D +10	6
D +12	8
D +14	10
D +16	13

What 15 counties should show

The margin each county should show for each statewide Senate result, calibrated to how each county actually voted in 2022.

County	2022 Senate	Statewide tie	Cooper +5	Cooper +10
Wake	D +26.6	D +31.5	D +36.7	D +41.6
Mecklenburg	D +32.5	D +39.3	D +43.6	D +47.7
Guilford	D +18.2	D +24.6	D +28.9	D +33.2
Forsyth	D +11.3	D +17.8	D +22.6	D +27.4
Durham	D +63.6	D +62.8	D +66.4	D +69.8
Buncombe	D +25.5	D +27.2	D +33.4	D +39.2
Cumberland	D +12.8	D +17.4	D +20.8	D +24.3
New Hanover	D +1.9	D +5.5	D +11.7	D +17.6
Pitt	D +2.8	D +7.9	D +12.5	D +17.3
Nash	R +6.7	R +1.2	D +1.9	D +5.3
Cabarrus	R +12.0	R +4.7	D +0.5	D +5.8
Johnston	R +23.8	R +19.8	R +14.6	R +9.2
Union	R +28.3	R +21.6	R +16.0	R +10.3
Brunswick	R +24.9	R +27.3	R +20.7	R +14.0
Onslow	R +35.4	R +31.0	R +26.0	R +20.7

What the early vote shows

Through Oct 27, 1.96 million ballots had been accepted, against 1.43 million at the same point in 2022. Early voters this year are also younger and more racially diverse than 2022’s.

The question is whether Republicans are voting later or voting less. People who voted in both the last midterm and the last presidential race turn out at very high rates in any climate, so a change in how fast they vote early mostly reflects a change in how people vote. Occasional voters are where turnout itself moves.

Reliable Republicans are voting early slightly ahead of their 2022 pace, so as a group Republicans haven’t shifted toward Election Day. Occasional Republicans are 14% behind their 2022 pace. Taken at face value, that’s a turnout gap. Democrats show the opposite: their occasional voters are up even more than their reliable ones.

The full table, with unaffiliated and new voters

Share who had voted, 7 days out	Reliable		Occasional		New voters
	2022	2026	2022	2026	2022	2026
Democrats	36.9%	54.7%	10.5%	21.3%	17.8%	27.3%
Republicans	32.3%	33.9%	10.6%	9.2%	12.7%	11.3%
Unaffiliated	32.1%	42.2%	9.1%	14.6%	9.5%	11.6%

Reliable: voted in both the 2022 midterm and the 2024 presidential race (for 2022, the 2018 and 2020 races). Occasional: voted in one of the two. New voters: registered since the last presidential election, and didn’t vote in it, the midterm before or the primaries between; those who registered after the registration deadline, only if they had no earlier vote on record in North Carolina since 2016. Party is each voter’s registration as of Oct 25, 2026.

Two ways to read it

• They’re voting later. Turnout moves the same way for every party; Republicans are just waiting for Election Day. With 2022’s patterns, Republicans end up 33.7% of voters and Democrats 30.6%, much like 2022.
• They’re voting less. Each party’s turnout moves on its own, as the chart above suggests. With 2022’s patterns, Democrats end up 33.7% and Republicans 29.5%.

If Republicans are voting later, the polls, shaded 2.4 points toward Whatley, require voters to have moved 5 to 11 points toward Cooper since 2022. If they’re voting less, 1 to 2 points. Either is plausible: in 2024, on nearly the same electorate, Stein won the governor’s race by 18 points more than Harris lost the presidential race (Stein +14.8, Trump +3.2).

Which is more likely? In both past pairs of elections I could check, the same comparison of reliable and occasional voters pointed the right way, and the gap between the parties this year is far larger than in either of those elections. But a turnout model that lets each party move on its own overstated Democrats by about 6 points when tested on 2024. So I weight “voting less” two to one, not all the way.

What to watch before Nov 3

1. Occasional Republicans. 9.2% had voted by 7 days out, against 10.6% in 2022. If they close the gap before early voting ends Sat, Oct 31, “voting later” gains. If the gap holds or widens, “voting less” gains.
2. Reliable Republicans. 33.9% had voted by 7 days out, against 32.3% in 2022. A drop-off would be the first sign of a turnout problem among Republicans’ most dependable voters.
3. Unaffiliated voters who usually take a Republican primary ballot. Whether they show up is the closest the voter files come to a persuasion signal. The early vote page follows them.

How the model works

1. Who will vote. Every registered voter falls into a group by party, race, age, past voting and when they registered. Each group’s turnout and early-vote pace in a past midterm, fitted to this year’s early vote, gives each group’s chance that a voter who hasn’t voted yet still will.
2. How groups vote. Voter history and precinct results show who voted in each precinct and how the precinct voted. A statistical model recovers how each group voted. It matches 2022 precinct results closely (R² 0.98) and county results within about 2 points on average, though a few small counties miss by much more.
3. Putting it together. Projected voters times group preferences gives the margin “if every group voted as it did in 2022.” The gap between that and the final result is mostly voters changing their minds, which only polls can measure, plus the model’s own error.

Limits

• Party registration isn’t a vote, and group preferences are estimated from precinct results, not individual ballots.
• Voters changing their minds is measured only by polls. If the polls miss the way they did in 2016 and 2020, this forecast misses with them; the shading covers only part of that.
• Voters added to the rolls after Oct 25, mostly same-day registrants at early-voting sites, aren’t included yet.
• About 146,000 voters who last voted in the 2018 general election were later removed from the rolls, about 67,000 of them as deceased. They aren’t counted among 2022’s occasional voters, because the voter file doesn’t say when they were removed. Those still registered in 2022 would make occasional voters in both parties compare better with 2022 than shown here, and the Republican gap smaller.
• Run on 2022’s early vote nine days out, the model put the drop from the early and mail count to the final at about 10 points; it was 12. The drop this year may run larger than the 11 to 16 points under Election Night.
• The General Assembly forecast reads the party vote from one race, the Supreme Court’s. Earls is the incumbent, and the court race with a Democratic incumbent ran almost 2 points bluer than the median statewide race in 2024, and about half a point in 2022. Seats also tend to swing together by region, which the simulations don’t capture, so the real ranges are probably a little wider.
• The voter file has no dates of death, so how many of 2022’s nonvoters later removed as deceased were still alive then is an estimate.

Data: NC State Board of Elections voter file and voter history (Oct 25), absentee files for 2018, 2020, 2022, 2024 and 2026, and precinct results for 2022 and 2024; the 2026 ballot. NC General Assembly statistical reports for the 2023 House and Senate maps. Polls: RealClearPolling, 270toWin and Poliwave (Senate); Harper Polling for Carolina Journal, Change Research, High Point University and Catawba College/YouGov (Supreme Court). Only group counts are used; no individual voter appears here.