NC early voting tops 1.2 million as Republicans narrow their 2022 gap

October 23, 2026

North Carolina has reported 1,269,919 accepted early and mail ballots after eight calendar days of in-person voting, 37.3% ahead of the same stage in 2022.

The lead has held beyond opening weekend. The bigger change since Tuesday is within the electorate: Republican participation is approaching its 2022 pace, while Democrats and unaffiliated voters remain well above theirs.

In-person voting keeps the midterm lead intact#

1,215,031 ballots came from early voting sites, up 44.1% from 842,986 at the equivalent 2022 cutoff. Accepted mail ballots remain 33.1% lower, at 54,888 versus 82,090.

The statewide count has grown by 536,620 since our October 20 report. In-person voting accounts for 525,898 of those net additions.

Registration growth does not explain the whole increase. Accepted ballots equal 16.10% of registered voters, versus 12.48% in 2022. Applying the old participation rate to today's registration total would produce about 984,350 ballots. The reported count is 29.0% above that benchmark.

Republicans are closing their 2022 participation gap#

The Republican count is now 296,478, up 2.9% from the matching 2022 count. But Republican registration has also grown, so participation remains slightly below its old pace.

The rates within each registration group are:

Democrats: 22.95%, versus 14.59% in 2022. Up 8.36 percentage points.
Unaffiliated: 13.71%, versus 10.24%. Up 3.46 points.
Republicans: 12.72%, versus 12.96%. Down 0.24 points.

In relative terms, those participation-rate changes are +57.3% for Democrats, +33.8% for unaffiliated voters, and −1.8% for Republicans. The Republican deficit was 9.3% in Tuesday's comparison.

Accepted early and mail ballots divided by each affiliation's registered voters at the matching eight-day point in 2022 and 2026

The latest additions are less Democratic than the opening count. Republicans account for 26.9% of net additions since Tuesday, compared with 20.7% of the earlier total. Their cumulative share has risen to 23.3%. The Democratic share has eased from 44.7% to 42.2%, while unaffiliated voters remain near 34.2%.

Democrats still added more ballots: 208,067, compared with 182,647 unaffiliated and 144,533 Republican. Their lead over Republicans in reported ballots grew from 175,960 to 239,494.

Unaffiliated ballot counts are 60.7% above 2022, but their participation rate is up 33.8%. The difference reflects a registration base that has grown 20.1%.

The presidential-year drop-off remains uneven#

Statewide participation has reached 54.7% of its rate at this point in 2024: 16.10% now, compared with 29.44% during that presidential election.

The midterm drop-off is much smaller among Democratic registrants:

Democrats are at 72.5% of their 2024 participation rate: 22.95% now, versus 31.66% then.
Unaffiliated voters are at 56.0%: 13.71% versus 24.48%.
Republicans are at 37.5%: 12.72% versus 33.89%.

Party participation at the equivalent eight-day point in the 2024 presidential election and 2026 midterm

Republicans had the highest participation rate of the three groups at this stage in 2024. They currently have the lowest. Their share of accepted ballots is 11.0 percentage points below 2024, while the Democratic share is 8.6 points higher.

The increase now reaches 97 counties#

97 of 100 counties have more accepted ballots than at the matching 2022 cutoff, up from 94 in Tuesday's comparison. Alexander, Alleghany, and Columbus remain below their earlier counts.

Wake and Mecklenburg together account for 99,895 additional ballots, or 29.0% of the statewide net gain over 2022. Their counts are up 54.7% and 53.7%, respectively.

The six counties with the largest increases in accepted early and mail ballots compared with the matching 2022 report

Black participation remains well above its 2022 pace#

Black registrants account for 279,688 accepted ballots. Their participation rate is 18.42%, up from 11.35% in 2022, a gain of 7.07 percentage points. White participation is 17.11%, up from 14.23%, a gain of 2.88 points.

Accepted early and mail ballots as a percentage of registered Black and White voters at the matching 2022 and 2026 cutoffs

Against the presidential cycle, Black participation stands at 69.9% of its 2024 rate, compared with 51.8% for White registrants. Black voters supply 22.0% of reported ballots, versus 17.6% at this point in 2024.

Women account for 52.9% of accepted ballots. Their participation rate is 17.32%, compared with 16.19% for men. In the matching 2022 report, those rates were nearly equal at 12.95% and 13.01%.

What we'll be looking for over the weekend#

Whether Republican participation moves above its 2022 rate. The gap has narrowed substantially since Tuesday.
Whether the party mix continues to change. Watch each group's participation rate alongside its share of the growing total.
Whether the presidential-year gaps narrow. The 2024 comparison still shows very different levels of early participation across affiliations.
Whether the statewide lead holds as voting expands. County schedules and reporting changes can affect the next few daily counts.

Sources and notes#

Published October 23, 2026, using NCSBE data reported through October 22.

The comparison windows match. Each covers the first eight calendar days of in-person early voting, Thursday through the following Thursday. All three cutoffs are 12 days before Election Day. The reports were published the next morning.

Election	In-person voting opened	Data through	Report date
2022 midterm	October 20	October 27	October 28
2024 presidential	October 17	October 24	October 25
2026 midterm	October 15	October 22	October 23

Current counts and registration: NCSBE's October 23, 2026 statistics report and official accepted-count CSV.
Historical counts and registration: the dated October 28, 2022 report and October 25, 2024 report.
Reader tables: participation and ballot-share calculations, daily counts and dated source reports, all 100 county comparisons, and net additions since Tuesday.
Definitions: NCSBE's absentee and early voting data guide.
Earlier installments: October 20 analysis and October 12 analysis.

Participation means accepted in-person and mail ballots divided by that year's reported registration count for the same group. Percentage-point changes use unrounded rates. Relative changes divide the new rate by the historical rate, then subtract one. Ratios against 2024 describe aggregate participation rates, not the share of individual 2024 voters who returned. Registration and affiliations change between elections.

Counts include accepted civilian, military, and overseas mail ballots, including mail accepted before early voting opened. Eight calendar days does not mean every site was open on both weekend days. Site hours, mail-distribution windows, contests, and voting-method preferences differ across years. Civilian mail voting began later in the 2024 cycle, as its report notes. The historical benchmark adjusts for registration only, not voting opportunities, and is not a forecast of final turnout.

Source reconciliation: the October 23 PDF and official aggregate CSV agree on total accepted ballots, party, method, and gender counts. The PDF's 100 county totals also sum to the statewide total. The separately updated voter-level export contains 1,269,920 accepted records, one more than those aggregates, and five repeated accepted voter identifiers. These are export inconsistencies, not evidence that a person voted twice. This edition uses the official aggregates and makes no age or individual-voter analysis from that export.

The current report's retrospective panel revises the matching 2022 total to 927,708, versus 925,076 in the dated 2022 report used here. Using the revised figure would put the lead at 36.9% instead of 37.3%. Daily curves consistently use each dated report's main totals. Some historical archive files were modified after their printed dates and may contain corrections.

Net additions subtract the October 20 report from the October 23 report. They can include processing, corrections, and changes to registration records, so they are not a verified cohort of people who voted on those three days. County comparisons are raw counts without county-registration or site-hours adjustments. Party registration does not reveal candidate choice, and lower early participation does not establish that someone will sit out the election.

The party charts show the three largest affiliations; another 2,931 ballots belong to Libertarian and Green registrants. The racial chart shows Black and White categories; other and undesignated records remain in the overall totals. Gender labels follow the official report, which lists another 62,733 ballots as undesignated.