The latest ECU Poll of North Carolina’s upcoming U.S. Senate election, conducted August 31 through September 3, shows Democratic nominee Roy Cooper leading Republican nominee Michael Whatley, 46% to 39%, with 14% undecided and the remainder supporting a minor-party or write-in candidate. On a forced-choice measure without an undecided option, Cooper leads Whatley, 52% to 44%, with the remainder supporting a minor-party or write-in candidate. (The poll’s credibility interval, which is similar to a margin of error, is +/- 4.4 percentage points.)
Based on the initial measure, which retains undecided voters, Cooper’s overall lead includes an 11-point advantage among self-described independents, 45% to 34%, with roughly 17% undecided. Cooper also receives support from about one in ten self-described Republicans, while support for Whatley among self-described Democrats stands at just 1%.
Additionally, Cooper leads Whatley among African American voters, 70% to 9%, while Whatley leads among white voters, 47% to 39%. A gender gap is also evident. Whatley holds a slim lead among men, 45% to 42%, while Cooper leads among women by 16 percentage points, 49% to 33%.
Voters also divide by education level and age. Among those without a four-year college degree, Whatley leads Cooper, 45% to 41%. Among those with a four-year college degree or higher, Cooper leads Whatley, 54% to 29%. Whatley leads among likely voters age 65 and older, 48% to 42%, while Cooper leads among those under age 65, 47% to 35%.
Beyond these demographic divisions, the poll indicates that voters have a more clearly defined view of Cooper than of Whatley. Cooper is viewed favorably by 47% and unfavorably by 40%, while 8% hold neither a favorable nor unfavorable opinion and 4% have not heard enough to offer an opinion. Views of Whatley are more evenly divided, with 33% viewing him favorably and 34% unfavorably. At the same time, 19% hold neither a favorable nor unfavorable opinion of Whatley, while 14% have not heard enough to offer an opinion.
When asked to identify the most important issue in deciding their vote in the upcoming election, 43% of likely voters cite the economy, including the cost of living. Other top issues include border security and/or immigration (12%), crime and/or public safety (10%), affordability of health care (6%), and the conflict with Iran and/or other foreign policy issues (6%). Among those who identify the economy and cost of living as their most important issue, Cooper leads Whatley, 58% to 25%.
The current electoral environment in North Carolina coincides with net-negative evaluations of President Donald Trump’s job performance. Among likely voters, 39% approve of his job performance, while 54% disapprove. By comparison, Governor Josh Stein receives a net-positive job approval rating, with 41% approving, 32% disapproving, and 27% unsure.
In assessing the full results of the poll, Dr. Peter Francia, Director of the ECU Center for Survey Research, remarked, “To paraphrase Jim Carville’s famous line, ‘It’s still the economy, stupid.’ Two years ago, the economy was an issue that worked to the Republicans’ advantage in North Carolina. In 2026, it is working to Roy Cooper’s advantage. Among voters who identify the economy and cost of living as their most important issue, Cooper leads Whatley, 58% to 25%.”
Francia added, “The poll also shows that voters have a more clearly defined view of Cooper than of Whatley. Roughly one-third of likely voters either hold neither a favorable nor unfavorable opinion of Whatley or have not heard enough to offer an opinion. By comparison, 12% say the same about Cooper. As the election enters the post-Labor Day period, these results indicate that voters’ impressions of Whatley remain less fully formed.”
Methodology
This ECU Poll was conducted August 31 through September 3, 2026. The results are based on 675 completed responses to the poll’s questionnaire from North Carolina registered voters who are likely to vote in the 2026 general election. Respondents were drawn from a random sample of registered voters statewide and, for those ages 18 to 29, an opt-in panel sample to supplement the data.
Respondents completed the poll in one of three ways: by cell phone via MMS-to-web text (N=392), by landline phone via Interactive Voice Response (N=184), or online through the opt-in panel (N=99). ReconMR, via L2, provided the cell phone and landline samples.
The sampling design intentionally over-sampled populations that are traditionally difficult to reach in public-opinion surveys, including non-white voters and those ages 18 to 29. The credibility interval (CI), which is similar to a margin of sampling error, is +/- 4.4 percentage points. (Please note that credibility intervals are larger for results based on subsamples, including demographic groups and party identification.)
The data were weighted to reflect the population of likely 2026 general election voters in North Carolina based on age, education level, race/ethnicity, gender, region, and party identification, as well as election modeling based on previous statewide elections in North Carolina. All percentages are rounded to the nearest whole number. As a result, percentage totals in some instances may be slightly higher or lower than 100%.
The ECU Center for Survey Research, which directed this poll, is a member of the AAPOR Transparency Initiative (TI). For additional information about the ECU Poll and its methodology, please refer to the documentation provided in the TI report.
SOURCE: ECU Center for Survey Research
