Cross Timbers Poll Nonpartisan public polling
CTP 2026-03 Final results · Day 1 of 3 Day 3 is out: the Inola smelter and voters in their own words → Released Thursday, Oct. 29, 2026 · Governor and statewide races

Mazzei trails his own ticket; estimate shows a race within 3 points, alternative weight an 18-point Munson lead

Corrected Oct. 30: the original headline also said Drummond voters break for Munson. That holds under the alternative weight but not under the estimate once respondents who could not have voted in the closed Republican runoff are removed. See the corrected section.

Cross Timbers Poll's third survey of the 2026 election, and the first drawn from the voter file. Today: governor and statewide races. State direction, issues and full crosstabs follow Friday; the Inola smelter and open-ended responses Monday.

How to read this poll. Our estimate for governor is Mazzei 46%, Munson 43% among likely voters (±5.7). We also published an alternative weight to show how much the result depends on one choice: correcting respondents’ recalled 2024 vote. It is a sensitivity check, not a second forecast. As far as we know, Cross Timbers has published the only public polls of this general election, and we would welcome others. Every response is in our public data file, and after Nov. 3 we will publish our error against the certified results.
Field dates
Oct. 24–28, 2026
Population
Oklahoma registered voters, voter-file sample
Sample
596 statewide / 1,019 all samples
Mode
Text-to-web via RumbleUp, from the L2 voter file
Election day
Tue., Nov. 3, 2026

Results · Governor

A race within 3 points under the estimate; a Munson lead under the alternative

Republican Mike Mazzei leads Democrat Cyndi Munson 46% to 43% among likely voters in the race for governor, with 9% undecided, in the primary estimate. The margin of error is plus or minus 5.7 points. With undecided voters who lean toward a candidate included, Mazzei leads 47% to 44%.

Under the alternative weight, which adjusts to party registration without recalled 2024 vote, Munson leads 54% to 36% (plus or minus 5.2 points; corrected Oct. 30 from 55%, a rounding error). Both are published. The first is the estimate, for reasons set out in the methods note. The two differ in one respect: whether respondents' recalled 2024 presidential vote is corrected to the state's actual result.

Estimate
Weighted to party registration, age, gender and region, and to recalled 2024 presidential vote among voters the file confirms voted in 2024.
Alternative
Weighted to the same party registration, age, gender and region targets, without recalled 2024 vote.
Governor, likely votersQ3 · weighted % · n=575
CandidateEstimateAlternative
Mike Mazzei Republican45.736.3
Cyndi Munson Democrat43.254.5
Independents (Brooks, Bush, Griffin) Independent2.01.6
Undecided9.17.6
Margin of error±5.7±5.2
With undecided voters' leaningsQ3 + Q3a · LV %
EstimateAlternative
Mazzei47.337.1
Munson43.954.9
Leans to an independent1.61.8
Still undecided5.34.6
Margin ±5.7 / ±5.2
Governor, registered votersQ3 · RV % · n=596
EstimateAlternative
Mazzei44.335.7
Munson43.854.4
Undecided9.78.0
Margin ±5.6 / ±5.1

Want more public polling in Oklahoma? Join the mailing list to hear when Day 2 and Day 3 are out.

Why two weights

Respondents recalled a far closer 2024 race than Oklahoma cast

Respondents who answered this survey reported a 2024 presidential vote much closer than the one Oklahoma cast. The estimate corrects that; the alternative does not.

Recalled 2024 vote, likely voters who reported oneTrumpHarrisOther
Unweighted46.749.73.6
Alternative47.146.86.1
Estimate65.033.11.9
Actual 2024 result66.231.91.9

Party registration alone does not correct the gap: under the alternative, which matches party registration, recalled 2024 vote stays close to even. The methods note explains why the recall-weighted figure is published as the estimate.

Results · Ticket

Mazzei runs behind his ticket

Under the estimate, Mazzei's 46% trails every other Republican running statewide: Kevin Hern for U.S. Senate at 49%, Attorney General candidate Echols at 50%, Corporation Commission candidate Boles at 55% and Superintendent candidate Franklin at 63%. The ordering is the same under the alternative weight, where Mazzei is again last at 36%.

Republican share, likely voters %EstimateAlternative
Mazzei Governor45.736.3
Hern U.S. Senate49.039.1
Echols Attorney General50.140.0
Boles Corporation Commission55.043.8
Franklin Superintendent63.458.7

Results · Runoff voters

Drummond's Republican voters: the answer depends on the weight

Corrected Oct. 30. As first published, this section was headed “Drummond voters break to Munson” and reported Munson 50%, Mazzei 38% under the estimate among all 181 respondents who said they voted for Drummond. That group included 61 registered Democrats and independents, who could not have voted in the closed Republican runoff; the voter file shows none of them changed party in the past year. The table now counts registered Republicans only. Details.

Among registered Republicans who said they voted for Gentner Drummond in the Aug. 25 runoff, which Mazzei won 50.27% to 49.73%, Mazzei leads Munson 52% to 33% under the estimate; under the alternative, Munson leads 54% to 36%. Mazzei's own runoff voters back him about 9 in 10 under both weights. Drummond received 184,584 votes in that runoff.

The voter file shows 62% of statewide respondents voting on Aug. 25, when the ballot also carried two state questions open to all voters, so the Drummond-voter table is published under both weights.

Self-reported Drummond runoff voters, registered RepublicansQ3 by Q15a · LV % · n=120
GovernorEstimateAlternative
Mazzei52.136.3
Munson33.253.7
Undecided12.28.0
Respondents120120
Margin of error±12.9±11.0

Results · Statewide

Other statewide races

Under the estimate, among likely voters:

  • U.S. Senate: Hern (R) 49%, Thomas (D) 30%, 16% undecided.
  • Attorney General: Echols (R) 50%, Coffey (D) 40%.
  • Superintendent of Public Instruction: Franklin 63%, Marshall 24%.
  • Corporation Commission: Boles 55%, Eastman 33%.
  • State Question 847: yes 54%, no 25%.
  • State Question 845: no 61%, yes 20%.
Likely voters, weighted %EstimateAlternative
U.S. Senate · Q4 · n=569 · ±5.7 / ±5.2
Hern Republican49.039.1
Thomas Democrat29.639.0
Stinnett2.72.4
White1.42.5
Meinhardt1.01.1
Undecided16.316.0
Attorney General · Q5 · n=567 · ±5.7 / ±5.2
Echols Republican50.140.0
Coffey Democrat39.750.6
Undecided10.19.4
Superintendent of Public Instruction · Q6 · n=564 · ±5.8 / ±5.2
Franklin Republican63.458.7
Marshall Democrat24.330.5
Undecided12.210.8
Corporation Commission · Q7 · n=560 · ±5.8 / ±5.3
Boles Republican55.043.8
Eastman Democrat33.043.4
Undecided12.112.8
State Question 847 · Q8 · n=556 · ±5.8 / ±5.3
Yes53.651.3
No25.429.1
Undecided21.019.6
State Question 845 · Q8a · n=550 · ±5.8 / ±5.3
Yes20.020.0
No60.663.2
Undecided19.416.8

Results · 5th Congressional District

5th District: Bice leads under the estimate, Nelson under the alternative

Among likely voters in the 5th Congressional District, Rep. Stephanie Bice (R) leads Democrat Jena Nelson 52% to 44% under the estimate (plus or minus 5.7 points). Under the alternative weight, Nelson leads 51% to 44% (plus or minus 5.4). In the governor’s race in the district, Munson leads Mazzei 51% to 42% under the estimate and 58% to 35% under the alternative.

The district table pools the 405 respondents who live in the district by the voter file (167 from the statewide sample, 172 from the district supplement, 66 from the runoff reserve), each with a base weight equal to the reciprocal of its probability of inclusion in any of the three samples. It is raked to the district’s party registration and county totals (Election Board, Jan. 15, 2026); the estimate also corrects recalled 2024 vote to the district’s result (Trump 58.0, Harris 39.8, other 2.3) among respondents the file shows voting in 2024, as statewide. It is not raked on age or gender, because no district-level counts were available. Design effect 1.26 (estimate), 1.14 (alternative), likely voters. Published at 3:48 p.m. Oct. 29; it belonged in this morning’s release, and the log says why it was held.

5th District likely voters, weighted %EstimateAlternative
U.S. House · Q9 · n=372 · ±5.7 / ±5.4
Stephanie Bice Republican51.743.8
Jena Nelson Democrat43.751.3
Robert Henri Independent1.61.3
Austin Nieves Independent0.50.7
Undecided2.42.9
Governor in the 5th District · Q3 · n=385 · ±5.6 / ±5.3
Cyndi Munson Democrat50.558.0
Mike Mazzei Republican41.734.6
Independents (Brooks, Griffin)1.31.2
Undecided6.56.2
Recalled 2024 vote, likely voters who reported one · n=335 (unweighted: Trump 40.6, Harris 55.5) · actual: Trump 58.0, Harris 39.8
Trump57.044.4
Harris40.950.4

Results · Approval

Stitt and Trump job approval

Among likely voters under the estimate, 40% approve of Gov. Kevin Stitt's job performance and 55% disapprove. President Trump's approval is 48%, with 52% disapproving. Under the alternative weight both are lower: Stitt 32% approve, 64% disapprove; Trump 36% approve, 63% disapprove.

Gov. Kevin Stitt job approvalQ13 · LV % · n=543
EstimateAlternative
Approve40.332.3
Disapprove55.563.9
Not sure4.23.8
Strongly approve 7.5 / 6.1Strongly disapprove 35.4 / 42.5
President Donald Trump job approvalQ14 · LV % · n=541
EstimateAlternative
Approve47.636.3
Disapprove51.563.0
Not sure0.90.7
Strongly approve 27.5 / 21.1Strongly disapprove 41.4 / 54.7

Sensitivity

Straightlining

The pre-registered quality rule — identical answers on every approval and support scale, completed in under three minutes — removed 65 respondents, 39 from the statewide sample (19 Democrats, 14 Republicans, 6 independents). The rule is applied as registered. Restoring the 39 and weighting them like every other respondent moves the registered-voter governor margin less than one point toward Munson under the estimate, leaving it about even, and about three points under the alternative (Munson 54, Mazzei 36 to Munson 56, Mazzei 34). Corrected Oct. 30: this paragraph first said about five points under the estimate; that figure did not apply the recalled-vote correction to the restored respondents.

Context

Earlier Cross Timbers polls

CTP 2026-01 and 2026-02, fielded in September and October on Prolific opt-in panels, found Mazzei ahead by 10 and 3 points among likely voters. Those surveys used a different frame and mode, and the three are not a trend line. The estimate is consistent with them. The alternative is not.

Field facts

How this poll was fielded

Dates
Oct. 24–28, 2026.
Mode
Text-to-web, sent through RumbleUp to a sample drawn from the L2 voter file.
Invited
60,000 sampled; 59,684 invitations sent (40,000 statewide, plus 5th District and runoff-voter supplements).
Delivered
51,557.
Included
1,019 respondents, 596 of them statewide; statewide results are from those 596 (575 likely voters).
Response rate
2.0% of delivered invitations (1.8% counting completes only).
Estimate weight
Party registration (Oklahoma State Election Board, Sept. 30, 2026), age, gender and region (L2), and recalled 2024 vote among voters the file confirms voted in 2024.
Alternative weight
The same, without recalled vote.
Design effect
1.95 (estimate), 1.61 (alternative), likely voters.
Pre-registration
Filed Oct. 24, 2026; every post-close decision is in the public log.

Release schedule

Three days of results

DayDateContents
Day 1 (today)Thursday, Oct. 29Governor and statewide races
Day 2Friday, Oct. 30State direction, issues and full crosstabs
Day 3Monday, Nov. 2The Inola smelter and open-ended responses

The respondent-level data file is released in three parts alongside.

Data and reproducibility

Respondent-level data, part 1 of 3

To reproduce the governor topline, keep statewide (sample = S) likely voters (lv = True) who answered Q3, and take each answer's share of the summed weight_estimate (or weight_alternative).

Methods

Methods note

Sample design, likely-voter definition, weighting targets and achieved marginals, margins of error, the choice of estimate and the limitations of this survey are in the CTP 2026-03 methods note.

Pre-registration

Filed before the first text was sent

The design, questionnaire, inclusion rules, likely-voter definition and weighting targets were pre-registered Oct. 24, 2026. Read the CTP 2026-03 pre-registration.

All polls

Release archive

PollSubjectField datesSampleStatus
CTP 2026-03 (this page)Oklahoma statewide and CD-5, voter-file text-to-web surveyOct. 24–28, 2026596Final results · Day 1 of 3 · pre-registration
CTP 2026-02Oklahoma governor, 2026 general electionOct. 7–12, 2026128Final
CTP 2026-01Oklahoma governor, 2026 general electionSept. 27–30, 2026118Final

Stay informed

Sign up for updates

Get new poll results and occasional invitations to take part in a Cross Timbers Poll survey.