A survey not commissioned by either governor campaign places both candidates at 44%. Pollster Pat McFerron identifies a key battleground: voters who view Trump favorably but still call themselves traditional Republicans.
OKLAHOMA CITY — Oklahoma’s Republican runoff for governor appears headed for a very close finish, with a new statewide poll finding Gentner Drummond and Mike Mazzei tied at 44%.
The Cole Hargrave Snodgrass & Associates survey found 13% undecided heading into the Aug. 25 election. Drummond held a slight edge, 42% to 40%, before initially undecided voters were asked which candidate they leaned toward.
Pat McFerron, president of the Oklahoma City polling firm, discussed the findings in an interview Monday evening with OK Politics Today. He said the dead heat reflects two competing Republican coalitions that are nearly canceling each other out.
Mazzei leads 62% to 22% among respondents who describe themselves as Trump Republicans. Drummond leads 60% to 30% among a larger group identifying as traditional Republicans.
But traditional Republican does not necessarily mean anti-Trump. McFerron said most Trump Republicans and most voters who dislike Trump have largely made their choices.
The likely battleground is the 32% who view President Donald Trump favorably but still call themselves traditional Republicans. Within that subgroup—not among all respondents—Drummond and Mazzei are also tied at 44%.
“That’s the group that’s conflicted right now,” McFerron said.
Trump endorsed Mazzei shortly before the June primary. Mazzei has made that endorsement central to his campaign and presents his tax and policy agenda as the conservative plan Trump supports.
Drummond has framed the runoff as a test of leadership and independence, pointing to his actions as attorney general and arguing that Oklahoma’s interests should come first.
Drummond narrowly leads among frequent primary voters
Drummond led 47% to 45% among respondents who had voted in at least three previous Republican primaries. The difference is small, but it gives him a slight edge within a group whose members have demonstrated repeated participation in Republican primaries.
McFerron said 55% of voters still undecided are 65 or older. As a group, those voters are divided almost evenly between self-described Trump and traditional Republicans.
Geography also makes a difference. The undecided share is 10% in both the Oklahoma City and Tulsa media markets, compared with 22% in the combined outlying markets.
“They just don’t have quite the same information flow,” McFerron said.
Two polls, different results
The CHS findings differ from a Pulse Decision Science poll promoted Monday by Mazzei’s campaign.
That survey, conducted for Club for Growth PAC, found Mazzei leading Drummond 44% to 38%, with 18% undecided. Club for Growth supports Mazzei in the runoff.
Both polls placed Mazzei at exactly 44%. The difference was in the percentage supporting Drummond and the number remaining undecided.
The surveys also overlapped substantially. Pulse questioned voters from Aug. 2-4, while CHS was in the field from Aug. 3-6.
Because the field dates overlapped, the results cannot be treated as a timeline showing that Drummond gained six points. Concurrent polls can produce different snapshots of the same race.
Both surveys interviewed 600 voters, used combinations of live telephone and text-to-web interviews, and reported margins of error of plus or minus 4 percentage points.
The CHS survey was not commissioned by either governor campaign. The firm’s quarterly polling is supported by private corporate underwriters that are not publicly identified, although McFerron said they do not determine which findings are released.
CHS also disclosed that McFerron polls for attorney general candidate Jon Echols, whose campaign did not commission the Sooner Survey.
McFerron declined to criticize the Pulse survey.
“I don’t ever want to criticize or talk about somebody else’s polling,” he said. “What I had before the debate is pretty much what they had.”
Pre- and post-debate interviews show different results
The CHS survey happened to straddle the Aug. 3 debate, presented by the Oklahoma Hospital Association and broadcast on Griffin Media platforms, allowing a comparison between voters interviewed before and after it.
Pollsters completed 149 interviews before the debate, with Mazzei leading that group by four points. The remaining 451 interviews were conducted from Tuesday through Thursday, with Drummond leading that group by one.
The results compare two different groups of voters; CHS did not interview the same people twice. They therefore do not prove that individual voters changed their minds.
“I think the debate and then the post-debate sentiment made a difference,” McFerron said.
The full CHS sample also shows a gender divide. Drummond leads Mazzei by 7 percentage points among women, while Mazzei leads by 6 points among men.
McFerron said Mazzei’s ratings changed most noticeably among women in the later interviews. His theory centers on a contentious exchange in which Mazzei continued speaking as moderator Amanda Taylor tried to move to the next question. The exchange ended with Taylor warning the candidates that their microphones could be muted.
McFerron said one hypothesis is that women reacted negatively to what he described as Mazzei trying to “hush Amanda Taylor.”
Mazzei did not literally tell Taylor to be quiet. The five-point margin shift is consistent with McFerron’s theory but does not prove it because it involved different groups of voters and a much smaller pre-debate sample.
Whatever effect the debate may have had, the broader findings point to a close race without a clear advantage for either candidate two weeks before the runoff.
