Jan Brewer verbally stumbled, went silent and mangled her grammar during last week's televised debate.
But a new statewide survey done Tuesday shows that apparently none of that matters.
Pollster Scott Rasmussen found that 60 percent of the 500 likely Arizona voters questioned in the automated telephone survey said they intend to vote for the incumbent. That's up three points from a survey taken a week before the debate, though well within the 4.5 percent margin of error in the survey.
Democrat Terry Goddard remains stuck at 38 percent.
And just 1 percent of those questioned said they had not yet made up their minds.
Goddard said he doesn't believe the survey.
"This is a Rasmussen poll that asks us to believe something that is almost impossible to believe, and that is she picked up support after her horrifyingly bad performance," he said. "I can't understand it. The poll itself has to be flawed."
But Jim Haynes, president of the Arizona-based Behavior Research Center, said that's not necessarily the case.
His company uses a more traditional method of polling, with live questioners instead of seeking touch tone responses to automated questions. And Haynes said he's not vouching for what Rasmussen's numbers found.
But he said Goddard is mistaken if he thinks Brewer's gaffes will help him win the election.
"There's a serious question whether debates change anybody's mind," he said. "They give rallying points to people that have already made up their minds."
Haynes also said a candidate tripping up in a debate is only half of what it takes to convince voters to rethink their support. He said they also need a reason to back the other contender.
"A stumble on the part of one candidate, in and of itself, doesn't necessarily mean there's going to be a movement," Haynes said.
That is reflected in other elements in the survey.
Rasmussen found that only 9 percent of Brewer supporters said her debate performance was important to them in making a decision. By comparison, how Brewer came across was listed as important by 43 percent of people who said they already intended to vote for the Democrat.
Goddard said he's not planning to do anything different in the wake of either the latest poll or prior surveys showing Brewer with a lead in the 20-point range. He said his focus will remain on the economy, "how we're being hurt and how she is precipitating that harm."
"I'm angry and I hope that anger reflects throughout the state because we have so many things to be angry about," Goddard said.
Rasmussen, however, said he's not sure that will work - at least not this year.
"You have the immigration law, you have the health care law, you have the fact that people are really angry at the federal government these days," he said.
Rasmussen said 95 percent of Brewer supporters disagree with the policies in Washington. And 90 percent of those who back her disapprove of the way Barack Obama is doing his job.
Goddard acknowledged Brewer got a boost for signing the state's new immigration law and defending the statute against legal attack by the Obama administration. But he is counting on immigration not remaining foremost in the minds of voters.
"It's a very volatile situation," he said of the race, saying there will be "twists and turns" in the weeks ahead.
But Goddard could be running out of time to change the discussion.
Early voting begins Oct. 7. And if last month's primary is a barometer, two-thirds of the ballots cast will have been filled out before people go to the polls on Nov. 2.
Rasmussen said, though, Goddard may still be unable to take the lead even if he is successful in reshaping the debate. The issue is deeper than that.
More than half told Rasmussen they have an unfavorable impression of Goddard, who has been state attorney general for the last eight years, with more than half of those saying they see him in a very unfavorable light.
Brewer, by contrast, has total negatives in the 37 percent range.
Rasmussen uses an unusual method of polling: Computers dial numbers and ask those who pick up to answer with touch-tone responses to questions about their registration status, party affiliation and specific issues. That enables him to do many different polls in a short period.
While the methodology is different than using live questioners, his Arizona surveys, including on the governor's race, have shown a certain consistency: Goddard's ratings have shrunk somewhat since March while Brewer's have skyrocketed after her decision to sign SB 1070.
And that gap Rasmussen has recorded between the two contenders has tracked closely with prior surveys done by Behavior Research Center.