New polling data suggests Andrew Knack is leading the race for Edmonton’s next mayor among decided and leaning voters. But the same poll suggests almost half of voters are undecided.
The exclusive CBC News poll, conducted by Janet Brown Opinion Research, suggests Knack and Tim Cartmell are the leading candidates when it comes to name recognition and voter intention.
Among people who were decided or were leaning a particular way, 41 per cent of those surveyed said they’d vote for Knack, and 21 per cent said they’d vote for Cartmell.
However, 46 per cent of respondents said they were still yet to decide.
The survey of 1,000 Edmonton residents through the Trend Research Online Panel was conducted between Oct. 1 and 8.
Pollster Janet Brown said that in any race, especially one without an incumbent, a large number of undecided voters this close to election day is surprising.
“This is much higher than I’ve ever seen this close to an election at the federal, provincial or municipal level,” said Brown.
She said the findings could indicate an electorate that is lacking energy amid a political environment oversaturated with issues like the state of Canada-U.S. relations and the Alberta teachers’ strike.
Many undecided voters might not cast a ballot at all, says political analyst Najib Jutt, adding that there is a lack of choice when it comes to candidates who are newer to politics.
The current crop of candidates includes Knack and Cartmell, who are current city councillors, Michael Walters and Tony Caterina who have served on previous councils and Rahim Jaffer who worked in politics as a federal Conservative MP.
Other public opinion research has indicated that Edmontonians have been looking for some new faces, said Brendan Boyd, an associate professor of political science at MacEwan University.
“Maybe people are looking for change, but maybe they just don’t know where that change is,” he said.
With this many undecided voters, there remains a lot of volatility in what election results could look like, Boyd said.
Most candidates trail 2 front-runners
While Brown’s polling suggests support for Knack and Cartmell is far higher than other mayoral hopefuls, there is also a noticeable gap in the level of support between the remaining candidates.
Omar Mohammad and Walters saw 10 per cent of survey respondents who identified themselves as decided or leaning voters indicated they would vote for either candidate.
Seven per cent responded Jaffer and six per cent said Caterina.
“I think the strength of incumbency is more and more diminished,” Jutt said.
The remaining candidates — Paul Bakhmut, Ronald Stewart Billingsley Jr., Abdul Malik Chukwudi, Vanessa Denman, Andy Gudanowski, Utha Nadauk and Olney Tugwell — all had one per cent support or less, according to the poll.
Jutt said it appears Knack’s campaigning has been effective.
“Knack has positioned himself as the only progressive choice,” he said.
“The rest of them are kind of cannibalizing each other’s votes in that centrist or right of centre kind of space.”
While Boyd wouldn’t go as far to say there is vote-splitting in this municipal election, he also noted Knack has shown a focus on progressive issues like homelessness and housing affordability at recent debates.
With just 10 days until the election and with advance voting already underway, Jutt said the winner could come down to who is the best at getting people out to vote.
Boyd agreed, saying in municipal elections where there aren’t millions of voters, door-knocking and getting people to the polls are sometimes enough to sway a race.
“I think we’ve seen lots of shifts in even the past week, and I still think we can see shifts between now and election day,” Jutt said.
CBC News’ survey of 1,000 City of Edmonton residents was conducted online between Oct. 1-8, 2025 by Edmonton-based Trend Research under the direction of Janet Brown Opinion Research. Quotas were set for age, gender and postal code. Minimal weighting was applied to match Statistics Canada population data. As a non-probability online survey, a margin of error does not apply. However, the margin of error for a comparable probability sample would be +/-3.1 percentage points, 19 times out of 20.
