Alberta referendum question phrasing could have major impact on results, poll finds
'When you're asking a double- or a triple-barreled question, it certainly leaves room for doubt'
Author of the article

evon Dekuyper
Published Sep 21, 2026 • Last updated 19 hours ago • 4 minute read
Elections Alberta had a technical media briefing and showed a simulated voting station for the Oct. 19 referendum on Thursday, Sept. 3, 2026.
Elections Alberta had a technical media briefing and showed a simulated voting station for the Oct. 19 referendum on Thursday, Sept. 3, 2026. Shaughn Butts/Postmedia
Many Albertans would have different answers if next month’s referendum questions were to be worded differently, a new survey found.
In an experiment conducted by the Angus Reid Institute, poll results were collected for the 10 existing referendum questions as they will appear on the ballot, and compared with results for the same questions, but phrased according to Angus Reid’s own best practices for polling.
The results showed Albertans’ answers would vary significantly, depending on exactly how the question is asked.
An earlier survey by Angus Reid found one in five Albertans (21 per cent) described them as difficult to understand, while 34 per cent said the questions are very clear, and 43 per cent described them as mostly clear.
“When you think about what one-fifth of the voting population of Alberta would look like, you’re already into potentially hundreds of thousands of potential voters,” said Shachi Kurl, president of Angus Reid.
When separated by age demographic, the data show more than a third of those aged 18-34 found the questions confusing or difficult to understand.
“When you’ve got a third of young Albertans throwing up their hands and saying this is either mostly or completely difficult to follow . . . I think there are big concerns about, first of all, are they going to vote, and do they feel confident in what exactly they’re casting a ballot around?” Kurl said.
With that in mind, Angus Reid researchers sought to find out how Albertans would respond if the questions were posed according to standard polling methodology.
‘There’s questions within questions within questions’
The first question on the referendum, for example, asks Albertans, “Do you support the Government of Alberta taking increased control over immigration for the purposes of decreasing immigration to more sustainable levels, prioritizing economic migration and giving Albertans first priority on new employment opportunities?”
When polled, 62 per cent of participants answered “yes,” 33 per cent answered “no,” and five per cent were unsure.
The existing phrasing asks Albertans multiple questions at once: whether the province should take increased control of immigration, whether immigration levels should be decreased, whether economic migration should be prioritized, and whether Albertans should have first priority for jobs.
When asked each sub-question individually, 56 per cent of those surveyed felt immigration to Alberta should be reduced, while only 46 per cent felt that people already living in Alberta should be given greater priority for new employment opportunities.
Overall, however, 80 per cent of participants disagreed with at least one of the original question’s elements.
“When you’re asking a double- or a triple-barreled question, it certainly leaves room for doubt as to what part of the question people are responding to or answering,” Kurl said.
Alberta referendum simulated voting station
Elections Alberta had a technical media briefing and showed a simulated voting station for the Oct. 19 referendum on Thursday, Sept. 3, 2026. Shaughn Butts/Postmedia
Mount Royal University political scientist Lori Williams said she’s found many people are “just completely overwhelmed” by the wording of the questions.
“The first four questions are very complex and difficult to understand,” she said. “That first question has four questions in it. Two of the (other) questions contradict one another . . . There’s questions within questions within questions.
“It essentially gives the government both control in terms of the question it asks, and control in terms of how it interprets the results,” she said.
Not all questions tested, as some ‘written fairly clearly,’ says pollster
Looking at question eight, which asks Albertans, “Do you support the Government of Alberta working with the governments of other willing provinces to amend the Canadian Constitution to allow provinces to opt out of federal programs that intrude on provincial jurisdiction such as health care, education and social services, without a province losing any of the associated federal funding for use in its social programs?” — those polled were split, with 44 per cent answering “yes” and 44 per cent answering “no.”
However, 58 per cent said they felt provinces receiving federal funding should generally be required to meet the conditions attached to that funding, and 52 per cent said the federal government should be able to reduce or withhold the funding if conditions are not met.
“We didn’t actually even test all of the official ballot questions because some of them were written fairly clearly,” Kurl said.
“But we certainly wanted to better understand and disaggregate some of the most obvious examples of questions that did not read as particularly clear in terms of what people were answering.”
The first survey, which asked about the referendum questions as they are written on the ballot, was conducted from Aug. 10 to 13.
The second survey, which parsed out various aspects of the questions, was conducted from Aug. 26 to Sept. 1. Each surveyed a randomized sample of around 1,000 Canadian adults who are residents of Alberta.
Angus Reid notes a margin of error of ±3 percentage points, 19 times out of 20.
Many Albertans would have different answers if the upcoming referendum questions were to be worded differently, a new survey found.
torontosun.com