Alberta Referendum Polls: Remain Leads With 11 Days to Go

Every recent public poll on Question 10 shows the option to remain in Canada well ahead. But with 11 days until voting day, the biggest uncertainty is no longer what Albertans tell pollsters — it is who actually turns out.

As of Oct. 8, AB Post could not find a public poll on Question 10 with fieldwork conducted in October. The newest widely reported numbers come from polling completed in late September.

Across six recent public surveys, support for Option 1 — Alberta remaining a province of Canada ranges from 61 to 74 per cent. Support for Option 2 — beginning the legal process required to hold a future binding referendum on separation ranges from 20 to 33 per cent.

The latest polling

The balance in each survey consists of undecided respondents, people who did not answer, or those who said they would not vote. Léger reported 10 per cent undecided in August, up from 6 per cent in its earlier wave. Janet Brown reported 4 per cent unsure in September.

Ipsos said the two-point decline in its Remain number compared with its late-May poll was not statistically significant. Janet Brown’s September survey put Option 2 at 23 per cent, compared with 27 per cent in an April survey, although the earlier question was worded differently.

A stable picture, with one notable outlier

Five of the six surveys place Option 2 between 20 and 24 per cent. The Angus Reid Institute’s August survey is the exception, at 33 per cent.

Angus Reid’s own work also shows why wording matters. In a later methodology experiment, the institute found that changing the wording and placement of the separation question moved the Remain response by several points. That does not invalidate the polling, but it is a reminder that Question 10 asks voters whether Alberta should begin a legal process toward another referendum — not whether Alberta should leave Canada immediately.

The official ballot wording is:

“Should Alberta remain a province of Canada, or should the Government of Alberta commence the legal process required under the Canadian Constitution to hold a binding provincial referendum on whether or not Alberta should separate from Canada?”

Option 1 is to remain a province of Canada. Option 2 is to commence the legal process toward a future binding referendum. Elections Alberta confirms that the result of Question 10 itself is non-binding.

The variable the polls cannot settle: turnout

Polling measures stated preference. The result on Oct. 19 will be determined by the people who vote.

Several surveys suggest the two camps may not be equally motivated. Angus Reid found Option 2 supporters were more likely than Remain supporters to describe the vote as very important to Alberta’s future. Léger also found large age differences in stated turnout intention, with older Albertans considerably more likely to say they intended to vote.

A recent Business Council of Alberta analysis, using a turnout simulator, illustrated how extreme the turnout gap would have to become to overturn a polling lead of roughly two-to-one. Starting from Option 2 support around 24 per cent, its scenario required about 92 per cent of Option 2 supporters to vote and only 29 per cent of Remain supporters to turn out.

That is not a forecast. It is an illustration of how large the turnout imbalance would need to be if the polling is broadly accurate.

Polling error still matters

The other uncertainty is whether the polls are systematically missing part of the electorate.

Most of the surveys in the table use online panels rather than probability samples, so conventional margins of error do not strictly apply to them. Janet Brown’s survey is different: respondents were recruited using randomly selected telephone numbers and could complete the interview by phone or online.

Pollsters also survey different populations. Research Co. surveyed likely voters; most of the others surveyed Alberta adults or residents. Weighting methods also differ.

None of that means the polling is wrong. It means a 20-point or 40-point lead should not be read as a guaranteed election result.

The other nine referendum questions

Polling on the rest of the ballot is thinner.

Research Co.’s August survey found majority support for all five immigration-related questions it tested, with support ranging from 58 to 66 per cent.

Angus Reid Institute later tested several of the government’s questions against alternatives that separated policies bundled together in the official wording. It found that 21 per cent of Albertans considered the referendum questions difficult to understand and that support could change when individual policy components were tested separately.

That matters when interpreting the eventual results. A “yes” vote on a multi-part question may not tell policymakers which individual element voters supported.

What we still don’t know

Whether opinion has moved in October. As of Oct. 8, no public poll located by AB Post had fieldwork conducted this month.

Who will actually vote. Special ballots are already in circulation, advance voting begins Oct. 13, and Election Day is Oct. 19. Poll averages do not determine turnout.

Whether the pollsters share a common blind spot. The surveys use different methods, but several rely on online panels and weighting. A systematic turnout or sampling error would not necessarily be visible before voting day.

For now, the public polling points in one direction: Remain holds a substantial lead. The unanswered question is whether that lead survives the transition from survey responses to actual ballots.

Voting dates

Advance voting runs Oct. 13–17. Election Day is Monday, Oct. 19, with polls open from 9 a.m. to 8 p.m. Elections Alberta says Question 10 will be counted first and the full unofficial count must be completed within 48 hours of polls closing.

Sources: Elections Alberta; Ipsos/Global News; Janet Brown Opinion Research/CBC News; Abacus Data; Research Co.; Léger; Angus Reid Institute; Business Council of Alberta.