Canadian federal polling averages

The following is a weighted average of the latest federal polls. A full description of the methodology used to weigh the polls can be found here. The federal polling average was last updated on November 24, 2014.

The aggregate has been updated with a new poll by Forum Research for the Toronto Star.

Forum was last in the field October 3-5. Compared to that survey, the Liberals dropped two points to 36% while the Conservatives were down one point to 33% (so, in Forum's and the Star's jargon, the Tories 'caught up' to the Liberals by losing less than them).

The New Democrats were down one point to 18%, while the Greens were up four points to 8%. That was the only shift in support that would be outside of the margin of error.

The Liberals have fallen in four consecutive polls by Forum. The party was at 44% in a mid-July survey.

It is worth noting that the Conservatives scored 22% in this poll in Quebec, the highest they have managed in any poll since September (which was another Forum poll putting the party at 22% there). In fact, these two Forum polls were the only ones that put the Conservatives that high in Quebec in the last 124 polls going back to 2012. So, that should be taken into consideration when looking at the overall results. If the Tories are bumped down to a more reasonable level in the province (and the NDP bumped up, as they are quite low at 23%), we get something more like 31% for the Conservatives nationwide and 20% for the NDP. That looks a lot more like other polls.

And this puts Forum's contention that the Liberals and Conservatives are in a 'virtual tie' in context. For the parties to be tied, one has to assume that the margin of error is working against the Conservatives' favour. But with the party scoring so unusually high in Quebec, is that more likely than the opposite?

The poll also highlights how a headline can put an odd angle on numbers. Forum titled their report "Conservatives catch up to Liberals", which the Star picked up as well. This suggests that the Conservatives have actively closed the gap with the Liberals. This is not the case - the Liberals have dropped support, putting them in range of the Tories. The active party here is the Liberals, not the Conservatives (the Tories were at 34% in the previous two surveys, and have now dropped to 33%).

Putting this kind of spin on the numbers gives a false impression of what is going on. The Conservatives have not caught up to the Liberals, they have held steady. Instead, the Liberals have lost their comfortable lead. That their lost support has, at least according to this poll, gone to the Greens does not seem to suggest that it is the Tories that have done the gap-closing.

With the aggregate levels of support, the Conservatives would likely win 136 seats, with 132 seats going to the Liberals, 67 seats to the New Democrats, 2 seats to the Greens, and 1 seat to the Bloc Québécois.

The aggregate had previously been updated on Nov. 19 with a new poll by CROP for La Presse (Quebec only).

The New Democrats were ahead in the poll with 34% support in Quebec, followed by the Liberals at 32%. The Conservatives were third with 16% support, while the Bloc Québécois had just 14%.

Compared to CROP's previous survey of October 16-20, the NDP was up four points while the Liberals were down five. The Conservatives and Greens were each up one point. Only the Liberal fall would be outside of the margin of error of a similarly sized probabilistic sample.

The trends have been steady in Quebec, however, with neither the NDP nor Liberals having any momentum. Over the last six polls by CROP, the NDP has wobbled between 29% and 36%, with the Liberals wobbling between 31% and 38%, with no discernible pattern.

The New Democrats had a big lead among francophones, with 39% to 27% for the Liberals and 17% for the Bloc. Thomas Mulcair was also ahead on who Quebecers preferred to be Prime Minister, with 29% support to 22% for Justin Trudeau.

---

The methodology used to weigh polls for the averages can be found here. By including polls in the average, no representation as to the accuracy or equivalency of the methods used is implied, nor should inclusion be seen as an acceptance, endorsement, or legitimization of their results. The weighting scheme takes reliability partly into account. See here for a complete rundown of the latest polls in Canada (external link).

The chart below shows average support in polls conducted in each month going back to January 2009.