Éric GrenierAkaun disahkan

On the latest episode of the Pollcast, @cath_cullen and @JPTasker join me for a little game of political speculation and forecasting. Stream, subscribe here! #cdnpoli https://www.cbc.ca/radio/podcasts/current-affairs-information/the-cbc-election-pollcast/ …

From Sept. 14 to Nov. 9, Nanos's trend line is -2 LIB, -3 CON, +5 NDP. The Campaign trend line is -1 LIB, -3 CON, +2 NDP. So, some similarities despite the overall very different picture of the race: both Liberals and Conservatives down, maybe Conservatives a little more, NDP up.

Campaign is showing a closer race than either Mainstreet (LIB +5) or Nanos (LIB +11). But Nanos and Campaign trend lines are similar. Campaign and Nanos were out of the field on Nov. 9 and also out of the field on Sept. 14, when it was CON +2 for Campaign and LIB +10 for Nanos.

Small correction, the MQO numbers in Atlantic Canada are 47% LIB, 33% CON, 10% NDP, 8% GRN, 2% OTH. The numbers posted earlier in the thread did not include leaning voters.

Nanos's "Best PM" polls, for example, show that Singh is less popular than the NDP as a whole, while Bernier is more popular than the PPC.

Mainstreet includes the leader's name in its voting intentions polls (Liberals led by Justin Trudeau, etc.). This could be one of the factors behind the NDP's poor results compared to other polls, and Bernier's better results for the PPC.

I suppose that's true if you believe people's opinions only matter once every four years.https://twitter.com/charlesadler/status/1062385211411881985 …

The Bloc seems to be suffering most from these unaccounted votes, as they are at just 8% in Nanos's polling. But in October 2017, it was the Conservatives who were taking the hit. They were at about 8% in the province while the Bloc was in the high-20s. Another thing to watch.

Could that be people who say they want to vote for the CAQ, Québec Solidaire or the PQ? Could be. This has happened before in Nanos's polling. In October 2017, about 10% was unaccounted for. I don't know what to make of it.

One odd thing about Nanos's polling is in Quebec. Support for the six parties (including the PPC) adds up to 85%. Presumably the remaining 15% is "Others". The polling was conducted after the provincial election was completed, so there shouldn't be too much cross-contamination.

But the house effects from different pollsters have been pronounced lately. We're seeing some pollsters with good Liberal leads and others with strong Conservative numbers. Hard to know where things actually are.

There isn't any regional result that is particularly strange for the Tories. They are low in BC (19%), Ontario (31%) and Quebec (13%), so it isn't one regional sample that is dragging the Conservatives downwards. I'll be looking for other data to see if this is real or not.