Note: bien que les billets concernant le Québec sont écrits en général en français, il existe un minorité anglophone dans cette province. Après tout, le journal The Gazette par exemple couvre la politique québécoise et semble lire ce blogue. Ainsi, par respect pour cette minorité (et aussi pour les anglophones hors Québec qui s'intéresseraient à la politique québécoise), ce billet sera en anglais. Ceci restera une exception pour les billets couvrant la politique Québécoise.
Political polls in Quebec are usually conducted by Crop (and published by La Presse) or Legermarketing (and published in Le Devoir). However, Forum Research is now conducting a monthly polls in La Belle Province. its polls are less known in the media but have two benefits. 1) Forum doesn't use an online panel as opposed to Crop/Leger who have switched completely. 2) The sample size is usually bigger at around 2000 respondents (although it's not the case in the last poll).
The latest Forum poll show the same big picture than the latest Crop, but the details are very different. If both polls agree that PLQ and PQ are neck-to-neck while the CAQ lags behind, they disagree on how low the CAQ is (or, alternatively, how high the other two main parties are). Indeed, Forum has the Liberals and PQ at 35% each and the CAQ at only 16%. Even by applying the margins of errors, the CAQ is statistically polled lower than in the latest Crop (25%). Forum has had the CAQ very low (i.e: under 20%) consistently for months. I already mentioned it, as well as a possible explanation (the non-online panel of Forum). We might not be able to determine who is right until the next election. And as we saw recently with the Alberta election, polls can be very, very wrong at times...
In term of seats, the Forum polls would give 61 seats to the PLQ, 58 to the PQ, 5 to the CAQ and 1 to QS (remember that whenever I miss or I'm slow at covering a poll, you can always use the simulator. Results are exactly the same with the exception of the inclusion of Option Nationale in my own projections). So as with Crop, a tight race to get a minority governement between the Liberals and the PQ. Interestingly enough, Forum has its own seat projection model and predict a PQ majority with 66 seats while the PLQ would get only 50. Why is that? Well, Forum doesn't really publish its methodology but my guess is that they use a linear model that doesn't take into accounts factors such as the region or the incumbency. Also, the inclusion of the CAQ is probably different from the way I did it. On top of that Forum doesn't take the risk of providing riding-by-riding projections. At the end of the day, I naturally prefer my method since this one is more advanced, but it doesn't mean I'm right. Also, remember that a lot of ridings are so close that a toss-up should be used in order to determine the winner. This could reconcile Forum's projections with mine. But I really think Forum overestimates the PQ (or underestimates the PLQ). Remember as well that at 35%, the PQ is actually losing some votes since 2008. How come? Well, I assume that the PQ got some votes from the ADQ when this party disapeared/merged. These votes mean that the PQ "starts" at 37% provincially (and not 35%, its official results of 2008). My guess is that Forum consider the PQ as stable since 2008. This assumption can explain much of the differences between the two projections.
By the way, the polls also include very interesting questions about the opinion of Québécois about anglophones in this province. For instance, we learn that as much as 29% think that anglophones shouldn't be allowed to publicly speak English!
So, who do you think is right? Crop/Leger with the CAQ at around 25% or Forum with François legault and his party at 15% only?