Wednesday, August 29, 2012

PQ still in majority territory, but less secure

A new poll by CROP for La Presse confirmed the results of the weekend's Léger Marketing poll to within a percentage point, giving the Parti Québécois a five-point lead over the (now) second place Coalition Avenir Québec. Though CROP believes the numbers would deliver the PQ a minority government, ThreeHundredEight still has Pauline Marois on track for a majority.

The changes from the last projection have been minimal, with the PQ and Liberals dropping one seat each to the benefit of the CAQ and Québec Solidaire. The Parti Québécois leads with a projected 34.1% support, followed by the Liberals at 28.5% and the CAQ at 25.8%.

Québec Solidaire (6.7%), Option Nationale (2.1%), and the Greens (1.8%) round-out the top six.

It is worth noting that though the projection gives the Liberals a second place finish, the unadjusted poll average now puts the CAQ narrowly ahead with 27.6% to 27.3% in voting intentions.

In terms of seats, the projection now awards the PQ 66, three more than is needed for a majority government. The Liberals take 32 and the CAQ take 25, with two seats going to Québec Solidaire.

The ranges have widened somewhat, however. The PQ could win between 48 and 79 seats, suggesting that a minority government is still very much a possibility. The Liberals could win between 24 and 52 seats, theoretically making a PLQ minority possible. But that is an extraordinarily unlikely scenario of Albertan proportions.

They should be more concerned with the CAQ, who has a range of between 13 and 34 seats. Their upper range is above the forecasted result for the Liberals. Considering the way the campaign is going (Liberals sliding, CAQ gaining), the race is truly on for the role of the Official Opposition.

Québec Solidaire should win one or two seats, while Jean-Martin Aussant can still hope to win his seat of Nicolet-Bécancour.

There have not been any major shifts in support as the CROP poll was very similar to the recent numbers by Léger and Segma Recherche. The Liberals did slip slightly on the island of Montreal and among non-francophones, but it did not cost them a seat. Their loss was in Quebec City, where they were down slightly and the PQ was up by even less. The CAQ made gains in the suburbs of Montreal, however, and are poised to start picking off PQ seats if things continue in this fashion.
CROP was last in the field Aug. 12-14, and since then there have been no statistically significant shifts in support. That does not mean that nothing has happened, as the confirmation of a third place finish for the Liberals does tell us something. 

The PQ was down one point to 33% while the CAQ was up three to 28%. The Liberals were down one to 26%, while Québec Solidaire was unchanged at 7%.

The lead the PQ holds is statistically significant (though just barely), while they also hold a noteworthy lead among francophones. Their support has dropped, however, by three points to 36%. The CAQ is up four to 30% among francophones, while the Liberals are down two to 19%.

The Liberals continue to lead among non-francophones with 57%, unchanged from CROP's last poll. The CAQ, at 20%, is also steady.

The PQ has the edge in the Montreal suburbs and in the regions of Quebec, while the CAQ is ahead in Quebec City. The Liberals are in front on the island of Montreal. The PQ saw negligible growth in the two regions where they have the advantage, but took losses in Quebec City and the island of Montreal. The CAQ was up across the board (except in Montreal), while the Liberals were down across the board (again, except in Montreal). The polls are unanimous on this question - despite their sagging support, Jean Charest's Liberals are still doing well on the island.

Riding polls

Two riding polls were also added to the projection. A poll by Cible Recherche for TVA in Trois-Rivières suggests that Djemila Benhabib of the PQ is in a strong position with 38%. Danielle St-Amand, the incumbent Liberal MNA, trails with 29%. The CAQ is further behind with 21% support. This poll echoes the one done by Segma early in the campaign, and is in opposition to the small-sample poll by Baromètre that gave the Liberals a wide lead. The projection has had the PQ in front for some time.

The other poll was in Hull. Conducted for Le Droit, the Segma poll found the Liberal incumbent Maryse Gaudreault leading with 39%. The PQ's Gilles Aubé trails with 30%, while Québec Solidaire placed third with 14%. The CAQ was in fourth with 12%. Though the projection had a slightly closer race, this is generally in line with what the model had. It also seems to contradict some of what Jean-Marc Léger has said about the Outaouais region, which he has called a three-way race. The ADQ has never been a factor in the Outaouais, and it does not seem like the CAQ is making inroads either.

A bevy of riding polls were released this morning for Taschereau, Brome-Missisquoi, Papineau, Orford, Saint-François, and Sherbrooke but they have yet to be added to the model. I will take a look at them in detail in the next update.

68 comments:

  1. http://www.redecoupage-federal-redistribution.ca/content.asp?section=on&dir=now/proposals&document=index&lang=e

    - Ontario proposed boundries are out

    ReplyDelete
  2. Éric, I notice that after polling at 6-7 day intervals from the start of the campaign, Forum has been silent ever since their "awry" poll from August 20th. Do you have any word on whether we can expect a new one from them soon? I'm very curious to see what their new numbers would be.

    Dom

    ReplyDelete
    Replies
    1. No idea what happened. They have had a poll in the NP every Wednesday, and even said they would have another on Twitter late last week. Not sure why it was changed.

      Delete
    2. So much for that theory, Forum has a poll in today's Post which is more or less in line with the latest CROP and Leger polls.

      Delete
  3. I also noticed the lack of any poll by Forum today after they had been tweeting that they would release something. I wonder if they came up with another embarrassingly unlikely result and decided to suppress it?

    ReplyDelete
  4. Eric,

    I was vacationing in the Orford/Sherbrooke Region during the 2nd week in August and noted not a single lawn sign. Is this the norm for Quebec or just indicative that people were not engaged?

    Peter

    ReplyDelete
    Replies
    1. I really can't say, I'm sure it depends on the riding.

      Delete
    2. In Quebec it is not traditional for people to put lawn signs on their property to show who they support the way it is in the rest of Canada...in Quebec almost all signs tend to be on public property

      Delete
    3. This seems peculiar to me! Why the difference? For a people so opinionated, I'm surprised that they would shy away from publicly endorsing a party in that way.

      Delete
    4. Peter,
      It's very much the norm for Quebecers not to have lawn signs. It's mostly due to the risk of vandalism (this has gone way down from the '60s and '70s when politics was a much more violent sport in Quebec, but that era shaped people's political behaviour).

      Of course, you will see a few here and there, but it's very rare compared to other parts of North America. The only time I ever saw them come out in Quebec was during the last 2 weeks of the 1995 referendum, and even then it was mostly just people attaching flags to their balconies.

      Delete
    5. It could be that they prefer to keep it private or perhaps the parties don't really give away signs that easily to the public. There are signs here and here on private property in Montreal (oftentimes on the apartment balconies), but it's not really common like DL just said.

      Delete
    6. Lawn signs are too confronting - just as conversations about politics among friends may be the norm in other areas, they're avoided in French Quebec 'cause:
      (a) there is a cultural bias towards social "consensus" - so much so that any 'serious' topic is usually avoided;
      (a) sovereignists despise federalism even more than federalists reject secessionnism - so, back to (a);
      (c) there is a tendency to extremism in thought and expression thereof and even sometimes even acting out (think: these hooded students raising havoc - for what?)- so, to stay out of trouble:(a)

      By certain standards, Quebec is still an adolescent society - yes, it contributes its charm... and annoyance!


      And, if Louis Hartz's theory of social fragments is any guide, then French Quebec is and will always remain adolescent - yet, pity us ;-)

      Delete
    7. 8e Samouraï, frankly, that'S a dubious assessment - and too reliant on stereotypes.

      I don't think that lack of signs is particularly meaningful, people here tend to discuss politics rather than merely cheerlead for them.

      Delete
    8. this is based on my own experience as a middle-class francophone... (1) it is too confronting and sterile to tell friends we don't support sovereignty (with non-francophones it's different);
      (2) in terms of discussions, everyone will agree to criticize "the system" or "the establishment", but very few will dare comme strongly in support of a mainstream political oganization (it's dirty, and it's serious... bah!);
      glad that you see things differently...

      Delete
    9. BTW if we want to know what's really going on in this province, let's check the participation rate on voting day...
      If it's just above 60%, what do we make of the remaining 40%? That's a huge
      problem; and, no, we cannot "allocate" the un-voting proportionnally... ;-) Today's society is much more fragmented that what we grew up with - I was flabbergasted by how many "connected" people don't even know the date of the election - many of us live in our little facebook bubbles nowadays...

      Delete
  5. The CROP poll in La Presse on August 28 showed 18% undecided, with a disproportionate number of non-Francophones in this category. Any idea as to how CROP distributed these potential votes among the parties, and any thoughts as to whether this could affect the outcome?

    ReplyDelete
    Replies
    1. I do not know for sure, but I assume CROP distributes those votes proportionately. As to outcome - my view is that undecideds either don't vote or break more or less the same way as the decided population.

      Delete
  6. I believe your voter projection model ascribes a lower performance to third parties like the CAQ; shouldn't that go the other way once they get into second (or first!) position, since they are no longer perceived as a "wasted vote"? Or is it more of a question of incubents being an advantage?

    ReplyDelete
    Replies
    1. Well, the why is just speculative. The analysis I have done merely shows that larger parties (government and official opposition) tend to be under-estimated while smaller parties (other opposition parties) are over-estimated.

      I took a look at whether there is any major difference when a party switches position in the polls, but nothing emerged.

      I believe that this occurs because larger parties have better organizations and have larger and more reliable support bases. Some parties will always buck the trends, but the model makes all the safe bets.

      Delete
    2. This has certainly been shown Federally, Eric, where the Greens are always shown getting 5+ seats and in fact have been lucky to elect one.

      Delete
    3. I'm not sure the Greens have ever been shown to be in a position to win five or more seats, but your point is a good one.

      Delete
    4. For the small parties, like the Greens, my speculation is that, because the party is small, a proportion of those who may have intended to vote for them end up voting strategically rather than lose their ballot. However, this logic does not quite work for the case of the CAQ in the present election, as they have true chances in many ridings, as suggested by Yannick in his question.

      Delete
    5. For the Greens, there is also the problem that a lot of people might say they are going to vote for the Greens only to find out there isn't a Green candidate in their riding (which is the case for almost half of all ridings in Quebec).

      Delete
    6. Plus Eric there seems to be a tendency for the smaller parties to be assigned seat numbers based on the projected percent of the votes and that is not a very accurate way of doing it as the results have consistently shown.

      Delete
  7. And how about a current prediction of 20mm of rain on super Tuesday? I think THAT is statistically significant!

    ReplyDelete
    Replies
    1. Heh, could be cool to correlate the vote participation with the weather

      Delete
  8. FWIW, I have a hypothesis. In past elections the PQ has often done a bit worse on election day than the final polls indicate due to its voter base being less likely to show up (ie: younger people), the "hidden federalist" vote in question (aka "the ballot box bonus"). On top of that I have a hunch that a disproportionate number of undecideds in this election are undecided between PLQ and CAQ and are not likely to vote PQ - this would be especially true of anglophone DKs. For all those reasons, i suspect the PQ may get a point or two less than what many are projecting.

    ReplyDelete
    Replies
    1. But in this election, more than any other, the young people who might normally vote PQ are likely leaning more towards QS and ON. So the penalty might end up going their way rather than the PQ's.

      Delete
    2. In 2007, Léger overevaluated the PLQ by 2 points (33 in the survey rather than 33 in the real result) while overevaluating the PQ by one point (29% in the survey rather than 28% in the election). In 2008, Léger overevaluated the PLQ by 3 points (45 in the survey rather than 42 in the real result). If my memory does not fail me, it has been since a certain number of years that the ballot advantage of the PLQ disappeared and since that time, the survey administrators began to simply distribute the undecided/no answer responses in the same proportion to the different parties as the decided vote. Immediately before that, they were aware of the PLQ ballot advantage and attributed a greater proportion of the vote to the PLQ in the undecided/no answer than to the decided vote.

      However, in this particular election, as mentioned by DL, when we look at the drop of the PLQ in the survey, the drop may have gone partly from PLQ to other parties (mostly CAQ), and partly from PLQ to "wandering whether I should vote CAQ" undecided persons.

      Delete
    3. Oops, in 2007, the PLQ obtained 31% - sorry for the typo (2nd line). Penultimate line: "wondering" rather than "wandering".

      Delete
  9. I am also sceptical that much of the QS, ON and PVQ vote will actually show up - they will be like the Greens federally - 7% in polls and 4% when the votes are counted.

    ReplyDelete
    Replies
    1. I disagree, at least in the case of the QS. I think the QS supporters are particularly motivated to vote this year, following from the protest movement that started this spring. I think the QS might actually do better than the polls would suggest (in terms of vote, if not in seats).

      Delete
  10. Eric how does "CONFIDENCE" work in Quebec?

    If its a minority could Charest stay on even if Marois had more seats ?

    ReplyDelete
    Replies
    1. The system is the same as in Ottawa.

      Delete
    2. I suspect if Marois had more, seats, she would become premier. Even if QS seat / seats aren't enough to get them over the hump, I dont see the CAQ propping up the liberals. I could, however, see the reverse: the liberals propping up the CAQ if they were to become the official opposition.

      Delete
    3. Technically, Charest could get a shot to form a government even if he finished behind Marois. In practice, unless the polls are badly wrong, I don't seem him trying to do that.

      But whether Marois can survive a confidence vote in the national assembly, that I don't know. If not,I could see some sort of CAQ/PLQ coalition.

      Delete
  11. As far as QS is concerned, i think their supporters will be motivated and will turnout in the 4 or 5 ridings where they have serious campaigns and some chance of winning, but in the other 120 ridings where they just have a name on the ballot I suspect they will be rattling around at "Green party-like" 3% levels

    ReplyDelete
  12. I don't agree with the way the undecided votes are distributed. IMO 18% undecided is not going to translate into 6% for the PQ. Out of the 3 leading parties the most fervent adherents are the PQ supporters. They are the ones whose minds are made up. Therefore of the 18% undecided I would conjecture that the bulk of those votes will go to either the PLQ or the CAQ. I will wager that they go to the CAQ. My prediction; a CAQ minority.

    ReplyDelete
  13. on the question of undecideds, is there any data or indications that a sizeable chunk of those people who are apathetic to the political process, and remain undecided all the way through the campaign, not bothering to vote?

    ReplyDelete
  14. Jean Lévesque30 August, 2012 08:24

    Im not a CAQ supporter, but I gotta say that 308's numbers seem to suffer from the same flaw we noticed during the orange wave; they are great when it comes to averaging multiple polls taken in a flat race or a short time frame but react too slowly to changes during a race caused by shifts in voting intention. Any sustained gains a party makes tends to build up big momentum, and its hard to argue the CAQ's got a little of that going on, but your mpdel clearly low balls it. Even 27.8%, the upper CAQ range, is lower than what they'd likely get were elections held today. This in turns screws up the whole numbers.

    The gap in francophone vote has narrowed too much for the PQ to expect a majority unless it can miraculously convince half of ON and QS supporters to vote strategically to block the CAQ.

    ReplyDelete
    Replies
    1. This is not about being slow to new trends - if you look at the poll average the CAQ is ahead of the PLQ.

      This is about an adjustment of the polls. Too many times the polls have been off and so my projections are off, but too few care to make the distinction.

      I am using a system which boosts larger parties and penalizes smaller ones, in this case the CAQ. Such a method would have predicted a Conservative majority in 2011 when all the polls said they were going to win a minority.

      Based on an analysis of past elections, this is the safe bet to make. The CAQ might prove to be an exception, but the model always makes the safest bets.

      Delete
    2. Eric's method is usually quite adept at predicting the results when the actual percentage results are used. I think this is a sign of an effective model. However, as Eric said, if the polls are off then the model will be off.

      Delete
  15. I understand and in some ways agree with the way you "penalize" small parties...I guess where the rubber hits the road is in how you define a "small party". Is CAQ a "small party" because it is a successor to the ADQ which had only 16% last election or has it crossed the threshold into being a "big party" by virtue of consistently polling in the high 20s and overtaking the ruling Liberals?

    ReplyDelete
    Replies
    1. I've been struggling with that as well, but two recent examples (NDP in Newfoundland & Labrador, Wildrose in Alberta) suggest that being a "small" party with a lot of support does not always help. There are counter-examples, of course, but the model has to go with the heavier weight of evidence.

      The CAQ is defined as a small party because it was in the opposition but not the Official Opposition. In terms of money and organization, it is certainly a "small" party when compared to the PQ and PLQ. Less so when compared to QS.

      We'll see - it would be safer to just go with the polls and blame them afterwards if it goes wrong, but I think I've shown that the seat projection model works well with the right vote numbers. So, I have to work on getting the right vote numbers.

      Delete
    2. I may be wrong, but in the last federal election didn't you often post two numbers (i) the seat predictions based on that raw results for a particular poll and (ii) the seat predictions based on your model.

      I don't know if that's still possible, but I found that to be a useful exercise. The first number (I think) tells us what that result might mean if it does accurately reflect the vote number, while the second tells us what you think the actual results will be given how you think that the polling number will translate into actual votes. Both numbers would be (were?) useful.

      A thought, for what it's worth.

      Delete
    3. I decided, during a campaign, not to do projections based on each individual poll. For one, it is a lot of extra work. And I also don't want to give too many mixed messages throughout a campaign.

      But when I post the final projection, I will do one based on the unadjusted poll average as well.

      Delete
    4. Fair enough, hey it's not like we're paying you! :) Will the unadjusted poll average be for polls over the campaign, or just the final set of polls?

      Delete
    5. The unadjusted poll average has been calculated throughout the campaign (it is in the left-hand column in the top chart). It uses the same methodology as the vote projection model, just without the turnout adjustments.

      Delete
  16. Maybe instead referring to it as a penalty on "small parties", you should also refer to it as a penalty for "new parties" (ie: Wildrose, CAQ) that are surging but which have no real electoral history...

    ReplyDelete
  17. Eric, what would be the prediction using the uniform and simple proportional methods?

    -Aalin

    ReplyDelete
  18. Eric, in your projections which average poll results for determining the percentage of votes, do you consider ignoring the last Forum poll?

    ReplyDelete
    Replies
    1. The results of today's Forum poll, even the regional results, look very similar to CROP and Léger so I will be adding them to the model.

      Delete
    2. I'm assuming Anonymous 30 August, 2012 12:28 is wondering if you'd use the previous, 'outlier' Forum results....

      Delete
    3. By the time the election rolls around, that poll will have virtually no weight in the projection anymore.

      Delete
    4. Eric, in your projections which average poll results for determining the percentage of votes, do you consider ignoring the Forum poll of August 20 which gave a lead to the PLQ (or do you already ignore it)?

      Delete
    5. Eric's point, that it will be discounted, is a good one. But even if it weren't, you can't just discard polls because their numbers seem off. Rogue polls are inherent in polling (within X% 19 times out of 20 means that occasion there's a 20th time), if you include all the polls the errors should cancel out.

      Delete
  19. It is very difficult to say who will be premier-elect of Quebec next week. I think it will be a case of all or nothing for the PQ, they either win a majority government or have the CAQ and Liberals form government in some way or form.

    If the PQ are short by one or two seats, I will be assuming the QS will prop them up for a while.

    Either way, I doubt any political party will get more than 35% of the vote, hardly a vote of confidence for any single party to form government without the support of other parties.

    ReplyDelete
  20. Éric, I see you've updated the federal averages... Which firm(s) published poll results most recently?

    ReplyDelete
  21. Anon 30 Aug 13:22:

    There will be no CAQ-Liberal coalition because neither party has the incentive to do so. No party seriously vying for power wants to be the junior partner in a coalition: if the coalition is popular then the senior partner gets most of the credit and if it's unpopular then the junior partner shares the blame.

    Only parties that have no realistic chance of winning power on their own would join a coalition as junior partner, figuring that any power is better than nothing (eg., the Lib Dems in the UK, the German Greens and FDP, or the Canadian NDP circa 2008).

    The Quebec Liberals will get spanked on Monday but unless they are nearly eliminated, they will still think of themselves as potential rivals for power with CAQ and would never want to join them in a coalition.

    ReplyDelete
    Replies
    1. GI, I hear what you're saying, and I've made that same argument about a coalition between the NDP and the Liberals federally.

      But there are some compelling counter-arguments. First, given that the CAQ and the PLQ are likely to end up with a similar number of seats, it's unlikely there will be a "junior" partner as opposed to a coalition of (more or less) equals. That changes the dynamic, they may still share the downside, but will likely get some share of the upside. And since the PLQ has significantly more experience in government than the CAQ, if they had some of their people appointed to the cabinet (which would likely be part of any deal), "their" ministers might impress relative to the CAQ.

      Second, setting aside the upshot for Legault and the CAQ (being part of the government) there are potentially significant advantages for the PLQ. For one thing, it would keep the PQ out of power. Given the craziness of the PQ platform, that's meaningful. "So what", you might say, it's a PQ minority. Sure, but if the Liberals lose badly (as polls show they will) and if Charest loses his seat and/or resigns (which seems likely) the Liberals probably won't be in a position to fight an election for another year or so (i.e., they have to pick a new leader, pay off debt, etc.) That, in turn, means that unless they're willing to prop up a CAQ government, they won't be in a position to bring down the PQ if they introduce crazy legislation in the next year or so. In that scenario, it might be better to form a coalition up front hammer that have to do it on the fly in 6 months.

      Indeed, if they put together a coalition with the CAQ that might finish off Marois, who has never been that popular with her party and has run a terrible campaign, throwing the PQ into confusion and in-fighting (could you imagine what the Tories would have done to Stephen Harper had the NDP/Liberal/Bloc coalition succeeded?).

      Finally, I think the PLQ will be desperate to avoid third party status, in light of Mulclair's proposal to create a provincial NDP. As the federal Liberals are learning, if you're not the government or official opposition, you have no visibility. I think the Liberals would have cause to worry that, outside of a coalition, they might have the lowest profile going into the next election (PQ in government, CAQ as the official opposition, and NDP getting support from their federal cousins). A coalition gives them an opportunity to rebuild without disappearing.

      I'm not saying that a coalition would be without risk for the PLQ (the CAQ could prove surprisingly competent and swallow them), but it may be the least risky alternative.

      Delete
    2. It doesn't matter how fancy you say it, the Liberals are dead one way or another. If they decide to prop up a CAQ government, they would lose more federalist support to the CAQ, because the people will no longer see them as an alternative to the CAQ or the PQ (just like what Ignatieff did before the 2011 election: Prop up the Tories, wait for the "right time" for an election, then gets reduced to third party status).

      If they decide to prop up the PQ, then federalists are going to drop them completely.

      If they dedide to continue to form a government in the case of a minority even though they don't have the most seats, then people will see them as an illegitimate government and hate them even more.

      If they do nothing, wait for the next election to happen, they would lose more support to the Quebec NDP, and more seats.

      The Liberals are over in Quebec, there will be a political realignment after this election. I can picture the main political parties in the next election to be the PQ, the CAQ, and the NDP.

      Delete
    3. "If they decide to prop up a CAQ government, they would lose more federalist support to the CAQ, because the people will no longer see them as an alternative to the CAQ or the PQ (just like what Ignatieff did before the 2011 election: Prop up the Tories, wait for the "right time" for an election, then gets reduced to third party status)."

      I'm not sure the analogy with Ignatief works. Remember, formally, the Federal Liberals opposed the Tories on pretty much everything they did, they just conveniently made sure that the government never fell. That's an untenable position, and the Liberals basically took it for 5 years. No wonder the anti-Tory vote abandoned them.

      It would have been very different, I suggest, had Ignatieff said "we'll support you on this, this and that, but only if you give is that, that and this". In fact, we know the NDP did precisely that in the fall of 2009 when Ignatieff threatened to bring down the Tories (remember "Mr. Harper, your time is up"?), and that didn't really hurt them, did it? In fact, I can think of other instances where the NDP agreed to support Tory legislation in exchange for concessions (immigration reform, after the Liberals double-crossed their own immigration critic and opposed the bill). I think that's what we'd see in a CAQ/PLQ coalition.

      And there's good reason to think that the PLQ might be able to outlast the CAQ, if they're smart (which their federal cousins have not been). The CAQ is still a pretty raw party, so may embarass itself in office (on the one hand, that's a bummer if their your coalition partner, on the other hand, if your people are competent, it may emphasize the competence of your party). While the Liberal brand is tarnished, they still have a machine that the CAQ lacks. They also still have a hard core of supporter among anglo/allophone Quebecers, which assures them a base of seats in Montreal. And the CAQ wouldn't be the first center-right political party to burst onto the scene only to disappear within a few elections (The ADQ? The Creditistes?)

      I agree that the Liberals probably couldn't form a government on their own right, but if the CAQ finishes second, I don't have a problem seeing a CAQ/Liberal coalition government (with Legault as Premier, but with senior Liberals in prominent cabinet posts). Would legitimacy be a problem? Not likely, they would represent close to 60% of the voters (and, unlike Stephen Dion in 2008, have not, to my knowledge, commited not to enter into a coalition with one another). Certainly, given that popular support for the PQ is only a few percentage points higher than that for either the CAQ or the PLQ, I don't think they'd have a better claim to form a government than a CAQ/PLQ coalition.

      You might be right, old, established parties do disappear (the federal PCs, Social Credit in BC and Alberta), but I wouldn't count on it yet.

      Delete
    4. I guess the possibility of a PLQ/CAQ coalition have disappeared. Legault has just said that he won't form an alliance with the PLQ.

      Delete
  22. I read your site regularly but I'm confused with some things.

    1) A couple of days ago, you had QS at 1 seat, maximum 2. Now you have them at 2, max 3. What happenned? Why the big changes in Laurier-Dorion? You went from a big PLQ lead to a close race with QS close behind.

    2) Still regarding QS, since you use a proportional model, and since this party is up almost 100% since 2008, why isn't QS doubled in most riding? Are you using a different model for QS?

    3) Finally, you mention that you use regional variations where it's possible. But how do you do that? I mean most polls have Mtl, Quebec and "rest". Do you use "Rest" for all the other regions?

    ReplyDelete
    Replies
    1. 1) QS had a boost in support on the island of Montreal and enough volatility in the polling to put their maximum range up in Laurier-Dorion, above the minimum ranges of the PLQ and PQ.

      2) QS is doubled in some parts of the province, other parts they are not. But it isn't as simple as doubling - the vote totals still have to add up to 100%. If a party was at 51% in a riding and the party doubled its support, the model wouldn't give it 102%.

      3) Leger and CROP provide details for the island of Montreal, the suburbs, and Quebec City. So those regions are covered. As explained in the write-up with the projection, the other regions of Quebec are projected using the Rest of Quebec numbers, as if each of the three regions were one big riding. The same system is used. But Léger often has that regional data available as well.

      Delete

COMMENT MODERATION POLICY - Please be respectful when commenting. If choosing to remain anonymous, please sign your comment with some sort of pseudonym to avoid confusion. Please do not use any derogatory terms for fellow commenters, parties, or politicians. Inflammatory and overly partisan comments will not be posted. PLEASE KEEP DISCUSSION ON TOPIC.