Thursday, April 7, 2011

Conservative majority

Four new polls this morning. Four! In 2008 this would have been nothing special, but in 2011 this is a deluge. However, new polls don't always mean new changes, and this set of polls has had little effect on the projection. But we do have a few seat swaps, and the end result is a Conservative majority government. But only just.
The four polls added this morning are from EKOS Research, Nanos Research, Angus-Reid (at least what could be culled from incomplete media reports), and Environics. That last poll was provided to me personally, and a release should be up on their site soon.

The end result is a seat gain for the Conservatives at the expense of the Liberals, and a Liberal gain at the expense of the Bloc Québécois. This keeps the Liberals at 71 seats and the NDP at 32 seats, while the Bloc has dropped to 50 and the Conservatives have risen to the magical 155-seat mark.

In terms of popular support, both the Conservatives and Liberals are unchanged at 38.6% and 27.6%. You'd expect some changes with these four polls, but they peg the Tories at 38%, 37%, 39.6%, and 38% and the Liberals at 25%, 27.8%, 30.4%, and 27%. Average them out and you can see why the national vote projection has not changed.

The New Democrats, however, are up: 0.2 points to 17%. The Bloc Québécois is down 0.2 points to 9.4% (thanks in large part to an exceptionally low result for the party in the EKOS poll), while the Greens are up 0.1 points to 6.4%.

Regionally, the Conservatives made some good gains in Alberta and the Prairies, regions where they are already mostly maxed out. They dropped in British Columbia and Quebec, and were relatively stable in Ontario and Atlantic Canada.

The Liberals jumped in British Columbia and re-took second place in Quebec from the Tories, but dropped by a significant amount in Alberta and Atlantic Canada.

The New Democrats were relatively stable, except in Quebec where they gained a full point. That's big in the projection. They've never been this high in the province before. The gain came mostly at the expense of the Bloc, which dropped a point to 37.8%, below their 2008 level of support. But the NDP is not yet projected to win more than one seat in the province.

The two seat changes took place in Ontario and Quebec. The Conservatives have edged ahead in Kingston and the Islands in the projection, and now Alicia Gordon is favoured over Liberal candidate Ted Hsu. In Quebec, the Liberals and their somewhat obscure MP Justin Trudeau are in front again in the riding of Papineau.

Though the Conservatives have been picking up a seat or two each day lately, I'm getting the impression they are currently riding the top of a wave. While it could, of course, go either way, I would not be surprised if this is as high as the Tories get in my projection. We shall see.

30 comments:

  1. I think this projection is fanciful, to put it mildly. The Conservatives are essentially at the same level as the 2008 election (38%) while the Liberals are GAINING significantly over their 2008 numbers. There's no way this can translate into more seats for Harper.

    ReplyDelete
  2. First, I commend you on maintaining an outstanding Blog. It si one that I am reading everyday now.

    I actually do expect a small Tory increase tomorrow in the NANOS Poll as one very heavily Liberal day of polling gets dumped.

    I do not know what has happened to Ipsos-Reid which had a pre-election poll and then nothing for two weeks now...

    ReplyDelete
  3. Anonymous,

    Looking at the national numbers is not the way to go. The Conservatives have gone from having a five-point lead in Ontario to having an eight-point lead. That is hugely significant.

    And combined with the much better numbers in Atlantic Canada, including a lead in Newfoundland after being swept from the province in 2008, this gives the Tories the 12 seats they didn't have in the last election.

    ReplyDelete
  4. And the Liberals are only up 1.3 points to the Conservatives' 0.9 points nationally, so I disagree that the Liberals have gained "significantly", especially considering you also said the Conservatives are "essentially at the same level".

    ReplyDelete
  5. Is it just me or do the projections by seat seem to be really the Cons "best case" results. They seem to be winning every close riding in the projections - meaning that everything would have to go perfectly on election night to win 155 (assuming of course that these projections do not change over the next few weeks). I could be wrong though, I have not looked at every riding to see if they are losing a lot of close races as well.

    ReplyDelete
  6. Eric why would you ruin decent analysis with a wishful thinking paragraph at the end. Just put out the numbers.

    And we already know the Conservative high is actually around 43%, not 38.7% so the numbers don't even back up you wishful thinking.

    Also interesting they didn't drop on the Ekos poll. Why is that interesting, the Conservative VERY good night of Sunday was dropped from the three day average for today's numbers. Since they stayed the same, that means they had another VERY good night last night.

    The Liberals very good Monday night on the Ekos drops ff tonight. We'll see what happens to the numbers tomorrow.

    ReplyDelete
  7. Eric, could you take a look at Elmwood-Transcona in Manitoba. It was really close last time when the Conservatives ran a star candidate (former Winnipeg Jet Thomas Steen) but I don't think it will be nearly as close this time. Steen got elected to city council last year and I've never heard of the Conservative candidate this time.

    ReplyDelete
  8. The Bloc are polling less than they got in 2008 and you have them gaining 3 seats.

    I have them winning 12 seats in 2008 by less than 15% (by far the most close wins in Quebec).

    They must be getting to a threshold where if they drop a bit more it will be a drop of a half dozen seats rather the pick up.

    Duceppe might have start campaigning as a separatist rather than a part of the Federalist coalition.

    ReplyDelete
  9. Anonymous,

    I don't know why you would think it is "wishful thinking". It is my observation based on my looking at polls every day. There are indications here and there that Liberal numbers are inching up. It's not wishful thinking - my only wish in this campaign is that my projection will be right!

    Brad,

    It would have been a good thing to study the effect in subsequent elections after a star candidate has run for office (and failed). But the Tories' numbers in the riding were still pretty good in 2006, so I think that they have a decent chance of holding on to Steen's numbers.

    This is one of the ridings I've been watching as well.

    ReplyDelete
  10. Hi Eric...multiple polls are the "fog of war" of election campaigns and they often don't give a clear or compelling picture. One thing I don't really like is NANOS tendency to take a poll result and then attribute reasons for it, often when it can just be static or movements within the MOE. Commentary suggesting that expelling a couple of kids from a rally can possibly move national numbers seemed inplausible to me regardless of party unless there were no issues of any kind in the campaign. In any event, there are a number of tight races that you still have placed in the opposition (Eglinton-Lawrence and Don Valley West for the Liberals and NDP seats in BC...I would not regard your take as "fanciful" as you are moving only tyhe exceptionally tight seats last time to the Tory column..

    ReplyDelete
  11. Anonymous said...
    I think this projection is fanciful, to put it mildly. The Conservatives are essentially at the same level as the 2008 election (38%) while the Liberals are GAINING significantly over their 2008 numbers. There's no way this can translate into more seats for Harper.

    07 April, 2011 10:06

    I reckon its because the support in Ontario has increased but has slipped slightly in BC. Tory support in BC slipping would not really affect seat numbers there for them so the increase in seats is clearly from Ontario.

    ReplyDelete
  12. BCVOR,

    The Bloc is down 0.4 points from 2008 in Quebec. The Liberals are down 3.4 points and the Conservatives are down 1.5 points.

    Yes, I have the Bloc picking up three seats, though one of them is the vacant seat that was previously held by a Bloc MP. One of the other seats was held by an independent, and the other is a Liberal seat.

    ReplyDelete
  13. Can you comment briefly on how the riding projections are worked out. For example, I am looking at my riding (North Vancouver) and while it was won by 5 points by Cons last time, your projection has Libs picking it up with an 8 point MOV. Clearly a 13 point swing is much larger than what has occurred at the provincial level in the polls, so what riding specific data is factored into these projections? Even in previous elections when a Lib won the riding it wasn't with that strong of a showing, and that is when Libs were forming government.

    ReplyDelete
  14. Colin,

    A full explanation can be found here:

    http://threehundredeight.blogspot.com/2011/03/methodology-of-projection-model-and.html

    ReplyDelete
  15. Eric, last time I looked you had no seats in Vancouver going to the Conservatives. I thought to myself at the time wow those numbers are low. I think they are going to win some seats.

    Full disclosure I live in Vancouver Kingsway. The Ekos poll for Vancouver has the Conservatives at 42% and the Liberals at 25%. I think that means the Conservatives are going to win a seat or two in Vancouver. Yes there must be a big MOE, and Vancouver can include North Van and West Van, but I think they mean Vancouver.

    Yes BC voters are the most fickle in the country and can swing on a dime in bizarre ways, aka from NDP to Conservative or back again, but I still think the Conservatives are taking at least two seats in Vancouver.

    Eliminate Vancouver East (NDP) and give the win to Joyce Murray in Quadra, and you have an even greater likelihood of Conservative wins in the other seats, Van South, and Van Kingsway. Van Centre is probably a toss up, not the romp you show.


    Also, as far as wishful thinking goes, Dion came within 3 points of the Conservatives at one point in the campaign last time around. Predicting that the Conservatives are at their ceiling, not even two full weeks in, is wishful thinking Eric, not statistical analysis. It's no less wishful thinking than saying the Liberals are at their ceiling, or low point. They aren't.

    The Conservative ceiling shown in any poll is 43%, without the actual turnout bias on election day that shifts to them demographically, as well as the inevitable third party vote collapse on election day we generally see in Canada. Which means their real ceiling is about 45%. I'm not saying that is going to happen, I'm saying that's the ceiling of their vote we have seen from the polls and historic voting day shifts we see in Canada.

    BTW if you believe the Greens will get anywhere near 9.8% on election day, 2% more than last time, with no profile, no debate, no environmental issues in the campaign, I have some swamp land to sell you in Florida.

    Why EKOS does this I don't know, but it really makes their polls lack credibility. Take away 5% from this number and distribute it as you see fit. Another reason why this "ceiling" for the Conservatives is wishful thinking.

    ReplyDelete
  16. Peter. I live in Van Kingsway and work in North Van.

    The Conservatives are taking at least two seats in Vancouver, plus North Van.

    The Ekos poll has the Conservatives at 42% and the Liberals at 25% in Vancouver and Frank Graves is no friend to Conservatives. He regularly underestimates their vote and overestimates the Greens by a wide margin.

    I just sent Eric a note about his projections for Vancouver.

    ReplyDelete
  17. Thanks Eric.....

    My interpretation of your answer:

    The BLOC is benefiting from the NDP federalist vote splitting and like anyone running against a fractured opposition has to hope that the vote doesn't coalesce against them.

    They are almost exactly in the same position as the CPC are in on the national basis.

    The NDP is helping them both.

    ReplyDelete
  18. EKOS city results are for the CMAs, not the city boundaries proper.

    As for wishful thinking, again, it is not. You are assuming that I am hoping I am right. That is not the case. You can call it idle speculation if you want, but "wishful thinking" is not accurate. I don't think the Liberals will improve because that is what I hope will happen, I think the Liberals will improve because that is what I think will happen. Hope does not come into it at all.

    As to ceiling, I never said the Conservatives hit their ceiling. I said I think they are at the top of their wave, meaning I think they will drop back down. Sure, they could hit 43% thanks to the MOE in a single poll, but that is something very different.

    ReplyDelete
  19. Woe, Canada...we're in a lot of trouble.

    ReplyDelete
  20. As per the projection for Kingston and the islands:

    Don't count on it going Conservative. It may be a close race now, but it will stay Liberal on election day. The Cons have always tried to buy a candidate from the local market there, but they never find a good one.

    DW

    ReplyDelete
  21. I know that it is difficult to argue the projections on individual ridings but I have to agree with the posters above who have said that North Vancouver will likely go to the CPC. In the last election the CPC candidate won over a popular Liberal who was a former MP and mayor. This time around the CPC incumbent is running against a lower profile newcomer.

    ReplyDelete
  22. @Anon (10:58)

    My count has 21 ridings where the Conservatives lead by under 5 points, and 10 where they trail by under 5 points.

    ReplyDelete
  23. Anon (10:58 Here). Thank you BT. That is a +11 ridings where the Conservatives would need things to go their "way" on Election Day. Obviously its possible that they get all 21 and even some of the 10 where they trial to push that number past 155 but that I think would involve a lot of momentum.

    Eric, I really like this blog and come every day. How do you arrive at the individual riding totals? Are those based primarily on the movement of the national poll numbers?

    ReplyDelete
  24. Obviously, the trend for all political parties is the same. No major changes could happen because the strategy from Cons is not debate with anyone and anything in this campaign. However, when the debate on TV is coming soon the results will be more realistic. Current findings in the polls are only taking into account people that know for whom is voting. Nevertheless, 15% of Canadians haven’t decided for whom they are voting. Moreover, if you add the margin of error (average) we are talking about 25% of Canadians that haven’t decided their vote. Another matter aspect, people voting for tories don’t change preferences but people voting for liberals and NDP yes. This means that they could vote for the best candidate able to stop conservatives. And this strategy is not in the polls. In short, conservatives have touched the ceiling and so they are not going over 38%.

    ReplyDelete
  25. This is getting interesting! Eric, could you perhaps expand a little on why you feel the Tories *might* be cresting? It's obviously sort of a hunch and I'm not one to dismiss that, having greatly enjoyed Malcolm Gladwell's "Blink," but that's obviously what's contributing to the "wishful thinking" meme, here.

    ReplyDelete
  26. Obviously the EKOS poll has huge numbers for the Greens (as usual), with them taking 2 seats in BC, but has anyone else looked at the Bloc numbers? 28% in Quebec? That's a horrible number for them. I know there's been polling from other firms (Abacus, for example) that showed Bloc support was very soft, but have any other pollsters been showing numbers this low?

    BCVoR - I don't think it's just a matter of the NDP taking federalist votes from the Liberals and the Tories. I think there's a fairly significant component of the NDP taking progressive separatist votes away from Bloc as well. That's a change I'd personally applaud, and I suspect you would too, eh?

    ReplyDelete
  27. Anon 10:58,

    The explanation for the riding projections can be found here:

    http://threehundredeight.blogspot.com/2011/03/methodology-of-projection-model-and.html

    In short, they are based on regional poll numbers, various individual factors, and riding polls when available.

    ReplyDelete
  28. What's with North Vancouver, Cons are incumbents, Lib have a no name candidate, Dippers have a known candidate but not hope of gaining ground.
    You have the Cons losing to the Libs. I'd be surprised if this is the case.

    ReplyDelete
  29. As far as I know, pollsters correct for different turnout rates for Conservative voters vs. Liberals, NDP and Greens, right? So if the raw numbers show the Tories at, say, 34, they'll tick them up a percent or two (or rather bring down the others) to correct for the difference in turnout. But what if the net effect of all of the scandals and bad press the Conservatives are getting is not an increase in support for the other parties (which doesn't yet look like happening in any great detail) so much as an increase in Tories staying home on election day? That might wind up skewing the election results in a wholly unpredictable fashion.

    ReplyDelete

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.