Tuesday, December 7, 2010

Conservative gains in new Nanos poll

Nanos Research has blessed us with a new poll, showing change within the margin of error that increases the gap between the Conservatives and the Liberals to seven points.At 38.1% support, the Conservatives have increased their score from Nanos' last poll in early November by one point. The Liberals have fallen 0.4 points to 31.2%, while the New Democrats are up 1.8 points to 17.2%.

The Bloc Québécois is at 10.2% nationally, while the Greens are down two points to 3.2%.

While I always look suspiciously at polls that put the Greens at over 10%, it would be hard to believe that the Greens are also below 4%.

In Ontario, the Conservatives are up 1.4 points and lead with 42.3%, followed by the Liberals at 35.4% (down 0.1). The NDP is up 3.1 points to 19.3%, while the Greens are down 4.3 points to 3%.

In Quebec, the Bloc is down 2.7 points to 40.1%, but still leads the Liberals, who are at 26.7% (+0.5). The Conservatives are down one to 18.3% and the NDP is up 2.2 points to 12.7%.

The Conservatives are up a huge (and MOE-influenced) 17.7 points, and lead with 49.4% in British Columbia. The Liberals follow with 25%, down 8.1 points, while the NDP is down 3.8 points to 22.3%. Dropping 5.8 points, the Greens won't elect Elizabeth May with 3.3% support in the province.

Nanos still has the Conservatives ahead in Atlantic Canada with 43.5%, up 1.1 points. The Liberals are up 3.3 points to 36%, while the NDP is down 4.1 points to 17.5%.

In the Prairies, which includes Alberta in Nanos' calculations, the Conservatives are down nine points to 46%. The Liberals are up three to 34% and the NDP is up seven to 16%. Assuming the Liberals are at 20% or so in Alberta, they must be doing very well in Saskatchewan and Manitoba to end up at 34% overall.

With this poll, the Conservatives would win 25 seats in British Columbia, 56 in Ontario, seven in Quebec, and 12 in Atlantic Canada. Giving them the 27 seats in Alberta and the 20 seats in the Prairies that I currently have them projected to win gives the party a total of 148 seats, well short of a majority, but 11 more than they were projected to win in Nanos' last poll.

The Liberals would win eight seats in British Columbia, one in Alberta, five in the Prairies, 39 in Ontario, 16 in Quebec, and 18 in Atlantic Canada for a total of 89, three less than last time.

The Bloc would win 51 seats in Quebec, unchanged from Nanos' last poll.

The NDP, despite national growth, would drop eight seats from the last poll due to the increase in support for the two main parties in British Columbia, Ontario, and Atlantic Canada. The NDP would win three seats in British Columbia, none in Alberta, three in the Prairies, 11 in Ontario, one in Quebec, and two in Atlantic Canada for a total of 20.

Aside from an improbable jump in support in British Columbia, there is not much in this poll that is different from Nanos' last report for the Conservatives. It does confirm a Conservative lead in Ontario and adds weight to their findings that the Tories are doing very well in Atlantic Canada. The Liberals shouldn't panic just yet - 31% is better than their 26% in 2008, and this poll has them competitive in every part of the country, including Alberta, Saskatchewan, and Manitoba.

34 comments:

  1. Eric said: "While I always look suspiciously at polls that put the Greens at over 10%, it would be hard to believe that the Greens are also below 4%."

    Actually, I thought that was one factor that made this poll particularly credible. I mean, let's face it, 2008 involved perfect conditions for the Greens. The Liberals made the environment one of the main campaign issues, they gave the Green leader a ton of free publicity (and an endorsement, of all things)and Liz May was seen as something new and interesting and got a ton of free (and often quite uncritical) coverage from the media. And they got less than 7% fo the vote.

    Now, in 2010, the environment is off the map as an environmental issue (witness (i) the absence of fuss when the Tories killed that Kyoto bill - as brain dead as it was - in the Senate without even a debate, (ii) the Government explicitly saying that it won't renew Kyoto, (iii) the fall in significance of the environment in the recent nanos polls). Liz May has basically no public profile (other that kurfluffle last summer about irregularities in the internal governance of the Greens - hardly the publicity you want) and the Liberals are unlikely to repeat the favour of endorsing her.

    Based on those factors I wouldn't expect them to repeat their 2008 success. 3.5%, while on the lower side, would be consistent with the pre-2008 level of support for the greens (i.e., 4-5%).

    ReplyDelete
  2. I agree that the low Green number is perfectly believable for all the reasons Carl mentions. On top of that - while we don't want to read too much into byelection results - normally a byelection ought to be a perfect opportunity for a small party to concentrate its resources and at least increase its popular vote. In fact in early 2008 when the Greens looked briefly like they might be establishing some momentum, the Greens actually had some decent byelection performances getting double-digit support in Toronto Centre and Vancouver-Quadra and beating the NDP in Willowdale. But since the 2008 election, we have now had SEVEN byelections and the Green percentage of the vote has been 3%, 2%, 3%, 4%, 1%, 1% and 5% - in other words an average of less than 3%.

    ReplyDelete
  3. It's pretty tough to take seriously any poll that so drastically samples out of touch with the demographic realities.

    Using Ontario as the benchmark, the west is over sampled by 40%, Quebec is oversampled by 27% and the Maritimes are oversampled by 235%.

    How anyone can argue these numbers are representative when they can't even follow the basic population figures of those regions is beyond me.

    This can only DOWNPLAY the will of the bulk of the voters in Ontario and OVERPLAY any parties that have regional blocs in the areas being bumped, ie. CPC in the West, BQ in QC. Even in Atlantic Canada, the CPC lead, so this poll is unnecessarily boosting their support figures.

    ReplyDelete
  4. Almost certainly, Nanos weighs the samples at the national level to give each region the correct weight.

    ReplyDelete
  5. Eric, the big problem with applying your seat projection model to this Nanos poll is that you essentially exclude the Prairies because Nanos lumps Alberta in with Man/Sask. - the trouble is that this is the one region where this poll indicates a major drop in Tory support compared to the last election.

    What was the popular vote in 2008 for the three Prairie provinces combined? I suspect that Tory support would have to have been in the high 50s since they had over 60% in Alberta and over 50% in Man/sask.

    ReplyDelete
  6. Eric,

    Gati actually raises an interesting point. It looks like the national numbers are just the averages of the regional numbers weighted according to their sample sizes. Maybe Nanos is doing something very subtle, but if there's a regional weighting in there it isn't obvious.

    However, weighting the regional results to reflect the actual population distribution of the regions doesn't appear to substantively change the national number as between the two big parties (it becomes 38.2 vs. 31.6).

    Moreover, since your seat analysis goes off the regional numbers, rather than the national one, it shouldn't affect you one way or the other. Indeed, because elections are decided at the local, rather than the national, level, its the regional data that is more relevant than the topline national number.

    ReplyDelete
  7. Carl, indeed. My national popular vote projection is just for informational purposes - it doesn't play a role in any of the regional projections, other than for the North.

    ReplyDelete
  8. Agreed. Nik Nanos knows what he's doing.

    by sampling a greater number of people in those other reagions, he;s reducing the margin of error in his regional numbers. I'm really happy about that.

    And, as always, I've always found Nik's Green numbers made the most sense. He's been very accurate with them in past elections.

    ReplyDelete
  9. I'm not sure what "gaki" means about all these regions being oversampled. If you look at the sample size by region - it strikes me as highly proportional - Atlantic Canada is a little less than 10% of the population of Canada and he has 76 interviews there out of 700 odd interviews with decided voters across Canada - so what's the problem?

    ReplyDelete
  10. The Green at 3% makes perfect sense but might be a bit high.

    Global warming, which was the sole platform of the the Green in the last election has become an embarrassing issue.

    The MSM, Movie stars, any scientist worth his grant money are all stumbling over each other to get off the band wagon. The Cancun/Kyoto Global warming party is something that politicians of all parties are running away from.

    In the by-elections, which are notorious as protect votes that the Green should do good:

    They got .7 % of the vote in Winnipeg North down from 4.8% in 2008 a drop of 85%.

    They got 1.2% of the vote in Vaughn compared to 6.9% a drop of 82%.

    In Dauphin they only went down from 6.5% to 5.4 %


    There should be no chance of having Elizabeth May included in any of the debates no matter what format it is.

    Well maybe she could debate Dion and Suzuki.

    In the polls were they are getting 10%, it would be cut in half if the Pirate party was added as an listed option. Cut in a 3rd if Pirate and marijuana were listed along side the Green.

    ReplyDelete
  11. Ira

    And, as always, I've always found Nik's Green numbers made the most sense. He's been very accurate with them in past elections.

    it is to laugh at feelings as opposed to facts.

    Oct 10 Nanos poll had the Green at 8%.

    The 2 panic polls he ran on the weekend after the campaign stopped he had them at 9%.

    They got 6.8 on the Oct 14 election poll. You know the one where the sample size is 13.8 million

    Nanos is pretty accurate with predicting the Green vote if you feel that over estimating their support by 24% is spot on.

    I guess you are comparing to Ekos who had the Greens at 11% on Oct 10, a 61.7% over estimation.

    ReplyDelete
  12. Unfortunately, climate change is not a problem that will be solved any time soon. The people driven to vote for the Greens in 2008 on that issue are likely to vote for them again, not to mention that the Greens are the only legitimate option for people wanting to vote for a non-traditional party. The apathy we see in politics today indicates that the Greens could get a lot of "none of the above" votes if those people still want to express their democratic voice.

    But the environment as an issue isn't going away. Just yesterday Nanos found that 8% of Canadians see it as the top issue. That looks like a good and reachable target for the Greens.

    ReplyDelete
  13. There are a lot of people who see the environment as the number one issue who will vote NDP or Liberal or even BQ. The Green party has to split that 8% with three other parties.

    ReplyDelete
  14. The environment as an issue would also include cutting down pollution and acid rain and creating more parks and cleaning up lakes.

    The CPC are declaring a new national marine park in Lancaster Sound on Baffin Island.

    The WWF and other environmental groups are ecstatic.

    The Green party does not own the 8% of the people concerned with the environment.

    They scare away many people who are truly concerned about the environment by focusing on the Global Warming wealth transfer scam as their reason for being.

    The recently elected CPC MP from Dauphin heats his house with wood he chops and plants the trees to replace the ones he uses. Sopuck, lives on a rural property in the Sandy Lake area, is a former fisheries biologist who also served as then-premier Gary Filmon’s environmental adviser for eight years. For the past two decades he’s been involved in a wide variety of environmental and outdoors initiatives, acting as director of environmental programs for the Pine Falls Paper Company and serving on the Canadian delegation at the Earth Summit in Rio de Janeiro in 1992.

    Sopuck is a former vice-president of the Delta Waterfowl Foundation, a former board member of the Manitoba Habitat Heritage Corporation and a widely published outdoors columnist.

    He has all the credentials to be a Green party candidate but he was actually elected and has a voice in government.


    The Green as a protest vote? It has the same impact as a ruining your ballot and not as much impact as voting for the Pirate or Rhino party.

    ReplyDelete
  15. DL, the Prairie vote in 2008 was:
    Con 59.2%, NDP 17.6%, Lib 13.2%

    ReplyDelete
  16. Thanks Brendan. To compare, if I grouped Alberta with the Prairies, my current projection would be about 54% for the Conservatives, 21% for the Liberals, and 15% for the NDP.

    So, for the Nanos projection it would likely mean an extra seat or two for the Liberals and less for the Tories.

    ReplyDelete
  17. http://www.torontosun.com/news/canada/2010/12/07/16456936.html

    On Tuesday, Abacus Data Inc., a new Ottawa-based polling firm, said the Conservatives enjoy the support of 35% of Canadians. The Liberals trail at 24% and the NDP are at 20%.

    Abacus conducted an online survey of 1,361 Canadians between Dec. 3-6. The pollster says its results are accurate to within 2.7 percentage points, 19 times out of 20.

    .............

    Abacus???

    oh well.

    Nov. 4
    http://abacusdata.ca/wp-content/uploads/2010/11/Abacus-Data-Vote-Choice-November-4.pdf

    33-25-21-10-10
    (8 point lead)

    And Dec 7s
    35-24-20
    11 point lead....

    ReplyDelete
  18. DL, small(est) parties tend to do worse in by-elections than general elections. This is because turnout is lower, and since the smallest parties also have the lowest percentage of strongly committed voters, their vote goes down in greater proportion.

    I have no doubt that if there was 90-100% turnout in the next federal election, the Greens would get over 10%. That is what polls that prompt party names are measuring. Of course, people who have trouble naming a party without prompting tend not to vote, and so polls that don't prompt tend to me more accurate predictors of an election.

    ReplyDelete
  19. I must be bored today....

    Found an alberta one by environics too.


    PC 34
    WRA 32 (+4)
    Lib 19 (-4)
    NDP 13 (+1)


    http://www.calgaryherald.com/news/Poll+Wildrose+support+grows+Albertans+worry+about+health+care/3931855/story.html

    ReplyDelete
  20. BC Said: "Nanos is pretty accurate with predicting the Green vote if you feel that over estimating their support by 24% is spot on"

    Actually, BC, the final green results were within the margin of error of the last nanos poll, so statistically, they were right on.

    ReplyDelete
  21. Eric said:
    "But the environment as an issue isn't going away. Just yesterday Nanos found that 8% of Canadians see it as the top issue. That looks like a good and reachable target for the Greens."

    Well, that's about the same number who ranked the environment as their top priority in 2006 when the Greens got less than 5% of the vote (and down from 20% in 2008, and even higher in 2007). That just highlights the point that the Greens are in much more hostile territory now than they were in 2008.

    ReplyDelete
  22. This poll has the Green at 3.2% The MOE for the sample of 747 committed voters is 3.6%.

    That would mean that the actual range for the Green party is 6.8 to -.4%

    The margin of error for the support of the Pirate Party who has 0 support is 3.6 % to -3.6%.

    So in theory the Pirate party is in a virtual tie with the Green Party.

    The Pirate Party may have a 4% point lead on the Green Party ahead 3.6% to -.4%.


    I hope the MSM takes this to heart and arranges a debate between Mr. Ignatieff and a Pirate representative like they forced Mr. Harper to do with Ms. May.


    The MOE breaks down on the fringes.

    ReplyDelete
  23. BC, the MOE doesn't "break down" at the fringe. It's just that, at the fringe, polling data isn't all that meaningful when the true level of population support is within the margin of error. It's not totally useless, though. We can, after all, rule out the possibility that the Greens are at 7% and that the Pirates are at 4%.

    ReplyDelete
  24. Carl: I mean, let's face it, 2008 involved perfect conditions for the Greens. The Liberals made the environment one of the main campaign issues...

    Actually, Greens are very happy--politically--that this isn't going to happen again. The impressive thing was how well the GPC did when the historically strongest party in the country's history was camping on a major part of the Green platform. The Green vote was artifically depressed.

    It's unfortunate that Dion delivered the worst of all possible worlds. He started with a good, simple idea, cutting taxes by charging for the external cost of carbon, and messed it up so badly that he made it toxic. We're seeing echos of this in Dalton McGuinty. Well-meaning, both of them, but their ability to execute is not only painful to watch; it's dangerous to bystanders, i.e., the planet.

    ReplyDelete
  25. Carl: That just highlights the point that the Greens are in much more hostile territory now than they were in 2008.

    No, as explained above, this means that Greens own this territory.

    This, by the way, is why we're seeing John Baird pronouncing on Lancaster Sound. That's tough sledding, though, when the entire globe is calling you out in Cancun.

    ReplyDelete
  26. "This poll has the Green at 3.2% The MOE for the sample of 747 committed voters is 3.6%.

    That would mean that the actual range for the Green party is 6.8 to -.4%"

    No, margin of error doesn't work that way. When people say the margin of error is - say 3.2% - that means 3.2% if the numbers are 55-45 or 50-50...the margin of error is actually much smaller if the results of a question are - say 95%-5% then the margin of error shrinks to about 1%.

    Its too complex to explain here - but there are tables in any stats textbook the explain it all

    ReplyDelete
  27. John,

    Hey, if the Greens are happy with their go-to issue not being on the national agenda, more power to them, though that probably says more about their lack of political sophistication than anything else. Typically you want "your" issue to be front of mind for voters. That was the case for the Greens in 2007-2008 (when Nanos had "the environment" as the top priority for 20-30% of Canadians), when, not coincidentally, the Greens had their best result ever. It isn't the case anymore, and for the Greens, that can only be a bad thing.

    As for the claim that the Lancaster sound announcement has anything to do with the Greens, that's just wishful thinking. The Conservatives have always been responsive to a core of "conservative" environmental issues, often driven by the rural wing of the party (particularly farmers and hunters and others who have close ties to the land) and linked to conservation of the natural environment (we saw that during their first term when they created a number of large national parks). Moreover the current Tories have a curious fascination with Canada's north. The Lancaster sound announcement flows from those sources and has nothing to do with the Greens or the self-described "environmental" movement.

    ReplyDelete
  28. The provincial NDP in NOva Scotia seems to be rebounding in popularity:

    http://www.cra.ca/en/home/Newsroom/SatisfactionWithNovaScotiaNDPGovernmentRebounds.aspx

    ReplyDelete
  29. Eric wrote:

    "not to mention that the Greens are the only legitimate option for people wanting to vote for a non-traditional party."

    I have to strongly disagree with this statement.

    Both independents and candidates for other minor parties contest elections. To suggest that they are not a "legitimate" option is insulting.

    ReplyDelete
  30. Nothing against independents or minor parties, but voting for them is not widely seen as a potential option. It also makes much less of a statement - few people notice votes for independents (aside from well-known independents) or minor parties. If the Christian Heritage Party, the Communist Party, and the Marijuana Party all increase their vote by 0.3 points, it will not be given as much notice as the Greens increasing their vote by one percentage point, for example. And if the Greens get to 10% without a seat, things could come to a head.

    ReplyDelete
  31. Eric,

    I don't entirely disagree with those observations.

    It is true that Green candidates received more than five times more votes than candidates for all smaller parties and independents combined.

    But that raises the question of how long the Green Party can sustain the interest of "people wanting to vote for a non-traditional party".

    The "protest vote" phenomenon does not last forever; a party that is recently on electors' radar screens does not seem "fresh" indefinitely.

    Without loads of free, uncritical media coverage and without their single-issue strength being at the top of voters' minds, it's not impossible to believe they could drop substantially.

    ReplyDelete
  32. I don't disagree with that either. The Greens could see a drop in support in the next election. There are plenty of reasons why that is possible. They might also see a growth in support, and there are plenty of reasons why that is possible, too.

    More than the other parties, the Green vote is more about speculation because they don't have comparable GOTV organizations or media coverage.

    ReplyDelete
  33. @ DL

    "I'm not sure what "gaki" means about all these regions being oversampled."

    Just look at the n values for yourself. Out of 822 total samples, pure sampling by regional population would break down in the various regions as follows (actual n values from Nanos in brackets):

    BC: 4.5/33.8 * 822 = 109 (119)
    Prairies: 5.9/33.8 * 822 = 143 (165)
    ON: 13.3/33.8 * 822 = 323 (254)
    QC: 7.8/33.8 * 822 = 190 (207)
    Atlantic: 2.3/33.8 * 822 = 55 (80)

    A little safer statement might have been to say that Ontario gets undersampled relative to the rest.

    All the same, I can't see how you can oversample the prairies and undersample Ontario and not give a boost to the CPC in some fashion, given their regional strength in the prairies.

    ReplyDelete
  34. Carl: Hey, if the Greens are happy with their go-to issue not being on the national agenda, more power to them, though that probably says more about their lack of political sophistication than anything else. Typically you want "your" issue to be front of mind for voters.

    Not quite. Typically, you want a "wedge" issue where you have the right idea and the other parties have the wrong one. Dion made it hard for the Greens to wedge on the environment. This time around, they own the territory.

    The environment is issue #3 in tough economic times. That puts it front of mind for a substantial number of voters and next in line for many more. I can live with that degree of sophistication.

    As for the claim that the Lancaster sound announcement has anything to do with the Greens, that's just wishful thinking.

    John Baird hasn't confided in me recently but you're probably right; Greens weren't the driver behind the Lancaster Sound announcement. Its main purpose was to divert attention from Cancun.

    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.