Tuesday, October 16, 2012

Liberals surge in Nanos poll

Yesterday was a big day for Liberals, with a poll being released showing the federal party in second place nationwide and Dalton McGuinty announcing his plans to resign as leader of the Liberal Party of Ontario. Both pieces of news were met with a lot of surprise and wild speculation. That McGuinty will run for the federal leadership is about as likely as this new poll marking a real change in voting intentions in Canada - though neither is impossible.
Nanos Research was last in the field Sept. 4-9, and since then the Conservatives gained 0.9 points to hit 33.3% support. The Liberals picked up 5.5 points to move into second place with 30.1%, while the New Democrats were down 2.5 points to 27.9%.

The Bloc Québécois was at 4.7% and the Greens slipped 2.1 points to only 2.9%.

This poll shows a real three-way race, as the edge the Tories have over the Liberals is within the margin of error, as is the edge the Liberals have over the NDP. The shifts in support towards the Liberals and away from the Greens are outside of the margin of error, but caution should be exercised before Liberals start taking measurements for the curtains at 24 Sussex.

There are many innate, unintentional methodological reasons why a polling firm might return better results for one party than the consensus opinion of other firms. And the consensus, thin as it is, does point to an uptick in Liberal support since Justin Trudeau's candidacy for the party leadership was announced. But Nanos Research has had most of the best results for the Liberal Party since the May 2011 election.
Of the top 12 poll results for the Liberals since the last election, seven of them were from Nanos Research, including the top six. No other firm has had the Liberals at more than 26% since that time. That is not to say that Nanos has it wrong - they only over-estimated the Liberals by less than two points in their final poll of the 2011 campaign and were off of the Liberal result by 0.1 percentage points in the Ontario provincial election.

All this does suggest is that Liberal support in this poll may be inflated relative to the polling consensus. Again, this does not mean they are certainly not in second place nationwide, but rather that this apparent surge is less informative than it would be if it was coming from a different polling firm that has had traditionally low Liberal results.

Nevertheless, the poll does hint that Trudeau is having a positive effect on the Liberal Party's fortunes. But other surveys will be needed to confirm if the party is indeed enjoying a surge in support.

Regionally, the New Democrats still hold a lead in Quebec with 35.7%, putting them ahead of the Liberals at 25% and the Bloc Québécois at 17.8%. The Conservatives, meanwhile, have the lead in the Prairies (which includes Alberta in this poll) at 48.9% to 28.9% for the NDP (a gain of 10.5 points) and 20.4% for the Liberals.

The Liberals have the advantage in two provinces: Ontario and British Columbia. Seeing the party up in Ontario is not too shocking as they have been flirting with the high-20s for some time and the margin of error is quite large. British Columbia, on the other hand, is probably an outlier result - in the 76 other polls taken since the last election the Liberals have averaged only 17.9% in British Columbia.

The Liberals picked up 9.8 points in Ontario to reach 36.1% and were trailed by the Conservatives at 33.8% and the New Democrats at 23.9%, a drop of 8.7 points. In British Columbia, the Liberals were up 19.6 points to 40.6% and were followed closely by the Conservatives at 36.5%. The NDP was down 11.5 points to 19.6% and the Greens were down 9.2 points to 3.2% support.

The Conservatives had the advantage in Atlantic Canada with 39.2% to 29.9% for the NDP and 29.1% for the Liberals.
With these numbers, the Conservatives would win 136 seats and have the first crack at a minority government. They would likely fail, as the Liberals would win 107 seats and the NDP 94 seats, giving them a combined 201 seats in the proposed 338-seat House.

The Conservatives would win a decent number of seats in every region except Quebec, battling it out with the Liberals in British Columbia, Ontario, and Atlantic Canada. The New Democrats win more than half of theirs in Quebec, with disappointing results in B.C.

There was not much change in Nanos's Leadership Index, the combination of scores on questions of trust, competence, and vision for Canada. Stephen Harper's total score increased by 2.5 points to 95.9 (i.e., an average of 32% on each of the three questions), while Thomas Mulcair's was down 0.3 points to 47.7 and Bob Rae's was down five points to 33.1.

The Conservative results in this poll are rather consistent with what we have seen in other surveys, suggesting that the real swing vote lies between the Liberals and New Democrats. We should expect to see more strange results like this one as the Liberal leadership plays out. So far, this one is having the same effect as the NDP race did: a spurt of support at the outset (update: which may or may not have been anything real - much like this poll). But the New Democrats lost support as the race wore on, only for it to be recovered by Mulcair (and now on its way back?). Will the Liberals manage to keep the momentum going through to April?

51 comments:

  1. I wonder how much of this change is Justin-mania ??

    ReplyDelete
  2. See? I was right they'd do better with JT running.

    ReplyDelete
  3. Your polling average shows the NDP as flat at the outset of their leadership race, not sharply up...

    ReplyDelete
    Replies
    1. The NDP's support picked up when the race began, dipped during the race, picked up dramatically with Mulcair, and now appears to be slipping again.

      Delete
    2. http://en.wikipedia.org/wiki/42nd_Canadian_federal_election#Opinion_polls

      The numbers don't back up that assertion. The NDP gained in the wake of Layton's death, and then slowly declined up until their convention.

      Delete
    3. Indeed it did decline during the race, but the NDP was still doing well in early September when the race was just beginning and Topp announced his candidacy.

      Delete
    4. The two polls with "good" numbers for the NDP were from August 29 - September 1 (Nanos) and September 12-15 (Leger). Both polls were the first from each pollster since Jack Layton's death. Here's your post for the September 12-15 poll from Leger:

      http://www.threehundredeight.com/2011/09/liberals-might-not-have-hit-rock-bottom.html

      Note that you didn't mention leadership once in your post. You did mention Jack Layton's passing twice though. Nor did you mention Brian Topp, who was the only candidate to announce at this point (he announced on the 12th). This was also Leger's first poll since the 2011 election. How can you conclude that the NDP support increased base on that?

      Angus Reid was the other firm to poll in September, and they actually showed a 2 point drop for the NDP.

      Delete
    5. I'll grant you that the passing of Jack Layton may have played a very large role in the NDP's good numbers at the end of August and beginning of September. Nevertheless, I think my characterization of this as a "spurt of support at the outset" is not unfounded.

      Delete
    6. For what it's worth, if you only take the poll numbers for after Topp announced his candidacy you'd have an average of 31% for September. The Nanos poll you included in your average actually had 2/3 of its data collected in August and was conducted only a week after Jack Layton's death.

      Speaking of which, Nanos has a nasty habit of timing the release of their polls to be as inconvenient as possible for your monthly averages don't they lol. :( The Nanos poll you put into your August 2011 average actually has most of its data from July. Not an issue when trying to compare month to month the changes, but it becomes an issue when you're trying to relate those averages to specific events.

      Delete
    7. If you take the numbers of polls actually conducted in August 2011 and compare them to September 2011, the NDP support declined from 32.3% to 31% (30.3% if you add in the Nanos poll that was taken at the end of September). So I wouldn't call a decline of 1-2 points a spurt of support. And given that Topp didn't announce until the 12th of September I don't think any polling data prior to that can be attributed to his "special" brand of charisma.

      Delete
    8. Alright, I'll relent. Update made.

      Delete
    9. Hahaha. Thanks. Cheers.

      Delete
  4. Fairly optimistic for the Liberals.

    ReplyDelete
  5. To get the true value out of any poll, I think it's best not to compare one polster to another at any given time before an election. Although there's a pretty good chance the Liberals have not vaulted as high as 30 per cent yet, this poll, when compared to the previous Nanos poll taken about a month before show a dramatic increase in Liberal support. And I think that's valid. I look forward to what other pollsters come up with in the coming weeks.

    ReplyDelete
  6. I must say 3 seats for the NDP in BC seems low. I know in the poll they're at 20% but, I still find it difficult to believe. 18 seats for the Liberals would be unprecedented although Trudeau won 16/23 seats in 1968.

    Eric,
    Could you please tell me the 3 ridings the NDP win?

    Thanks,

    ReplyDelete
    Replies
    1. I don't think it is unusual - the party only won five seats in BC with 27% of the vote in 2004, and three with 18% of the vote in 1997.

      The seats I have them winning with these numbers are New Westminster - Burnaby East, Skeena - Bulkley Valley, and Vancouver East.

      Delete
    2. Yes, but Trudeau in 1968 was a completely different Trudeau. BC was a different BC, too, and I speak as a British Columbian. 3 seats is a little low, knowing my own province and its traditional swing.

      Derek, You've got something there.

      Charlie

      Delete
    3. Thanks Eric,

      Delete
    4. The projection looks consistent with the input but as they say, "garbage in, garbage out". I don't believe those BC numbers either for a second. It's not like there is a particular reason that anti-NDP sentiment would have suddenly swept the province. The last time core NDP ridings swung over to Liberals was in 1993 when even Vancouver East went red and that was in the face of a deeply unpopular BCNDP government.

      That being said, polls are just polls. A major factor in BC federal NDP support come 2015 will be how the public feels about their presumptive first two years under Adrian Dix. If he follows his implied plan of incremental change and doesn't ruffle too many features along the way, he will prevent another 1993 from happening. But if the brand is tarnished by then, it could well give way to a result like this where seats swing away to other parties.

      Delete
    5. Irrespective of this poll, one thing to keep in mind is that next May, 2013, BC potentially will have an NDP provincial government.

      And as past history suggests, BC'ers turned on previous BC NDP governments at the federal ballot box.

      1972 - NDP government (1975)
      1974 federal election - 2 BC NDP MPs

      1991 - NDP government (2001)
      1993 federal election - 2 BC NDP MPs
      1997 federal election - 3 BC NDP MPs
      2000 federal election - 2 BC NDP MPs

      The next federal election is 2015 and if history repeats itself, BC will again only elect 2 - 3 BC NDP MPs. Just saying.

      Delete
    6. Jonny, you are forgetting one federal election: 1972. It was held only two months after the BC election that swept Barrett's NDP to power, and the federal NDP gathered a record 11 seats in BC, up from 7. And it's the *only* time the federal NDP saw significant gains in BC with the provincial NDP in power.

      Was BC perhaps still honeymooning with the NDP? Dix is a smart man and he knows it's in his best interest to avoid reflecting negatively on the federal party leading up to 2015.

      Delete
    7. Charles Harrison17 October, 2012 12:50

      Yes, but once is not much. 75 per cent of the time it was under 4. Never over 11. It's all-time record was 12. That's what he's trying to say.

      Delete
  7. NDP are still slated to win 50+ seats in Quebec with an 11 point spread. How much does the gap need to close before the Liberals start picking off a significant amount of Quebec seats?

    ReplyDelete
    Replies
    1. Not too much, the Liberals are within five points of the NDP in six Quebec seats on these numbers. But the Liberals vote in Quebec is unlikely to be very efficient, since in many ridings they are working from a tiny base.

      It wouldn't take much to turn Montreal red, but outside of the city the Liberals need to be over 35% to start winning a lot of those seats.

      Delete
    2. I'm not sure. His father won 74 of 75 seats (~98.67%) in 1980, but Justin's his father's-not his mother's-son. However-offense not meant-some Quebecois may not know that.

      Delete
    3. Unlike the recent Forum polls too, this one didn't ask how people would vote if Justin was leader. So him actually becoming leader could still swing more voters.

      Delete
  8. When comparing current Nanos regional breakdowns to their previous poll, it looks like there were as many Green supporters as NDPers who switched to the Libs. I wonder if this has something to do with the fact that AFAIK Nanos is the only pollster who doesn't prompt for party names, hence Greens normally do much worse in their polls.

    ReplyDelete
  9. Nanos is going to need to provide some kind of explanation for what they think they are doing that none of the others are doing that results in such a high Liberal number. I don't doubt that the Liberals are up somewhat, and it would be nice if we had new national numbers from anyone else in the last little while to compare to, but I really don't think that the Liberals are at 30%.

    ReplyDelete
    Replies
    1. Charles Harrison17 October, 2012 12:41

      Oh, they are doing pretty well. I think they might be @30%, and also, I just think Nanos somehow polls lots of Liberals. I trust those numbers as showing something, whether or not they show the correct positions of the parties. I do trust that poll to show something, as all polls do.

      Delete
    2. Nanos doesn't prompt for party names. They just ask who you will vote for. Most other parties ask "will you support party A? B? C?"

      Now, why this methodology produces a different result is a very interesting question.

      Delete
  10. It not surprising that Elizabeth May will not get re-elected with such low support. This is probably a statistical flop I guess. Can you tell me how much she will lost by thou?

    ReplyDelete
  11. Eric, this is almost certainly a bit of an outlier. But the possibility also exists that this poll reflects the first glimmer of some real change in the electorate. Trudeau was already known to be popular, despite a media assigning him the role of charismatic, but untested, inexperienced, insubstantial child trading on the family name. That BC and Ontario should demonstrate the biggest impact is consistent with what a Trudeau-led Liberals might represent. So while your repeated criticism of the Nanos record underscores one likely explanation, I think a truly non-partisan observer might at least contemplate the one other clear possibility.

    I have to say that as you become better known, your analytical skills have started taking a back seat to "punditry". It doesn't serve you well, and is disappointing, since we need at least one cogent, critical and objective go-to source. Please exercise more of your critical faculties when writing these things for the HuffPo and G&M.

    ReplyDelete
    Replies
    1. Nowhere was I criticizing Nanos's record. In fact, I praised Nanos's results in the last federal and Ontario provincial elections. As a truly non-partisan commentator, I explained that this high result for the Liberals is in keeping with Nanos's recent history, which is something that a cogent, critical, and objective analyst would do when looking at the results of this poll.

      As I took pains to point out, this does not mean that the Liberals have not had a surge in support due to Trudeau's candidacy, but rather that it is unlikely that the Liberals have had the sort of boost that 30% represents, at least in relation to other polls from other polling firms.

      Delete
    2. Thanks, Eric. I suspect that the Conservatives have been doing well because there is a perception that there are no real alternatives (apologies to NDP partisans, but that has been the case. Whether it deserves to be or not is a different question). I suspect that, as the next election is well over 2 years away, and as the leadership of the Opposition has been an issue throughout the last year, Conservative fortunes have most likely been somewhat overstated. I am prepared to accept the Nanos data, therefore, as an accurate expression more of aspiration than concrete intent. As for accuracy in the numbers, your guess is probably better than mine.

      Delete
  12. Charles Harrison17 October, 2012 12:44

    Éric,what way did Okanagan-Shuswap go? I like to know how my own riding is leaning.

    ReplyDelete
    Replies
    1. Well, your riding may not be leaning in this way, but with Nanos's numbers the model would give North Okanagan-Shuswap to the Conservatives.

      Delete
    2. Charles Harrison17 October, 2012 15:33

      Sounds about right.

      Delete
  13. The Liberal brand will be attempting to go through some sort of a renewal in 2013.

    The major leadership races are at the federal level, Quebec, Ontario and Newfoundland. New Brunswick will have a new Liberal leader later this month, while Liberal party in British Columbia, Saskatchewan and Manitoba may have a leadership change in the coming year.

    It's true some of these parties are not aligned with the federal Liberals (BC/QB), and some of these parties are minor parties (SK/MB).

    This is the last great hope for big Liberalism in Canada. We already see former Liberals finding home in the increasingly moderate NDP or various center-right parties.

    The Alberta PC party and Sask Party found success by becoming a loose coalition between traditional Conservatives and Liberals. The same way the Manitoba NDP is a loose coalition traditional NDP and Liberal voters.

    ReplyDelete
  14. There is absolutely no way I can believe those numbers out of BC or the Atlantic are reliable.

    ReplyDelete
    Replies
    1. Charles Harrison17 October, 2012 15:34

      They actually are quite reliable, and I speak as a British Columbian.

      Delete
    2. I have had trouble believing the numbers out of Alberta for about 3 decades, but I've learned by now that they are generally accurate based on election results. Still can't explain them using models that work anywhere else in the country, but completely in keeping with Alberta's electoral record.

      Delete
  15. Charles Harrison17 October, 2012 16:08

    I looked at the PDF and this is the question they ask:
    QUESTION: For those parties you would consider voting for federally, could you please rank your top
    two current local preferences? (Committed voters only - First Preference)

    That is why they have such odd results. The link to the PDF is http://www.nanosresearch.com/library/polls/2012-10-BallotE.pdf.

    ReplyDelete
  16. I live in BC as well and I can tell you the Fed Libs are invisible.

    ReplyDelete
    Replies
    1. Invisible is good....better than the visible CPC mired in financial and electoral scandals and arrogance.

      Delete
  17. There is no hope for the non-Conservative supporters in Canada unless the NDP and the Liberals unite into a single party of the centre-left.

    ReplyDelete
    Replies
    1. Charles Harrison18 October, 2012 15:13

      That I strongly doubt.

      Delete
    2. Two parties polling at 30.1% and 27.9% respectively need to merge in order to beat a party that is polling at 33.3%?

      Absurd to say "there is no hope".

      Delete
    3. Why would the NDP or Grits think a merger necessary? At 30.1% and 27.9 both parties have a good chance winning a plurality of the vote come 2015. If anything the poll hints Canada could support two major centre-right to centre-left parties.

      A merger would be terrible for Canada if for no other reason than the new entity would win 60% of the vote. Such a landslide would effectively lead to a very weak or possibly no opposition. A strong opposition benefits Canada by holding the Government accountable and raising important policy issues.

      Delete
  18. Charles Harrison18 October, 2012 15:15

    I actually think that poll shows low Liberal results. I think they have minority ability.

    ReplyDelete
  19. what province concentration allows the Liberals to get more seats at 30% than when the NDP polls at 30%?

    I find that fascinating, must be lots of resistance of traditional Liberal supporters that split NDP votes I guess.

    ReplyDelete
  20. Let the by-elections begin!

    http://ca.news.yahoo.com/prime-minister-stephen-harper-announces-three-byelections-nov-143947015.html

    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.