Monday, September 27, 2010

NDP plummets in latest Ipsos-Reid poll

In the latest poll from Ipsos-Reid, the New Democrats have shed four points and have sunk to 12%, lower than they've ever been in the last seven months. But it is the Greens who take advantage of the NDP fall.Compared to Ipsos-Reid's last poll, taken between September 8-12, the Conservatives have gained one point and now have 35% support. The Liberals have dropped two to 29%.

The NDP drop has put them in a tie at 12% with the Greens, who have gained three points. The Bloc Québécois is up one point to 11% nationally.

This is a horrendous number for the NDP - but it is impossible to say whether this is a product of the recent long-gun registry issue or simply a matter of the margin of error. Considering that Harris-Decima has also shown NDP weakness of late, perhaps it does have more to do with the LGR.

The Liberals and Conservatives are tied in Ontario with 37%, as the Liberals lose four points and the Conservatives gain as much. The Greens have gained five points and stand at 15%, while the NDP is down four to 11%.

The Bloc is up five points in Quebec and dominates with 44%, followed by the Liberals at 22% (unchanged) and the Conservatives at 16% (down one). The Greens have claimed fourth spot with 11%, while the NDP is down nine (!) to 7%. The Bloc seems to have been boosted by the NDP's slip.

The Conservatives lead in British Columbia but have dropped eight points, standing at 33%. The Liberals are up one to 26% and the NDP is up four to 25%, one of the only good bit of news in this poll for Jack Layton. The Greens are up two to 15% here.

The Liberals lead in Atlantic Canada with 38%, down ten points. That drop is almost certainly due to the unreasonable 48% Ipsos-Reid had the party at in this region a few weeks ago. The Conservatives follow with 34%.

They lead in Alberta with 57%, while the Liberals are up eight to 25%. This is not the first time we've seen the Liberals riding high in Alberta.

The Tories are up nine to 63% in the Prairies, followed by the NDP at 15% (down eight) and the Liberals at 14% (down six).

With this poll, the Conservatives would win 69 seats in the West and North, 48 in Ontario, nine in Atlantic Canada, and six in Quebec for a total of 132.

The Liberals would win 48 in Ontario, 21 in Atlantic Canada, 15 in the West and North, and 14 in Quebec for a total of 98.

The Bloc would win 55 seats in Quebec, an all-time best for them.

The NDP would be reduced to 11 seats in the West, nine in Ontario, and two in Atlantic Canada for a total of 22. Their relatively strong showing in British Columbia saves the party from eradication.

The Greens win one seat in Ontario.

This is a good poll for the Conservatives, undoubtedly. At 35%, the party is within striking distance (and the MOE) of their 2008 electoral result. They have good numbers in Alberta, the Prairies, and Atlantic Canada, and would probably gladly take this result in Ontario.

It isn't bad for the Liberals either, who are doing well in British Columbia and Alberta, as well as in Ontario and Atlantic Canada.

It's a great poll for the Greens and also very good for the Bloc, but it is a disastrous one for the NDP. They would lose their seat in Quebec, many of their seats in Ontario, and be reduced to a rump caucus with extremely little influence. They wouldn't even have enough MPs to help the Conservatives pass legislation, if they were so inclined.

It's too early to panic for the NDP, but a few more polls like this and Mr. Layton will have to start working hard to regain the support of the old CCF base and the new urban social democrats.

29 comments:

  1. It looks like Harpers strategy in persuing the LG issue is to push northern ont/Western voters away from the NDP. It seems to be attracting western voters to the tories and Ontario voters to the greens. Hopefully the NDP drop especially in Ontario due to the LG issue will fade by spring. Its a delicate balance-they need the NDP strong in Ontario to split the vote with the Libs. The libs don't seem to be gaining too much though. Perhaps Harper's strategy(assuming there was one) is to swing enough western/northern ont voters away from the NDP to scare them away from supporting a fall election. As long as voters dont go Lib in Ontario that seems to be the effect. I'm sure they would be happy with this result as you say.

    ReplyDelete
  2. Greens tied with Dippers? Not real.

    Yet.

    ReplyDelete
  3. eric rw: Perhaps Harper's strategy(assuming there was one)...

    Methinks you make large assumptions.

    ReplyDelete
  4. It is early days yet I think to say the LGR issue caused the NDP collapse.

    We will need another month or so to see if this is really permanent or only a small anomaly.

    ReplyDelete
  5. 1) These results are true.
    2) These results won't last.

    As everyone says, polling is a snapshot in time. Progressive anger at the NDP will fade, especially since the registry survived.

    ReplyDelete
  6. Gun control is a ballot issue for pro-gun voters. This NDP drop is real.

    ReplyDelete
  7. Eric any idea where a Southern Ontario resident could find the NB election results on TV tonight. What time do the polls close? TIA.

    Earl

    ReplyDelete
  8. Let's sanity-check these results. Applying Éric's Infallible House Effects to this poll and rounding each correction to the nearest per cent, we get:

    Party: Ispsos-Reid+/-Correction=>Corrected Result
    Conservative: 35-3=>32%
    Liberal: 29+1=>30%
    NDP: 12+2=>14%
    Green: 12+1=13%
    Bloc: 11-2=>9%

    The rounding adds a bit more dirt to the numbers but at the top, they're generally consistent with what we've seen for a while. The Dipper numbers are not good, but much of that could be variation. The Green numbers are good, but ditto. The Bloc number could be low from applying too large a correction due to rounding.

    There's less to this poll than meets the eye.

    Even if the numbers were truly accurate (and that we will never know), we should be careful in attributing the entire NDP drop to the long-gun registry. That's a popular media take because it fits in a sound bite, but "tidy" does not mean "right".

    ReplyDelete
  9. Earl,

    If you have a satellite or cable package, you can probably find coverage on CBC's Atlantic stations.

    Otherwise, you won't find it on TV. I think CBC will be streaming coverage online via the NB Votes section of their site. Coverage should begin at around 7pm here in Ontario.

    ReplyDelete
  10. Eric your piece in the G&M made the National Newswatch!

    Congratulations,

    Earl.

    ReplyDelete
  11. Ha, so much for Ipsos' pro-CPC slant. In BC , the CPC sees an 8% drop! That said, it's only ~140 sample size, which doesn't mean much in any event.

    And the first time that the Prairies (SK & MB) have exceeded AB in the CPC vote at 63% that I can recall. Again, an even smaller sample size that doesn't mean much.

    BTW, Earl - out here in BC, CBC NewsWorld always carries various provincial election results live. Have watched the ON, QC, and NS election results live thereon in the past. Ergo, I'd suggest that you tune in to CBC NewsWorld for live NB election results - doesn't matter whee you reside in Canada.

    ReplyDelete
  12. CBC Newsworld almost always has live coverage of provincial election results. Check them out.

    ReplyDelete
  13. John your "sanity-check" isn't helpful.

    You've taken these numbers, applied Eric's house effects, and then discovered that this poll is "generally consistent" with other polls.

    This is, by defintion, true. Its some of a tautology that tells us nothing about the world.

    In fact its true for ALL polls since Eric's house effects is a measure of how individual polling firms differ from the AVERAGE of all polling firms. Taking polling data back to average will always give us "generally consistent" numbers.


    #1 rule of reading polling data is to ignore the top line numbers and look at the change.

    -4% for the NDP is REAL and is SERIOUS.

    "Even if the numbers were truly accurate (and that we will never know), we should be careful in attributing the entire NDP drop to the long-gun registry."

    Normally i'd agree.

    However, the only mention of the NDP in the news during this polling period was in regards to their dillema over the LGR.

    We are perfectly safe, reasonable, and sane to attribute their 4% drop to the LGR.

    No "sanity-check" needed here John.

    ReplyDelete
  14. This looks like a specific strategy on the part of the CPC.

    They can't win wholly urban ridings. They just can't.

    And rural Québec always votes Bloc.

    Since the CPC wants a majority, the way for them to do that is to sweep rural English Canada. If there's a wedge issue they can use to capture a rural riding outside Quebec, they will do it.

    ReplyDelete
  15. CBC News Network will not be covering the election results tonight, if the CBC journalist who told me that is to be trusted.

    ReplyDelete
  16. Newsworld isn't going to cover a provincial election ballot count live? Burn! I remember when they covered Northwest Territories elections live.

    The politics of the Atlantic provinces and the territories have had a much lower profile in the national media since we stopped having constitutional crises and first ministers' meetings all the time. Maybe that early 90's type of politics wasn't good for the country, but it was a lot less boring than what we have now. Mulroney as PM and Chretien as Opposition leader-- who knew what crazy stuff they would do/say?
    Nowadays, If it wasn't for Danny Williams, we wouldn't have a single unpredictable
    politician. Harper and Ignatieff have a daily boringness competition. Other than the non-Coalition, not a single interesting thing has happened in Canadian politics since the '95 referendum.
    That's why I like to read about these polls, it's the only interesting thing about Canadian politics these days in that it changes from time to time.

    ReplyDelete
  17. John, methinks the Harpies spend million$ on polls for a reason-to see what wedge issues they can exploit! Right now they want to avoid a fall election after a couple of months of Lib resurgence. The best way is to cripple the NDP temporarily. Hope Springs eternal if the economy is improving by next April.

    ReplyDelete
  18. The economy is improving right now. The economy has been improving for over a year.

    In some parts of the country, this improvement is obvious (Saskatchewan). In others it is not (Windsor).

    ReplyDelete
  19. Shadow: You've taken these numbers, applied Eric's house effects, and then discovered that this poll is "generally consistent" with other polls.

    This is, by defintion, true. Its some of a tautology that tells us nothing about the world.


    Those are refreshingly unique semantics to apply to the word "tautology". The mathematical understanding of the situation also displays some originality.

    The result would indeed be tautological if the house effect sample space was the poll in question. In fact, the sample space doesn't even include that poll. It might be helpful to apply a forensic analysis to the quoted statement and work back to where logic and reality parted ways.

    #1 rule of reading polling data is to ignore the top line numbers and look at the change.

    By contrast, this is a valid statement, so investigation of longer-term changes is in order. The August and September Ipsos-Reid polls had the NDP at 15%, as did the April survey. The party is down 3% from those. Is that drop real? Part of it probably is. However, if I were Jack Layton I wouldn't be looking over my shoulder for the Four Horsemen. Not yet.

    ReplyDelete
  20. Ira wrote:

    "... CPC. They can't win wholly urban ridings. They just can't."

    Come on now. That's just silly.

    The Tories can and do win wholly urban ridings. Most obvious is their total domination of Calgary and near dominaton of Edmonton.

    The Tories have also proven to be quite competitive in numerous smaller cities across the country.

    "Urban ridings" is not equivalent to "Toronto and Montreal".

    ReplyDelete
  21. If the NDP had of just let their MP's vote on the LGR the urban folk would have forgotten it come the next election. Rural folk however will not...
    The NDP will be reduced to inner city.

    ReplyDelete
  22. John this jovial condescension is putting you in the ranks of Glen Pearson (one of the worst MPs in Canada, I cannot stand preening, saccharine, sage types.)

    Correcting for the typo the line "its sort of a tautology that tells us nothing about the world" is perfectly reasonable and I stand by it.

    Taking polling results and "fixing" them with Eric's house effects is always going to make them more consistent with other polls.

    The fact that this poll wasn't included in the sample set used to develop the house effects only UNDERMINES your argument.

    (If they were included the average NDP number would be lower, this poll would show less deviation, and your "correction" would be a slightly LOWER (ie. less "sane") NDP result.)

    "By contrast, this is a valid statement, so investigation of longer-term changes is in order."

    That's not necessary. Consulting long term trend lines is NOT a must, you may think it prudent and wise but its not helpful in this situation.

    By defintion wedge issues like the LGR capture the imagination of the country and for a time move numbers outside of their usual holding pattern.

    4% from the previous poll is a LARGE drop and statistically significant in its own right.

    Jack Layton indeed SHOULD be looking over his shoulders for the four horsemen if he cannot come up with a better message track to sell to the public on this issue.


    John a quick word of advice. Not every situation calls for a prudent, patient, and measured response. What you call "sanity" is the political equivalent of sticking your head in the sand.

    ReplyDelete
  23. Shadow: John this jovial condescension is putting you in the ranks of Glen Pearson

    I'm honoured by the comparison, but I rank many levels below Glen Pearson in my contributions to this country and its democracy. I do generally strive for joviality and beyond a certain level of misunderstanding I feel free to make observations.

    Correcting for the typo the line "its sort of a tautology that tells us nothing about the world" is perfectly reasonable and I stand by it.

    A typo is still there (I personally find proofreading helpful) but that wasn't the point. This poll, corrected for past persistent differences between firms, is not wildly inconsistent with recent results. Couple that with the concepts of "never believe any single poll" and "the last poll ain't gospel either" and the results are interesting but very preliminary. The 4% drop looks closer to a 3% drop. Some of that is probably real. Much of it probably isn't.

    There's clearly a comprehension problem here, so try this exercise: take any poll from last December and apply Éric's house effect corrections. See if the results match polls today. Hint: they won't. That's because polls have changed since then. And because there's nothing remotely tautological about the original observations. Even if it's a cool word.

    ReplyDelete
  24. Shadow: John a quick word of advice. Not every situation calls for a prudent, patient, and measured response. What you call "sanity" is the political equivalent of sticking your head in the sand.

    Or the political equivalent of building for the future. There is life beyond the shrill blogosphere and there is an electorate that is not galvanized by hate. I spent the weekend reaching out to voters at a rural fair. With one single exception, the people I spoke to were polite and willing--even eager--to engage in dialog. At this fair and elsewhere, I'm meeting a groundswell of citizens who are tired of flamethrower politics and who are looking for a reasoned and reasonable alternative.

    The exception? One would-be blogger whose first words were "Forget it!" and who carried on by vehemently asserting an error of fact. There's always a small minority whose fuel is rage. That's not the basis of a healthy political system. It's not even the kind of base that will lead to the next government.

    Try getting out some time. Contact your local federal and provincial riding associations. Volunteer to canvass and help out at events between elections. You'll meet people--real voters, not blog commenters--from outside your immediate circle. You might be surprised at what they want to see in Ottawa.

    ReplyDelete
  25. John we are indeed at an impasse, especially after you dispensed more Glen Pearson style nonsense about me getting out and meeting non-bloggers (presumptious much? I live about as far away as you can get from the Ottawa bubble.)

    BTW IR polls regular Canadians, not bloggers. BTW regular Canadians care about the LGR. BTW opposition to the LGR isn't hateful or unhealthy.

    I happen to think the NDP will set aside your advice. What you consider "healthy" is irrelevent. We're talking about politics as it is, not as you think it should be.

    The Green party being an unqualified disaster tends to undermine your arguments here.


    Back to the debate over the facts. We are indeed experiencing a lack of understanding on your part.

    "This poll, corrected for past persistent differences between firms, is not wildly inconsistent with recent results."

    You still don't get how this is automatically true eh ? What matters is internal consistency, when you make an apples to apples comparison between IR polls the 4% drop is real, significant, and reason for concern.

    "The 4% drop looks closer to a 3% drop. Some of that is probably real. Much of it probably isn't."

    This is an assumption. Its actually quite likely that the last poll WAS accurate and that this poll is too. Polls move, going back to months ago and assuming that's the baseline is ridiculous.

    The distribution of likelyhood in a confidence interval is curved around the numbers given, with smack dab right on being the most likely result.

    "so try this exercise ... See if the results match polls today."

    Hint: that's not the claim you're making. You're saying the latest batch of polls are consistent, which is true when you lump them all together by averaging them out.

    ReplyDelete
  26. Henry - Calgary and Edmonton are suburban ridings. Have you ever seen an Albertan city?

    The city of Calgary (the city proper - no metropolitan area) has a lower population density than Switzerland.

    ReplyDelete
  27. Canada is remarkably low in truly urban areas. There's Toronto, Montreal, Vancouver (the city, not the GVRD), and arguably central Winnipeg.

    That's pretty much it.

    ReplyDelete
  28. Looks like the CPe will have to start dusting up their dormant coalition with the Socialists and Separatists in order to survive.

    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.