Wednesday, July 18, 2012

Can NDP cobble together a plurality of seats?

Late last week, iPolitics.ca released the details of the newest EKOS Research poll. The survey showed little real movement at either the national or regional levels, but the result pointed to the potential for the New Democrats to tie the Conservatives in the seat count, even with a national lead of only 2.1 points.
After being quiet for some time on the federal scene, EKOS is quite active again. They were last in the field Jun. 21-26, meaning this poll of Jun. 27-Jul. 5 did not miss a beat.

Since that last poll, there have been no statistically significant shifts of support at the national level. This is unsurprising considering we are in the midst of a very hot summer and little has been happening that could bump the numbers around.

The New Democrats decreased by 0.1 point to 32.3% while the Conservatives increased by 0.9 points to 30.2%. The Liberals were up 0.3 points to 19.5%. Another 4.9% (-1.6) would vote for the Bloc Québécois while 3.1% said they would vote for other parties.

Then there are the Greens who registered 10% support, up 0.5 points from EKOS's last poll. That puts them out of step with what other pollsters have recorded, and EKOS recognizes that. In their analysis that accompanied their polling data, they give two reasons for the discrepancy: firstly, that some other firms are not prompting for the Greens, and secondly, that some other firms are not calling cell phones.

I don't doubt that these two issues explain some of the higher levels of support recorded by EKOS (something that has traditionally been the case for the firm). But most firms do prompt for the Greens. The only firm that I am aware of that does not prompt for them (only, some polls prompt for none of the parties) is Ipsos-Reid. And as for cell phones, one would expect that the firms using online panels would also be able to capture a high degree of Green support. Sometimes they do, sometimes they don't, but none are consistently so much higher than the norm. But perhaps this has something to do with how they weigh their panel with an eye towards turnout.

Is Green support rising? Probably not to 10%, but it does seem to be on a slow increase. The party scores more 6s and 7s than it did shortly after the 2011 election, when they managed more 3s and 4s. But I don't think there is enough evidence to spark talk of a Green Invasion just yet.

The Conservatives led in Ontario in the poll with 35.3%, up 2.9 points since EKOS's last survey. The NDP trailed at 29.6% (-3.2) while the Liberals were at 23% (-0.9). These shifts are all within the margin of error.

In Quebec, the New Democrats led with 39.1% (+5.7), putting them well ahead of the Bloc Québécois, which slipped 6.2 points to 20.8%. That is a statistically significant drop for the Bloc. The Liberals were down 0.9 points to 16.4% and the Tories were up 1.7% to 13.2%. At 7.7%, Green support appears considerably above what others have recorded in the province.

In Alberta, the Conservatives were up 5.5 points to lead with 59.8%, while the NDP stayed stable at 19.5% and the Liberals were down 5.1 points to 10%.

The New Democrats were up 1.8 points to 39.4% in British Columbia, while the Tories were down 2.3 points to 26.3%. The Greens were up 3.1 points to 19.4%, a score that puts them well above what other firms have found to be the case. The Liberals were down 1.7 points to 12.9%.

In Atlantic Canada, the Liberals made a statistically significant gain of 14.7 points to hit 32.5%, while the NDP was down 5.5 points to 28.4% and the Conservatives were down 3.7 points to 25.8%.

And in the Prairies, the New Democrats stood at 31.9% (-1.6), the Conservatives at 29.8% (-11.5), and the Liberals at 24% (+7.6).
With these levels of support, the New Democrats and Conservatives would tie with 126 seats apiece on the current 308-seat map. The Liberals would win 51 seats, the Bloc Québécois three, and the Greens two.

At almost 20% support, the Greens would be able to win a second seat in British Columbia. Would they actually get 20% of the vote in B.C. if an election were held today? Probably not, but they do have the potential to pull off another seat win if they manage to make major gains in the province, particularly at the expense of the Liberals.

The New Democrats are able to tie for the most seats in this survey in large part because of their strong performances in the West. Depending on the poll, the Conservatives are usually able to win a plurality even if they trail by three or four points because of their lock on Ontario and the West. In this poll, however, that lock is broken in British Columbia and the Prairies. With 35 seats in the four western provinces and 59 in Quebec, it is not difficult for the NDP to win enough seats in Ontario and Atlantic Canada to put them over the top. If the New Democrats cannot manage to make major inroads in Ontario, they could by-pass the province if they win a majority of seats in B.C., Saskatchewan, and Manitoba.

But the Conservatives have the advantage, due in large part to the six new seats going to Alberta and the 15 new seats in Ontario. The Conservatives should be able to win two-thirds of these 21 seats on these kind of support levels, which swings the pendulum back in their favour. However, a few extra seats in Atlantic Canada (where the NDP have been leading in most polls) could tie things up again. This points to something that has been apparent in poll after poll: British Columbia is looking like a real problem for the Conservatives and a key province as we head towards 2015.

25 comments:

  1. It seems like the anomalous Green number is being fed by that anomalous result from BC and the Prairies. In both places, the Greens score about 80-100% higher than they sit in the poll averages.

    As I recall, in the middle of the 2011 election, EKOS started applying a likely voter screen to their results, and reporting both the screened and unscreened numbers. I would be very interested to see what result that would produce now. Just based on the demographic cross-tabs, which show a big CPC lead among seniors, and the NDP and Greens 1 and 2 among the under-24 group, I would imagine such a screen would produce markedly better CPC numbers.

    ReplyDelete
    Replies
    1. ^ Agree. My understanding though was EKOS hasn't been good at providing break downs for their likely voter numbers?

      Delete
    2. TS,

      They actually started doing that during the Ontario election; I imagine they were pretty badly embarrassed last federal for being the pollster who most severely underestimated Conservative support. See Éric's article:

      http://www.threehundredeight.com/2011/05/ranking-pollsters.html

      Dom

      Delete
  2. High Green support in-between elections should be good for the NDP or Liberals. During the election time, many of these Greens would vote NDP/Liberal or stay at home.

    I can't see the Green party garnering more than 4% of the vote, especially if the progressive vote rallys around Mulcair's NDP in 2015.

    I'm sure the NDP would focus on a GOTV scheme as the election nears by. The youth vote is crucial for a future NDP government. The NDP cannot win decisively without an increase in voter turnout.

    ReplyDelete
  3. FYI: Many other polls show the NDP gaining a lot of ground among seniors. It looks like even if the Tories win a plurality among the elederly, it won't be anywhere near the margin they had last time. Guess what? Elderly people don't tend to like parties that want to take a wrecking ball to pensions and public health care and that are widely viewed as being cheats who don't play by the rules.

    ReplyDelete
  4. What would the Greens' second seat be?

    ReplyDelete
  5. Is Vancouver South the second Liberal seat?

    ReplyDelete
  6. The biggest threat to the NDP IMHO is the possibility of an NDP government in BC. If Dix gets elected and proves and unpopular premier...

    You can see the same thing happening with the Liberals too. I may be reading too much into the regional numbers given the MOE, but it seems like the Liberals are having the most growth in the places where "Liberal" parties are the farthest from power. The same was true for the NDP in 2011...

    ReplyDelete
    Replies
    1. South Parkdale Jack18 July, 2012 12:26

      What if Dix gets elected and proves to be a popular premier?!!

      Dexter in Nova Scotia seems to be doing fine although his personal numbers now are apparently one of the lowest among premiers. Yet look at most of the federal poll numbers there and in the rest of Atlantic Canada.

      Selinger in Manitoba certainly doesn't appear to be a drag on the federal NDP in the Prairies, in fact, they're on the rise there.

      Mind you, many Canadians prefer a different party heading the federal and their provincial governments.

      Of course, these polls are only a snapshot in time, anything can happen between now and election day .... but keep in mind that many governments are voted out rather than in.

      Delete
    2. Retired in BC19 July, 2012 10:24

      Yes how can you not be popular when you have rising unemployment, rising taxes and 10% increases for public sector unions and Teachers.

      Dix will be the NDP poster child.

      Selinger isn't a drag the NDP in Sask and Manitoba... they can count on 2 and as many as 3 seats out of the 28.

      Nothing compared to the Glory days of the Manitoba NDP when they could generate 4 seats out of 28.

      Delete
    3. There also hasn't been a "popular" government in BC in what, 30 years?

      Delete
    4. @Retired-- Eric's current projection has the NDP at 13 seats in the Prairies!! We shall see.

      Delete
  7. Derek Strelioff18 July, 2012 12:40

    I'm not putting a lot of weight into the Saskitoba number. The sample size is only 55, a quarter the size of Nanos' usual sample, and a third the size of EKOS' own previous poll. So, a "surge" in Green support or a "collapse" in CPC support came from a difference of about 5-10 respondents. It's about as reliable as an Atlantic sample; the trendline fluctuates like a seismograph during an earthquake.

    ReplyDelete
    Replies
    1. Nanos just put out their numbers and they back up EKOS' numbers in Saskitoba. Still a small sample size though.

      Delete
  8. I'm an American with an intense interest in Canadian politics who has been reading this blog almost daily for over a year. I've never posted before, but this poll seems strange enough to me that I wanted to ask some questions.

    First, while I saw it was noted that the Greens are almost certainly not at 10% nationally, it was their 19% share in B.C. that really caught my attention. That number really seem inordinately high to me, and I also wonder how feasible it would be for both the Greens and NDP to each poll as highly as they did in B.C. at the same time, because it would seem to me that the Greens would pull voters from the NDP before they would affect the Conservatives or Liberals.

    Secondly, the Prairies numbers seem strange to me. I can see the NDP making gains in Manitoba, as they have an NDP government under Selinger that as far as I know is relatively popular. But, I question if the Conservatives could have fallen as far in Saskatchewan to only win 9 seats in the region, considering the overwhelming mandate Wall and the Saskatchewan party earned just last November.

    Finally, I know it was stated after last May's election that EKOS was supposed to begin using a new system to combat their tendency to underestimate the Conservative vote. They also, if I recall correctly, overstated Green support in the last election fairly severely. Did they ever implement that? If not, would that be an explanation for some of the more unusual features of the poll?

    Eric, I very much enjoy your blog, as it's a good way for me to get my Canadian political fix as our media doesn't cover it. Keep up the good work!

    ReplyDelete
    Replies
    1. I don't think EKOS has implemented any sort of weighting mechanism for over or under estimation of various parties. There are a few wonky results, the ones I mention are the ones that are likely outliers since they are not within the MOE of the aggregation.

      Delete
    2. I think EKOS actually put a statement somewhere that they wouldn't weight the numbers by likelihood to vote unless an election was imminent. I'll see if I can find it.

      Worth noting that the NDP has done as well as 12% in BC provincial politics though, so even if 19% is a high number they could still do pretty well out here.

      Delete
    3. Their candidate in Vancouver Centre (Adriane Carr) is the former leader of (and founding leader of) the BC Greens, and the first ever Green member of Vancouver City Council. It was under Carr's leadership that the BC Greens got 12% of the vote. I personally would really like to see her parliament - I just wish it wasn't at the expense of a Liberal MP!

      Delete
    4. Actually the BC Greens only got 8% of the vote in the last two provincial elections. The only time they got 12% was in the once in a millennium 2001 election when the NDP government was annhilated and only got 22% of the voted and 2 seats. The Greens were the only place to put your "x" if you were pissed off at Glen Clark and didn't want to vote for the extreme right under Campbell. Now the BC Greens are on their last legs. They didn't even run candidates in the two provincial byelections in April and they are on the verge of seeking banruptcy protection. Their provincial leader Jane Sterk has to be the most colourless, dull, uninspiring leader of any party in Canada. RIP

      Delete
    5. I never said they got 12% in the most recent election DL.

      I'd argue the circumstances for Green growth federally are the same as the ones you describe provincially - ie people becoming increasingly dissatisfied with the established political parties. Seeing something like that happen going foward seems plausible to me. Though yah, like I said, 19%... yah no.

      Delete
    6. I don't think people are all that dissatisfied with the established parties federally. People who oppose the Harper government seem quite enthusiastic about the Mulcair-led NDP and once they pick a leader there will be lots of "too smart to vote Tory and too rich to vote NDP" - type snobs who will be happy with the Liberals and if you are a rightwing conservative you are probably delighted with Harper! There is something for everyone in the Canadian party system.

      As for the Greens, they had their high water mark in 2008 when they won 6.7% of the national popular vote. In 2011 they collapsed to 3%. Elizabeth May clearly has no interest in building an actual party - she just wants to be a "Green Independent" MP who is not responsible to anyone and who is her own boss.

      There is really no reason for the Green Party to exist in Canada. If you care about the environment - what's not to like in Mulcair's NDP...and if you are an environmentalist and you hate unions and don't want to "slum it" with the NDP - you can join the federal Liberals. What does the Green Party offer that is unique? NOTHING. Its basically just a vehicle to help one person (May) get a seat in parliament. Once you leave Saanich-Gulf Islands - the party is non-existent and vies with the Marijuana Party and the Marxist-Leninists to see who can come in dead last in every seat.

      Delete
    7. Greens aren't much more than a lobby group masquerading as a political party for media attention and easy political money for the cause. (which they won't get as much of now that the per vote subsidy is going away)

      May brought to mind again this week that they are just a protest party with the suggestion that if the Etobicoke Centre result is overturned and a by election is called the greens probably won't run a candidate.

      One more spot where her voice and cause carries more trying not to get elected than by actually trying to get people elected. Protesting... lobbying... failing upwards again.

      Delete
  9. I noticed during the last election campaign that Ekos consistently over estimates the amount of Green support (relative to other pollsters, and indeed, relative to the final results).

    When I first heard of Ekos polls, I thought they were an environmental organization. Is it possible that disclosing their name at the beginning of phone questionnaires makes people more likely to identify as Green?

    Unconscious bias sometimes slips into polls. In Quebec, poll results on language and unity issues differ based on whether the person being questioned thinks the pollster is anglophone of francophone. It stands to reason that if people think they are being questioned by an environmental organization, some (say 2-5%) would be more prone to select the Greens.

    Perhaps if Ekos commissioned another polling company to perform the same poll in parallel to their own, they could find if they need a name change!

    ReplyDelete
  10. Court challenges to 2011 Elections !!

    http://www.cbc.ca/news/politics/story/2012/07/19/pol-council-of-canadians-lawsuit.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.