Friday, August 28, 2015

2015 Federal Election Link Round-up, Week 4

The beat goes on. Just seven weeks to go! Seven! After this one!

Friday, August 28, 2015

- Are the Conservatives headed for third place? I ask this question in my CBC analysis this afternoon. The trends are pointing in that direction, and two of the last three polls are showing it. Could be a blip, or could be something really unexpected.

- The Poll Tracker has been updated with the Ipsos and new EKOS polls. It is getting to the point where we can only call this a three-way race within the historical context. The riding projections have also been updated.

- It's poll day. Ipsos Reid was out last night, showing the same close race but with the Liberals narrowly moving up into second place. It's a wobble within the margin of error, but since the end of July Ipsos has definitely recorded a movement from the Tories to the Liberals. An interesting parallel to yesterday's Forum poll, which showed some of the same trends (though with the NDP up) but might have also suffered from a little elephantitis. EKOS will be out later today, and I'll be looking to see if they are showing some of the same movement. The next Poll Tracker update will come after EKOS is out.

Thursday, August 27, 2015

- The Poll Tracker has been updated with the Forum poll, pushing the NDP's likely seat range above that of the Conservatives. The riding projections are now up-to-date as well.

- I joined Chris Hall on Power and Politics last night to talk about my piece on the leaders' tours. We also went over where the numbers stand today. Speaking of which, Forum broke the mold this morning. Harbinger of things to come, or the kind of quickly-corrected swing we've seen from Forum before?

Wednesday, August 26, 2015

- Two Pollcast episodes are up today! A special treat, with three this week after yesterday's podcast with Mario Canseco. Today, a discussion with Facebook's Kevin Chan about what they are doing in this campaign, and another discussion with the Angus Reid Institute's Shachi Kurl. What are her numbers showing? You can find both episodes here. You can also subscribe on iTunes or with the iPhone podcast app.

- The Poll Tracker has been updated, reflecting the latest Angus Reid numbers as well as those for Alberta from Insights West. The riding projections are also up-to-date. And watch for the next episode of the Pollcast coming later today. Today's guest is Shachi Kurl from Angus Reid Institute to talk about their new poll.

- I tried to read the tea leaves of the leaders' tours in my piece for the CBC this morning. My conclusions? The Conservatives are playing the long game, the NDP is consolidating gains, and the Liberals are focusing on ridings that are on the bubble now.

- The latest episode of the Pollcast features Mario Canseco of Insights West. We discuss their latest B.C.-only poll.

- An Angus Reid Institute poll was released this morning, showing numbers broadly in-line with other surveys. But the numbers for the Liberals were a little lower than the norm.

Tuesday, August 25, 2015

- The riding projections have been updated to reflect the latest numbers on the Poll Tracker.

- An interesting poll from Insights West on the race in B.C., with detailed regional breakdowns. Watch for the latest episode of the Pollcast, in which I discuss the poll with Insights West's Mario Canseco.

- A lot of talk about the economy today, so I took a look at how the parties and leaders rate on the issue. The verdict? The economy is not the Conservative trump card it once was.

- Nanos's four-week rolling poll is out, showing everyone still clumped together. The big bone of contention with the other polls, though, is in Ontario. Nanos has the Tories doing much better than other polls, and I don't think we can chalk it up to old data.

Monday, August 24, 2015

- This weekend, I joined Chris Hall on The House to talk about Battleground Quebec: the recent CROP, the Conservatives' chances in Quebec City, and how the Bloc Québécois is doing.

- Quiet weekend and Monday morning on the polling front. There have been 11 polls conducted during the campaign so far after 22 days. At this point of the 2011 campaign, there had been 46 polls. Of course, that was a shorter campaign. In the same period this far out from the 2011 vote, there had been eight polls conducted. We're beating that mark, at least.

121 comments:

  1. What will the stock market crash and the possibility of a global economic slowdown do to this campaign? Is this Harper's ticket to another majority? Inquiring minds, etc. etc...

    ReplyDelete
    Replies
    1. I expect the stock market weakness will help Mulcair. People are tired of economic uncertainty and might try something radical in order to deal with it.

      Delete
    2. I'd guess it hurts Harper. Even with right wing parties being branded "good for the economy" Harper is still the incumbent and incumbents get the blame.

      Worse there's a bit of legitimacy to that charge this time since Harper doubled down on oil and that's what's really hurting us relative to other economies.

      Delete
  2. Why do Canadian pollsters never obtain religion of party supporters as in the USA. Some Cdn Pol Sci surveys seem to, but they come out perhaps a year after the election.

    ReplyDelete
    Replies
    1. Because Canadians are vastly less religious than Americans are, so it doesn't matter as much.

      Roughly half of all Americans attend religious services at least once a week. Fewer than 10% of Canadians do.

      Most Canadians have a religious denomination in name only. They don't actually practice.

      Delete
    2. Forum occasionally asks religion.

      Delete
    3. For example: https://www.forumresearch.com/forms/News%20Archives/News%20Releases/46791_Fed_Horserace_(Forum_Research)_(05232013).pdf

      Delete
    4. Pollsters ask questions that are relevant to election outcomes. In the US, religious denomination has a strong correlation to the Dem/Rep divide. In Canada, Evangelical Christian are strongly Conservative (similar to Republican support in the US), but in Canada, there are about equal in total size to all non-Christian religions combined (8% each), and their total share is shrinking, while in the US they are over quarter of the population. In contrast, the US has less Catholics (Liberal/Democratic inclined, but half support non-Liberal parties) than Evangelical, while they are over a third of Canadians.

      Delete
  3. Éric,

    Angus Reid has provided the questionnaire for their recent poll on the impacts of the Duffy Trial, and I noticed it starts with a straightforward ballot question. In the detailed tables of their report you can even sort of back-calculate what their % results were for party support, albeit without whatever weighting they may normally apply to their regular federal party support polls. Any idea if they intend to release the ballot question results separately?

    Dom

    ReplyDelete
    Replies
    1. I'm not sure if they will be releasing horse race numbers. They apparently will have something out this week, though. Whether or not it will have VI numbers, we'll see...

      Delete
    2. How about the question they pose in the survey as there are only five "Likely" outcomes to this election based on current polls? None of the five possible outcomes posed offer an outcome of a Liberal Government. With all three parties still in a very tight race it may be wishful thinking on Angus Reid's part but still a very unethical question to pose in a supposed non-partisan poll. Seems they may have a hidden agenda and given the weighting to that poll here it leads to some question on this sites partisan viewpoint? Angus Reid missed the last three Provincial Election calls by a football field of wrong! They were in another galaxy - they were so wrong - but yet they are weighted heavier here than any other poll based on what? Their polling size or the fact it is a recent poll? It certain;y cannot be on their track record!

      Delete
    3. Well, a Liberal victory just isn't likely based on their results. So why would they include that possibility? If things shift and their poll finding shift as well, then they will presumably consider the possibility of a Liberal victory.

      Delete
  4. Fwiw Angus Reid is currently in the field (Aug 21 - 26).

    ReplyDelete
  5. Right off the top let's stop kidding ourselves.
    This election is all about which parties can work together ??
    Because if you look at the numbers in essence what you have is

    1/3,1/3,13

    Which means at minimum NO majority and not coalition but cooperation between two parties. Which two remains to be seen ??

    ReplyDelete
    Replies
    1. There's more than 7 weeks left in this campaign, and with vote-splitting (any way) we could well see a majority with a party getting only 35% of the vote.

      All 3 of the major parties are within reach of 35%. The Liberals could get there if voters get cold feet about Mulcair. Mulcair could get there if he consolidates his recent gains. Harper could get there just on voter turnout alone.

      So that's not necessarily what this is about, no.

      Delete
    2. Harper is tanking. The trend lines show this clearly. He has no growth potential and will not get a majority.

      Delete
    3. The NDP already has British Columbia and Quebec. They have the ability to take Ontario (they are usually second or third in a tight three-way race). They are also competitive in Atlantic Canada. They have an easy road to a plurality but an almost impossible road to a majority without Ontario.

      The Conservatives have the Prairies and have more ability than any other party to open a wide lead in Ontario. It will not be enough to secure a majority but it will be enough to be a plurality.

      The Liberals have a tendency to be second place, with the exception of Atlantic Canada. This puts them in a great position to consolidate their support and push themselves over the top in certain areas, but it also results in much seat inefficiency where they don't consolidate. They do usually beat out the NDP in Ontario, though, and if they can do that the voters in other areas may also switch out of strategic voting.

      The Bloc has the ability to concentrate their votes and stay in the House of Commons, maybe even increasing their seat count, but they will not return to Official Party Status.

      The Greens have an opportunity for expansion in British Columbia and Ontario, and to a much smaller extent Quebec (as a result of one MP's defection to the Greens) but not much else.

      Delete
    4. Both the NDP and CPC have more second place ridings than the Liberals. The LPC would need to win all their first and second place ridings plus a handful of third places to form a majority.

      The Bloc has 2 ridings second-place within 10% of the NDP support, then a large gap to a 20% gap. They might rise to 3 seats, but the Greens aren't polling in second place anywhere.

      In Ontario, with seats under a 70% probability, the potential shifts to second placers are: 15 CPC, 10 NDP, 16 LPC.

      Delete
  6. Not forgetting the Duffy Gate effect !!!

    ReplyDelete
    Replies
    1. Let's remind people that Duffy trial concerns $90,000 that wasn't paid by the government or by the party, and didn't cost the ordinary Canadian anything.

      A rich guy, Nigel Wright, paid a $90,000 bill that was owed by his colleague, Mike Duffy.

      What exactly is the big deal here?

      Delete
    2. "What exactly is the big deal here?"

      Ira try LIES, obfuscation, lies and denial along with "talking points" and more outright lies. They are still lying by the way as today's trial evidence shows !!

      Delete
    3. Ira, the big deal, apart from the lies and cheating involved with the cheque, is that Harper was trying to overturn the Constitution to stack the composition of the Senate for his own benefit.

      Delete
    4. Corruption and bribery. We shouldn't tolerate any of that. Anytime government does that it should be punished!!

      Delete
    5. I'll agree the optics are terrible. The government couldn't give a straight answer when being asked what I think were innocuous questions with innocuous answers.

      So they're being terribly deceitful about something that doesn't really matter...

      What's this constitution angle? I've not heard that one.

      Delete
    6. chirumenga,

      Harper can be accuswed of many things relating to "L'affaire de Mike Duffey" but acting in an extra-constitutional way to stack the Senate is not one of them. The Constitution gives the Crown the ability to appoint senators, Harper has done so on many occasions-exactly where doers this unconstitutional or extra-constitutional use of power arise? It doesn't because it is mere partisan mud-slinging.

      It does however, provide an intimate glimpse into the mind of the NDP and should remind people the NDP does not understand the Constitution and has no intention of following the Constitution or the rule of law in Canada should they ever form Government. We see this everyday with Mulcair's contention he can abolish the Senate when every first year poli-sci student along with the Supreme Court of Canada knows and has ruled that Senate abolition will require a formal constitutional process and negotiations.

      The NDP motto: One rule for the NDP another for Canadians!

      Delete
    7. Ira, the Constitutional aspect is that, legally, Senators are determined by residency and not merely by where they happen to spend money. The infamous cheque was an attempt to prove Duffy's residency in PEI by showing that he'd spent money on his accomodations there, but that is not sufficient per the legal standard - as, in fact, Harper's own legal counsel advised him. The emails offered in court outline this whole debate within the PMO on this point. By lowering the standards applyingto Resdiency, Harper could effectively stack the Senate regionally in any way he wanted. One of Duffy's claims in court is that he did not want to be a Senator for PEI, because he knew by legal standards that he wasn't a resident of that province.

      Delete
    8. Pete, see my reply to Ira.

      Nothing to say in response to your random musings about the NDP.

      Delete
    9. This is the crux of things really....take any law for example, which laws are enforceable, the ones in which most people comply with. If everyone is breaking a law, its not really enforceable since its rejected by the population at large and would be considered an unjust law. Conversely, if the public at large does not much care about the constitutionality of abolishing the senate, they only care about getting rid of it, arguments about how it cannot legally be done will not resonate much.

      Delete
    10. I disagree Carl. People don't like the Constitution or the Senate being manipulated for personal gain/vested interest - even if the abused institution is not something they have a great affinity for.

      Delete
    11. I'd like to return the conversation to the political effect...The trial is now postponed to November. Will the Duffy effect just continue when newspapers and other media need new stories each day or will public opinion be drawn into how other issues take over?

      Delete
    12. In the mainstream media, because the trial is on hold till November, they will mostly ignore the issue (unless new information somehow arises). The public, on the other hand, may well continue to look into the scandal through whatever means available.

      Delete
  7. Why does Nanos continue to be the odd one of the bunch? Numbers don't seem to look credible.

    ReplyDelete
  8. Do the CBC editors make up your Headline or are you personally biased against the CPC?

    The polls have not changed and are basically the same as they were when the election was called. Here is an analysis of the Headlines under your name at CBC

    Party Neutral headlines 3
    Anti CPC headlines 5
    Anti Liberal Headlines 0
    Anti NDP headlines 0
    Pro NDP headline 2
    Pro Liberal Headline 1
    Pro CPC headline 1

    The headlines in Question in most recent published order:

    Conservatives have lost edge they had on the economy, polls suggest (anti-CPC)

    Conservative swing voters could drift because of Duffy trial (anti-CPC)

    Parties to parents: We want your vote (neutral)
    NDP continues to lead subtly shifting national race (Pro NDP)

    Quebec currently Conservatives' only hope for seat gains (Anti CPC)

    Ontario behind tightening 3-way national race (neutral)
    Liberal fortunes improve following 1st leaders' debate (pro Liberal)

    Conservatives stand to suffer most from lack of incumbents (anti CPC)

    NDP lead widens slightly over Conservatives (pro NDP, anti CPC)
    Harper, Mulcair in close race as campaign begins (neutral)

    How will Joe Oliver, Olivia Chow navigate Toronto's shifting electoral tide? (neutral)

    Conservatives move into tie with NDP (pro CPC)

    ReplyDelete
    Replies
    1. The CPC is doing poorly in the polls. They are down 10 points from 2011. The NDP is up two or three, the Liberals are up eight or nine.

      If I was writing the same number of pro/anti headlines for each of the parties, that would suggest a pro-CPC bias.

      Delete
    2. @ BC voice of reason...Your comment is dull!!!

      Harper calls the longest election writ in Canadian history, all in an attempt to silence third party campaigns, all to gain an edge with "more money for attack ads than the rest"

      The budget is in deficit, the economy flat, in recession.
      Harper in my opinion is lying his face off about the Duffy scandal.

      National Post is in the tank for Harper, Globe n Mail too..You watch, those propaganda rags will endorse Harper.

      Meanwhile Stephen Harper threatens CBC with a scalpel....

      And you whine about headlines????

      Stephen Harper is done and you are NOT a voice of reason for British Columbia..you are alone in the wilderness.

      British Columbia is saying to Harper..

      Go #$%$% yourself

      http://www.vancouversun.com/news/voters+planning+push+Harper+Tories+office+election+poll/11316003/story.html


      Delete
    3. Doing even worse in today's Forum poll. CPC is third at a dismal 23%, Harper is third as best PM and a terrible -41 approval rating. The wheels are falling off the Harper bus.

      Delete
    4. Grant G,

      Harper called the 3rd longest election campaign in Canadian history-not writ!

      A writ is the formal piece of paper that is signed to initiate the campaign. Please take note.

      Delete
    5. There is a difference between being "anti-CPC" and reporting on the outcome of polling. "NDP lead widens slightly over Conservatives" is factual, not biased.

      Delete
  9. Will you be updating the riding projections today to reflect the new Nanos poll and your new poll tracker update?

    ReplyDelete
    Replies
    1. I will try, but running up against the clock.

      Delete
  10. Will there ever be a statistical measure of strategic voting, and how it might impact riding outcomes? The question is, "SV, yes or no." Assumption: the "S" is getting rid of Harper's Conservatives.

    I personally know of a couple of scenarios: Green supporters go SV a) to preclude a Con win, b) to focus donations onto the couple ridings Greens have a real shot at, c) to have a shot at the balance of power in a hung parliament---or a cabinet seat in a coalition, d) to elect a government committed to electoral reform which would ultimately result in more Green seats; Traditional Liberal voters who broke the mould in the last election ( in protest of their illegitimately appointed---instead of elected by the membership--- leader) go SV to a) preclude a Con win, b) to continue rebuilding the party in the meantime, c) increase odds of coalition.

    I've only met Dippers from one riding whose SV candidate isn't already NDP (Vancouver Granville)---and they're still pretty confidant their candidate will overcome the "early lead" of the current Liberal SV choice; in fact the NDP is gaining here. Otherwise, in most BC ridings the SV choice is the same as the runners-up last time: NDP.

    ReplyDelete
  11. Question about the Nanos poll is a rolling four week poll... do they ever publish the results of just one week.

    Eric while you stated you don't think this difference is due to old data; are you saying then that Nanos is catching something no one else can see? Or are you saying that their weekly sample size or methodology is a problem.

    Personally I think the data over time is less a problem then the amount of data being added. While the rolling poll has a total of 1000 people the weekly input is only 250 people, that seems like a very low number. What would be the MOE on a poll that small, even more so what would be the MOE on the provincial sub-groups?

    It would be interesting to see what a rolling poll of Forum's or EKOS's last four weeks would produce. I have a feeling they would not replicate the same trends as Nanos.

    Then again my concerns could be nothing and Nanos might well be catching the NDP in decline before anyone else... but I don't think so.

    ReplyDelete
    Replies
    1. No, Nanos doesn't publish one-week-only data.

      I'm saying that it can't be old data because the CPC was never that high in Ontario. So it could be they are catching something or it could be they are wrong.

      Delete
    2. The striking thing is that the Nanos charts show a downtick in the chart points for the Conservatives in every region except Atlantic Canada - including Ontario.

      Delete
  12. The rolling average from Nanos seems so odd. Does anyone else do that kind of poll? Even Gallup's rolling averages have only three days in the sample.

    Such small samples per week (250) over such a long period of time (4 weeks) seem unusual at least and the smaller weekly samples could be more susceptible to variations from week to week. Could explain why it shows such different results from other polling firms

    What would the projection look like without the Nanos rolling polls?

    ReplyDelete
  13. Today struck me as an interesting day. I feel that the Conservatives have to be breathing a sigh of relief at the announcement of the suspension of the Duffy trial until after election day.

    I also found the Trudeau "austerity" charge against the NDP interesting. This is purely speculation on my part, but it makes me think the Liberals are more interested in being the top left-leaning party than sharing power with the NDP. I wonder if Trudeau fears that a second consecutive third place finish would seriously damage the party.

    All in all, I think Harper has to be happy right now, though he probably still has to step perfectly for a chance at even a minority.

    ReplyDelete
    Replies
    1. There's a chance Harper has an extra advantage. The Fair Elections Act prevents Elections Canada from encouraging people to vote. Normally in elections, we'd see ads from Elections Canada telling them when the election was and how to find out where to vote, but this election there will be none of that.

      So, if that has an asymmetrical effect on the electorate, it will likely affect the outcome. To what degree? And presumably Harper thinks it will tilt things in his favour, else he wouldn't have done it. But we need to see one election with that rule in place to find out what it has actually accomplished.

      Delete
    2. Not being Canadian, I wasn't aware of that. Methinks that could be an advantage to the Tories because of the Big Blue Machine. Maybe the Conservatives have more margin for error than I thought. I'm going to have to remember to follow up on that after the election.

      Delete
    3. "The Fair Elections Act prevents Elections Canada from encouraging people to vote."

      How can any political party in good conscience want fewer people to vote? Even in discussions I have with people that don't share my political views, I encourage people to vote.

      Delete
    4. There was a pollster who wrote an op-ed (published in, I think, the Calgary Herald) arguing that lower voter turnout improved the fidelity of the public opinion data. By weeding out disengaged or less enthusiastic voters, a lower turnout has less noise in the signal.

      I don't think there should be barriers to voting, but I also don't think universal voting would be a good idea. Having the people self-select might be the way to go.

      But that was already true. Now Elections Canada isn't allowed to educate people about when the election is or where and how to vote. Sure, that information is on their website, but they're not allowed to advertise that. So the self-selection is no longer made by a well informed populace.

      Delete
    5. Because....winning elections matters more than having a well informed and engaged electorate....

      Delete
    6. The proof will be in the results.

      If it turns out that Elections Canada's ads had little effect, then there was no reason to stop them.

      But if they did, then that suggests that lots of people who didn't otherwise care much about the election were inspired to vote by the ads. If that's true, then the ads do create noise in the signal, and there's an argument for stopping them.

      Again, assuming we buy that fidelity argument.

      Delete
    7. Wow, the CPC is getting desperate. Openly promoting voter suppression. Can they get any worse?

      Delete
    8. Jimmythedeke, they can get worse if they're re-elected.

      Delete
  14. Would the greens likely gain any seats with the support levels the insight west poll gave them on Vancouver island?

    ReplyDelete
    Replies
    1. Absolutely.

      Even within the island, voter support isn't evenly distributed. If these numbers are real and hold, we'll see a lot more Green seats than Éric's model is currently predicting.

      Beyond that, Vancouver Island is obviously a major focus of Green efforts. The key is increasing voter turnout; the biggest Green opponent is not the Tories, Grits or Dippers, but "Didn't Vote". Elizabeth May's landslide over a former cabinet minister came mostly from increasing voter turnout, not converting voters. You can expect to see that replicated across the island this campaign. Green fortunes are intimately tied to engaging the electorate and strengthening Canadian democracy.

      Get-out-the-vote observations apply to decided voters, but also those who are still making up their minds. The 19% Vancouver Island "Undecided" vote is three times the Green/NDP difference, so there are massive Green opportunities.

      So yes, this poll points to a healthy Green caucus in the next Parliament. With over six weeks to go, it will be interesting to see how the momentum builds.

      Delete
    2. Keep dreaming...

      unless there is a major shift May will keep her seat that's it.

      But heck I never believed the NDP could win a majority in Alberta and now I'm on the hook for a bet that the NDP will win 209 seats Oct. 19th.

      Delete
  15. I think that the NDP is basically a Quebec regional party that will have a hard time winning in the rest of Canada. In the GTA the NDP is unpopular and I think that the Liberals will win big there. I think that the Liberals have a much better change than the polls suggest they do. (Full discloure: I intend to vote Liberal).

    ReplyDelete
    Replies
    1. Have you looked at any polls in the last 5 years? STicking just to the most recent history. In BC, the NDP has been leading by double digits in most polls for months now. In Ontario, there is a three-way tie going back for some time now as well. The NDP is in second place in every other region, and polling first in a couple of polls in Atlantic Canada. If the NDP is "a Quebec regional party", then it seems the whole country has been absorbed within Quebec's borders.

      Delete
    2. I wouldn't make any bets if I were you.

      Delete
  16. Where did you find the NDP 37, CPC 30 and LPC 24 eligible voter numbers in the Angus Reid poll? I can't find them.

    Thanks!

    ReplyDelete
    Replies
    1. Hi William - see my response to chimurenga below.

      Delete
  17. Éric, On your CBC Poll Tracker, why did you use Angus Reid's "eligible voters" data (and where did you find those numbers) rather than the "likely voters" numbers provided in the linked article? Your chart gives Con 30; NDP 37; Lib 24, but Angus Reid's "likely voters" results are Con 32; NDP 36; Lib 23.

    ReplyDelete
    Replies
    1. When I have the option, I use eligible vs. likely tallies. The Ontario election is why - it convinced me that, until proven otherwise, pollsters have yet to settle on a winner 'likely voter' model.

      As to where I got the numbers, you could calculate them by removing the undecideds from the eligible voter tally in the report. I did that, and also asked ARI for the numbers. They provided them to me, and were virtually identical to my original estimations.

      Delete
    2. YOU could calculate them, I prefer to think of math as a spectator pastime...

      I recall the discussion about eligible vs. likely following the Ontario election... No "eligible voter" data for the other election period polls?

      Delete
    3. Do you have the regional breakdowns of the eligible voters for that Angus Reid poll?

      Delete
    4. ARI published those.

      Delete
    5. I only see the regionals for likely voters, not eligible.

      Delete
    6. Nazar, I believe Éric contacted Angus Reid directly for the eligible voters figures - after calculating them himself from the published data. There's a comment from him somewhere on this page to that effect.

      Delete
  18. Today's Toronto Star Forum Poll !!

    The New Democratic Party has the backing of an unprecedented 40 per cent of Canadians, a level of support that could secure Thomas Mulcair a majority in the House of Commons, a new poll has found.

    The Forum Research poll for the Toronto Star projects the NDP with enough support to win 174 seats in the Oct. 19 election. Justin Trudeau’s Liberals now sit in second place with 30 per cent support, while Stephen Harper’s Conservatives are losing support and have the backing of just 23 per cent of the 1,440 Canadians surveyed.

    ReplyDelete
    Replies
    1. One of the Forum and Angus Reid polls is wrong.

      It remains to be seen which.

      Delete
    2. At this point, this number seems exceedingly too low for the conservatives. I'd say outlier for now, but if the next few polls have them at less than 26% or so it could be plausible.

      Delete
    3. Not unexpected. With all the dirty dealing on the Duffy trial, Carson's fraud trial coming up and the CPC supporter today reported to be fraudulently charging Liberals credit card donations as CPC donations, I'll be surprised if its not Kim Campbell Redux.

      Delete
    4. And it looks like EKOS might be backing up those results

      Delete
    5. Hi Chimurenga, where did you see the EKOS results?

      Delete
    6. Jimmy,

      The context in 1993 was so completely different. The country had just emerged through a number of heated rounds of constitutional negotiations. Constitutional wrangling that both Justin and Mulcair wish to re-open. We were still adapting to the reprecussions of free trade with America and the French-Canadien psyche (if I can use that term) was both more fragile and uncertain than it is today. These factors cause the Canadian political system to fracture, the same criteria or catalysts are not in play in 2015.

      Delete
    7. Nazar,

      When I posted, i hadn't seen the results but Frank Graves had hinted on Twitter that the results would show a big shift, and suggesting further strength in the NDP lead. As it turned out, it was less dramatic then he intimated. Go to this Wikipedia article for all the poll results (it's gets updated quite promptly, largely thanks to 308 regular, Undermedia), including a link to the latest EKOS figures :

      https://en.wikipedia.org/wiki/Opinion_polling_in_the_Canadian_federal_election,_2015

      Delete
  19. How much influence does the Angus Reid Institute (eligible voters) poll currently have on the poll tracker? (It's a bit difficult to tell exactly on the bar chart on the website, I'm not sure if it is possible to clean it up or put the percentage somewhere in the table?

    I am a bit concerned that with a reasonable amount of polling in the field one poll is able to have so much impact (it looks to me like it is making up 66% of the polltracker). The NDP at 37% may be a start of a trend or it may be an outlier but including it as 2/3s of the model means the NDP sit at 35.4% despite the previous ten polls having the NDP average at about 32%. I know the readers of your blog recognize the numbers can bounce around a bit based upon the influence of a recent poll but given the prominence that media are giving your numbers, I think you should re-consider the influence one poll can have, particularly during an election campaign. This one poll is at worst only a minor outlier but my fear is a poll that is a large outlier could easily falsely suggest a large change in your model. Some media would then use your data along side the poll to demonstrate a dramatic shift in the likely composition of the next parliament. .


    I'd just like to close saying I think your work is excellent and to thank you for the positive impact you have had on media coverage in Canada. Your analysis is serving as an important stabilizing influence on the media's portrayal of polling data. I hope you will take the very pedantic nature of my feedback as constructive suggestions on how to improve the model when you do your next review.

    ReplyDelete
  20. I don't know about you Eric but I find these numbers a little "suspicious" as the shift in a short period of time seems extreme ??

    ReplyDelete
    Replies
    1. Duffygate. The obvious cover up of this mess is killing them.

      Delete
    2. OK check these Forum shifts out

      Aug 17-19

      CPC 29

      NDP 34

      Libs 28
      -----------
      Aug 23-24

      CPC 23

      NDP 40

      Libs 30

      So the CPC has dropped 6 and the NDP has gained 6 and the Libs have gained 2. Bit odd if you ask me ???

      Delete
    3. Yeah Jimmy it could well be. Let's hope !!

      Delete
    4. We will see if other polls support or challenge this one.

      Delete
    5. Ipsos has CPC in third too.

      Delete
    6. We'll it is Forum, so take that as you will. However it's not impossible that the CPC have a major drop due to the economy and Duffy.

      It's a bit shocking to see them drop lower then the LPC, however I would envisions a lot of progressive moving over to the NDP from the LCP and the right-wing vote going from CPC to LPC.

      Delete
    7. Ipsos shows a tight race.

      It is a mischaracterization of their data to say that it shows the CPC in third, as the gaps are not large enough to have confidence in the ranking.

      Delete
    8. I don't see why the right-wing vote would go to the Liberals. They're the only party promising to run deficits.

      Delete
    9. As opposed to Harper, who has always run deficits (if you exclude asset sales to balance the books)? At least Trudeau is being honest about the extent of our recession, but we all know how well honest works in politics. I think regardless of party, we're going to be running a deficit in 2016.

      Delete
    10. Either way, Harper is losing ground fast. It is a race between the Libs and the NDP. Harper is done.

      Delete
    11. Harper almost released a balanced budget in late 2008, you might recall, but the opposition threatened to topple him over it (Harper was insisting that there was no need for stimulus - he was right).

      The real reason for the opposition's actions was, of course, Harper's plan to cancel (with immediate effect) the per-vote subsidy, but the opposition's PR won the day. After prorogation, Harper released a budget in 2009 which wasn't close to balanced (because of the extensive economic stimulus which he'd insisted only months earlier wasn't necessary) and with a much slower phase-in of the end of the per-vote subsidy.

      It was Harper's worst moment as Prime Minister, I think, when he blatantly abandoned his own good judgment for reasons of political expediency.

      And he hasn't been the same since. I liked Harper before that. Until that 2009 budget, Harper was the guy I always hoped he'd be. Since then, not so much.

      Delete
  21. Good thing I put money on "NDP Majority Government" two weeks ago when the odds were 17:1. I figured there was no way Tory support would stay constant through an entire month of the most negative coverage the party has had since the Stockwell Day era.

    As for the National Post's analysis that NDP numbers would stay low in Ontario due to voters' memories of Bob Rae, that sounded highly implausible, considering that Rae was turfed out more than 20 years ago (meaning anyone under 40 barely remembers him) and anyone who cares enough about Bob Rae to let that decide their vote will know that Rae has been a major figure in the federal Liberals for the past decade.

    Harper's electoral strategy has been horrible: (I) calling the election before the Duffy trial, which guaranteed to amplify its importance (ii) agreeing to an early MacLean's debate that nobody would watch, giving the debate rookies Mulcair and Trudeau a free practice session before higher profile debates coming up (iii) extreme limits on reporter questions and giving memorized-by-rote answers, which was only going to annoy the reporters into giving negative coverage; (iv) now, playing up economic uncertainty as a reason to vote Tory -- surely the last thing that he should be reminding voters about is that there is economic uncertainty after he has been in power for a decade. I know this is a long campaign, but the momentum is already in motion and there won't be a lot of happy Tories on October 19.

    ReplyDelete
    Replies
    1. I don't disagree with your analysis, the NDP winning is certainly possible and flagging Conservative support over the campaign is possible but they do have a very committed core...but there is a lot of time for more movement in the polls and with numbers so close between the parties a couple % points is enough to make a difference. I am far less certain compared to you on the outcome.

      Delete
  22. HI Eric
    I always loved your work but I feel your new tracker a bit off lately. With the update today you made the liberal go below 26% when three of the latest pools have Liberals at 30%. Only one poll have the liberals very off and earned way too much credit. I understand it's was the biggest poll since the beginning of this campaign… but still weighing more than double others polls do not make sense. Even in July Angus had the liberals really low more than normal. Hard to trust a model when in one day riding projection change so much only for one poll.

    ReplyDelete
  23. Looks about right for a party that has gone through the amount of scandal Harper's party has. The wheels are falling off the Harper bus.

    ReplyDelete
    Replies
    1. Still alot of hockey to be played Jimmy, you may end up eating your words!

      Delete
  24. Hi Eric,
    I have a small problem with your weighing method. You are putting to much emphasis on the recentness of the polls. The result is that usually only one poll dominates your numbers and with each new poll we see a sudden shift in your numbers. In my opinion, in a true aggregation, at least 4-5 polls should be weighted somewhat equally. At least, you can treat any poll within the last week as a recent one.

    ReplyDelete
    Replies
    1. The age weighting was determined by past elections - it is the rate of decay that works best, and is necessary to capture trends occurring in the final week of a campaign, of which we've seen many examples.

      But the decay rate is based on a short, normal campaign. It is fluctuating a lot at this stage because there aren't many polls. This will not be an issue when we get deeper into the campaign, and we will start seeing 2-3 polls per day. When we reach that part of the campaign, the most recent 5-10 polls will be carrying a lot of weight.

      Delete
    2. I think the method you used to develop the rate of decay makes sense and I have seen in past election it works well to blend polls late in the campaign. It's the 2/3's cap on any one poll that I think needs some further consideration. Obviously you don't want to change things this election but after the election I'd encourage you to graph a few simulations at different maximums (i.e. 1/3, 2/5,1/2) and see how they look. I think ideally the model should look closer to a smooth line in picking up trends. I suspect with some polls taking up 2/3rds it will lead to some choppy spikes on the graph.

      Delete
    3. I did! That is why I settled on the 2/3 cap.

      Delete
  25. Hmm, I guess I am just looking for a reason to be skeptical of the recent NDP bounce - I am sold, it looks like the NDP are experiencing a bounce.

    ReplyDelete
  26. What an odd group of polls! There's obviously plenty of reason to think the NDP are leading since they're leading most of the polls in the last two weeks, but I agree with those who have expressed their doubt about the 40% mark of the Forum poll.

    By the same token, I find some of the numbers in the Ipsos poll strange as well. I hardly see how the Tories could be leading in BC and with the way the Dippers have been dominating the province and be in third place nationally.

    There's a very odd dynamic to those polls, imho.

    ReplyDelete
    Replies
    1. If you look at BC's election results historically, they have a recent tendency to move in the opposite direction of the rest of the country.

      Delete
    2. True between the NDP and then Reform, but it's the swing the polls have shown in such a rapid fashion that seems so odd. I'm not saying it couldn't be happening, but it looks odd. I wouldn't question it at all if it had been a longer term trend

      Delete
  27. Eric, in ridings where a party has not chosen a candidate yet, or appears to have decided not to run one (e.g., CPC in Churchill - Keewatinook Aski in Manitoba) how do you assign percentages to a non-existing candidate? Are the 2011 election results factored in the predication?

    ReplyDelete
    Replies
    1. All projections are based on a generic candidate, unless they are an incumbent or a star candidate. If a party does not run a candidate, they are removed from the calculation and the vote is distributed proportionally to the other candidates.

      Delete
    2. Do you find that there's a difference in accuracy in the predictions between ridings which have all the major parties represented and those which don't?

      Delete
    3. I haven't looked into it, I'm afraid.

      Delete
  28. The way your methodology downweights Nanos, the most accurate pollster from 2011, is an issue.

    ReplyDelete
    Replies
    1. That won't be the case once Nanos moves to a 3-day rolling sample.

      Delete
    2. Are they planning to?

      Delete
  29. Also, do you still apply a star candidate factor? If so you should probably apply it to the Liberals in Delta, where the former President of Canada's Paralympic Committee, Carla Qualtrough, is running.

    ReplyDelete
  30. Eric could you elaborate on your view "It is getting to the point where we can only call this a three-way race within the historical context."

    I personally have been thinking that calling the campaign thus far a three way race is more media spin for ratings then facts.

    ReplyDelete
    Replies
    1. Within a historical context, this is one of the closest three-way race we've ever seen.

      But objectively, the NDP is holding a pretty good lead - at least in the aggregate.

      Delete
  31. Notley, Duffy and Trudeau have all been nibbling down CPC numbers.

    Notley upgraded the NDP to potential winner status.

    The Duffy trial has not helped the CPC brand. The PMO guy in the red hoodie attending the trial is enormously disrespectful of the court. He's masquerading as somebody homeless. Would he show up at the PMO dressed like that?

    Trudeau in the first debate outperformed the low expectations set by the CPC attack ads and got a nice boost in subsequent polls.

    We see the NDP and CPC statistically equal in Ontario. But when the CPC polls below both of them, it ain't gonna help the CPC.

    ReplyDelete
    Replies
    1. Correction: NDP and LPC statistically equal in Ontario.

      Delete
  32. I don't think the Tories will end up in third place but, it sure looks as if they may lose the election. I don't think they'll end in third simply because of the rural vote in Ontario and the Prairies should give them roughly the Liberal seat count average. The shy Tory factor is also likely to boost their popular vote number come voting day. Finally, as we saw in BC those who answer polls and those who vote are not necessarily the same. Those who vote tend to be older and more conservative. These factors will likely save the Tories the humiliation of third place.

    ReplyDelete
    Replies
    1. I don't think they will lose the election. First of all the quality of polling continues to decrease and there are far too many online panels in these samples. Second of all unless the polling is done in the major language groups for the various ridings, it's also not going to be accurate. Everyone screams about cell phones, while ignoring that a riding may be 1/2 Mandarin, and receiving polls in English or French.

      Third of all, Alberta is controlled by the NDP, Ontario by the Liberals. Electing either of those parties Nationally won't sit well with people who show up to vote. And there is a majority of people who will not want the NDP in power. if it looks like they will win, these people will vote Conservative even if they said Liberal to every poll.

      And in response to the but in Alberta response to that meme, the day Trudeau quits the Liberal Party and joins the Conservatives for a cabinet position , I'll give the win to the NDP. That is the analogy that fits. NDP polling data fits the Broadbent crash and burn or the Adrien Dix crash and burn not the Notely protest vote against a party leader crossing the floor, which is virtually unheard of.

      Delete
    2. Shoshana,

      Sorry but you're blowing smoke, this election is nothing like 1988, and even less then nothing like Dix in BC; for two simple reasons, Quebec and being Her Majesties Official Opposition. You can complain about the quality of polling till you're blue in the face (pun not intended) but there is ZERO evidence for any of issues you bring up.

      Likewise there is ZERO evidence to prove that people will switch from LPC to CPC simply to stop the NDP. There is no anti-NDP or even anti-LCP movements, there are however multiple campaigns, groups and activists pushing ABC/anti-con/anti-harper. On the other hand there are big groups of NDP/CPC swingers in the west and NDP/LPC swingers in Ontario and the Maritime which does not bode well for the old line parties.

      Now will the CPC come in last... I don't think so, not unless they have a major collapse in Alberta (or Ontario) as what happened in May, The LPC on the other hand, at this moment are, in dire straits, they need a big improvement to stay out of third however they are well within reach of doubling their seat count at least. In any other election that would be worthy of celebration, but I think the Liberals are making a last stand. Anything less then Official Opposition to a Majority Conservative Government would be a death Knell. Their whole Raison D'etre disintegrates if the NDP win or even place ahead of them, I would think the would suffer a similar far to the UK Liberal party.

      Delete
  33. Harper is losing ground. Its a drip drip drip slow decent that he cannot seem to stop. The frustration and desperation shows on his face.

    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.