Friday, February 25, 2011

Abacus has Conservatives still well in front

It appears that the pollsters now disagree once again. Abacus Data has come forward with a new set of numbers, showing that the Conservatives hold a massive 15-point lead, and arguing that the gap has not narrowed as EKOS Research has recently found.

However, we should not compare apples to oranges. EKOS conducted its poll over two full weeks, while Abacus's survey was conducted on one day. Without daily tracking of the EKOS poll, we cannot say with any real confidence that these two polls are contradictory. For all we know, in the last few days of EKOS's poll the Conservatives could have had a double-digit lead.

Consider also that a poll conducted on one day would be far more susceptible to individual events and how they have been portrayed in the media than a poll conducted over two weeks. Imagine a poll conducted today, with news that high-level Conservatives have been charged with breaking electoral law emblazoned on every front page.

In other words, this poll needs to stand alone and not be compared to the recent EKOS survey. If other polls show that the Conservative lead is still in place, then Abacus's poll will simply provide further evidence. If, on the other hand, they show that the Conservatives have lost support, there is a very good chance that Abacus was accurately recording the voting intentions of Canadians on February 23 only.

Enough with the preamble, now to the poll. We don't often get one-day polls, so it is an interestingly precise snapshot in time.Since Abacus's last poll in late January, the Conservatives have gained three points and now lead with 38%. The Liberals, on the other hand, have dropped four points and are down to a woeful 23%.

This poll was conducted online, and a random sample of similar size would have a margin of error of +/- 3.1 points, 19 times out of 20. But according to this paper, it is simplistic and inaccurate to say that, based on the 3.1 MOE, only the Liberal shift is statistically significant.

The fact of the matter is that properly calculating the margin of error for comparing two different sets of polls is not the same as just looking at the margin of error of a single poll. For the Conservatives, comparing their result in Abacus's last poll to their result in this poll would have a margin of error of +/- 5.2%, assuming that this would be a random sample. The Liberal margin of error over these two polls is +/- 4.3%. That means that neither shift is statistically significant.

The permutations of calculating a margin of error are complex. There are different margins of error for the different parties due to their level of support (parties with less support tend to have a smaller margin of error). For example, the margin of error for the New Democrats in this survey would be +/- 2.4%, rather than 3.1%. There are different margins of error for gauging the statistical significance of a gap between two parties (in this case, it is 4.7% for the Liberals and Conservatives), and as noted there are different margins of error when comparing one party's results in one poll to another.

Based on this paper it appears that virtually everyone talks about the margin of error incorrectly, including in the recent Canadian Press piece on polling. I'm not sure how to approach this in the future, but I will be taking more care before announcing any shift as statistically significant or not.

Today's post is quite tangential, so let's get back to it. The New Democrats are up one point to 19%, while the Bloc Québécois is up one to 11%. The Greens, meanwhile, are down one to 8%.

Regionally, we see a lot of the same things that other polls have shown. In Ontario, the Conservatives are up one point to 39%, while the Liberals are down seven to 29% (though apparently even that is not a statistically significant drop). The New Democrats are up two to 18%, while the Greens are up three to 12%.

In Quebec, the Bloc leads with 44% (+3), followed by the Liberals and Conservatives at 19% (+1 each). The New Democrats are down one to 15%.

In British Columbia, the Conservatives lead with 40% (-1), followed by the New Democrats at 30% (+1) and the Liberals at 21% (+3). The Greens are down four to 9%.

In Atlantic Canada, the Conservatives are leading with 39%, up nine points. The Liberals are down 15 to 29%, while the New Democrats are up four to 27%. Again, even these wild shifts appear to not be statistically significant.

The Conservatives are leading in Alberta with 70% (+15), while the Liberals are down 13 points to 11%. In the Prairies, the Tories have gained four points and stand at 48%, followed by the New Democrats at 26% (-1).

With this poll, I would project 19 Conservative seats in British Columbia, 28 in Alberta, 21 in the Prairies, 58 in Ontario, eight in Quebec, and 12 in Atlantic Canada for a total of 147, a gain of 14 seats compared to the projection for Abacus's last poll.

The Liberals would win six seats in British Columbia, none in Alberta, three in the Prairies, 31 in Ontario, 13 in Quebec, and 15 in Atlantic Canada for a total of 70, down 18.

The Bloc Québécois would win 53 seats in Quebec, down two.

The New Democrats would win 11 seats in British Columbia, none in Alberta, four in the Prairies, 17 in Ontario, one in Quebec, and five in Atlantic Canada for a total of 38 (+6).

For regular readers, note that the new projection model is in place for British Columbia and Alberta.

Abacus also helpfully asked the question Angus-Reid neglected to on the topic of Bev Oda. Do you know who she is? Amazingly, 46% of respondents said they did. It would appear that the 15% of Canadians who pay attention to politics are joined by another 31% who aren't exactly ignorant of what is going on in Ottawa.


  1. From the Sun:

    "Abacus also asked Canadians why Oda was in trouble and found that 46% of respondents correctly identified some aspect of the Oda controversy."

    From here:

    "Do you know who she is? Amazingly, 46% of respondents said they did."

    Not quite the same thing. Was it a direct question "Do you know who she is?" that would ask people to pump themselves up as to how smart they are and say "yes". Or did they ask a few questions looking for someone to respond... and show a little bit of knowledge?

    Personally I am just as concerned about Rae's Jihad language, Graves calling for an election, Iggy refusing to help make parliament work... etc etc

  2. Eric, it's not "would project." It's project. You are projecting. You are not possibly projecting. To project is what you are doing. It sound like you are in fourth grade making a presentation to the class when you write, "I would project."

    Based on your model you project the number of seats any given party (would) win (if) an election (were) held today. The parentheses indicate the correct placement for the conditional tense.

    Sorry but its been driving me crazy for a while.

  3. For all of 2010, CPC support sat in a very consistent band:

    29% - 37%

    So far in 2011, CPC support has fallen inside a similarly broad band, but three points higher:

    32% - 40%

    I'd say they look stronger this year than last.

  4. Its interesting that even with a 15 point lead - the seat projection still doesn't yield a Tory majority. I have to think that if the vote spread in Atlantic Canada actually was CPC 39, Libs 29 NDP 26 - you would NOT see the Liberals with 15 seats and the Tories with 12 in that region. The Tories would pick up a few more.

    Also, I think that if the Ontario numbers were correct and the Liberals dropped from 34% to 29% while the Tories and NDP stayed put at 39% and 18% respectively - the NDP would gain a couple of Liberal held seats in Toronto (ie: Parkdale-High park and beaches-East York and maybe Davenport).

    IF the vote in BC was actually Tories 40% (down 5% from '08) and NDP 30% (up 4% from '08) - you would probably see more like 12 or 13 NDP seats with ridings like Kamloops, Newton-North Delta and Esquimalt Juan de Fuca being high on the list of medium hanging fruit.

  5. Anonymous,

    Duly noted. I use "would" because I don't consider my poll-specific projections to be very accurate, as the results have a lot to do with the vagaries of MOE and methodology. I don't project with any great degree of confidence that this is the actual situation, I'm just translating a poll result into seats.

    My "official" projections, however, are done with confidence. That's why when describing a projection update I say "I project" or "X party is projected to"...

  6. DL,

    With this poll, I have the NDP within 3.2 points of the Conservatives in Kamloops-Thompson-Cariboo, within 8.8 points of the Liberals in Newton-North Delta, and 7 points of the Liberals in Esquimalt-Juan de Fuca.

  7. I've found Abacus polls to be unusual since they started a few months ago. I don't agree with their metod of polling for one day either. Whatever happens in the news that day will effect the results.

    It's similar to the poll that was conducted in NL a few weeks ago. It was conducted over two days and the first day was when Harper was unveiling the new ferry for Marine Atlantic. NTV who commissioned the poll even said it most likely effected the results.

    I'd say if a poll was conducted today in NL on federal politics you'd see a big difference because people are pissed with a few things that have happened this week with the CPC.

    Polling over a period of a week would obviously give a better idea of intentions. CBC had shown the daily numbers of support for the parties but don't anymore.

  8. Eric where do you show your seats in the East? Just wondering if you could break it down by province?


  9. Earl, I don't have riding-level projections for Atlantic Canada yet.

  10. "Whatever happens in the news that day will effect the results."

    Different from "whatever happened this week", How??

    "I'd say if a poll was conducted today in NL on federal politics you'd see a big difference because people are pissed with a few things that have happened this week with the CPC."

    We did.... 2 days ago. Abacus. 39-29-27. But maybe that artificially lowered the 39 that the tories got this time because of the single day during the "people are pissed with a few things that have happened this week with the CPC."

    Then there is Ekos that finished on Wednesday... 37-30-14-17. .. But then,,, maybe stuff was affected by events the week?

    But that isn't affected as much as the day... which is better or worse for the tories, huh

    I um , no wait, I don't get it... wait... Which whatchamacallit was the effect with made the tories higer, but lower where the liberals went up, or down...?? huh?

  11. "This poll was conducted online, and a random sample of similar size would have a margin of error of +/- 3.1 points, 19 times out of 20. But according to this paper, it is simplistic and inaccurate to say that, based on the 3.1 MOE, only the Liberal shift is statistically significant."

    Stop. Stop right there before you go into your whole tangent about the proper way to calculate/report margin of error. Because frankly, it's irrelevant.

    The Abacus poll was conducted online, with people who signed up to participate in a panel. The 1,005 participants in this particular may have been "randomly selected" from that larger sample, but the fact is that it is not a random sample, and therefore you CANNOT ascribe a margin of error. Period. Full stop.

    Since there's this whole meta-debate roiling in the newspapers and on blogs about the state of polling and how the media report political polls, we should really be discussing and highlighting this disingenuous practice.

    Pollsters do it because it gives their online polls an extra veneer of credibility (even though the MRIA still says they shouldn't), and then media outlets erroneously put this in their stories because the journalists writing them either don't know any better or don't care.

    Credit to you, Éric, for pointing out that a "random sample of similar size" would have a particular margin of error, but if you really want to be accurate, you shouldn't be using the term at all when describing these types of polls.

  12. Barcs,

    On the question of knowing who Bev Oda is, the full Abacus report makes clear that those polled had to give more than just a 'yes' answer -- they were looking for answers akin to: 'federal politician, Conservative cabinet minister, or federal minister'.

  13. This is an online poll and in case people haven't noticed, online polls almost invariably have very short times in field (ie: 1 or 2 days). This is because the survey gets e-mailed to however many people in the panel - and almost everyone responds within a day - so bingo the poll is done. If you look at the ARS polls (haven't heard from them in a while) they are usually done on two days.

  14. DL,

    Angus-Reid's polls are usually over three days, while Léger's span four days and CROP as much as a week.


    Yes, I am aware of the debate on this issue, which is why I always point out that the MOE is in terms of a random sample size, and I identify online polls in my polling model (they have the asterisk).

    I've also mentioned your point about the MRIA in a past poll done by Angus-Reid.

    I will take greater care to be more explicit in the future.

  15. Anyways, my point is that having a poll conducted in one day as opposed to over the course of a week could be a strength or a weakness. It certainly is "hot off the press".

    Once things we have certainly been seeing from Ipsos, Nanos and now Abacus has been a consistent pattern of Liberals decline with the Tories and the NDP benefitting.

  16. No, there isn't anything wrong with a poll taken on a single day. But comparing it to a poll taken over two weeks, as is being done with Wednesday's EKOS poll, is not exactly fair.

  17. Ah, Abacus, and here the last poll you did didn't seem so bloody nuts. This, however...

    I do find it hard to believe that nearly 50% of people know who or what Oda is, or even "some aspect" of the controversy surrounding her. Seems a tad bit too high, unless Canadians really are paying attention to politics and the rest of pollsters who say they aren't, are simply getting it wrong.

  18. Technically, wasn't it only one pollster who said they weren't paying attention to politics?

  19. Well, it was only one who went out into the field to do it, and came back with that 15% number, which was I think the Pollara guy - Marzolini? I dunno, I can't remember.

    But there's been commentary from others, Allan Gregg for sure, saying something of the same thing, that Canadians aren't paying attention. The person who I've seen diverge from this conclusion, at least recently, was Frank Graves on CBC in trying to explain the Con drop.

  20. Different from "whatever happened this week", How??

    A week is longer and more things happen during a week. You do have a point, just not a good one.

    For example Harper and the Conservatives may look really good in the media one day and then the next day it may be all over the news that Conservative officlas are facing charges. The next day everything could be wonderful again.

  21. PT your views on this are incorrect.

    One week polling is NOT superior to one day polling, just different.

    Imagine this scenario:

    Conservatives are having a great week until on wednesday an epic scandal hits and everyone hates them.

    A one week poll would be an average of MTWRF, released the following monday.

    The M and T results would be inflating actual support.

    Where as a one day poll taken on F would be the most accurate and one taken on M would be the least.

    1 week polls offer a more stable view but they are not automatically better (as the above scenario proves a 1 day poll taken on friday would be superior).

    1 day polls are high risk, high reward.

  22. Here's another factoid for why we should take this Abacus poll with a giant grain of oilsand:

    Five days BEFORE the day whose results they ran with, one of the co-owners of the company, the very partisan Tim Powers, said they were "currently" sampling on whether Cndns actually care about the Oda affair, but it's clear from the whole tone of his piece that he didn't think they would... or should.

    Which raises the possibility: maybe they did run a poll that first week, right after the Oda Ado broke, but didn't like the results. So they waited 'til Wed, when almost all the stories had petered out, & Libya was crowding it out in the news, and ran it again, and, whew, it was negative, just like the client AND owner wanted.

  23. If 46% know who Oda is that means 54% don't know who she is.
    And how does it affect polls with all leadership stuff going on in AB and BC.

  24. Pollwatcher: They're not randomly sampling an online panel. They're sampling an online panel such that their sample mimics a random sample.

    So, if women are more heavily represented in their online panel than they are in the real world, their votes get weighted less heavily when tabulating the results.

    All polls have to do this, because there's no way to get an actual random sample.

  25. Weighting fixes some of the problems, but the major issue with online panels is that their motivations are different than those who respond to a random telephone survey, and since they are all internet users you can't catch the "don't use the internet" demographic. Many of those are older voters.

    You might be able to get enough older panel members to have a correct weighting, but are older people who use the internet different from older people who don't?

  26. WhigWhag you're suggesting people knew who Oda was and then a week later forgot her ?

    Even after being prompted with her name ?

    That's pretty far fetched.

    It takes longer than 1 week to forget a fact.

  27. The number of people with no internet access at all is getting to be very small and is now heavily concentrated among people over 75 years of age with very low incomes. One of the dirty little secrets of polling these days is that telephone polls now face the challenge of the fact that almost no one under the age of 30 has a land line anymore and cannot be reached through a phone poll!

  28. Éric - You're absolutely right, but the same is true of polling by telephone (how many people don't answer calls from people they don't know - I know I don't).

    My point is there's no way to get a random sample, so whichever method they use needs some weighting to fix it.

  29. Well, it looks like AR has a massive new poll out with similar result. The Star has the results on their website. It seems the NDP takes a beating in this poll.

  30. no... I'm suggesting they may've tried at least twice and got a different set of people on the second or last time, who hadn't just heard it, and didn't have as strong a lingering impression in the calm after the first week of storm as the 1-day respondents would have had almost every day BEFORE that Wed., since there was no new news on it from Monday forward.

  31. DL,

    Not a dirty secret at all, it's the major issue facing pollsters. EKOS calls both land lines and cell phones. I'm not sure of the others.

  32. Whigwag,

    Or, the more plausible explanation is that Abacus poll Rob Silver was referring to on the 18th was the poll that was done on the 23rd. On the surface, that's a five day gap, but Abacus is based in Ontario so monday was a holiday, there's only a couple of business days between the Silver article (at which time he clearly had no polling data one way or another - if Abacus is partisan, as you suggest and a "missing" poll was favourable for the Tories, he would have mentioned it. If it was negative, he wouldn't have written the article) and the 23rd when this poll was done.

    But why look for a simple and straight-forward answer, when you can invent a conspiracy.

  33. "I'd say if a poll was conducted today in NL on federal politics you'd see a big difference because people are pissed with a few things that have happened this week with the CPC."


    Sorry Progressive Tory but you may be letting your dislike of Harper cloud your vision.

    The new AR poll shows the Tories at 42% in Newfoundland compared to the Liberals at 36% and the NDP at 21%.

    I'll say it one more time:

    The rift between the federal CPC and the provincial PCs has been 100% healed.

    Harper is poised to win 4 seats in the province.

  34. "Well, it looks like AR has a massive new poll out with similar result. The Star has the results on their website. It seems the NDP takes a beating in this poll."

    hmmm...the NDP took 18% of the vote in the 2008 election and they have 18% of the vote in this Angus Reid poll. You have a strange definition of "taking a beating"

  35. How are the Greens so strong despite never winning a seat? Canada seems to have the spoiler "effect" in operation, but the voters dont seem to react to it.

  36. New AR poll:

    CPC on the cusp of majority.

  37. Angus Reid on Feb 26
    39 / 26 / 18 / 9 / 6

  38. Eric said:

    "EKOS calls both land lines and cell phones. I'm not sure of the others."


    Nik Nanos has previously stated that land lines, cell phones, and VOIP (Voice Over Internet Protocol) phones (cable-system phones) are all part of the Nanos mix.

  39. JQ,

    Thanks, will remember that.

  40. Quite impressively, AR's new report actually has numbers for each of the ten provinces. It's interesting actually being able to see if there's a difference between Sask and Man (there is - in the relative NDP/Lib numbers) and the Atlantic Provinces (Libs ahead in NS and PEI, Cons in Nfld and NB). And miraculously it manages it all with an MOE of less than 6% everywhere except PEI. The amount of detail here is impressive and useful.

  41. Yes, I intend to make full use of it. It will be a handy guide to distribute the Atlantic Canada poll numbers more accurately.

  42. I believe the Green vote in Atlantic Canada id a undecided vote. Especially in NL and NB. They have been continually told to not trust the Tory's for the last decade that while they don't want the Liberals or the NDP, they are just hesitant to push the blue button and give them a majority

  43. The Oda affair is a 2009-2010 issue the opposition had stored in there war chest for a scandal opportunity in the future such as now. I don't think Canadians buy it or care. It's a foreign office portfolio dealing in donations with foreign countries. Canadians with the where with all to afford donations tend to go with the United Way or the Cancer Society and others that address concerns dear to them, not spreading Christianity to people that already have a religion


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.