Monday, March 5, 2012

Conservatives hold steady

Nanos Research reports today that support for the Conservatives hasn't changed a bit over the last month, despite allegations of impropriety during the last federal election.

Just how successful might this impropriety have been? My article for The Globe and Mail today shows that it might not have been very successful.

You can also read my article for The Huffington Post Canada on this Nanos poll and the recent EKOS poll here.

Before getting to the Nanos poll, a few words on my article for the Globe. I've already seen a lot of comments emphasizing the point made in the headline, that turnout was higher in allegedly robo-called ridings than in other ridings. But the most interesting aspect of my analysis, at least in my view, is further down the page, seemingly confirming my fear that too many people only read the headline and the first paragraph before coming to a conclusion about what an article says.

The point I make is that turnout was higher, yes, but likely because the ridings allegedly targeted are competitive. The margin of victory in the robo-call ridings was roughly half of the margin of victory in other ridings. So while this might not tell us whether this alleged tactic was successful or not, it does appear to suggest why these ridings were allegedly targeted.

That might not come as a surprise, but I always find it interesting to provide numbers to what is a generally perceived notion. It is more interesting to disprove that generally perceived notion - early reports of the scandal said that turnout was lower in targeted ridings. That does not seem to be the case.
Now to this Nanos poll. It shows that the Conservatives are unchanged since Nanos was last in the field January 20-23. They still have the support of 35.7% of Canadians.

The Liberals are up 1.9 points to 29.5% while the New Democrats are down 0.2 points to 25%.

The Bloc Québécois stands at 4.9% national support. The Greens are down 1.1 points to 3.4%.

As I mention in the Huffington Post article, Nanos has often registered Liberal support higher than other pollsters. That does not mean they are wrong (the others might be wrong, or the truth somewhere in the middle), but it does reduce the significance of the Liberals placing second and the NDP third. The polls do not consistently show this to be the case, so at the very least we can say that the race for second is probably tight.

The Conservatives lead in British Columbia with 41.1% (+0.1), trailed by the Liberals at 28.2% (+7) and the NDP at 23.1% (-5.4). The Greens are fourth with 6.2%, down 1.9 points since the end of January.

The Tories also lead in Atlantic Canada with 45.6%, up 16.1 points. The Liberals are down 0.3 points to 33.3% while the NDP is down 14.9 points to 20.3%. Wild swings like this, combined with the small sample, make the meaning of these numbers relatively low.

In the Prairies, the Conservatives have 48.2% support (+0.2), well ahead of the NDP (24.4%, -1.5) and the Liberals (20.6%, +0.1).

The Liberals lead in Ontario with 37.8% (+2.7), thanks to a big 6.2-point slip by the Conservatives to 35.9%. The NDP is third with 21.9%, up five points.

The New Democrats lead in Quebec with 32.6%, up 3.6 points. There is some inconsistency among the pollsters as to whether the NDP is on the uptick or the downswing. This suggests that the NDP is generally stable somewhere around 30%. The Liberals are second in Quebec with 26.8% (+0.3), edging out the Bloc Québécois at 20.9% (-3.2) and the Conservatives at 15.8% (+0.7).

With these numbers, the Conservatives would win 136 seats. They would likely be toppled by the Liberals and the New Democrats, who combine for 167 seats. But in contrast to some other polls, this one results with the Liberals at the top of the duo with 86 seats to the NDP's 81. The Bloc Québécois would win four seats and the Greens one.

The Conservatives win 22 seats in British Columbia, 25 in Alberta, 17 in the Prairies, 46 in Ontario, seven in Quebec, 18 in Atlantic Canada, and one in the north.

The Liberals win eight seats in British Columbia, two in Alberta, six in the Prairies, 43 in Ontario, 18 in Quebec, nine in Atlantic Canada, and one in the north.

The New Democrats win five seats in British Columbia, one in Alberta, six in the Prairies, 17 in Ontario, 46 in Quebec, five in Atlantic Canada, and one in the north.

With the House increasing to 338 seats by 2015, a rough estimate ups the Conservative total to 152 seats, with the Liberals winning 95, the NDP 86, the Bloc four, and the Greens one. The Conservatives and Liberals benefit from the extra seats, with the Conservative portion of the House increasing from 44.2% to 45% and the Liberal portion from 27.9% to 28.1%. The NDP portion falls from 26.2% to 25.4%.

In Nanos's Leadership Index (a combination of percentages on the questions of trust, competence, and vision), Stephen Harper tops the list with 102.4 points, compared to 54.4 for Bob Rae, 24.6 for Elizabeth May (rocketing up to third!), 20.3 for Nycole Turmel, and 11.7 for Daniel Paillé.

If we change those into percentages of the total, we get 48% for Harper, 25% for Rae, 12% for May, 10% for Turmel, and 5% for Paillé. If we took those as "Best PM" results, it would be pretty plausible at this point.

This poll points to stability - not exactly the replacement of the NDP as the Opposition by the Liberals. Nanos's last poll also showed that. In today's volatile climate, this lack of change is somewhat surprising. It will be more surprising if this persists straight through to the NDP leadership convention and the budget at the end of March.

41 comments:

  1. 3000 votes difference in the 12th closest ridings list.

    In other words. If wrongdoing is ever proven (and most of the riding complaints outside of Guelph amount to just annoying calls. Or "They must have been impersonations because they were annoying") then we can go to a judge and see if we can come up with enough proof that more than 3000 were affected.

    Not just that 3000 were affected in that riding though... that enough on top of that were affected that the result might have been changed. And with only 31000 national complaints, (some? Many? most? of which won't pan out...) how will that translate to turning over those several ridings? Enough to overturn the election??

    .

    This is why I have trouble believing (most) of the allegations. Because they happened in ridings that were out of reach either for the tories to win or for them to lose....

    If there isn't a close race, then there is nothing to gain by taking the risk.... If there is nothing to gain.... then why do it???

    If tories have shown us how targeted every measure they take is. Then why take a scatter gun approach to something like election fraud ???

    I don't see how one can talk about how tight and controlled the tory actions are in one breath and then talk about how sloppy they were in targeting ridings in which to perpetrate this alleged scandal.

    It doesn't make sense for them to have done it in most of the ridings which are alledged.

    It will however be interesting to see how the Guelph investigation pans out.

    ReplyDelete
    Replies
    1. For me, it isn't whether these misleading calls, directing people to the wrong or non-existant polling stations, affected the outcome of the election. What upsets me is they may have tampered with a Canadian's right to vote. There's a good chance that someone's consititutional right was violated. And that just won't do. If that's the case, forget new elections, someone needs to go to jail.

      Delete
    2. Absolutely.

      When they find the one responsible for the calls directing people to a non existent polling station, that person should go to jail. 100%.

      But I am waiting for the RCMP/EC investigation to tell me who that is.

      As to the other just annoying phone calls? enough arm waving hysteria.


      But on the other hand... Guelph is the riding that had a polling station authorized after the fact too... That is also outside the rules.

      Delete
    3. Point taken Barcs. But there is another aspect to this "percieved" scandal that could cause the CPC dearly. It may not matter what the EC or RCMP come up with in the end. If the perception by a huge chunk of voters is that the CPC trod on their rights or others rights, that'll hurt them. Now of course it'll depend on how large the group of voters feel this way and if their anger carries on to the next election. However, a canny opposition could build on the myth, if you will, that the CPC is dirty and that could cost them for generations. For example, see what the Conservatives in Alberta and to some extent Saskatchewan have done to the Federal Liberals.
      If this thing grows legs, logic won't matter. It also won't matter what the Liberals or NDP might have done a long time ago. The federal CPC, in some areas such as Ontario, BC and Quebec, could become like the federal Liberals in Alberta.

      Delete
    4. So what you are saying is that an orchestrated campaign of voter suppression done now by the opposition parties.... even if there is no evidence found in the end... might work??

      Thought that's what you were fighting against the tories about this time.

      Delete
    5. To put it bluntly, no. I didn't say anything about voter suppression. I merely suggested the opposition take a page out of the Conservative playbook and paint them with this scandal. Look Barcs you can't support the CPC tactics then complain when other parties do the same thing. The reality may well be that we have a new way of doing politics in this country, as practiced by the Tories for more than two decades. If that's the case, it's fair game for the other parties to do the same thing.

      Delete
    6. One more point of clarification. I'm not suggesting anyone supress voters' rights to vote. I'm suggesting other parties, such as the NDP and Liberals, use the Tories tried and true tactics such as I have already described. If the reality is that there is a political NWO, then so be it. But breaching a constitutional right, no. The Tories can have that one all to themselves.

      Delete
    7. "Barcs you can't support the CPC tactics then complain when other parties do the same thing."

      Nope.

      But I can question the high-mindedness of those that decry what the tories do while simultaneously supporting others doing it (as some have done) or suggesting that they do it the same (as you are doing)



      "I didn't say anything about voter suppression." .... My apologies, I thought that was the scandal (or perceived scandal) that we were talking about when you suggested that the opposition paint them with the scandal. (Thus trying to get conservative voters to not vote conservative...)

      Delete
    8. Again, what I was advocating the opposition do is not voter suppression. You're comparing apples to oranges here.
      The CPC are artists at hanging other parties' dirty laundry out to dry, and even smearing some dirt on it first. But that's not voter suppression. Ugly politics maybe but not voter suppression. Phoning someone up, or getting someone else to do it, and telling them their polling station has been changed in an attempt to frustrate them to the point they don't vote or don't have time to vote when they learn the truth is something altogether different. Apples and oranges.
      I was suggesting the NDP and Libs follow the former, as it is now accepted political practice, and not the latter, which is in my opinion a crime.
      Good twisting though Barcs. Here's to you.

      Delete
  2. Can you please provide evidence that the small percentage differences you note between targeted and non-targeted ridings is not within sampling error.

    ReplyDelete
  3. Eric wrote: "Just how successful might this impropriety have been?"

    Just being a stickler for structure, but the "impropriety" may have been successful. The allegations (according to the poll) were not successful.

    Love the site!

    ReplyDelete
  4. I know you talked about "more competitive" ridings so higher turnout Eric.

    Might there be more reasons for the disparity, not just whether or not there was allegations?

    Could differences be explained by regional differences in voter turnout for example?

    ReplyDelete
  5. Barcs, no. Most of the allegations come from Ontario ridings, and the relative TO numbers are no different if we just look at Ontario. I also compared some ridings to their unaffected neighbouring ridings and did not discern any pattern.

    ReplyDelete
  6. This Nanos poll seems like a bit of an outlier. The numbers seem very skewed in favour of the Tories and Liberals in different regions of the country when compared to the trend.

    I'm not saying it's wrong but it does seem unlikely.

    ReplyDelete
    Replies
    1. I'd be interested in seeing what Angus Reid has to say. Nanos and Angus Reid were the two most accurate pollsters in the last three elections, but both use very different methodologies. If they disagree, I wonder what that tells us?

      Delete
  7. I'd just point out that the point of the fake Liberal/NDP robocalls wasn't to keep people away from the ballot box - it was to piss them off so they go and vote against the Liberals and NDP.

    Eric - I've the same question as Anonymous 2012. How statistically significant are the differences in turnout and margin of victory? How about when you control for region? Or turnout in 2008?

    ReplyDelete
    Replies
    1. To your first point, that is why I separated the harassment calls from the poll-moving calls.

      To your second point, I don't how statistically significant the difference is but it is a fact that turnout was higher in the robo-call ridings than other ridings. It could be a meaningless statistical fluke, but I did not assign any importance to the fact that turnout was higher, aside from it being an interesting counter-intuitive.

      It is difficult to control for regional variations outside of Ontario, as in some regions we're talking about one or two ridings. But the pattern is consistent in Ontario and for 2008.

      Delete
    2. Ryan I have a feeling most of those "annoying" calls were actually made by the liberals themselves.

      This was the first election that the party unveiled their new voter tracking/id/contact system.

      The gun registry disaster teaches us that giant new databases have a lot of kinks that need to be worked out.

      Also a lot of those nuisance calls were stuff like "they called on saturday! or "they called just before 9!" or "they called at dinner".

      That basically describes all calls that come from call centers. Does anybody like these calls ? There's a reason why people wanted a do not call list.

      No surprise that there have been complaints with the new Liberal call system.

      Delete
    3. I hear you on that Anonymous, and I suspect that is a decent chunk of what's being reported. Not all though - Liberal callers don't use spoofed North Dakota phone numbers, for example. I don't think there's any question that at least some low-level volunteers did something unethical. What remains to be seen is if any higher ups were aware of it.

      Eric - I believe this would be the appropriate test to use (http://en.wikipedia.org/wiki/Paired_difference_test). Wouldn't have to get too fancy in controlling for external factors to at least get a rough idea.

      Frankly, I don't think you'll see any statistically significant differences. You wouldn't need all that large an effect to change the results in ridings where the margin was only 12 votes. The goal was probably only to swing a few hundred votes per riding - not enough to show up in the overall turnout.

      Delete
    4. BTW if anyone wants to read a really fascinating paper on detecting fraud in elections with models like the one here, I highly recommend the link below. It's an analysis of the 2009 Iranian presidential election by Prof. Walter Mebane at the University of Michigan.


      www-personal.umich.edu/~wmebane/note22jun2009.pdf

      Delete
    5. You seem to be confused

      The Phony "EC phoning to change polling stations" (of which there was several hundred stations where EC did change), are in my opinion fraud, intended to keep people from voting. The perpetrator should be prosecuted.

      But there was no party identifier with those. Thus no way to make someone more pissed off at one party or another. The caller falsely identified as an EC agent, nothing else.


      But the bulk of the complaints are about annoying or rude behavior by call centers. (who ever heard of that). As Anon says, calls late, or early, or on Saturday or religious holiday.

      These aren't illegal in themselves, unless you are mischief making from another party. Which all parties do some of, volunteer taking it upon themselves or directed by campaigners.


      Was the liberal campaign with a new database so organized as to preclude their own call center from making the mistakes??

      ya right.

      There is probably some mischief calls out there from several directions and parties, but I doubt all the complaints are legitimate problems and not just a volunteer screwing up.

      It is these calls that you are talking about being designed to piss off people and get them to not vote.

      ... (if it wasn't just volunteers screwing up... 5 times in the last week dealing with 3 different government departments I have had the pleasure of dealing with a bureaucrat in Winnipeg or Ottawa that can't tell time where I am. Why should I expect a call center of volunteers to be better??)



      If it was people mischief making then it is a problem. Not unlike the NDP campaign directing calls to a Conservative MP to shut the office down. Or Joe Volpe caught on camera as a volunteer removes green party literature and replaces with liberal.

      Delete
    6. I'm not confused Barcs. I stand by what I said. There is direct evidence that some Conservative operatives directly misrepresented themselves as being from another party. Under the Canada elections act that is just as serious as pretending to be from Elections Canada.

      Delete
    7. Haven't seen the evidence you talk about (misrepresenting as another party). That the opposition parties claim they have but only talk about.

      Maybe they did give it to the police. And something might come of it.


      But you were talking about the "EC polling district changed" robocalls and the "annoying phone calls that must be part of voter suppression campaign" like they were the same phone calls.

      You can't identify yourself as both from EC and from a party at the same time, that would mess up the masking you were trying to do.

      Delete
    8. "There is direct evidence that some Conservative operatives directly misrepresented themselves as being from another party."

      Actually, there really isn't. There have been a few statements to the effect that people first identified themselves as being from the Liberals, and then when challenged, as being from the Tories. One of those statements is a lie, but which one?

      Which scenario is more likely, a Conservative lying about being a Liberal, annoying the voter, then admitting to being a Conservative, or a Liberal calling as a Liberal, and when they realized they were annoying the voter, pretending to be a conservative. I'm partisan, but the latter scenario at least makes sense. In either case, it sounds like the sort of thing that would be done by the low-level frat rats who operate on the ground in campaigns, not part of a broader strategy.

      Delete
  8. At least Eric looked at the 2008 results when looking at the affected ridings, I'm sure that's what the tories looked at because nobody predicted the outcome in 2011. If we look at them, we see that Guelph was a target riding (very close in 2008, with a lot of Green support). Others like Kitchener were very close in 2008 as well.

    I think our mentality of "Did it change the outcome?" is the wrong question. The right question is, "Did they try to affect the outcome of elections?". Attempted murder is a crime, the fact they didn't succeed is another issue in my mind. They should be tried, and put in jail.

    ReplyDelete
    Replies
    1. Whether they affected the outcome or not I think every agrees that there should be criminal prosecutions. The question on what the net effect was is still relevant, as it will determine how many by-elections will be called.

      I agree completely though that these primarily targeted ridings that were perceived to be close, regardless of whether or not they were in fact close. I'd be interested in seeing what Eric's projections were for each of the targeted ridings were - may the perpetrators used a website like this to decide where to suppress votes?

      Delete
    2. The person who did it/ordered it certainly should be.


      As to "did it change the outcome", I only pointed that out in regards to a motive to do it. ... If it wasn't going to change the outcome, or provide some benefit that was worth taking the risk of a mega scandal and possible jail time... then why do it?

      If there is no far reaching benifits, then there probably was no campaign to do it and it is probably very limited in scope to only a couple people, not a national campaign directive.


      And yes, when they find the person/group of persons involved, they should be prosecuted.

      And "31,000 complaints"... prosecute the ones that turn out to be fraudulent complaints too. If you are looking for justice.. it has to work both ways.

      Delete
  9. Any idea why every two months someone puts out a poll with ridiculously high Liberal numbers in BC? or is this just my bias as a liberal in BC who pays attention to that number?

    Or is that 25-30% of BC really is Liberal and the Liberal party just fails at providing credible, non-ubc-student candidates for them when elections roll around?

    ReplyDelete
  10. two liberal seats in Alberta? Where? haha

    ReplyDelete
    Replies
    1. And the NDP will win 59 seats in Quebec! Haha! Hoho! Stop it, you're killing me!

      Edmonton Centre and Calgary Northeast. Please note that this poll bumps up Liberal support in Alberta by almost double.

      Delete
  11. Quite frankly I don't believe a lot of the complaints are legitimate. I do believe that there are a lot of fanatical supporters of the NDP and Liberals who are lodging complaints simply because they want to hurt the Conservatives.

    We know that there is online campaigns out there to get more complaints made and by no means do 1% of them have any evidence about what happened coming up on 1 year ago.

    ReplyDelete
    Replies
    1. Something happened. Where there's smoke, there's fire. And there's a hell of a lot of smoke coming from this one. You and Barcs can fan the smoke in the other guys' direction all you want, but that doesn't change the fact the Tories are dirty on this one. It's their fire and no matter what EC or the RCMP come up with they are going to get burned by the flames of public opinion. Remember, popularity is like an onion, one layer at a time on and one layer at a time off. Guess which way the CPC is going now.

      Delete
    2. Brian Lilley claims Avvaz.org, leadnow.ca, etc are involved in the "spontaneous" 31,000 complaints.

      http://blogs.canoe.ca/lilleyspad/general/robo-complaints/

      He's right, there is templates and instructions on their sites for mail-in's and phone-calls.

      But you can't call this an American dirty tricks campaign... Because only the tories are capable of that. lol

      Delete
    3. Pink

      CTV has announced the Liberals are going to release their phone data.

      Of course the Tories have refused to release theirs !!

      What does this tell us ??

      Delete
    4. That the liberals were faster with the delete key??

      That soon Rae will ask Elections Canada to open up the investigation so that we can see how many of the complaints are not legitimate and prosecute those that made false complaints??


      But seriously. It means that the liberals think they have more to gain then they have to lose by opening up their records. It's too bad we won't have the list of complaints that EC got to compare. I think that little fact factored into the calculation.

      Delete
    5. As the Anonymous poster who made the first comment I would like to state for the record that I didn't vote for my Conservative candidate in 2012.

      Second, I'm not fanning smoke anywhere. Frankly I'm choking on all the second hand smoke coming from the NDP and Liberals.

      Third. RackNine has released their phone data which is the investigation in question and they have been cooperating with the authorities.

      Fourth the nature of any part of a political campaign could be construed as harassing. Phoning (automated and cold calling), Door Knocking, Glad handling. The most effective means of political campaigning all involve the intrusion into someones life usually at a time they don't choose.

      Fifth Auto Dials (leftists only use the term Robo call) are a very effective and proven way to quickly identify a vast number of your supporters. Anyone who has sent out an auto dial knows that you're going to piss some people off because they're mad at hearing your message to your supporters. Its a proven way to identify your supporters and opponents in a very short amount of time. Auto dials also are proven to increase voter turnout and can be very effective at reminding people to go to vote.

      Sixth. Don't BS me about the Conservatives being the only ones to harass people by Auto Dial. The NDP harassed the constituents of Saint-Maurice—Champlain by Auto dial after Lise St-Denis left the caucus. The NDP have always played dirtier politics than the Conservatives.

      Seventh - If you don't want a political party to call you, don't give your phone number when your enumerated to vote. That's where all parties get their phone numbers from

      Delete
    6. I'll tell you a story about "harrassing" phone calls. I was involved in a provincial conservative campaign in Ontario in 1999. As part of our strategy, we contracted with a company to send out a mass voicemail to every phone number in the riding associated with a voter. The voicemail was completely innocuous and perfectly legitimate, along the lines of: "Hi, I'm so-and-so, you conservative Canada in this riding, please vote for me". The logic was that this would be a message that would go out in the middle of the day, and people would get it when they checked their messages when they came home. So far, so good.

      Unfortunately, there was some sort of glitch in the voicemail software, so instead of sending one message to every phone number associated with a voter (i.e., if you have three voters in your house, you'd only get one message), it send one message to every voter associated with a phone number (so if you have threee voters in your house, you'd get three messages). Again, this was an innocent mistake, but understandably, it might be annoying.

      But the reaction, oh my god, it was incredible. We received hundreds of angry phone calls accusing us of tapping people's phones, hacking their voicemail box, harrassing them, you name it. People were threatening to call the cops (we told them that they should feel free to do so, we'd done nothing wrong). And remember, this was a message we sent out trying to get people to vote for us!

      I tell this story to emphasize two points. First, because it emphasizes the point that for a certain segment of the population, any phone call from a political party will constitute harrassment or some sort of criminal act, no matter how innocuous. There was a story in the Globe last week where a fellow described the "harrassing" call from a puported liberals as being along the lines of "I'm calling from the Liberal party of Canada, yada, yada, yada, on election day we're counting on your vote." That's a harassing phone call? Maybe, but what does he think a legitimate Liberal phone call sounds like?

      Second, it emphasizes that incompetence on the part of political parties is a more likely explanation than maleavolence. We weren't trying to harras voters, quite the contrary, but I have no doubt that some voters felt harrassed and many more were annoyed. But it wasn't because we were trying to play dirty or to harras voters, it was because someone screwed up (not me).

      Finally, it emphasizes the point that a political party can, completely innocently, and inadvertently, annoy the people whose support it's trying to get. I have no doubt that a good chunk of the voters who claim to have been "harrassed" by people indentifying themselves as Liberals, were, in fact, called by actual Liberals. They probably weren't harrassed, or at least not intentionally, but I have no trouble imagining that either (i) they felt like they were being harrassed (and that may be a comment on their state of mind) and/or (ii) the Liberal callers annoyed the hell out of them, without even realizing it. I mean, most telemarketers ARE annoying, it's the nature of the beast.

      And that isn't to suggest that the Liberals are being disingenuous now in claiming that they didn't harrass their own voters, they probably sincerely believe that THEIR calls weren't annoying (just as, I didn't think that the voicemail that we sent out was annoying. Heck, it wasn't), although I suspect they're not trying all that hard to confirm that theory.

      Delete
  12. Hey Eric.

    When is the official monthly update coming?

    The suspense is killing me.

    ReplyDelete
    Replies
    1. You're in luck - this morning! Just hope no polling firm will release week old polls tomorrow.

      Delete
  13. Pinkobme: "It's their fire and no matter what EC or the RCMP come up with they are going to get burned by the flames of public opinion."

    Your faith in the goodwill and intellect of the Canadian public is inspiring.

    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.