By the numbers - Macleans.ca
 

By the numbers


 

Elections Canada says the majority of the 31,000 complaints it reported two weeks ago originated with online forms.

David Akin notes a study that attempts to determine whether there is any statistical evidence of voter suppression in those ridings that were reportedly targeted. Stephen Gordon explains.

If you look at riding-level data, there’s not much to see. But Simon Fraser University’s Anke Kessler has dug deeper into Elections Canada’s poll-level database and uses information that is available at the poll level.  Outcomes at polling stations differ in turnout and in vote shares for particular candidates; this makes each riding look like a smaller copy of a country-wide election. In a first step, she finds that polling stations with predominantly non-Conservative voters generally experienced a decline in voter turnout from 2008 to 2011. In a second step, she asks how the extent of this decline varies with reported robocalls. She finds that it was larger in the former, meaning that in ridings where robocalling was reported, polling stations that voted predominantly non-Conservative in the 2008 election saw a greater-than-average decline in voter turnout.

Gordon also notes some important caveats and has posted Prof. Kessler’s paper here.


 

By the numbers

  1. ” In a first step, she finds that polling stations with predominantly non-Conservative voters generally experienced a decline in voter turnout from 2008 to 2011. In a second step, she asks how the extent of this decline varies with reported robocalls.”

    xkcd ~ Correlation: 

    Man: I used to think correlation implied causation. Then I took a statistics class. Now I don’t.
    Woman: Sounds like class helped.
    Man: Well, maybe.

    • To Tony, on the topic of correlation:

      “You keep using that word.  I do not think it means what you think it means.”

      -Inigo Montoya, ‘The Princess Bride’

      • ” The calls appear to have targeted voters with a Liberal Party or New Democratic Party (NDP) affiliation.”

        Ottawa Citizen ~ March 9 2012:
        On the day “Pierre Poutine” activated the burner cellphone used to launch his robocall blitz on voters in Guelph, a young Conservative campaign worker sent out a message on Twitter warning of “voter suppression calls” aimed at his party.

        Weekly Standard ~ Fat City:
        Neil Gross of Harvard found that 87.6 percent of social scientists voted for Kerry, 6.2 percent for Bush. Gross also found that 25 percent of sociologists characterize themselves as Marxists, likely a higher percentage than members of the Chinese Communist party. I would guess that if Lenin were around today he would be teaching sociology and seeking grants to fund the revolution.

        Andrew Lang ~ An unsophisticated forecaster uses statistics as a drunken man uses lamp posts – for support rather than for illumination.

        • Let’s take this load of bull one claim at a time:

          1) A completely unsupported claim by a partisan actor who may have been pre-muddying the waters, knowing that the Conservative Party of Canada itself was actively pursuing a campaign of voter suppression. No other claims of voter suppression targeting Conservative voters has been made to date. But thousands of complaints from across the country have been made to Elections Canada, from non-Conservative voters.

          2) A meaningless statistic from the US, tweaked to sound like a meaningful insight about “social scientists.” One can safely assume that most people educated enough to be called a scientist of any sort voted against George W. Bush. Nice crack about Lenin though – again, meaningless.

          3) Aaaand a meaningless quote just for shits and giggles.

  2. Did the researcher take into account changes in poll locations between the 2008 and 2011 elections?

    From EC’s website:

    Advance polls: For the 41st general election, a total of 4,706 advance polls were set up – 665 more than for the 40th general election.

    Election day polls: For the 41st general election, returning officers set up 64,477 stationary polling stations, an increase of 1,041 (1.6 percent) from the 40th general election.

    Not only were new polls added, but existing poll boundaries most likely changed between the two elections.

    Can we easily compare turnouts per poll location if their boundaries were changed, and new poll locations were added?

  3. Oh dear. I wonder how the people touting the sheer number of complaints as being somehow significant are going to feel about this.

    If I were conspiratorially-minded and also supported one of the opposition parties (but I repeat myself), I might suspect that releasing the number one week, to admit it was a number largely made up of BS robocomplaints the next, would be awfully convenient for discrediting the hysteriamongers…

    • So your position is that all 31,000 complaints were made up?

      • If a majority of that 31,000 were generic letters generated automatically by websites, and don’t allege or substantiate any specific incidents, they’re pretty much worthless by definition, sweetie. That makes them a petition in multiple parts, not complaints.

        •  Imagine 2004 Stephen Harper crying “Haha, there were only 15,000 actual complaints about Conservatvie Party election fraud! Don’t you feel pretty stupid now!”

        • Just because the complaints were easy to complete doesn’t make them dishonest.  I’m sure that Conservative agents filled out a number of bogus complaints in order to muddy the waters of their investigation, but I’d be surprised if your team could flood EC with 31,000 false complaints.

      • More to the point it has nothing to do with the article being discussed whose data seems to not be dependent on the mythical 31000 figure. However, muddying the water and obscuring the debate seems to be AVR’s rhetorical weapon of choice.

  4. I thought that this was common knowledge. It was an exceedingly poor tactic insofar that it makes any concerns much easier to dismiss as partisan hysteria in the eyes of the general public. I’d say that this is a problem that the generalized-lefty class in Canada has, however. Their most common tactics seem to be street protests, (e)mail, and internationally-sourced shame – all tactics that seem to alienate a majority of Canadians, even those who are sympathetic to a cause. 

    The generalized-righty class have their own stale tactics too. Browbeating, moralizing, false-equivalencies, and long-term residency upon crosses are all used to make their points – also alienating to those who would otherwise be sympathetic. Personally, I’d like to see our politics evolve from these tactics. They’re juvenile, they’re stale, and they contribute to people tuning out.What little attention the general public was paying to this will now vanish, and it really will come off as little more than a partisan concern (sustained Harper Derangement Syndrome) and subsequently forgotten. Sustained Chretien Derangement Syndrome kept the CPC in the woods as well. Voting may have becoming more polarized along some left-right camp, but Canadians themselves sure as heck haven’t.

    If we were all so concerned about the integrity of our elections, we could have done without the whole “OMG STOPHARPER!!!!” and the “RoboCON HAHAHAHA! Get it? CONservatives? HAHAHA” crap. It may ignite the base, but it distracts entirely from (what should be) the main concern: the fair play in the running of elections. 

    • It sounds like you’d advocating for a sort of post partisan politics among the political class – good luck with that. If it ever happens I’d be among the first to jump aboard. Meanwhile the best we can hope for is the general common sense of the public causing the pols to exercise some restraint where it really matters.

      • I’m apt to believe that it’s unlikely to happen any time soon. 

        On the other hand, I also retain the hope that partisan tactics can become a little more skilled, effective, nuanced, and ultimately less alienating in the meantime while I’m waiting for my post-partisan unicorns to arrive.

        • Far as I’m concerned we need lots of new blood like Nenshi.

    • Wow.  CR, I actually agree with most of what you said.  That’s a rarity.

  5. Could this not be simply attributed to the fact that Liberal voter turnout was the lowest in history? I mean, we know that the Liberals received fewer votes in the last election than they ever have, would most of those missing votes not naturally have come in ridings that were predominantly non-Conservative? I hate to say the obvious, but I think that the most successful vote suppression strategy in the last election was Michael Ignatieff just being Michael Ignatieff.

    • From Kessler’s abstract “… I show that those polling stations with predominantly nonconservative voters experienced a decline in voter turnout from 2008 to 2011, and that this
      effect was larger in ridings that were allegedly targeted by the fraudulent phone calls. The
      results thus indicate a statistically significant effect of the alleged demobilization efforts: in
      those ridings where allegations of robocalls emerged, turnout was an estimated 3 percentage
      points lower on average.”

      • Thanks. I guess what I’d like to know is: is that 3% lower than average being an average of all ridings, or an average of all “predominantly non-Conservative” ridings.

        • Reading Prof Gordon’s summary, I think the 3% is in “predominantly non-Conservative” ridings, not all ridings.

          I’m just starting to read the full paper myself http://worthwhile.typepad.com/robocalls.pdf, for those into that kind of thing I would actually suggest reading it – it is actually a fairly accessible read for an economics paper (i.e. not as technical as these often get to be).

  6. I couldn’t agree more. I’m a CPC supporter, but they have done some things in recent years that have come off as just plain stupid and have actually bothered me. But every time something like that happens, the opposition parties go into hysterics and start using hyperbole that is completely over the top, and I realize that the only alternative to CPC stupidity is a completely different kind of stupidity, and I stop caring all together.

    Or I’ll be reading some comments at the G&M, or often here at Wherry’s blog, and I realize that most of the criticisms that are thrown at the CPC are meaningless and baseless. How many people on these boards oppose perfectly sensible legislation, just because it’s come from the government side? I would say that makes up 90% of the “discourse” on any issue.
    Using prorogation to avoid accountability? Unacceptable if the CPC does it, but it’s totally alright when Chretien did it.
    Bill C-30, as Coyne so aptly pointed out, was basically a reboot of old LPC policy, but it’s only worthy of outrage when the CPC introduces it.
    Wafer-Gate was a whole different level of idiotic tail-chasing.

    *IF* there turns out to be a systemic attempt to suppress the vote (and that is still a very large if), I will be very disappointed. But I suspect that the opposition parties will squander any good-will from people like me by comparing Harper to a dictator, or insisting that the entire election is then suspect, and will go into over-react mode yet again.

    • If there turns out to be a systemic attempt to suppress the vote of course the whole election will be called into question; that’s the whole enchilada, the crux of the argument when a dozen swing ridings can get you a majority. But thanks for making it clear you aren’t a partisan and seem to see yourself as among the sole arbiters of what is acceptably sensible. A decent book on cognitive dissonance should help with that.

      • cognitive dissonance

        Well, you are an expert at it, admittedly.

        • Not an expert but I have taken the trouble to read something about the theory.

          I try to follow Orwell’s dictum, to be aware of the personal tendency toward bias and to resist it. Obviously I’m as susceptible as the next person. You’re welcome to point it out; It might make a change from your usual diet of raw cynicism?

    • Rick, don’t forget that old Vulcan proverb: Only Nixon can go to China.

    • Okay. That was an interesting observation.  Putting it into context, there is two sides of it.  You see, I think part of the reason that there’s so much yelling about this stuff is that nobody — on the CPC side, anyway — really seems to care. You say you get turned off because the complaints are so over the top, but the reason the complaints are so over the top is because unless they’re over the top the reaction seems to be that you don’t give a damn at all.  

      However, some of your assertions seem to be unsupported.. such as it being okay when Chretien did it. (It wasn’t) Or that Bill C-30 was basically a reboot of old LPC policy (it isn’t.. it has some similarities but the differences.. specifically the warrant-less information provisions make it a problem).  The wafer-thing I’m not going attempt to defend.. I’ll agree, that one was simply stupidity.

      Also, I think some of this comes from your comparing a time before widespread use of internet forums, so you couldn’t *see* the reaction to the LPC activities, but you’re assuming because of this that there wasn’t any.  Consider if we’d been without internet during the last 5 years.  Most of the arguments and anger against the CPC you simply wouldn’t have known about, and all we would have really seen was them getting elected over and over again — exactly what you saw with the LPC and Chretien.  So I guess what I’m getting at, is perhaps some of your view that people are only angry when the CPC does something wrong and not when the LPC does something wrong is more a result of simply not being able to see when people were angry during the LPC years.

      • I’ll certainly agree that you’re third paragraph is quite possible. 

        As for your first paragraph, I think you’re partially right that the over-the-topness is to get a reaction. I just don’t think that it gets a reaction from the government side at all. All it does is fire up the base, so that they too are screaming about the end of democracy as Canada knows it, and it just makes them look crazy to those who are disinterested in politics in general. In fact, I’d even suggest that this might be the only reason why the CPCs share of the vote has been consistently growing over recent years. Policy wise, with the exception of managing the economy, they’ve done plenty of things that could have alienated voters, but to anybody who leans even marginally right there is just not a single viable option out there.

  7. All I can say is wow – Prof. Kessler must be confident of her tenure to be putting her time and effort into such a ridiculous jumble of quasi- and non-scientific research.

    First off, she states as follows: “First, since I can only rely on self-reported incidences of misleading or harassing calls, which are made available through the Canadian media as my data source, there might be a considerable amount of noise present in the data as the actual occurrence of misconduct is obviously not observed, ”

    “The Canadian media as my data source!”  Now there’s a sacrosanct and unbiased source for research data!    I respectfully suggest the decibel reading of the “noise” inherent in such a data source is akin to a 747 at 50 feet.

    Then she categorizes ridings into “those that predominantly supported the CPC candidate” and “those that predominantly supported the main opposition parties”.  Given the CPC took around 40% of the vote, and other than the odd Alberta riding, was there a SINGLE riding that didn’t “predominantly support the main opposition parties”?  What calculus did the good professor employ to do her categorizations?

    Finally, she doesn’t seem to feel it necessary to compare and contrast the effects she purports the data reflect with, for example, the difference in voter turnout in other elections that resulted in a change of government, i.e. compare voter turnout in the 1992 election to the 1988 one.  I suspect what such a further analysis would demonstrate is that lack of enthusiasm of voters who perceive their party of choice is about to lose is a far more likely explanation for election-by-election voter turnout changes than “robocalls”.

    In case she stumbles upon this little forum about her reearch, here’s some advice for Professor Kessler – rather than spending dozens of hours producing a paper founded upon crap data and faulty premises, perhaps a better use of your time is to find a single actual “suppressed voter”.  Lord knows the media hasn’t come up with one among the 31000 complainants a month into this story – perhaps you’ll have better luck.

    • If there is any mainstream media bias, it logically falls to the right. The reasons are simple.
      Most media relies on private funding from corporations through advertising. Leftist causes tend to be anti-corporate and urge people toward anti-corporate activities. Boards of Directors are not going to empower editors that piss off their advertisers. Therefore, editors will tend to lean center or right, and thus will edit stories/hire journalists/contract freelancers accordingly.

      The only way it might be to the left is if there is such a significant majority of people who feel that way that corporations feel it is worth their time to advertise in a left-leaning publication regardless of the message it sends. However, I put to you that if the majority of people who feel that way is that significant, then what you are seeing is not a leftist bias in the media, but rather a rightward bias in yourself from that of the mainstream public, one which the media is thus obviously not going to reflect because doing so would take them away from appealing to the mainstream.

      As for your comments, learn statistical regression analysis. She took into account that the Liberals lost votes generally, and still found a correlation between where reports of voter suppression have occurred, and a decline in the number of voters greater than the average.

      Finally, while the actual numbers of voters suppressed is as irrelevant as whether you hit or miss someone when you attempt to murder them, there is also this story: http://www2.macleans.ca/2012/03/01/add-five-more-to-the-list/

      • “As for your comments, learn statistical regression analysis. She took into account that the Liberals lost votes generally, and still found a correlation between where reports of voter suppression have occurred, and a decline in the number of voters greater than the average.”

        Learn it yourself.  A basic concept is “margin of error” increases as the integrity of data decreases.  The learned professor, after all her concerted efforts, proclaims the difference in voter turnout between “robocall” ridings and others is – wait for it – 3%.  She then suggests this equates to 2,500 voters, which is accurate only if one assumes 100% of eligible voters to show up to vote.  Since this has never happened in a Canadian election, one ponders why Professor Kessler would make such an obvious error.

        “Finally, while the actual numbers of voters suppressed is as irrelevant as whether you hit or miss someone when you attempt to murder them, there is also this story: http://www2.macleans.ca/2012/0… ”

        A Liberal operative describing a tired old guy who turns down an invitation to be driven to the correct polling station is all you got?  Anybody track him down and interview him to to ensure they got his story straight?  Didn’t think so.

        • Perhaps, before slagging her, you should take the effort to read the next paragraph after that 2500 number.  She includes a fair number of caveats in there, acknowledging potential sources of error.  

          So really, you’re the only one trying to frame it as if she said it was this way without a doubt, and then trying to tear her down based on that straw-man.

          As to the second point, seriously? You asked if someone’s been confirmed as not voting. I presented a story doing such. You’re now asking for someone to confirm the confirmation? I expect next you’ll be needing someone to confirm the person confirming the confirmation and so on?

          And even despite all of this, again it doesn’t matter whether a single person was deterred from voting. It is the attempt that is reprehensible.   This is not a “no-harm, no-foul” situation, no matter how much you hate democracy.

          • “She includes a fair number of caveats in there, acknowledging potential sources of error. ”

            The caveats and acknowledgments of potential error make her conclusions worth about as much as the paper she wrote them on.  So why did she bother, other than to throw more sh*& on the robo-call fire.

            “So really, you’re the only one trying to frame it as if she said it was this way without a doubt, and then trying to tear her down based on that straw-man. ”

            One would have thought quoting from her report would preclude accusations of “trying to frame it” a certain way and creating straw-men.  Apparently not.

            “You asked if someone’s been confirmed as not voting. I presented a story doing such. You’re now asking for someone to confirm the confirmation? I expect next you’ll be needing someone to confirm the person confirming the confirmation and so on?”

            I know the rules of evidence are relaxed when the object of hatred is Harper but seriously – you consider a story about a guy told by a Liberal apparatchik to be acceptable on its face, without need of at least confirmation from the guy the story’s about to be “confirmation”?.  If that’s your standard, I’ll be happy to inundate you with third and fourth and fifth hand accounts about Liberal/Dipper/Green/Rhino party transgressions.

            “And even despite all of this, again it doesn’t matter whether a single person was deterred from voting. It is the attempt that is reprehensible.   This is not a “no-harm, no-foul” situation, no matter how much you hate democracy. ”

            And despite your apparent view that robo-calls are the unpardonable sin against democracy, you’re a voice crying in the wilderness.  WIth the exception of Wherry, Professor Kessler and you and your cohorts, Canadians have moved on.  I submit this is because they understand that Canadian democracy – like all others – is far from pure as the driven snow, that any system that attracts lawyers to the extent that Canadian elections do is bound to result in bending and twisting and spinning of the rules and that while it is important to be diligent about challenging these schemes, Canadians are quite comfortable that their elections are fair and produce the right results, as did the last one.

               

          • Oh. So you won’t accept confirmation about someone being denied the right to vote, but have no problem in asserting Canadians as a whole have moved on from this?

            Does your hypocrisy even register with you anymore?

          • Thwim said:

            “Oh. So you won’t accept confirmation about someone being denied the right to vote, but have no problem in asserting Canadians as a whole have moved on from this?” 

            Vancouver  Sun:

            “When asked about the robocalls, 54 per cent agreed with the statement that “this is just politics as usual and something that all parties do in one way or another.” By comparison, 45 per cent disagreed with that statement.”

            Globe and Mail

            “A month of controversy over pensions, privacy and Pierre Poutine has failed to dent support for the Conservative Party, according to a new poll by Nanos Research.Support for the Tories remained exactly the same – at 35.7 per cent – compared to a month earlier.Support for the Liberals climbed slightly to 29.5 per cent from 27.6 per cent, while the NDP’s numbers were essentially unchanged at 25 per cent.”

            In your case, I suspect the moving on of Canadians while you hang back obsessing about just how evil Conservatives are occurred a considerable time ago.

            Thwim said:

            Does your hypocrisy even register with you anymore?

            My hypocrisy is within 3 percentage points of my hatred of democracy, notwithstanding neither is the subject of learned SFU prof treatise.