09.08.2012 11:33 PM

In Sunday’s Sun: lies, damn lies, and polls

A joke.

That’s what media political polls have become in this country. And, if the news media continue to trumpet the results of polls, they risk becoming a joke, too.

By the time you read this, the pollsters will have hit the TV panels, trying to fool everyone into thinking that they didn’t really get the stunning Quebec election results wrong. But they did, and dramatically so. And they’re doing it all the time now.

“PQ headed to comfortable majority: Final poll before Quebec election.” That was the actual headline on a National Post story, published one day before Quebec trooped to the polls.

The PQ had “a large lead over the Liberals,” the Post declared, relying on a poll by an outfit called Forum Research. Forum claimed to have polled nearly 3,000 adult Quebecers, and that the PQ would win 36% of the vote, with the Liberals eking out only 29%. Forum also claimed their poll had a margin of error of less than 2%.

Well, they — and the Post — were off by a hell of a lot more than that. In the end, the Quebec Liberals and the separatist Parti Quebecois both received 31% of the vote, with the PQ a paltry .7% ahead.

The Post wasn’t alone in getting it wrong, however. A political website called ThreeHundredEight.com, one that is relied upon by many reporters, analyzed a number of polls, and stated that the Parti Quebecois could win as many 75 of the National Assembly’s 125 seats on Tuesday night, with the Liberals winning as few as 25.

When all the votes were counted on election night, however, the PQ had won only 54 seats, and the Liberals — who too many had suggested were as good as dead — captured 50 seats.


  1. Jeremiah Gallinger says:

    Eric Grenier of threehundredeight has got Québec 2012 wrong, Alberta 2012 wrong, Newfoundland & Labrador 2011 wrong, Ontario 2011 wrong, and Canada 2011 wrong.

    Why people want to know what he thinks is beyond me; he’s just a lame rip off of fivethirtyeight with no statistical training or methodology. The fact that the Globe and Mail actually pays him to come up with the crap he does is amazing.

  2. Don Johnson says:

    Hi Warren,

    Truly an ongoing problem. What can be done about it? I mean, other than the media breaking their addiction (like that will happen!)

  3. reformatory says:

    and ABACUS leads the pack in unreliability. The bottom line, why aren’t polls banned. Experts know the nonsense published sways elections. In a true democracy, there should be no place for the manipulation that polls facilitate.

    Polls should be used for internal strategy only and the media should not be allowed to publish them. During the writ period it ought to be considered influence peddling.

  4. Pete says:

    Eric Grenier, who writes the 308.com stuff is a rank amateur playing with numbers he really doesn’t understand in the context of real voter intentions. He plays entirely on some theory of his that has gotten him a columnist gig with the G&M.

    I noticed the major pollsters stayed out of the limelight on the Que. election front except for Leger who has separatist leanings and also forecast a PQ majority.

  5. Cam Prymak says:

    Good column and it highlights the polarizing effect of the media.

    ‘Analysis’ from media outlets with a focus on their own narrative where readers on the left are just as convinced they’re correct as the readers on the right.

    But where is the middle ground and the fact checkers and the investigative reporters? Was it a coincidence they went the way of the dodo when reality TV paved the way for ever easier profits for the media?

  6. I really think that when people get calls from these polling companies they lie about who they intend to vote for. Voting to me is a private issue, and guys like Eric Grenier, who’s track record is very bad should just stay out of guessing elections and report on the party’s platforms instead. This is not just restricted to Canada, as some of the major media outlets in the States get it wrong all the time.

  7. Graham says:


    Mr. Kinsella and I actually agree on something. Maybe the Mayan’s were right: The world will end this year.

    Seriously though, Forum Research and EKOS are tied for the most pathetic polling companies on the planet.

    Forum Research refuses to publicly release any information about their polling, just the results. They refuse to release who PAID for any polls. They don’t release the questions asked, age and geographical breakdowns or any data tables. NOTHING. If you want to see them, you need to pay a yearly “membership fee” of several hundered dollars. Pay hundreds of dollars for inaccurate data? Where do I sign up?!?

    As for Frank Graves and EKOS – I’ve lost count how many polls/seat predictions he has gotten wrong. Just off the top of my head:

    2004 federal election
    2006 federal election
    2008 federal election
    2011 federal election
    2010 Toronto Mayoral election
    2011 Ontario provincial election (he got that one wrong at least 4 times!)
    2012 Alberta election
    2012 Quebec election

    He has even submitted a rather lengthy document to the courts supporting the ultra left leaning Council of Canadians legal action regarding alleged Conservative election fraud via robocalls.

    Frankie Graves’ proof the Conservatives committed election fraud? The results on election night were too far different FROM HIS COMPANY’S ELECTION POLLING.

    Seriously. That’s his reasoning. The fact he got his polling data on the May 2011 general election so wrong is proof the Conservatives committed election fraud.

    I can’t wait to hear his excuse for his grossly inaccurate Quebec election data.

  8. Joe says:

    Didn’T Forum Research get the prediction right in Kitchener-Waterloo?

  9. DanR says:

    I would take this column more seriously if the Liberals, provincially or federally, didn’t themselves commission polls and discuss, internally at least, seat projections. All parties do the same and pay a great deal of money on this junk.

    Oh, and the media is already a joke. There is no risk that they may become a joke. The fact that you are only waking up to this now demonstrates what a joke it is.

    Worse still, you write about this critically serious issue, ignoring the plight of billions around the world, not the least of millions in Canada. Ah, the problems of the 1% …

    • Tim Sullivan says:

      You’re a loon just like that Ross Perot clown a few years ago. He refused to engage pollsters, preferring to rely on publicly available ones instead. If you think the parties are polling “who would you support if an election was held today” type ones, well, you simply don’t know things.

      Your comment makes you look like one of those stupid conservatives (not the ones running the party because they know what to poll) who are scientifically proven to be that way.

  10. Derek Pearce says:

    So, is it a problem of people who answer are lying, or is it that pollsters should spend more money and get a damn bigger sample size each time they do a poll? Because it seems that US polls are more on target, they predict with more accuracy.

    • Kelly says:

      They have two major parties. We have 5 parties in Parliament. Our phony electoral system only works in a two party system but a two party system is a sham democracy at best. In he US your only real choice now is between a party that is pretty right wing and One that’s really right wing. That’s a false choice. The cons and NDP are trying to create this same phony system here. We need the Liberal and Green parties … And if youre a true democrat, even the Bloc. Just because one may not like the Bloc’s agenda doesn’t mean the many folks who vote for them should be denied the option. PR would be the best way to neutralize sovereigntist parties. They would never hold a majority of Quebec seats again unless they somehow managed to get a majority of the vote. Something that is highly unlikely.

  11. Ronald O'Dowd says:


    A minor and yet increasingly significant piece of the puzzle: people DELIBERATELY lie to pollsters…

  12. William says:

    At the risk of re-running the election, I believe Charest didn’t nearly play up the economy more during the campaign. I can’t believe the amount of new buildings going up in Montreal….it had been a good 15 years since we’ve see it like this.

    I’m not overly surprised that pollsters got it wrong…there was a large amount of undecided and those people were not happy with what was going on in the streets by the students.

    Interesting to see if Marois’s budget gets taken down whenever it is presented (prob. in the Spring) and if we return to the polls or if the LPQ & CAQ try to convince the Lt. GG that a coalition can govern.

    Interesting times!

  13. William says:

    Bit OT, but thought you’d like this photo


    Best part, the guy’s a registered Republican

  14. Kelly says:

    We have a phony electoral system. Popular vote has almost no bearing on seat count. Our system is total bullshit and as a result, the polls are bullshit. It’s always this way. We have at least 4 competitive parties that poll in the double digits at various times in various ridings. The Cons pretend they have a majority when they’re not very popular at all (sorry 33% just isn’t a majority no matter how you spin it.) None of the parties are very popular. The only fair system is a proportional system. You want more accurate polls? Get a more accurate electoral system.

  15. Dan says:

    I think they have to revist their methodology. Most of the polling companies have been missing the mark by a substantial margin which should be alarming to those who are actually running the polls.
    Switching away from land-lines to mobile phones, the anti-SPAM and do not call lists are the new realities of today. Make up a model that accomodates for that.

Leave a Reply

Your email address will not be published. Required fields are marked *