The Indifferent Pollster

Frank Graves of EKOS Research, whose last poll underestimated the Conservatives’ popular support by a whopping 5.6 percent:

“We were almost spot on about the popular vote, but where we really screwed up was with the Tory majority. We need to go back and do some work to find out what went wrong, but nobody has the resources to do this. This is done on a shoestring budget, and while we do our best to try and understand our failures it requires analysis and time and nobody is really interested in that.”

Perhaps EKOS should print that on their business cards.

54 Replies to “The Indifferent Pollster”

  1. EKOS is the CBC’s in-house pollster. They always — without fail — prop up left-wing electees. Their polling criteria is much like the CBC’s “Vote Compass” leading questions.

  2. Exactly the same thing EKOS did with the Toronto mayoral race earlier this year, when they showed the conservative Rob Ford “neck-and-neck” with the Liberal George Smitherman right up to voting day. But reality was there was huge percentage difference and the conservative Ford in fact won by a landslide.
    EKOS seems to only under-estimate re: conservatives, because the rest of their polling is usually dead on.

  3. All the polls did was give false hope to all the Libs and Dippers. There is nothing the MSM can say that I will ever believe.

  4. Guys, guys. EKOS works fine. You just need to perform the “conservative adjustment”: take whatever EKOS’ number is for the Torys and add 6%. Bang on.

  5. slaw at May 3, 2011 4:18 PM:
    “…take whatever EKOS’ number is for the Torys and add 6%. Bang on.”
    ‘Cept for the Toronto mayoral race it was closer to 20%.

  6. Huffpo (or was KOS?) fired their pollster for doing such a lousy job (or were they scapegoated?) in 2010. Will the CBC fire Ekos and maybe hire Compas which when you read their analysis you will see they were bang-on.

  7. I’m reminded of this famous quote from Albert Einstein:
    “Insanity: doing the same thing over and over again and expecting different results.”
    The only question is, who’s the insane one, Graves or the public for listening to Graves?
    I have some professional history with pollsters and it’s not very positive. A lot of them are exceptionally arrogant and have an endless list of excuses when their methodologies and/or results prove to be flawed/wrong.
    Further to Graves’ comment, imagine if an engineer or mechanic or accountant were ever to say, “it requires analysis and time [to understand our failures] and nobody is really interested in that.” They would be immediately fired and never able to get work again. Not so with pollsters. Much to our detriment. 🙁

  8. The Liberanos cannot ‘give’ Frank a job anymore so he should start working on his resume. Maybe he can get a job in the Georgie Porgie spin room in Chicago.

  9. I wouldn’t mind having a look at Frank’s ‘shoestring budget’. But that’s difficult to do since CBC blatantly violates the Access to Information Act, preferring to spend our tax dollars fighting in court for their right to keep their books closed. Fire.Them.All.

  10. It was great to see that 40% + of Canadian voters were screaming at CTV, CBC and other lamestream media:
    “You don’t speak for me!!”

  11. How’s that culture war working out for ya Frank, didn’t help the liberals did it buttercup?

  12. My understanding is Frank Graves as President of ECHOS would spend a day or two in PMJC’s office each week determining the best way to phrase the questions the LPC wanted their poll to say.
    I am reminded of the old joke; An engineer asks what do you want me to do? A lawyer, or in this case a pollster from ECHOS asks what do you want me to do or tell you that will help your distortion. Cheers;

  13. EKOS should never be paid another cent of taxpayer money,they’re either biased or incompetent.
    Either way, why in hell should WE pay for their highly inaccurate services.
    I’ve seen more accurate results from a monkey spinning a “wheel of fortune”.

  14. as my Re-Elected Conservative MP Larry Miller said:
    “Polls are only good for hydro lines & dogs”

  15. “We need to go back and do some work to find out what went wrong, but nobody has the resources to do this. This is done on a shoestring budget, and while we do our best to try and understand our failures it requires analysis and time and nobody is really interested in that.”
    I am still laughing… is this the excuse of the Liberal Party or Frank “the dork” Graves?. EKOS – pronounced echoes…because it’s methodology and ethics are hollow.

  16. ” … it requires analysis and time and nobody is really interested in that.”
    That is a stunning statement from this shyster. He doesn’t care what went wrong with his polling apparatus — which had the potential to change people’s votes — and says that nobody’s interested??????
    Kate is! SDA readers are! I am! I imagine the CPC is!
    From my past: “Is Frank Graves ever a dumbcoff (dummkopf).” (I never really knew what a dumbcoff was, except that whatever it was, was really bad.)
    (I wasn’t putting much store in the polls, anyway, remembering that ever since PM Harper and the CPC have been in the mix, the pollsters have constantly under-“estimated” (uh hum) support for the CPC.
    I was pretty sure there’d be a majority for the CPC — I predicted there’d be lots of votes from “secret Conservatives,” people who would vote CPC holding their noses — but I’m absolutely delighted it’s such a whopping one!
    YIPPEE!!!!

  17. ” Geeze . We don’t have the resources or time to
    to figure out why we screw up—but—Yes!
    We can do a poll for you. What results would you desire?”

  18. “Perhaps EKOS should print that on their business cards.”
    Thanks for the best laugh in a while.

  19. The difference between the left and the right is that when the left sees low poll numbers they stay home on voting day and say “what’s the point ?” so they need to see the right polling low to get them off their unemployed arses. The right on the other hand, and this is what a left-leaning pollster can never understand, see those low numbers and instead of becoming apathetic as expected, say “I’d better get that fixed” and make time to vote and drag everyone they can along with them…notice how the moonbat ‘vote mobs’ only happened during school hours….the only thing that could get a student moving is a chance to skip classes and be on youtube at the same time.

  20. I belong to a number of on-line polling groups. I get invitations to complete surveys several times a day. I find that when I indicate that I am a white male from Alberta I am invariably excluded from the poll. Three strikes and you’re out. I wonder if some of these polls that were incorrectly predicting the election results have the same kind of exclusion algorithm. It would explain why they were wildly inaccurate.

  21. Maybe Frank should change the name of his firm to:
    “WeGuess” Polling.
    “We guess so you don’t have to think”

  22. No doubt the CBC’s pollster will make the same mistake anytime there is a possible threat from a conservative politician.

  23. Well I thought the Cons would get a small majority.
    Here’s why.
    All the polls I looked at showed the Cons pretty flat — up a bit, down a bit — but more-or-less where they started.
    All polls showed the NDP up and the Libs and Bloc down.
    The clear deduction was that the NDP was gaining Lib and Bloc support.
    The simple deduction was that this vote splitting would favour the Cons on close ridings.
    And so it was.
    Easy-Peasy deduction.
    I am a bit amazed that the polling groups couldn’t figure that out.

  24. CBC (yeah, I know) has a good article on their web site about political polling. Turns out, its a money loser for the polling companies, and they only do it for exposure and because various networks throw some money at them to give them some broadcast noise. They’re not interested in spending any money on procedural review and investigation and refinement, because its a losing proposition for them.

  25. ‘We need to go back and do some work to find out what went wrong’
    Frankieee, Fraankieee, don worree boudit…
    prob’ly you was just tired, eh Frankie.
    Listen, housabout we start by washin’ you mouth out with soap, eh Frankiee.

  26. conservatives generally poll slightly low. In my opinion its because there are a lot of us that won’t take the time to answer a telephone poll or we’re not available to answer a phone because we’re disproportionately AT WORK! or we just don’t like revealing our intentions to a stranger.

  27. Actually there is one other factor that shows up pretty well if you look at the Presidential approval polling at realclearpolitics.com. Its the difference between polls which include Adults, Registered Voters and Likely Voters. Adults always lean really left, Registered Voters are left of center and Likely Voters tend to be Right of Center. Been watching that for about 5 years and it never seems to stray from that pattern.

  28. As Mark Twain said, there are three kinds of lies;
    Lies, damn lies and statistics.

  29. Mr graves is essentially saying “it’s good enough for government work”

  30. Hey Skip, try Shaw Direct.
    Posted by: schlomo at May 3, 2011 7:59 PM
    I get SunTV direct off air – They’re running a crawler tonight mentioning that they’ve been dropped for Bell subscribers…

  31. EKOS is sadly a bit of a joke. Mr. Graves proves it in spades by his comments. Bye Frank. Compas had it right and Nanos did too.
    I watched the CTV coverage (sorry use rabbit ears) and Bobby Fife seemed quite upset. If nothing else, it was great to see Fife et al have their hat handed to them.
    My guess is that we will be seeing a few bi-elections in Quebec over the the next few months as the freshly elected crop of NDP members realize that parliament cuts into their skateboarding time.

  32. CanuckinMI (above at 7:16PM) has nailed it.
    Polling is ALL about who is defined as the target being polled. “Adults 18+” includes illegals who cannot vote, and they tend to be left … “Registered Voters 18+” gets closer to reality … “Likely Voters” is always and everywhere MOST accurate ….. BUT …. the cheapest, fastest, easiest polling to do is “Adults 18+” because they can slide a lot of OTHER ‘product useage’ questions into the same poll to spread the cost.
    EKOS isn’t interested in accuracy, just VOLUME.
    And to thinki it’s OUR tax billion that’s going toward their CBC polling … grrrrrrrr!!!!!

  33. “We were almost spot on about the popular vote”
    They were. There was one.
    Bet ya they get that right next time, too.

  34. ‘as the freshly elected crop of NDP members realize that parliament cuts into their skateboarding time.’
    Heh… good point. I can’t wait for the teenagers currently attending McGill or the bartenders and babysitting communists start opening their mouths on camera. I am sure what comes out will be very er.. enlightening to Canadians. At the very least watching them all entering and exiting the House of Commons with their ball-gags ought to be pretty entertaining.

  35. They might need a whole new wing at CSIS just to keep track of the NDP now.
    Of course, there was already one planned for us bloggers, so I’m just saying put it to better use.

  36. “Well, folks, when you’re right 52% of the time, you’re wrong 48% of the time.”
    /Simpsons (Lisa the Greek)
    “This is done on a shoestring budget”. Then maybe you should get out of the polling business.

  37. Now that you mention it, the NDP in Quebec is sort of like the small-pox-on-the-blankets thing

  38. the NDP in Quebec is sort of like the small-pox-on-the-blankets thing
    Indeed it is.
    Jack and his caucus are going to be playing the Bloc’s tune now and it’s going to be rather obvious to the RoC.

  39. Sorry but laughing this sorry episode off should not be in the cards. Once again polls have been used to try and sway voter sentiment. I guess that doesn’t say much for those that would be swayed but all the same… AT the bare minimum it should be public knowledge who commissions the pollsters. Political parties should have the ability to caution the public on their accuracy over time.

Navigation