Thursday, January 28, 2010

It's Ekos Being Green

I decided to do a mathematical investigation into polling numbers in the month before and the month after the last time Canadians actually voted. Now that EKOS is all the rage at the CBC, it made sense to investigate their performance around the last interval where Canadians physically went to the polling booths and voted. I gathered all the polls for Nanos, Angus Reid, Harris Decima, Ipsos Reid, and even Donolo's alma mater Strategic Council. I bundled them together with a 3 point moving average and then juxtaposed that over the EKOS polls.

The first thing I noticed is that when people polled for EKOS, it was really easy being Green. In the month prior to the election, the Green Party's moving average for all the pollsters was 8.9%. Nanos averaged 7.7%. EKOS averaged the Green Party at 10.9%. On October 14, 2009 6.8% of Canadians voted Green. You may have noticed last week that the Green Party was boasting that EKOS was projecting them to win a seat if an election were held today. On Jan 7 2010 EKOS had the Green Party at 13.4% nationally. The Green Party cites Jane Taber writing about it, thus it must be accurate.

I did a Google search for "EKOS Green Party" and of the 9,000 hits most were blog posts boasting about the Green surge at various points in history. I would like to know if the Green Party has ever hired EKOS to do market research. I can't prove whether the pool of people they query are more likely to be Green than other firms, or if EKOS deliberately manipulates their results. What can be proven is that EKOS favours the Green Party significantly more than the population of Canadian polling firms. On their Wikipedia page, there is a claim that EKOS is the number 1 pollster for the Government of Canada, but a citation is needed.

On the matter of the Conservative Party, from Sept 23 to Oct 3 2009 EKOS undervalued the Tories by an average 2%. From October 4th to October 12 they overvalued the Tories by 1%. The day before the election, they had the Tories at 34.8%. On Election Day 37.6% of Canadians voted Tory. Draw your own conclusions.


  1. Originally I was going to use "Ekos, it's easy being Green", but decided less is more.

    It's Ekos Being Green works, right? Sorry, the Muppet Show had a magnetic influence over my early years. My favourite characters were the grumpy old men in the balcony...

  2. Where was I on October 14, 2008? Working a 14 hour shift as a Registration Officer for Elections Canada.

  3. Elizabeth May will ride her majestic unicorn to glory! Just you wait and see!

  4. All these pollsters have been calling up people for ages and they know who to phone in order to get the result they want. Nine times out of ten anyways.

    I'm sure it's as easy as clicking on the appropriate link in the computer program they use.

    The fact is, Polling companies are just a different variety of advertising and public relations businesses. Propaganda is what they make and sell.
    Nine times out of ten anyways.

  5. Marx: Care to put money on it? Stupid is as stupid does!

  6. The problem with seat projections is that they ignore the trend of recent election--it's getting harder to knock off incumbent MPs. Obviously, incumbents have many huge built-in advantages.

    How many successful, high profile Opposition candidates will put their careers on hold, working to build up a riding, etc etc, while not knowing when the next election will be held ? Danielle Smith was commenting along those lines a couple of days ago. But she is a very hard worker, compared to Iggy the diletantte. Has he recruited any high profile candidates since smokin' snowboard Ross ? Not that I've heard. And recruiting/fundraising should be the top priorities for an Opps leader.

    So all these people prediciting an Iggy minority gov't, show me the 40 or so seats that he will win. How many seats have the Libs not even got a candidate yet ?

    Calgary Junkie