Wednesday, 29 August 2012

Right on, Leger Marketing

Democratic Socialist CCF appealing
to the prairies with a leftwing populist
message during the middle of the 20th

Image Source: Next Year Country
Earlier I wrote about the ideological self-identification of people in Saskatchewan and Manitoba. I was particularly interested in dissecting an EKOS poll which gave respondents the option of small-l "liberal", small-c "conservative", and "neither". I viewed this as problematic for a region with many left-leaning Dipper partisans who may not quite like a term like "liberal" that's so strongly associated with a particular party (the Liberal Party).

Leger Marketing, a few months after the EKOS survey, conducted their own survey with an ideological self-identification component. The Leger Marketing survey gave the options of "Left wing", "Centre-left", "Centre", "Centre-right", "Right wing", and "Don't Know". This, especially alongside the EKOS survey, is a very useful tool for making educated guesses about the political psyche of the prairies.

Like the EKOS survey, the Leger Marketing Survey offered regional breakdowns. At least according to the the release I've linked to, Leger Marketing hasn't broken down ideological self-positioning by other measures (partisanship, gender, age) like EKOS did.

The margin of error (MOE) for Manitoba/Saskatchewan isn't given in the Leger Survey. For 1, 506 respondents the MOE is 2.5% and the sample size for MB/SK is 125, so I'd assume it's big like the EKOS survey, a good 8-10%. Results should be taken with a considerable spoonful of salt. 

With all warnings obligatorily given, let's dive in. 

Results from EKOS and Leger Marketing surveys compared, click to make larger. 
For EKOS the "Don't Know" response is surprisingly smaller than in the Leger Marketing Survey. Perhaps "liberal-conservative-neither" politics is really easier to grasp in Canada than "left-right" politics. A Compass poll back in 2004, after all, showed that many Canadians didn't position the parties on an left-right scale the way political scientists would (only 47% of respondents positioned the then-recently defunct Alliance Party to the right of the NDP, for instance). Still, there's a great risk that  Canadians are just confusing small-l liberalism or small c-conservatism with the Liberal and Conservative parties in the EKOS survey. Neither guarantees that the public has a firm grasp of the ideologies talked about.

When given the option of the "centre" it looks like many respondents hover there. Over a third of people in the MB/SK choose the centre, showing that our region has the greatest portion of centrists in all of Canada. The total right is greater than the total left, but is nothing like the whopping 54 to 23 gap in the EKOS poll for small-c conservatives vs small-l liberals.

Also interestingly, it seems that the centre-left and centre-right are rather evenly matched, allowing for a competitive fight between left-leaning and right-leaning parties come election time.

Percentage of respondents in each ideological category
according to the Leger Marketing Survey.

The prefix free "left wing" (3%) still is smaller than the prefix free "right wing" (8%), but a high MOE should make us weary of interpreting too much into this.

All in all, a rather interesting look at the two prairie provinces of Manitoba and Saskatchewan. It indicates a rather centre-hugging electorate, with enough centre-left and centre-right voters to make elections competitive. Which is exactly what we see tend to see, "NDP dynasty" talk aside.

Would you like to support this blogger? Consider making a donation or checking out their shop!

Also consider liking this blog on Facebook!  


  1. Marketting surveys are done by polling folks over the phone. Here's the funny part. Only landline phones are to be called (Something about cellphones paying for minutes that don't want to be wasted on solicitors). Well, you know who still has a landline these days? your grandparents! And that's about it. So their results may be skewered from polling only a small (& old) section of society.

    1. Apparently the sample was weighted.

      From linked press release:

      "Final data was weighted by age, gender, language, level of education and household composition (with or without children under the age of 18) in order to obtain a representative sample of the Canadian population."

    2. That's what they tell the public. But meanwhile...