I have always taken publicly released polls with a grain of salt. Anybody with an ounce of sense would realise that pollsters have political affiliations. Polls have increasingly become partisan, spin doctors tools, and while polling companies won’t like their names being put to erroneous data, there’s no problem with selectively releasing the data that serves their masters purposes best.
All that said, properly constructed surveys are the life blood of public opinion research. Yesterday, EKOS released the first of a series of really massive public opinion polls on behalf of the CBC. To quantify ‘massive’, they surveyed about 10,600 people, from coast to coast. Put into perspective, I start to trust poll results when the sample sizes get up to maybe 800+ respondants.
This poll is important because it actually has significant regional sample sizes, with 1,254 in BC, 917 in Alberta, 3,480 in Ontario, and 2,555 in Quebec. The poll respondants are sorted geographically, but also by gender, incomes, and ages. These sub-groups have been statistically weighted to ensure they are accurately sampled subsets. Again, these smaller subsets are still statistically significant sample sizes. You can contrast aggregate results, and regional results with a pretty high degree of confidence that you are comparing apples to apples, and draw meaningful conclusions from the comparison.
I won’t draw any conclusions from this poll at this time. The purpose of this post is basically to ensure that my Green Party comrades are alerted to the fact that this poll is not an ordinary poll. It is probably the most accurate poll ever published in Canada, and should be scrutinised carefully. My thanks to the CBC as well. There is no way in hell that anybody outside a major political Party could ever have access to this quality of data, and to make this public puts our researches on a par with the ‘big boys’ for once.