Very short post, because I need to sleep for my night job.
1CalgaryCentre, that "unite all the progressives behind one candidate!" movement in Calgary Centre, has released its "polling" data.
This poll was done among those registered to the website. Essentially, people who went on the site and signed up. A "self-selecting" set of voters. Presumably, "progressive" voters, or more likely, "progressive" voters that know/care about this website enough to take part in its poll.
What the poll found: Harvey Locke isn't good enough to win, Chris Turner has more attraction value. Ergo, Green candidate Green Turner is the person we should unite behind. Or so the implication will go (they aren't going to endorse anyone), at least that is what I'm betting.
Problem: ITS A BIASED SAMPLE! This was a poll done among people who went to the site and were interested in the site, in no way actually representative of a) the voting population of Calgary Centre; and b) the "progressive" voters of Calgary Centre. There was no weighting done to the 454 respondents, as far as I can tell, and I checked. Its just the collected opinions of 454 people who signed up to do this poll - technically you can't even say this is a "random sampling."
Not a representative sample, so utterly useless for what we need it to do. Sorry, do proper polling or do nothing.
Return on Insight's poll is a better, if not still the best, way to do this - random sampling with a proper methodology. Not that.