Thursday, July 31, 2008

Panic, MASSIVE POLL-style

After her long-winded defense of phone-spam polling on Monday, SUZANNE's been thinkin'. Today, rather than defend the KLRVU poll with backup data on its methodology, she tries to poke holes in the methodology of the Ipsos-Reid and Angus Reid polls that found 2/3 of Canadians were good with the Morgentaler OC. Too much weighting, or not enough. Over-representation of pro-abortion demographics. Then she concludes:
""The media and the left-wing bloggers accepted these polls without question, without looking at the methodology.""
Gee... I wonder why that might be. Ya think the reason we "accepted these polls without question" might be because we could find Ipsos-Reid and Angus Reid in the phone book, or on the internet, or on some roster of polling companies, or through Industry Canada, in the event that we did have questions? Or because Ipsos-Reid and Angus Reid are proven pollsters with years of experience? Or because the questions asked were unbiased? Or... I could go on, but you get the gist.

SUZANNE still doesn't seem to get that, as John Counsell said on his radio show last week, the stats mean nothing if it's not a legitimate polling company. And that has certainly not been proven. Although a website was cobbled together under duress, it still doesn't even have its own phone number, still sharing an unpublished number with LaserCut Engraving and who knows what other little cottage-industry businesses.

There's nothing inherently wrong with hiring a guy to do some kind of a poll using phone spam technology. Just don't trumpet the results as if they were anywhere near being on the same playing field as those produced by legitimate pollsters with long-standing reputations and business portfolios. They're not. That's been put to bed. KLRVU itself does a great job of explaining why we shouldn't believe their polls.

But while we're still on this increasingly tiresome subject, something that hasn't been put to rest is the question of how CLC hooked up with KLRVU in the first place. It took us days to come up with information on the company -- how did CLC manage to find them, with no phone number, no website, no listing in any business roster? Or did they just dial random phone numbers until they came up with a "polling company" that just happened to be run by the brother of one of their pet anti-abortion MPs? Serendipity?

Canada is an adamantly pro-choice country. Get over it and stop trying to prove otherwise with "polls" that are no more than lame-ass propaganda.

UPDATE: Someone from Angus Reid weighs in at SUZANNE's:

Please don't fall into the "bigger is better" trap. Angus Reid has conducted surveys on a wide variety of topics (including two that showed support for Bill C-484) and called 21 of 22 parties within the margin of error in five provincial elections using its online panel.

It's not a quesiton of contacting 13,000+ Canadians or dialing a million numbers. It's about the quality of the sample and the options available to respondents.

From a methodological standpoint, the problems with what we have seen from the KLRVU poll are:

a) No breakdowns for age, gender, household income, education or federal political preference. All we have are percentages of respondents who said "No". We can't tell if only men, only women, or even if minors answered this poll.

b) Apparently, no "Not sure" option. A poll must give a respondent the option to state that he or she does not know, and the pollster must register these answers.

c) Misleading press release. The question used is: "Do you believe abortionist Henry Morgentaler deserves the Order of Canada?" Yet the press release headline reads: "56% Oppose Morgentaler Order of Canada." A person may feel that Morgentaler does not deserve the Order, but that is not the same as actually voicing opposition to the fact that he will get it.
Sounds reasonable. Shorter SUZANNE's response:
"LALALALALALA I can't heeeeeeear yooooooou, Angus Reid Pollster!"
Are we done here? Yep, I think we're done here.