A new poll conducted exclusively for Keeping an eye on Fox News North indicates 100 per cent of Canadians believe Sun Media/QMI is horrible at reporting poll results. This is based on sample of one (ourselves) and has a margin of error of +/- 98 per cent, 19 times out of 20. (Which means its possible 198 per cent of Canadians think they’re horrible, right? — .ed)
Of course, we jest. But so does QMI with their poll-related polling. For reals, how could they not be jesting when they report a poll that’s three weeks old and pretend its new, as we discussed yesterday? And we must say, Kory Tenecyke is nothing if not frugal with Quebecor’s money, because they got no less than three stories out of that one Leger poll.
Besides using it to belatedly trying to change the narrative on the Liberal Express, the same poll generated a story on hiring quotas by David Akin (just eight days after the polling period) and a Brian Lilley piece that claims that Canadians believe “by a margin of five to one” that all the Tamil refugees on the MV Sun Sea should be sent back to Sri Lanka.
Interestingly, not only was that story published 16 days after the polling period, the poll was conducted while the ship was still at sea, some nine or so days before the ship would arrive on Canada’s West Coast. That seems to be somewhat putting the cart before the horse. One wonders, what information were the respondents prompted with? The story does give us a hint:
Asked which statement best described their own opinion on what should be done with the ship, which may include members of the banned Tamil Tiger terrorist group, 60% agreed with the statement: “They should be turned away – the boat should be escorted back to Sri Lanka by the Canadian Navy.”
Just 17% agreed with the statement: “They should be accepted into Canada as political refugees.”
A significant number, 20%, said they did not know which answer to choose and 4% did not answer.
So when told a boatload of probable terrorists is on its way to Canada, 60 per cent of Canadians were concerned? Well colour us surprised. We’re surprised the number wasn’t higher, when you prompt them like that. And before we move on…
“There is that worry that this is potentially people coming in, not just jumping the queue, but coming in and falsely representing where they are from,” said Scholz. “We don’t often hear about other refugee claims that are done on an individual or family basis, but when we see lots of them at the same time we tend to get a little nervous.”
… this whole “jumping the queue” thing is an oft-repeated talking-point, because it plays on our innate sense of fairness. A responsible journalist though would take the time to note there actually is no queue for refugees. It’s entirely different from immigration. Refugees are processed as they arrive, and they arrive every day by plane, train, automobile and yes, by boat. These Tamils aren’t jumping any queue. And speaking of other refugee claims as Lilley does, it would have been illuminating for his readers if he’d mentioned here that, since January, 85 per cent of refugee claims from Sri Lanka have been accepted by Canada, compared to 38 per cent from other countries.
Let’s move-on from that one Leger poll, though. Having squeezed all the “news” it could from it, QMI turned to another “poll” on the Tamil issue (we’re sensing a pattern here — ed.). At least Leger is a known professional pollster (and they can’t help how QMI plays with their research). Lilley filed a piece today that claims polling information shows, shockingly, that some Sri Lankans go back to visit the country for the holidays.
A secret government survey reveals the majority of successful Tamil refugees travel back to Sri Lanka, raising questions about the legitimacy of their refugee status.
While government officials refused to release the controversial survey they did confirm the top-line figures to QMI Agency. The survey of Sri Lankan nationals was conducted in early August. A total of 50 people were surveyed, 31 of them had successfully obtained refugee status and 22 had returned to Sri Lanka.
And don’t miss the accompanying column from the always-entertaining Ezra.
First of all, just how “secret” is this “government survey” if we’re reading about it in the newspaper, and government officials are “confirming the top-line figures” for QMI? The answer is it’s not really secret at all, the government just wanted it reported as news and found a compliant news organization, run by their former communications director, to “leak” it to.
And just what do we know about this “poll” anyways? How was the sample selected? Was it weighted? What’s the margin of error for a poll of just 50 people? What questions were asked, and what were the findings below the top-line figures? How many of those surveyed just got here, and how many have been here for many years? When did these people go back, and what was the situation in Sri Lanka at the time? How many are refugees, and how many are immigrants?
These are all things a reputable media organization would want to know before reporting on a poll, because they’re central to knowing if those top-line figures should be taken with any seriousness or not. Yet, the story seems to make it clear QMI only had the information that the “leaker” wanted them to have. Which was good enough for Sun Media, apparently. (You’re not being played if you’re in on the game. — ed)
Kady O’Malley of the CBC obtained more details on the methodology of this “secret” survey from the government that Lilley didn’t report (she e-mailed them, and they e-mailed her back) and it provided some much-needed context that was missing from QMI’s story. Go give it a read. Here’s one key line though from the government:
Numbers were counted within the sample group and no statistical modeling or statistical process was applied. As such, CBSA has no ability to state that this sample is statistically representative of any pattern that may or may not exist beyond this sample. Rather, these numbers are solely indicative of an observed pattern within the small sample reviewed.
Since there has been no analysis beyond what was on the immigration file in Colombo, these findings should be considered as context only.
Additionally based on the size and statistical validity of this sample, it should be noted that no conclusions should be reached on the comments/observations made on this file.
Or, in other words, these numbers are essentially meaningless and probably shouldn’t be used to support flimsy conclusions to drive sensationalist media coverage.
Shooting polling blanks…
Moving-on to our third example of questionable “poll” reporting by QMI, we come to this story yesterday by Jeff Cummings that claims a new survey shows a majority of police officers oppose the gun registry. This would, of course, be newsworthy because it’s an odds with what most police chiefs and what most organizations that represent both management and front-line officers say, which is that the gun registry is an invaluable tool that helps them do their jobs.
So, what is the story’s thesis based on?
An Edmonton police officer believes an overwhelming number of law enforcement officials across the country are in favour of scrapping Canada’s long-gun registry through a survey in a national police magazine.
Kuntz, a 22-year EPS veteran, says 2,410 of the 2,631 officers from across the country he surveyed in Blue Line magazine since last spring believe “inaccurate” data from the registry is affecting police safety in every province and territory.
Well, that is about five-times the sample that Lilley and Levant based their Tamil vacation story on, so that’s something at least. But just how seriously can we take a survey some random person put in a magazine, following no statistical methodologies, with an unknown and self-selecting sample of respondents?
The answer is, not seriously at all. Unless you’re Sun Media, apparently.