Well if Tom Lukwiski is any indication you might be right, but the day is early.

Did he really think they would take offense and refuse to answer such an intrusive question?

You got that right…

90 minutes of Clement debating Sheikh would tell us more than 90 hours of this posturing bs.

Heck, 30 minutes would probably do it.

Yeah, if only there were some forum in which public policy decisions could be debated.

I seem to recall something like that, but it's used so rarely and so poorly these days, it's all kind of foggy. Anyway, I'm sure King Harper will do the right thing.

That's a good question. I would say I am wavering somewhat.

I still think that much of the long-form census data is of questionable valuable, but I think the solution is not the wholesale slaughter of the long-form that I've advocated (sometimes strongly, sometimes weakly), maybe just a crash diet.

In the immediate case, I think the 2011 census should be kept mandatory because there probably is not enough time to fix it properly and doing a half-assed job is probably the worst solution.

Watching this gov't up close gives half-assery a bad odour.

Once again, your personal info is never sold. Aggregate information is sold.

It seems you are assuming that means they will go after the high income neighbourhoods. It doesn't necessarily mean that.

I read somewhere that some group (not sure if it was United Church or other) used the Census data to demonstrate the need to corporate donors.

The info sold is data not persional, not ided to anyone.

Seems to me that conservatives were gung-ho to have stay-at-home moms receive pay. re: housecleaning?

Would statistics not help in pushing that, Wilson?

"One would think."
That's not even an argument. You're just claiming something can be done about an idenifiable problem without any proof.

"StatsCan has identified the groups least likely to fill out a voluntary form,
if they can do that, they should be able to take historical data, find the factor difference and apply it."
Sheikh also pointed out that response rates have been dropping. So we would have to extrapolate from historical trends. Thereby using old, out-of-date data to pile onto the errors already intrinsic to extrapolation processes, and then add on the errors inherent in surveying, and weight according to that? MAYBE you can do that for one year? And then what, keep going? Assume that data from 2006 and older censuses are applicable, say, 20 years from now?

You're illustrating a scenario where, under a responsible, prudent government, a policy is tested, problems identified, and rectified. This scenario, as you've pointed out, would involve massive public education. Meanwhile, in the real world, a voluntary survey is plopped out in the middle of summer through a Gazette publication. Nobody's heard of the idea before that day, and we end up in a maelstrom where a ton of the government's own statements seem to suggest that filling in the survey is a bad idea.

The actual number of respondents isn't the key here. If you can do a decent enough ad campaign to target expected low respondents, 20% may be just fine. (Drummond is saying this RIGHT NOW.) It's the well-known fact that the people who don't respond to a survey are different from those who do, and the differences are EXACTLY what you're trying to measure.

And how would you benchmark a successful targetting campaign? You can't be expecting the resultant ratios to fit the old data, since that's the whole point of measuring them in the first place! Maybe survey pre- and post-campaign response rates, but the survey itself would again be subject to sampling bias. There MIGHT be some ideal world scenario where we can fix everything, but honestly, all evidence suggests that we're way off that scenario.

And another note: how much of this talk about the government itself dismissing the value of its own
NHS/LFC questions is going to distort things? Ironically, my guess is that all this talk will work the "right" way and rebalance the usual surveying gaps against lower incomes, new immigrants, and such. But what a horrible, disgusting way to do it.

No, he said he did not care what Clement said, had nothing to do with his resignation,
it was the media reports and the perception they left with Canadians and the integrity of StatsCan.
(my words) that the 2011 survey has been deemed unreliable by the media reguritating select opinions…

Spector started slandering Sheikh's motives from the get-go… with squat for facts and a bizare interpretation of the timeline. Now he is simply spinning.