I was going to write a long ranty post about the federal Conservative party's asinine plan to scrap the long form from the next Canadian census in 2011. But Beth did a much better job. What's puzzling is that Statistics Canada is renowned around the world for its excellent work, and I've never heard it accused of misusing or leaking census information.
Despite opposition from groups of all political stripes across the country, the government is being intransigent. I hope that if they do manage to screw up our next census by removing the long form, the Conservatives get turfed next election and the new government will reinstate it. I'm dismayed that senior ministers today seem either not to understand how demographic statistics work and must be used, or are letting a poisonous ideology override what understanding they might otherwise have.
Regardless, cancelling the mandatory long form (available from StatsCan, if you want to see the "controversy") is a stupid move, and a disservice to everyone in the country.
![[RSS syndication feed]](https://www.penmachine.com/images/rssfull.gif){kind=small}
![[RSS comments feed]](https://www.penmachine.com/images/rss_comments.gif){kind=small}
![[Penmachine Podcast]](https://www.penmachine.com/images/rsspodcast.gif){kind=small}
Of course, the government isn't cancelling the long form. They are moving it from mandatory to voluntary. The census will still be taken. Stats Can will still send people en masse to poll people missed when the mailings are completed.
Mischaracterizing this issue will get you nowhere. When the long form census starts arriving in people's mailboxes next year they're going to look hard at you and all the others in the chattering class who went apeshit about this and wonder why the hell they got a form that was supposedly cancelled.
The only thing that will change is that when people don't fill out their long form census, and the Statistics Canada representative who shows up at their door insisting that they comply, the non-compliance doesn't lead to legal action.
Besides, who says the threat of jail improved the quality of the responses anyways?
I don't believe I'm mischaracterizing. From the point of view of statistical analysis as I understand it (both from some stats courses many years ago and an amateur interest in the subject since), a mandatory long form and a voluntary survey or poll containing the same questions are two different things, because they produce different data sets from different respondents—so I'd personally still call the long form "cancelled." And replaced by an inferior alternative.
However, you do raise a good point that some people lie on the long form (and, presumably, the short form as well), so the planned survey might very well simply produce a differently skewed set of data. There will be errors and distortions in any census or survey, yes. But statisticians can account for and attempt to correct for the small distortions that arise in the real world. The article you linked to about people who intentionally try to skew the data is an anecdote without context to show how widespread it is, which is precisely the sort of potentially misleading information that real statistics are designed to counteract. Do we even have any information about how many people complained about the long form to the government, for instance, and whether they are representative of the population more generally? It doesn't seem so.
It is clear that people who actually understand and work with statistical information every day—and not just those working for StatsCan and various other levels of government—think the change is a bad move, for the reasons I described, among others. Is it more likely that the expert statisticians and data analysts (of many political stripes) are all wrong, or that the officials who made and announced this change unexpectedly, and also "without notice and in the absence of any obvious outcry against the census' being excessively nosy," are?
The statements so far from Minister Clement and others do not reassure me that they really understand how census data collection and analysis work. That is worrisome.
And that's before even mentioning how significantly changing the nature of the long form's respondents makes inter-census comparisons (and thus analysis of the effectiveness of programs based on that data) much more difficult.
As many have pointed out, governments (at least theoretically) compel us to do many things under threat of penalty: pay taxes, shovel snow in front of our houses, not sell tobacco and alcohol to or purchase it for minors, drive with proper care and attention, follow building codes, obtain business licenses, and so on. Some people might consider several of those requirements intrusive and coercive, as Mr. Clement has described the long form census. I disagree, and the past month has shown that many other Canadians do too.
To my mind, the change is still stupid, and it's still a disservice. It also seems profoundly unnecessary.
Damn, I ended up with a long rant anyway.