Saturday, July 03, 2010

In those days Caesar Augustus issued a decree that a census should be taken of the entire Roman world...among those who felt like it

I know this isn't a big issue. I don't expect facebook groups and protests in the streets over it. But, as someone who worked on the Census in the past, this really grates me:

Former StatsCan head slams census decision by Tories

OTTAWA - The former chief statistician of Canada says he would have quit his job if the Conservative government had tried to axe the long census form, as they're doing now.

Ivan Fellegi, a veteran public servant who spent 51 years at Statistics Canada before retiring in 2008, says he's alarmed by the decision to replace the long census form with a voluntary survey next year.

He joins a chorus of groups, including the Federation of Canadian Municipalities, the Canadian Council on Social Development and others who rely on the detailed data to make major public policy decisions.

"It would have been a heck of a lot better if this long-form census was cancelled because at least we would have saved $100 million — that would have had a rationale," Fellegi said in an interview with The Canadian Press.

"To come out with something (voluntary) that has uncertain quality, and certainly for some groups it will be unpublishable quality, is not something that I can understand."

The mandatory short census, with its basic questions on the ages and sexes of people in a household remains unchanged. But one-third of Canadians, up from one-fifth, will now receive a voluntary "national household survey" with detailed questions on ethnicity, education and income similar to those on the long census.

The cost of the change could reach $30 million, says Statistics Canada: $5 million for the additional mailout, and $25 million in case there is a major problem in getting people to respond.


The problem is that once you make the form voluntary, the sample becomes self-selecting, losing its randomness and, by consequence, its representativeness. The bottom line is that certain types of people will be more likely to fill out their form, skewing the results.

You'll get far more accurate data from 1/5 of all households than you will from the 1/3 of all households that bother to complete their form. It's the same reason why you get better data from a scientific telephone poll of 1,000 Canadians, than you do from the 10,000 people who vote on the Globe and Mail's web poll.

Census data is used for everything from municipal governments deciding where to build schools, to entrepreneurs deciding where to open up new restaurants. Making the long form voluntary will compromise the data used to make thousands of public and private sector decisions.

Yeah, I know data collection isn't sexy. But it's important. This decision needs to be reviewed.

Labels:

35 Comments:

  • Good call, grit.

    This is just a brain-dead move. It should be reversed.

    By Blogger Gauntlet, at 12:46 PM  

  • Will the census be part of your upcoming book being released before the new year?

    By Blogger James Bowie, at 2:35 PM  

  • Sorry, but no.

    There is evidence to suggest that at least some of the long-form data was not reliable, because people resisted filling it out and therefore did not fill it out accurately. This was particularly true of the "ancestral heritage" questions.

    And as I look through the long-form questionnaire, I really don't notice anything absolutely required for policy decisions which isn't available through other means. For example, Transit Operators know how many people take public transit: I don't need census data in order for government to make an informed choice.

    By Anonymous Anonymous, at 2:51 PM  

  • Do you realize what you're saying, Anon? You're saying that the entire science behind census-taking is flawed. You're saying that all those billions that have been spent on census-taking around the world has rendered useless information.

    Like I'm going to believe that.

    By Anonymous Anonymous, at 4:37 PM  

  • Boris at Galloping Beaver explains some of the consequences of the stupid ignorant Harper government decision:

    http://thegallopingbeaver.blogspot.com/search/label/Census

    The various liknks also explain more consequences.

    By Blogger Holly Stick, at 4:38 PM  

  • @Anon #1

    The long-form data may not be perfect, but data from a self-selected sample is imperfect in a much more unpredictable way.

    By Blogger Daniel, at 5:53 PM  

  • Exactly. If the problem Anon 1 says is true - that the data is flawed because some resist filling out the long form - then this will worsen the situation even more. Because those who resist won't fill it out.

    They either need to axe the long form completely or keep it mandatory. Switching it to a self selecting form is absolutely pointless.

    By Blogger calgarygrit, at 6:20 PM  

  • Having worked as a "Census Commissioner" in the past, I can say that a lot of extra time and effort is put into getting reluctant respondants to comply with what is actually law. So apparently the "law and order" Conservatives are selective about which laws they want to enforce.

    As Felligi states, if the rationale for axing the long-form was saving $100 million, that would actually make sense.

    StatsCan does lots of surveys and censi (?), and has a well-earned INTERNATIONAL reputation for reliable information that is (in aggregated form) available to the public and business. That would be lost with a "voluntary" long-form, but the need for that information remains, and it is likely that StatsCan would just have to spend more money getting that information through other surveys.

    So, is the current government of Canada acknowledging that it would prefer to make decisions on the basis of incomplete and inaccurate information? That's hardly a commitment to competence, is it?

    By Blogger Party of One, at 1:17 AM  

  • "...Transit Operators know how many people take public transit..."

    Yeah, but they have a vested interest in overstating their case, as do many others vying for public funding. The point is that StatsCan is a well respected IMPARTIAL collector of reliable information.

    And anyway, I've never seen questions on the Canada Census about public transportation, although there may be other surveys that measure this.

    By Blogger Party of One, at 1:23 AM  

  • Data collection isn't sexy??? Says who???

    I tend to be a CPC supporter, but I gotta agree with CG, this isn't a very good move. Better to scrap it or leave it alone. My preference would be to scrap it.

    By Anonymous mecheng, at 9:24 AM  

  • "...Transit Operators know how many people take public transit..."

    No, they don't. TTC knows that it collected 500 million fares last year. It can't tell you if that was 500 million people riding it once or 1 person riding it 500 million times.

    By Blogger leonsp, at 10:16 AM  

  • Have any of you ever had to fill our the farm census, lots of room for lies to distort facts. And who cares how many bathrooms I have in my house. Under livestock we even included the rabbits and deers that came to eat. Refused to answer any question dealing with income or crop sales. Census taker threatened us with jail if we did not answer them. I asked her how she knew and was told, they open all forms to make sure they have been filled out properly. I threatened her with a lawsuit for breaching my (and many others ) privacy. I told her to put in figures if she must. Oh, can't do that, it would not be legal.
    Also Canadian was not accepted for nationality, but we insisted.

    By Blogger maryT, at 11:06 AM  

  • Mary, they check for completeness, not accuracy. Yes, you can lie like a rug if you choose. I wouldn't assume that everyone else does just because you might. And in any case, statistical methodology does handle anomolies and outliers...

    Those who complain about "invasion of privacy" better not be using credit cards or "affinity" cards. A hell of a lot more personal information is collected from those sources

    By Blogger Party of One, at 2:04 PM  

  • I did not lie on my census form but it does state that only stats Canada will see them A census taker opening all forms prior to submitting them was wrong, and an invasion of privacy. Refusing to answer a question is not lying. As for info collected via credit cards etc, that is why I advise my clients to pay cash for things like ammunition for their rifles and shotguns, and other articles they don't want anyone to know about. But, even paying cash is difficult as the receipt want you name etc.
    A few years ago, when involved in getting a federal boundry changed and I told them how many people were involved one person said I was wrong. Their info had the population several 1000 more. They were wrong, and I pointed out that their info from a census was taken on a weekend of a rodeo in the area and everyone was counted. I asked them if that entitled our town to a huge increase in grants etc based on population. Won the argument and the boundry was changed. Garbage in garbage out.

    By Blogger maryT, at 2:29 PM  

  • " A census taker opening all forms prior to submitting them was wrong, and an invasion of privacy"

    Sorry, Mary, but you're just plain wrong with that statement. Having managed the Census(at a low level, responsible for directly managing the actual census takers), I can tell you that the process REQUIRES all submitted envelopes to be opened in order to determine that all questions have been responded to.

    Frankly, nobody involved in the process at this stage CARES about your information or the accuracy of that information!

    Furthermore, the StatsCan Census takers are bound by federal law with respect to FOIPP. There's a very stiff penalty involved for transgressions of that law, far more than any possible benefit to an individual using that information for their own purposes.

    It's entirely possible that a census was taken while the rodeo was in town, but one of the very first questions on a Canada Census form is "Is this your normal place of residence?". If you respond "no", the next question is "Where is your normal place of residence?" If you provide that information, any data collected at that point is "assigned" to your normal place of residence.

    Yes, it's possible that there was some sort of screw up where data collected during the rodeo was assigned to your area, leading to increased population levels. That's an administrative problem, that can be easily remedied, and typical of ANY large organization. It doesn't invalidate the entire Census process, however.

    By Blogger Party of One, at 5:13 PM  

  • How would you know if a form was filled out correctly. How would you know if a person answered honestly. That is the first time I have heard a census taker (which I have been in the past) had to open all forms. In the year I am talking about I did investigate and was told NO they can't open them. There was quite a discussion on the Dave R show, when he told me I was wrong, then had to apologize when he was faced with the same problem.
    Could you please post a link to somewhere that proves your point.
    We will fill out the next form whenever it comes, but will not give any financial info, or info on how many tvs, phones, or bathrooms we have. And we will again put down Canadian on ethnic origin.

    By Blogger maryT, at 5:44 PM  

  • "And anyway, I've never seen questions on the Canada Census about public transportation, although there may be other surveys that measure this."

    Then you may wish to take a look at the Long Form Census, which I linked above.

    By Anonymous Anonymous, at 6:53 PM  

  • Not "correctly", Mary..."completely". They check to see that every question has been answered (i.e. that there is SOME response). Even if you make things up, as long as you respond, they don't care. From the Census taker's p.o.v., they only get paid for "completed" forms (it's piece work, really).

    What year are you talking about? I worked Census in 1990. The process was as I described above. It actually saves money doing it this way, since if Statscan waited until it was time to actually enter the data, you would have to pay permanent staff to follow up.

    To anon653, yes, it's possible that there were questions about public transportation on the LONG form. I didn't get one this time. Every Census there are requests for specific information from fed government departments, provincial governments, business groups, academics, etc. StatsCan doesn't ask all the questions proposed for the long form, but do pick the ones that seem relevant.

    By Blogger Party of One, at 8:28 PM  

  • This comment has been removed by the author.

    By Blogger Red Tory Liberal, at 9:02 PM  

  • What are you on maryT?

    By Blogger Red Tory Liberal, at 9:03 PM  

  • i'm a bit confused why people are shocked that Team Harper and the CPC are ignorant about science/math...

    By Blogger Disgruntled Liberal, at 11:33 PM  

  • There are a few duties that one can argue should be mandatory for citizens, like paying taxes, jury duty and voting (the last of which actually isn't mandatory).

    It's reasonable to include participating in census as one of those mandatory duties.

    HOWEVER many of the questions on the census aren't anyone's business and a respondant should have the legal right to decline.

    Which questions? Well, that to me is where the debate comes in.

    P.S. Remember the debates when Dief the Chief wanted to remove ethnicity?

    By Blogger Robert Vollman, at 11:24 AM  

  • The bottom line is you either axe the long form completely (or edit it) or you keep it mandatory.

    By Blogger calgarygrit, at 12:00 PM  

  • Incidentally, "Canadian" has been an acceptable answer on the ethnic identity question since 2001 (i.e. for both the 2001 and 2006 censuses). Both times, it was the most common answer.

    By Blogger ajbeecroft, at 12:50 PM  

  • I am obviously talking about pre 2001. Just talked to Stats Can and my MPs office and guess what, census will be voluntary next time, and NO workers can't open forms to see if they have been filled out. They can assist a person if requested.
    Hope they can read shorthand for my next one.

    By Blogger maryT, at 4:55 PM  

  • Glad to hear that your story was at least 14 years old, MaryT! I wouldn't want you feeling defiant about identifying your ethnicity as Canadian when that's been the most common response for a decade.

    Personally, I don't have any trouble with the question, but I do think that if "Canadian" is an allowable answer then we might just as well cut the question -- I don't think it reveals anything useful as it stands.

    And I agree with the original post -- making the long form optional is a huge waste of taxpayer dollars.

    By Blogger ajbeecroft, at 6:11 PM  

  • At one time Canadian was not an acceptable answer, but I think a lot of us just answered that and refused to put down ethic origin. If a refugee can have a baby in Canada and be Canadian, why couldn't those of us born in Canada years ago be Canadian, regardless of where our parents or grandparents were born.

    By Blogger maryT, at 6:33 PM  

  • Hope they can read shorthand for my next one.

    maryT, why bother filling it out in the first place if the census is going to be voluntary?

    By Blogger sharonapple88, at 7:51 PM  

  • Stubborness would be the answer to that. I do think we should be counted, name, age, SIN place of birth, citizen to be answered yes or no, Province and city/town. Nothing else.

    By Blogger maryT, at 8:08 PM  

  • "Stubborness would be the answer to that. I do think we should be counted, name, age, SIN place of birth, citizen to be answered yes or no, Province and city/town. Nothing else."

    Why not say, gender? To cite but one example, men and women have different healthcare needs and one would think the health system would benefit from being equipped to serve the communities around it.

    Just because you can't think of a potential use for particular information doesn't mean there is no use. This is a terrible idea that will make government far far less effective in delivering services and allocating resources (the result of that will be more government, as it will take more dollars to get the same result as before).

    By Blogger hosertohoosier, at 9:25 PM  

  • You see, some people PREFER that policy be made in a vacuum. Then, when it doesn't work, they can complain about THAT!

    By Blogger Party of One, at 9:41 PM  

  • I knew I left out a pertinent question and I do prefer gender to sex.

    By Blogger maryT, at 10:22 PM  

  • The problem is that once you make the form voluntary, the sample becomes self-selecting, losing its randomness and, by consequence, its representativeness. The bottom line is that certain types of people will be more likely to fill out their form, skewing the results.

    While I agree that this is a bad move, I don't know that moving to a voluntary survey necessarily implies a "loss of randomness". There will be a component of non-response bias, but that should not in principle eliminate the possibility of precise population estimates. Doesn't make this any less of a dumb idea. Fortunately the Harper government is exceptionally well-honed at backtracking on such issues.

    By Blogger Josh, at 12:18 AM  

  • This comment has been removed by the author.

    By Blogger Josh, at 12:18 AM  

  • "There will be a component of non-response bias, but that should not in principle eliminate the possibility of precise population estimates."

    The census has a lot more in it than population estimates. It also has a lot of data about individuals - like income, gender, age, etc. If individuals with certain characteristics respond (or don't respond) disproportionately, they will be under or over-represented.

    By Blogger hosertohoosier, at 2:05 PM  

Post a Comment

Links to this post:

Create a Link

<< Home