We Don’t Govern on the Basis of Statistics

In the past three months, the Harper government has said some things about statistics that are blatantly wrong and received some media attention, which is a bit unusual. What things did the government say, and were they actually correct?

First, in June the government announced that it would be replacing the mandatory long form census with a voluntary one, and sending the voluntary one to more people.  Industry Minister Tony Clement implied that sending the long form census to more households would be an acceptable substitute for the mandatory census, and that Statistics Canada agreed. This wasn’t the case, and on July 21, Munir Sheikh, the head of Statistics Canada, resigned over the issue. The government is sticking to its initial plan.

Can a voluntary census sent to more people replace the mandatory one with an almost 100% response rate? No. The main problem with a voluntary survey is that there will inevitably be biases in the responses. As an example, adults with lower literacy rates in French or English will be underrepresented in the census because of the additional difficulty they face in completing it. The mandatory census reduces this sort of bias as much as possible, and then additional filters developed by Stats Can remove the joke responses. For the surveys that Stats Can completes most of the time, the census is the benchmark used to determine how representative the respondents were of the general population. If the census becomes, essentially, a particularly large survey, there is no benchmark to know how close the results are to reality.

Second, at the beginning of August, Treasury Board President Stockwell Day announced that the government would be spending more money on prisons and justified it by claiming that surveys show that unreported crime rate has increased alarmingly.  The latest Stats Can information from 2004 shows that unreported violent crime has remained stable. The unreported household crime rate (break & enter, theft, vandalism, etc.) for 2004 increased slightly, but the main reason given for not notifying the police was that the crime wasn’t important enough.

August 10, 2010 Globe and Mail Editorial Cartoon

August 10, 2010 Globe and Mail Editorial Cartoon

So, we have two cases this summer, where the Harper government has taken controversial issues and tried, unsuccessfully, to back them up with statistics. Obviously, they have other justifications for their decisions, but it’s disappointing to see them try to misuse facts. I think a quote from Justice Minister Rob Nicholson sums things up nicely:

“We don’t govern on the basis of statistics. We govern on the basis of what we hear from the public and what law enforcement agencies tell us.”

Statistics Canada General Social Survey Victimization information:
1999, Catalogue no. 85-553
2004, Catalogue no. 85-002, Vol. 25, no. 7

8 Responses to “We Don’t Govern on the Basis of Statistics”

  1. “We govern on the basis of what we hear from the public and what law enforcement agencies tell us.”

    And yet law enforcement agencies have spoken out against changing the census and the scrapping of the long gun registry. This government doesn’t appear to listen to anyone.

  2. Raimund says:

    Good article! You do have to hand it to the conservatives though, as far as cleaning up Canada’s image, this should be good. After all, the goody do rights will have full representation while the stoner crowd will fall entirely away. Suddenly Canada won’t be representing minorities of any sort, subcultures and odd-balls will cease to exist and the good Christian right will finally be large and in charge. Who couldn’t say no to that? Oh right! Those who care about objectivity and proper representation! That’s okay though, Harper was planning on continuing to ignore them anyway and what better way than letting them vanish entirely on their own?

  3. itzac says:

    Good article, Marion.

    It does give me a shred of hope that the Harper government at least still feels the need justify it’s actions, even they fail at it so reliably.

  4. AndrewK says:

    Me thinks the Conservatives are expecting to gain support on the basis of tenacity alone.

  5. Alex says:

    Right! In the famous words of the Great Jean Chretien:

    “A proof is a proof. What kind of a proof? It’s a proof. A proof is a proof. And when you have a good proof, it’s because it’s proven.”

  6. As a follow up, Stats Can did an analysis on the 2006 census data by assuming that anyone who returned their mandatory long form after Stats Can started following up (May 2006) would not have returned it if it were voluntary. The analysis was for the Toronto area.

    As you might guess, some of the simulated results are quite different than the real ones.


  7. Alex Murdoch says:

    After hearing Mr. Clement state that a single complaint justifies a change in public policy, I have sent emails to both Mr. Clement and my MP. I would encourage everyone else to do the same. Perhaps an organized campaign to push for evidence based decisions rather than ideological ones could be started.

    I’m in the mood for a scrap!

    • I’d be less annoyed if they just said up front that they’re ideological choices. This whole “We have statistics…I’m sure we do…no, wait, forget we said that” routine is getting old.


  • Marion Kilgour

    Marion is a mechanical engineer, and also works to promote critical thinking and scientific literacy through local skeptical and atheist activism in Edmonton, Alberta. Marion especially wishes to encourage girls to consider science or technology-based careers, and is involved in the University of Alberta's Women in Scholarship, Engineering, Science and Technology (WISEST) project.