The 2011 Census came in the mail yesterday as another reminder of why the Harper government is bad for science. What am I talking about? No, I’m not referring to the restrictions that now limit environmental scientists from speaking freely to the media, or my frustration with the current resistance to supervised injecting sites or exporting generic medicines to other countries.
I’m talking about the science of statistics and how the Harper government did away with the mandatory long-form Census last summer. If you filled out the Census this year, you may have noticed that there are now two parts to the questionnaire: a 10-question short-form Census (which collects basic information such as age, gender and language) and the National Household Survey (which collects more detailed information about income, employment, ethnic origin, housing, educational attainment and many other topics). All of this information is incredibly valuable to all levels of government, as well as academic researchers and many not-for-profit and for-profit organizations. The data is made available by Statistics Canada and used to plan the delivery of services, target markets for business opportunities, and identify demographic trends, among many other potential applications.
What’s changed is that while previously it was mandatory to fill out the whole set of questions (the long-form Census), now the National Household Survey is voluntary and only the 10-question short-form Census is mandatory. The issue is that the data collected by a voluntary survey is of lower quality than if it were collected by a mandatory survey. A bias has been introduced such that some Canadians will opt-out of completing the voluntary survey, while some Canadians will still take the time to fill out the additional questions. In other words, the data collected is less representative of the Canadian population… which is the whole point of the Census.
The Harper government went ahead with this change last summer despite criticism from many individuals and organizations, across sectors and the political spectrum. The head of Statistics Canada resigned in protest and stated publicly that the move to a voluntary survey was a mistake. If that doesn’t tell you enough right there, anyone who has taken a statistics course should be able to tell you the same thing. Industry Minister Tony Clement said the decision was based on concerns that the mandatory questions were coercive and intrusive. Does anyone care right now? Probably not – we need to implement that super-bill to wipe out crime, and should probably make sure the economy is still happy. But please don’t tell me that crime is increasing (or decreasing) or that unemployment is rising (or decreasing) without the data to back it up.