This past week, Statistics Canada released some findings from its 2011 Canadian census data, showing that Canadians are speaking a greater variety of languages at home. Justifiably, this news got a lot of attention, because of the political and societal implications of linguistic diversity in a country with two official languages. However, what was almost completely unreported that Statistics Canada also issued a warning about the quality of its own information.
According to this Canadian Press story, the Statistics Canada information on linguistic trends was accompanied by this blunt warning from the agency:
Data users are advised to exercise caution when evaluating trends related to mother tongue and home language that compare 2011 census data to those of previous censuses.
As the story indicates, it’s very unusual for a statistical agency to attach this sort of warning to its own data. Several statisticians quoted in the story (some of them former Statistics Canada employees) suggest that the warning is likely due to “anomalies” in the data, which they attribute not only to the 2010 cancellation of the mandatory longer form of the census questionnaire, but also to the resulting change in the placement of language-related questions in the mandatory shorter form of the census questionnaire.
I’ve written before about the importance of some of the other data that will now no longer be collected as a result of the Harper government’s interference with the work of Statistics Canada. And now we are starting to see the negative impact of some of those decisions. Reliability, sampling, cross-period comparisons, and questionnaire design may seem like wonky details that only statisticians and researchers get excited about. But this story shows how crucial these questions are – because, ultimately, they end up affecting the accuracy of our knowledge about ourselves and our country.
We should be concerned that Statistics Canada’s warning was underplayed or ignored in this week’s reporting. It’s not always easy to explain statistical stuff in straightforward terms, but that doesn’t mean it shouldn’t be explained at all. And we should also be concerned that the Harper government’s manipulations at Statistics Canada are already starting to affect the quality of data about our country – because it’s only going to get worse.