Stephen Harper’s 1991 master’s thesis used census data to make his case about “political business cycles” and he even noted how disruptive changes in methodology could be for long-term analysis in understanding how Canadian political behaviour changed over time.
He ran a model to show the links between variables such as unemployment and changes in government.
The unemployment data is based on the Labour Force Survey in more recent years, but originally came from the decennial census
The census prior to the Labour Force Survey was much longer and “intrusive” with regard to questions on employment than the short form is now, and was mandatory for everyone to answer.
The two-stage questionnaire didn’t come into use until the 1971 census, and up to that point the Census questionnaire became longer and longer. Statistics Canada provides a fascinating description of the history of the Census. Even enumerations back in the 1700s, prior to census-taking, asked highly personal questions regarding the amount of stock in a household, the number of swords and guns, and religious beliefs – highly political given the Protestant/Catholic clashes of that era. On top of everything, given widespread illiteracy, the questionnaire was filled in face-to-face, and likely by people who knew you. We’ve come a long way baby.
I thank Tracey Lauriault of the awesome project datalibre.ca who provide a reference to Harper’s thesis via the french story that ran in cyberpresse on Saturday “Quand Stephen Harper aimait le rencensement”.
We all change over time. It is doubtful that if the Stephen Harper of the 1990s were the one running the show today this whole census fiasco would have occurred.