Guest Post: Why we need the long form census – a data librarian’s views

Ted Hsu
Comments Off on Guest Post: Why we need the long form census – a data librarian’s views

We All Count graphic finalVince Gray, who has been a data librarian at Western University for over 29 years, talks about why we need the long form census.

Western is a member of Statistics Canada’s Data Liberation Initiative (DLI), which provides us with licensed access for institutional planning, academic research, teaching, and course work to all publicly released data from Statistics Canada. I’m not a user of data: I provide it to administrators, researchers and students for the purposes above. The lack of reliable data from the 2011 National Household Survey has had negative consequences for meeting the needs of this clientele. Let me give an example.

This past week, I was requested by Don Kerr (King’s University College, London: http://www.kings.uwo.ca/academics/kings-faculty/faculty-biographies/don-kerr/, in his role working with the New Poverty Research Centre at King’s) to provide census-tract level data on low-income status for the city of London from the National Household Survey. Not finding these data in the DLI holdings, I checked with Statistics Canada to determine whether these data are available publicly, through DLI files which I’d missed, through controlled access services (Statistics Canada’s Research Data Centres or Real-time Remote Access, which require considerable time and effort to use on the part of the researcher), or available as a cost-recovery tabulation from Statistics Canada. The answer that I received was that the data would be available through either the RDC or by custom tabulation. By contrast, for 2006 Census data, Don was able to simply go into an online retrieval system and download the required data to his desktop as needed. After informing Don of the availability of NHS data, he wondered:

“Yet on even more fundamental issue, is this worth doing? … is it possible to draw meaningful inferences at this level of disaggregation in terms of making comparisons over time on low income? Would Statistics Canada be confident that we are capturing “real change” in terms of the incidence of low income in making these comparisons (2011 NHS; 2006 Census; 2001 Census) on low income?”

These are precisely the questions I confront when asked  by students and researchers to provide data so that they may compare the situations in 2006 and 2011. I feel that I must refer users to the NHS User’s Guide, which warns “Caution must be exercised when NHS estimates are compared with estimates produced from the 2006 Census long form, especially when the analysis involves small geographies. [etc.]”.  Further, if they are working with income data, I must refer them to the NHS Income Reference Guide: “Given the sensitivity of most income indicators to such methodological differences, users should use caution when comparing income estimates from the NHS to other household income surveys, administrative data or 2006 Census data or earlier censuses”. It is then up to the individual student or researcher to decide for themselves whether this fundamental type of analysis – comparing now to then – is feasible with the data provided.

Ever since the cancellation of the mandatory long-form census was announced, whenever I have spoken to classes, I have encouraged them to lobby their MPs for the restoration of the mandatory long form census. I explain that their work and life when they graduate will be affected negatively by the unavailability of reliable data:

  • if they are city or social planners or community activists, how can they plan without reliable demographic characteristics? How do they plan, locate or advocate for programs without knowing the need or the optimal site for them?
  • If they are academics, how will they conduct research without reliable, comparable data on the nation’s population?
  • If they are private citizens, how will they (and their children) receive appropriate public services if the planners don’t know what is appropriate?

I generally finish by explaining that I anticipate being retired by the time the next census data are released: I won’t have the ongoing problem of not being able to provide users with what they need. Instead, I’ll be in the category of a member of the general public, hoping that crystal balls and/or best guesses can take the place of evidence-based planning based on reliable data from a mandatory long-form census, and dreading that political expediency and ideology will instead become the default planning tool.