Preliminary Report on Methodology Options for the 2016 Census
Prepared by Don Royce
for the 2016 Census Strategy of the Census Management Office.
The conclusions and opinions expressed in this paper are those of the author, and do not necessarily reflect those of Statistics Canada.
As part of the 2016 Census Strategy Project, I was retained by Statistics Canada in January 2011 to perform an initial assessment of methodology options for the 2016 Census of Population. This report presents the results of this initial assessment, which was completed in June 2011. It is intended to serve, along with the results of other work at Statistics Canada, as a basis for a more detailed assessment of options, to be completed by Statistics Canada by December 2011.
The assessment began with a wide-ranging review of census-taking approaches in countries around the world. It also considered international principles and recommendations for censuses of population and housing, in particular those from the United Nations, the United Nations Economic Commission for Europe, and the Conference of European Statisticians (Canada is a member of all three organizations). International census-taking approaches were then classified into four basic types: the traditional census, a census based on existing administrative registers (with or without additional data collection), the traditional census with yearly updates of characteristics, and the so-called 'rolling' census. The latter two approaches involve some form of continuous measurement, in addition to or in place of a census taken at a specific point in time.
The necessary conditions for using each approach and the main strengths and weaknesses of each approach were identified and documented. The use of these methods internationally in the 2010 census round was examined, and 11 case studies representing the approaches of 14 different countries were developed to gain a better appreciation of the potential applicability of each census-taking approach in Canada. This was followed by an assessment of the extent to which Canada currently meets the necessary conditions for each approach and, where it does not, what the experience of other countries can tell us about the necessary steps and the likely time frame for these conditions to become a reality.
As a result of this assessment, it is my conclusion that the only possible option for the 2016 Census is some variant of the traditional census approach. In the case of a census based on administrative registers, none of the necessary conditions are currently met. In particular, Canada has neither a Central Population Register nor a universal Personal Identification Number (PIN) that could be used to link such a population register (even if it existed) to other registers. Even in those countries that do have population registers and PINs, it generally takes several decades for such an approach to be implemented. In the case of approaches involving continuous measurement, for small-area data to be released in the same time frame as the 2016 Census or the National Household Survey (NHS), data collection would have to start in 2012, which is not possible. Continuous measurement might be a possible approach for the 2021 or later censuses, but further user consultation, research and testing would be needed to determine whether it would have any advantages over the current approach of a census taken every five years.
Within the traditional census category, a key unknown at this point is the accuracy of the data from the 2011 NHS, as well as the NHS' collection costs. Until further information becomes available later in 2011, it is recommended that the boundary between what census content should be mandatory and what content should be voluntary be left open.
Another key consideration in developing the options for 2016 is the appropriate balance between collecting some content from all households and other content from only a sample of households.
Based on these two considerations (mandatory or voluntary, 100% or sample), it is recommended that the potential options for 2016 be based on the possible combinations of the three following types of data content (building blocks):
- content that is required by legislation or that is otherwise sufficiently important that it must be collected on a mandatory basis from 100% of the population
- content that is required by legislation or that is sufficiently important that it must be collected on a mandatory basis, but only for a sample of the population
- content that can be collected on a voluntary basis and only for a sample of the population.
The first type of content would be part of any option that could be considered for the 2016 Census; consequently there are four possible options, depending on the presence or absence of the second and third building blocks. Three of these four options correspond to approaches that have been used in the past (including 2011); only the approach that would include all three building blocks has not.
Other considerations for 2016 that should be considered for further assessment and possibly pre-2016 testing include sample design considerations, the potential for closer integration of data collection for mandatory and voluntary questions, and the expanded use of administrative data within the traditional census approach.
Finally, the study identified three potential approaches for the 2021 Census or beyond. These are a census based on a new Central Population Register created by the federal government, together with the creation of a universal PIN; an approach whereby the 2016 Census would be kept up to date through a combination of existing administrative records and surveys; and some form of continuous measurement approach. The report outlines the issues that would have to be addressed for any of these approaches to be pursued. Considerably more work would be needed to develop and assess these approaches in detail before decisions about the census methodology for 2021 and beyond could be taken.