Combined census research

What is a combined census, and how does it differ from a traditional Census?

The first census after Confederation in Canada was conducted in 1871, and the methods of collecting census data have evolved incrementally with each passing decade. During a typical or traditional census, the population is enumerated through self-response using either an electronic (online) or paper questionnaire, or by a Statistics Canada representative collecting information directly from respondents. In contrast, a "combined census" merges the collection of information directly from respondents with administrative data that individuals have already provided to other government organizations and agencies.

A combined census approach reduces the burden on Canadians by avoiding the collection of the same information multiple times, enhances the efficiency of the census, and can lead to improved data quality. Information is securely shared with Statistics Canada, with the privacy and confidentiality of personal information safeguarded. Both traditional and combined censuses aim to provide a complete count of the population and dwellings on a specified day.

What is administrative data, and how does it benefit Canadians?

Administrative data refers to information collected by government or private sector organizations as part of their routine operations. Examples include birth and death records, taxation records, data on the flow of goods and people across borders, and more.

The 2021 Census serves as a recent example of how the use of administrative data can benefit Canadians. The 2021 Census faced unprecedented challenges due to the onset of the global COVID-19 pandemic. To adapt, Statistics Canada drew upon its methodological innovations, experience with linked administrative data, and research into a combined census to develop a robust contingency plan for the 2021 Census. This plan was based on the secure, responsible, and appropriate use of administrative data.

Fortunately, despite the pandemic, the 2021 Census was a success, achieving a 98% collection response rate. Moreover, the administrative data contingency plan allowed for the imputation of data for approximately 12,000 dwellings in areas impacted by lower response rates. This led Statistics Canada to consider further developments in administrative data for future censuses.

Statistics Canada is dedicated to ensuring the highest quality of data collection. The administrative data undergoes rigorous checks and validation processes to ensure its accuracy and currency. Statistics Canada is also committed to ensuring the privacy and security of all Canadians' personal information.

Census Futures Research

Census Futures serves as a coordinating unit dedicated to fostering innovation and research for future censuses. Drawing on several years of research into the use of administrative data by the census and other programs at Statistics Canada, Census Futures is exploring the expanded use of administrative data in upcoming censuses. Simulations and testing indicate that administrative data could potentially act as a high-quality replacement for traditional census enumeration in certain situations, as well as enhance the current Census imputation methods.

Building on previous research

Statistics Canada has decades of experience in using administrative data to enhance its statistical programs and reduce the burden on Canadians. For example, records of births, deaths, and migration are integral to the Demographic Estimates Program. Similarly, the Census has, for several cycles, replaced certain questions with administrative data. For example, the Census program now accesses income information from the Canada Revenue Agency's tax and benefits records, eliminating the need for income-related questions.

More recently, as part of its continuous research into administrative data, Statistics Canada is developing innovative and robust methods for linking administrative data files. It is also evaluating the quality of population counts at both the individual and household levels and devising techniques to measure census concepts, starting primarily with administrative data. Simulations of census outcomes using linked administrative files and modeling techniques are yielding promising results. With each simulation, the methods are refined to enhance accuracy, striving to match the precision of traditional census results. Moreover, these datasets proved indispensable as a contingency plan for the 2016 and 2021 Censuses during collection challenges.

See these examples of Statistics Canada’s innovative research on the use of administrative data for the Census:

1. Lundy, E. (2022). Predicting the quality and evaluating the use of administrative data for the 2021 Canadian Census of Population. Statistical Journal of the IAOS. 38: 1177-1183. https://content.iospress.com/articles/statistical-journal-of-the-iaos/sji220082
2. Yoon et. al. (2022) Modernization of the Canadian Census: An Administrative Data-Driven Approach to Defining Households. Proceedings of the Survey Methods section of the Statistical Society of Canada, May 2022. URL: https://ssc.ca/sites/default/files/imce/yoon_ssc2022.pdf
3. Royce, D (2017). First report on the Census Program Transformation Project: Researching a new approach to Census taking. URL: https://www12.statcan.gc.ca/census-recensement/fc-rf/98-506-x/98-506-x2017001-eng.cfm

Simulating various models of a combined census

Statistics Canada has developed statistical models to assess whether individuals in administrative data are in-scope for the census, to determine their usual place of residence, and to group individuals into households. These models are reliable for a significant portion of the population and incorporating additional administrative sources could enhance their performance. Further simulations and testing will determine the feasibility of using administrative data in future censuses.

Several versions of a combined census are under consideration, each incorporating administrative data to different extents. Should a recommendation be made to adopt a combined census approach in the future, the chosen model will be informed by the results of simulations and testing, as well as the social acceptability of using administrative data among internal and external stakeholders, data users, and the Canadian public.

These simulations, tests, and assessments of public acceptability form part of a multi-year Census Futures research initiative. Increasing the use of administrative data for population counts also necessitates researching alternative methods to evaluate census coverage, while ensuring the continued relevance and high quality of the census counts and the Demographic Estimates Program.

Next steps

Statistics Canada has benefited from years of research and the experiences of other international statistical agencies that have made, or are in the process of making, the transition toward greater use of administrative data for a combined census. The agency is nearing a point where it can effectively estimate population counts in certain geographic regions using administrative data. Consequently, it has initiated research to determine whether new census collection methods would be acceptable to the Canadian public, as well as to internal and external stakeholders and data users.

The next steps for the Census Futures project are to:

• research whether a combined census could meet the high standards of data quality required for a Census of Population, how a combined census would be operationalized, and whether changes to census collection are acceptable to the public and to data users
• determine how the quality and coverage of a combined census would be assessed
• assess potential impacts on the Demographic Estimates Program.