Supplement to Statistics Canada's Generic Privacy Impact Assessment related to the 2019 Census Test

Date: April 2019

Program manager: Marc Hamel, Director General, Census Program

Reference to Personal Information Bank (PIB)

Personal information collected through the Census Test is described in Statistics Canada's "Census of Population - Census Program Content Test" PIB.

The "Census of Population - Census Program Content Test" PIB (Bank number: StatCan PPU 007) is published on the Statistics Canada website under the latest Information about Programs and Information Holdings chapter.

Description of statistical activity

The objectives of the Census Test are to determine whether new or revised questions under consideration for the 2021 Census of Population can be easily understood and correctly answered, to assess public reaction to these questions by testing them on a smaller scale basis, and to evaluate behaviour of staff when using new systems and procedures. This quantitative test, mandatory under the Statistics Act and involving a sample of dwellings across the country, will be conducted from May to July of 2019.

The Census of Population's purpose is to provide statistical information, analyses and services that measure changes in the Canadian population and, demographic characteristics. It serves as a basis for public and private decision making, research and analysis in areas of concern to the people of Canada.

The Census of Population is a reliable basis for the estimation of the population of the provinces, territories and municipal areas. The information collected is related to federal and provincial legislative measures and provides a basis for the distribution of federal transfer payments. The census also provides information about the characteristics of the population and its housing within small geographic areas and for small population groups to support planning, administration, policy development and evaluation activities of governments at all levels, as well as data users in the private sector.

To ensure the continued relevance of the census to Canadians, Statistics Canada conducts a formal consultation at the start of each census cycle. During that time, Statistics Canada invites data users, stakeholders and the general public to provide feedback on what information they use, for what purpose and what, if any, data gaps Statistics Canada should consider addressing in the next census cycle. A Consultation Report, 2021 Census of Population Consultation Results: What we heard from Canadians, was published on Statistics Canada's website on April 8, 2019.

As in past censuses, extensive consultations on the questions to include in the 2021 Census of Population have been held with Canadians. New and modified questions, developed to reflect new needs identified in the consultations, have been qualitatively tested by Statistics Canada in 2018 by using methods such as individual interviews and discussion groups.

Reason for supplement

While the Generic Privacy Impact Assessment (PIA) addresses privacy and security risks related to statistical activities conducted by Statistics Canada, this supplement describes potential new risks associated with the collection, processing and use of data related to new census content being considered, and the possible concerns of Canadians about the intrusiveness of the collection. Key additions to content being tested are:

  • a new gender question that allows respondents to report a non-binary gender to respond to new Statistics Canada standards for sex and gender and Bill C-16 (Canadian Human Rights Act);
  • new language rights-holders questions to identify children with right to attend French schools outside Quebec, English schools inside Quebec, in support of Article 23 of the Canadian Charter of Rights and Freedoms;
  • a new veteran identifier question to support the mandate of Veterans Affairs Canada (evidence for service delivery, policy development, program evaluation);
  • a new self-reported health question to provide a strong predictor of health service needs, including health status of vulnerable populations and for small areas;
  • a modified list of examples for ancestry and religion; modified immigration and place of birth questions (the list of examples has significant impacts on response patterns);
  • two modified Indigenous Identity questions; new Métis and Inuit questions to update terminology, and to respond to needs for more precise information on Métis and Inuit, and to address issues with census data highlighted by land claims' agreements;
  • new questions on most recent credential and its field/location of study and year of completion; digital skills of recent graduates; modified questions (i.e., credentials and school attendance) – to increase relevance by reflecting changes in the delivery of college education in Canada; improve quality; assess if recent graduates have the digital skills they need for the digital economy; and
  • new questions on labour market experiences; modified questions on journey to work to increase relevance by better capturing labour market experiences such as underemployment, and involuntary part time work; better capture commuting to better meet data user needs.

This supplement also describes the procedures in place for the fingerprinting process. In compliance with the Government Security Policy, all personnel recruited for the census must be granted Reliability Status, which requires the conduct of a fingerprint based Criminal Records Check (CRC) in accordance with the Treasury Board Standard on Security Screening. The Agency is working with a third party provider to test an automated work flow for fingerprinting and credit checks as part of the 2019 Census Test.

Necessity and proportionality

The collection and use of personal information for the 2019 Census of Population Test can be justified against the four-part test proposed by the Office of the Privacy Commissioner of Canada:

  • Necessity: Statistics Canada's mandate is to ensure that Canadians have access to a trusted source of statistics that meet their highest priority information needs. The efficient production of relevant, accessible, high-quality statistics helps to ensure that the Canadian economy functions efficiently and our society is governed effectively. As part of this mandate, Statistics Canada is responsible under the Statistics Act (R.S.C., 1985, c. S-19) for conducting the Census of Population every five years. By law, the government (by an Order-in-Council) prescribes the questions to be asked in the census. By the same law, each person is required to provide the information requested in the census and Statistics Canada must protect the confidentiality of the personal information provided by respondents. The next census will be conducted in May 2021. Being a major statistical operation, the census presents a formidable challenge for Statistics Canada and requires careful planning to ensure its successful completion. An essential element to ensuring this success is the conduct of a census test, which traditionally is scheduled two years prior to the census. The content proposed for the census is the result of a Census Content Determination Framework that begins with consulting external stakeholders on their uses of the data and related information needs including the requirements for relevance and quality, followed by an assessment of the Canadian context which encompasses respondent burden and societal privacy concerns, and an assessment of Statistics Canada considerations including costs, operational factors, data quality and safeguards against loss of efficiency and/or quality in other Statistics Canada programs.

    The specific content under consideration responds to high priority data requirements identified through consultation, for example:
    • A new gender question, and modified sex and relationship to Person 1 questions, allow census respondents to report non-binary gender, in support of new Statistics Canada standards for sex and gender and Bill C-16 (Canadian Human Rights Act).
    • New language rights-holders questions and modified language questions identify children with right to attend French schools outside Quebec, English schools inside Quebec, under Article 23 of the Canadian Charter of Rights and Freedoms.
    • A new veteran identifier question supports Veterans Affairs Canada's mandate, with evidence for service delivery, policy development, program evaluation.
    • A new self-reported health question provides data that is highly correlated with the person's actual health, strong predictor of health service needs, and would provide health status of vulnerable populations and for small areas.
    • A modified Indigenous Identity question, and new Métis and Inuit questions, update terminology, and respond to a need for more precise information on Métis; a need for more precise information on Inuit; and address issues with census data highlighted by land claims' agreements.
    • New questions on most recent credential and field/location of study and year of completion; digital skills of recent graduates, and credentials and school attendance, provide increased relevance by reflecting changes in the delivery of college education in Canada; improve quality; and assess if recent graduates have the digital skills they need for the digital economy.
    • New questions on labour market experiences, and modified questions on journey to work, provide increased relevance by better capturing labour market experiences such as underemployment, and involuntary part time work, and better capture commuting information to meet data user needs.
  • Effectiveness: The collection and use of this data during the 2019 Census Test will be effective in meeting Statistics Canada's objectives because it is necessary to ensure that the 2021 Census of Population produces relevant, high priority, statistically meaningful information.

    A considerable number of content changes are being tested in 2019. Some affect core concepts such as sex and who should be enumerated in the household (coverage). Many of the content changes proposed for 2021 affect smaller population groups (transgender, non-binary, same-sex couples; ethnic groups; residents with work or student visas; Indigenous populations, etc.). In order to be effective, the content test needs to accurately discern statistical differences between 2016 content (benchmark) and different versions of modified content, in order to ensure that proposed content can produce relevant and statistically meaningful information.
  • Proportionality: Data from the Census of Population Program provide a comprehensive socioeconomic portrait of Canada's population over time, that supports key requirements for policy areas, including labour market conditions, immigration, Indigenous peoples, education, mobility, skills development, official languages, housing and income. This information is vital to all levels of government, the private sector, academia and non-profit organizations for decision-making and for developing and monitoring of programs and policies. To ensure the efficient and successful conduct of the 2021 Census of Population, Statistics Canada is conducting the 2019 Census Test, which will comprise a sample of 258,000 households across Canada. The test serves multiple purposes, from testing improved or new questions to testing improvements to collection tools and procedures. The sample size and the complex test design were determined in such a way as to ensure that all the test objectives are met, while minimizing as much as possible the respondent burden and the cost. Smaller qualitative tests were previously conducted to help minimize the sample size of the 2019 Census Test. Without such a content test, the success of the 2021 Census of Population would be at significant risk.
  • Alternatives: The Census Program is one of the only sources of information for small geographic areas, based on the same statistical concepts for the entire country, and the only source of information for many socioeconomic characteristics.

    Prior to conducting the 2021 Census, extensive testing will be done to assess changes to content, communication materials, field procedures, and collection methodologies. The 2019 Test is designed to accurately measure respondent behaviours to changes in content, field and collection methodologies. Voluntary tests in 2019 could yield inaccurate or inconclusive findings for many of the proposed changes to questionnaire content. Literature on survey collection shows that mandatory census tests provide the most accurate impact measurement when studying small sub-groups of the population.

    If responding to the test was voluntary, non-response bias could occur, specifically for smaller population groups. Similar concerns were raised in the context of the 2011 National Household Survey (NHS). The 2011 NHS demonstrated that in a voluntary context, some respondents tend to skip questions or abandon more often, amplifying non-response for certain questions, especially those closer to the end of the questionnaire. These questions could be more impacted, leading to wrong conclusions being drawn, and, as a result, recommendations for new or changed content for the 2021 Census being based on incomplete test results. In addition, without reducing the risk of non-response bias, a voluntary test would require a larger sample size to account for total non-response. Based on a preliminary design, at least 60,000 additional dwellings would need to be part of the test, and with anticipated higher proportion of non-response, there would be an increase in the number of reminders and contact attempts for selected households.

    Further, although Statistics Canada continues to explore the use of new administrative sources to replace data collection from respondents, the research is not yet sufficiently advanced to consider the use administrative data to replace new content under consideration for the 2021 Census of Population.

Mitigation factors

The 2019 Census Test will adhere to all Statistics Canada policies on collection, processing and dissemination of information.

This section defines the safeguards in place to address privacy risks that are identified in the Generic PIA as they relate to this specific activity.

Storage

Respondents' personal information is transmitted directly to Statistics Canada via secure, web-enabled electronic questionnaires, or, in the case of paper questionnaire formats, through the mail directly to a Statistics Canada secured processing site, where the information is captured and an electronic record created.

Records are stored in a response database containing household responses, including personal information to perform record linkages to the 2016 Census of Population response files and to tax files.

All electronic records are stored within the secure Statistics Canada information technology environment. Paper questionnaires are securely stored within the Statistics Canada access-controlled processing environment.

Records are retained for a period of five years after the completion of the Census Test and then are destroyed.

Processing and analysis

The Agency's longstanding linkage experience has shown that particular combinations of personal information elements can be used to identify individuals in different data sources with a very high level of confidence, and thus link individuals across various files.

The following personal information will be collected and used to perform record linkages:

  • First and last names (collected)
  • Date of birth (collected)
  • Sex/Gender (collected)
  • Full civic address (from frame and collected)
  • Phone number (from frame and collected)
  • Mobility one and five years (collected)
  • Country of citizenship (collected)
  • Place of birth (collected)
  • Immigrant status (collected for respondents getting Control questionnaire or Test questionnaire version 2; added through linkages for respondents getting Test questionnaire version 1)
  • Year of immigration (collected for respondents getting Control questionnaire or Test questionnaire version 2; added through linkages for respondents getting Test questionnaire version 1)
  • Frame identifier (from frame)

The linkage of Census of Population Program data and 2019 Census Test data will be used in statistical analyses to evaluate the impact of new or modified questions of the questionnaire by comparing the 2019 Census Test to other existing sources or to the data collected in the 2016 Census. The linkage file will be used only within Statistics Canada, for methodological research, development and processing. No statistical estimates will be published from the linked data.

Personal identifiers such as name, phone number and address (excluding postal code) will be removed from linked files, and replaced with an anonymized statistical identifier.

Access

Access to any confidential data held by Statistics Canada is closely monitored and restricted to designated individuals as per the program's operational requirements. Employees are required to provide the justification for access and obtain the necessary approval. Furthermore, all access permissions are only applicable for a set duration of time and must be regularly renewed including justification for re-approval.

The response database for the 2019 test has not yet been created. The number of employees with access to the 2019 Census Test response database will vary depending on the timing of 2021 Census preparation activities and the need for staff to have access for their job function. Access to the identifying information (names, address and phone numbers) within this database will be controlled, and limited to a small set of staff requiring such access. Most staff who will analyze the data will not have access to any identifying information within the database.

Dissemination

Data collected during the 2019 Census Test will be used for Statistics Canada's analytical purposes only, and will not be publicly disseminated. Statistics Canada will not disclose personal information from the 2019 Census Test without the consent of the respondent.

Openness

It is the policy of Statistics Canada to provide all respondents with information about: the purpose of a survey (including the expected uses and users of the statistics to be produced from the survey), the authority under which the survey is taken, the mandatory or voluntary nature of the survey, confidentiality protection, the record linkage plans and the identity of the parties to any agreements for sharing of the information provided by those respondents, where applicable.

For the 2019 Census Test, this information is provided in the letter of invitation to complete the Census Test questionnaire, in the electronic questionnaire itself and in Frequently Asked Questions (FAQs) accessible through the Census website.

This supplemental PIA will be publicly available on the Statistics Canada website as an addendum to the Generic PIA. In addition, Statistics Canada has prepared FAQs and speaking points to assist its internal staff to deal with inquiries from the public regarding the Test.

Other factors

This section identifies other privacy considerations related to the 2019 Census Test.

Use of administrative data for collective dwellings

Statistics Canada attempts to collect the same information from residents in collective dwellings that is collected from other Canadians and has used various collection approaches in the past. The 2019 Census test will repeat the methodology used for the 2016 Census whereby institutions' administrative records were used as a collection instrument when available, replacing direct enumerations of residents.

Cell phone text reminders

In 2016, as a final reminder before starting field follow-up activities with non-respondents, the Census of Population used Voice Broadcasts to remind Canadians to complete their Census. These reminders resulted in increases in response. As more and more Canadians are using cell phones, Statistics Canada is exploring a new approach using text reminders in the 2019 Census Test and whether this method could prove effective for 2021 as a complement to the voice broadcast.

On June 3, 2019, all non-responding dwellings with cell phone numbers in scope for the content portion of the 2019 test (approximately 27,000 dwellings) will receive one of three treatments: one third will receive a voice broadcast reminder, one third will receive a single text-based reminder, and the remaining third will not receive any reminders. Note that, when respondents in scope receive a third mail reminder to complete their census, they are informed that they could subsequently be contacted by Statistics Canada by phone or in person. Non-responding dwellings without a cellular phone number will receive a voice broadcast reminder.

Statistics Act employee recruitment

Prospective candidates for Census Field Operations are required to complete an application through the secure, web-based Census Online Recruitment Application. The personal information associated with the application is stored in the Collection Management Portal (Statistics Canada's secure corporate tool for field staff), which is addressed in Statistics Canada's Generic PIA. This information includes name, address, contact information, citizenship status, experience, abilities and references. This information is used by recruiters to screen, test and interview candidates, as well as to initiate the security clearance process.

In compliance with the Government Security Policy, all personnel hired for the census must be granted Reliability Status. This includes a fingerprint based Criminal Records Check (CRC) and a credit check, in accordance with the Treasury Board Standard on Security Screening. Statistics Canada expects to conduct approximately 80,000 Reliability Status security assessments as part of the staffing process for the 2021 Census of Population.

Statistics Canada is working with a third party provider (as a deemed employee) to test an automated security screening work flow for fingerprinting and credit checks during the 2019 Census Test. The third party provider will not have access to candidate information, except if access to the system is required to trouble-shoot critical issues, in which case the request must be approved by the Departmental Security Office following consultation with the Census Program. The access will be supervised and subject to access controls. Two departmental security officers will be authorized to access the security screening results during the production and retention periods.

The security screening work flow begins with the candidates' written consent to the fingerprint-based criminal record check and credit check. Subsequently, their fingerprints are taken using accredited mobile fingerprinting devices at recruitment test sessions administered by Statistics Canada staff. The security screening solution being used for the test is accredited by the Royal Canadian Mounted Police (RCMP). It includes fingerprinting scanning devices and a web-based application to capture the candidate's digital fingerprints, along with the candidate's name, address and date of birth. The solution securely interfaces to both RCMP computer systems – to perform the criminal record check – and to a credit rating service – to perform the credit check. The results are sent via secure electronic transmission directly to Statistic Canada's Departmental Security Office (DSO) as part of an automated workflow. DSO officers complete a further assessment of all candidates who are non-Canadian citizens, have declared a criminal record or have been out of the country more than 6 months in the previous 5 years. Reference checks are performed by Statistics Canada staff concurrently. Candidates who pass the security assessment screening (criminal record and credit checks) and the reference checks, are granted a Reliability Status and are considered eligible for census employment.

A review of the proposed procedures was conducted to ensure compliance with the Library and Archives Canada's Multi-institutional Disposition Authority 98/001 on the General Administration Function. As per ongoing operations at Statistics Canada and according to the disposition, the criminal record result sheets and credit check results are stored on a secure network drive in the candidate's electronic personnel file and retained for the required retention periods:

  • Criminal Record result sheets (which include fingerprints) and Credit Check results of candidates who qualified for position and are in the pool of candidates (not hired) are kept for the duration of the pool and then securely destroyed.
  • Criminal Record result sheets (which include fingerprints) and Credit Check results for candidates hired are kept for a period of 2 years after the end of employment.
  • Fingerprints of candidates who did not pass the test and were not hired have no business value and are considered transitory in nature. They are securely disposed of as soon as they are confirmed as no longer being required.

Conclusion

This assessment concludes that, with the existing Statistics Canada safeguards, any remaining risks are such that Statistics Canada is prepared to accept and manage the risk.