Date and location:
November 7, 2023, 9:00 a.m. to 12:00 p.m.
Hybrid meeting
- MS Teams
- The Ivan P. Fellegi Boardroom, 26th floor, R.H. Coats Building
100 Tunney's Pasture Driveway, Ottawa
Members:
- Chantal Bernier, Co-chair, Global Privacy and Cybersecurity Group, Dentons Canada LLP (Chair)
- Anil Arora, Chief Statistician of Canada
- Len Garis, Researcher and Adjunct Professor, University of the Fraser Valley
- Ali Ghorbani, Cyber Security Chair, University of New Brunswick
- Robert Gordon, Executive Director, Canadian Cyber Threat Exchange
- David Robichaud, Ethics Professor, University of Ottawa
Regrets:
- Mary Dawson, former Conflict of Interest and Ethics Commissioner, Lawyer
Statistics Canada guests/support:
- Antonio Bakopoulos
- Martin Beaulieu
- Josée Bégin
- Miguel DaCostaeSilva
- Pierre Desrochers
- Stéphanie Dion
- Tom Dufour
- Rebecca Kong
- Tony Labillois
- Lucie Léonard
- Guillaume Maranda
- Patrice Martineau
- Nicolas Rahal
- Eric Rancourt
1. Introductions and opening remarks
Anil Arora, Chief Statistician of Canada
Chantal Bernier, Co-chair, Global Privacy and Cybersecurity Group, Dentons Canada LLP (Chair)
Mr. Arora opened the meeting by welcoming members to the seventh meeting of the Advisory Council on Ethics and Modernization of Microdata Access (ACEMMA). He also informed the Council that Mary Dawson had decided to step down from the Council and thanked her for her participation over many years. He added that she contributed greatly with her pertinent questions and always judicious interventions, and that she will be greatly missed.
Mr. Arora went on to talk about the changing environment in which Statistics Canada must evolve, and the various impacts on the organization. In particular, he discussed the growing need for disaggregated data to inform decision-making, the evolution of individual and collective rights, notably in terms of privacy, and new technologies which offer undeniable opportunities, but also present certain threats.
Mr. Arora pointed out that Statistics Canada remains very active, both nationally and internationally, to adapt quickly to change and maintain the trust of its partners, including Canadians.
He then introduced the members of Statistics Canada present at the meeting, as well as his team of Assistant Chief Statisticians.
Ms. Bernier went on to say how nice it was to have the chance to see each other in person again. Virtual meetings have their advantages, and were very useful during the recent pandemic, but face-to-face meetings offer undeniable intangible benefits.
Ms. Bernier also reiterated the importance of quality statistics and organizations like Statistics Canada for a better understanding of the world's challenges. She concluded by outlining the agenda for the meeting and introducing the members of the Council.
2. Management of the continuum between privacy protection and the production of statistical data to inform Canadians
Pierre Desrochers, Director, Office of Privacy Management and Information Coordination, Statistics Canada
Mr. Desrochers spoke of the importance for Statistics Canada of managing relationships with partners to maintain their trust and reiterated that this represents specific challenges and made a presentation to the Council on Social Acceptability in the Context of Microdata Access and Privacy.
Social acceptability includes maintaining direct and constant contact with Canadians, who inquire about how Statistics Canada uses and protects their information. Drivers towards change and social acceptability include the context of misinformation, the general level of trust in government, citizens concern about data protection in reaction to significant breaches as well as legal and technological developments in the public and in the private sector. Keeping abreast of technological and potential legislative changes is important.
Council members were pleased with the presentation and with Statistics Canada's concern with social acceptability. In reply to Statistics Canada's specific requests for guidance, with respect to necessary laws and regulations, the Council members raised the challenge for legislation to keep pace with technological advances, and therefore insisted on the importance of adopting guiding principles that can be adapted to new technologies, rather than strict policies that can quickly become outdated. Second, they underscored the renewed importance of Statistics Canada in quality statistics production, in the age of fake news and disinformation. Thirdly, it was recognized that trust remains a cornerstone to produce quality data, and an organization like Statistics Canada should maintain relationships with all Canadians.
In conclusion, the Council recommends that Statistics Canada continue to pay particular attention to social acceptability, since it materializes into trust. To maintain this trust, Statistics Canada must continue to represent all Canadians, explain its approaches and objectives transparently, and remain consistent and predictable.
3. Use of nudge to encourage participation in statistical programs
Guillaume Beaulac, Manager, Behavioural Insights and Impact, Shared Services Canada
Mr. Beaulac presented the concepts of nudge and choice architecture to the Council. He explained that our choice architecture decisions always have an impact. We should therefore opt for a structure aligned with nudge principles. He then defined the concepts underlying nudge and distinguished it from sludge, whose intentions are not necessarily benevolent.
Mr. Beaulac then provided several examples of nudge. He pointed out, however, that impacts can be difficult to predict and differ from one person or group to another. He therefore recommended to use experimentation and randomized controlled trials to assess impact.
Council members were pleased with the presentation. They support the use of nudge to encourage participation in statistical programs, especially in a context where response rates are declining. However, the use of nudge must respect ethical principles: first, before implementing a nudge, its necessity should be justified; second, the use of the nudge must be transparent; third, the option to withdraw should always exist (except for mandatory surveys); finally, it was suggested to offer, in return for accepting to participate in statistical programs, to receive the results. The use of experiments should also be considered.
4. Do no harm when publishing data on more vulnerable populations
Rebecca Kong, Assistant Director, Canadian Centre for Justice and Community Safety Statistics, Statistics Canada
Guillaume Maranda, Lead Analyst, Data Ethics Secretariat, Statistics Canada
Ms. Kong and Mr. Maranda shared the guiding principles used in different contexts including ethical reviews and the publication of disaggregated data:
- Benefits for Canadians
- Fairness and do no harm
- Transparency and accountability
- Quality
- Trust and sustainability
- Privacy and security
- Nothing about us without us
They also presented concrete, recent examples to illustrate each of these principles, and show how they can pose significant challenges. Their suggestion to help manage these situations is to establish guiding principles, guidelines and best practices for Statistics Canada employees and external users.
Council members were pleased with the presentation. They agreed with the challenges presented by the publication of disaggregated data, and in particular the concept of do no harm. There are sometimes several elements to consider, and no single solution. They also supported the use of guiding principles, guidelines, and best practices, although Statistics Canada must remain realistic about their use by external users.
The concept of nothing about us without us also raised questions. For example, could a group object to data concerning them (do they have a right of veto)? And who makes the decision, given that several groups may be involved? The Council reiterated two important principles when it comes to statistics: unless there is a major problem, the data collected should be published, and the rules should be established in advance and apply equally to all. Statistics Canada will need to clarify the scope of this principle and how it intends to manage its application.
The Council would like Statistics Canada to present the concrete results of the application of the various guiding principles at a future meeting.
5. Familiarizing statistical agencies and other organizations with data ethics
Walter Radermacher, Chair, Advisory Board on Ethics of the International Statistical Institute
Mr. Radermacher presented to the Council the tools his Advisory Board has put in place to promote the use of ethical principles in organizations, particularly statistical organizations. For example, they promote the establishment of values and guiding principles such as respect, integrity, and professionalism. He also shared tools for promoting data ethics, such as a periodic table of open data elements and a stakeholder analysis matrix for ethical purposes.
Council members were pleased with the presentation. Interestingly, the presentation bridged the three previous presentations, and provided ideas for solutions. In the Council's opinion, the whole presentation is an excellent reference tool.
6. Roundtable and closing remarks
Anil Arora, Chief Statistician of Canada
Chantal Bernier, Co-chair, Global Privacy and Cybersecurity Group, Dentons Canada LLP (Chair)
Mr. Arora and Ms. Bernier thanked members and invited them to share their final thoughts.
Advisory council action items and considerations
Action Items:
- Next meeting: Before April 2024.
- Find a replacement for Mary Dawson.
- Present the concrete results of the application of the various guiding principles related to the publication of disaggregated data at a future meeting.
Recommendations for Statistics Canada:
- Management of the continuum between privacy protection and the production of statistical data to inform Canadians
- Adopt guiding principles that can be adapted to new situations.
- Remain a model when it comes to quality data.
- Continue to pay particular attention to social acceptability.
- Continue to represent all Canadians, explain Statistics Canada approaches and objectives transparently, and remain consistent and predictable.
- Use of nudge to encourage participation in statistical programs
- Use nudges to encourage participation in statistical programs.
- Consider the use of experimentation to build knowledge.
- Do no harm when publishing data on more vulnerable populations
- Produce guiding principles, guidelines, and best practices to help users.
- Clarify the scope of the nothing about us without us principle and how Statistics Canada intends to manage its application.