Why is ensuring anonymity and voluntary participation important?

Navigating data protection principles when collecting JEDI data

• Ensuring voluntary participation and anonymity in practice
• Design for voluntary participation
• Establish the infrastructure and rules that protect anonymity 

Case Study 1: Small companies

Case Study 2: Large companies

Under the Justice, Equity, Diversity and Inclusion (JEDI) Impact Topic, companies with more than 10 workers are expected to collect and use data to shape their JEDI approach in the workplace. Smaller companies do this through discussions or surveys; larger companies also have to record quantitative worker‑related measurements. In the B Lab Standards V2, JEDI 1 is designed to create a clear link between data and action:

• JEDI 1 requires companies to gather feedback and data on diversity and inclusion in the workplace.

• JEDI 2 then asks companies to use that information to choose and implement concrete JEDI actions.

In other words, certification is not just about stating a commitment to justice, equity, diversity, and inclusion; it is about showing that you are using real workforce data to guide decisions, set priorities, and measure change.

• If you have more than 10 and fewer than 50 workers, you must have facilitated at least one discussion or survey about JEDI principles in the workplace in the twelve months before Year 0, and then do so again in each subsequent year (JEDI1.1.1)

• If you have 50 workers or more, you must collect gender identity or sex at birth related data for at least five worker‑related measurements (such as hiring, promotion, pay, or retention) and then break down (disaggregate) that data to report internally on the measurements you have chosen (JEDI1.1.3). Workers must provide this data voluntarily, and they must always have the option to remain anonymous (JEDI1.1.4). Read JEDI 1.1: How to track gender identity or sex at birth for more guidance. 

In addition to JEDI 1.1, JEDI 1.2 requires companies (when applicable) to collect data on one additional social identity (other than sex at birth or gender identity) for at least five worker-related measurements, based on worker feedback and context. Social identity refers to a person’s own sense of why they are, based on the social group or groups they belong to. Social identity terms vary by country, culture, and language and examples can include, but are not limited to, race, ethnicity, religion, caste, or sexual orientation. 

Worker-related measurements relate to all of the touchpoints and experiences across a worker’s journey with your company, often starting with recruitment and hiring practices.

Why is ensuring anonymity and voluntary participation important?

When you need to collect data on sex at birth, gender identity, or other social identities to meet the standard, this information should be gathered using methods that are not linked to individual employee records.The standards clarify that companies are not expected to hold sex at birth, gender identity, or additional social identity data in employment records and should take precautions where asking about certain identities is illegal or may put workers at risk (e.g. recording LGBTQIA+ identities in countries where same-sex relationships are criminalised). 

Across all the JEDI requirements, you must ensure that workers provide their data voluntarily and can choose to remain anonymous when doing so. Taking this approach means you:

• Demonstrate compliance with JEDI 1.1.4.

• Align with data protection laws.

• Prevent the identification of individual employees and protect workers from potential bias or retaliation.

• Foster the psychological safety workers need to provide honest, high‑quality data that can drive genuine impact and change.

Navigating data protection principles when collecting JEDI data

The JEDI requirements were created factoring in the core principles, ethical considerations, and local laws relating to data privacy. However, before you collect any JEDI-related data, you should still consider the specific legal and ethical principles, and applicable local laws to support worker’s rights and foster psychological safety in your company’s context. Regardless of where your company is based, these four pillars can help keep your process compliant and trust-based:

1. Clear purpose: Explicitly state that JEDI data will be used only to identify inequities and patterns in the workplace, and inform JEDI2 actions.

It must not be used for performance reviews, disciplinary procedures, or individual monitoring. This is essential both for trust and for aligning with the intent of JEDI1, which focused on using data “to inform its JEDI actions (JEDI 2)”

2. Data minimisation: Avoid collecting data for the sake of collecting data or “just in case” you might need it at a later point. Limit yourself to what is needed to comply with the relevant JEDI requirements and understand and address issues in your workplace.
   
   

3. Data storage and security:  You are not expected to store gender or social identity data in employment or personnel records; in many cases it is better to use separate JEDI-specific tools and storage. When possible, use external, third-party survey platforms with robust privacy and security features that do not automatically attach names, work emails, or IP addresses to responses. If you must use internal tools: 

   1. Implement strict rules on who can access the data;

   2. Limit raw data access to the smallest possible group responsible for high-level JEDI analysis;

   3. Ensure that direct managers and senior leaders never see individual-level responses, unless the responder has given their permission.

4. The right to opt out: For participation to be truly voluntary:

   1. Workers must be able to decline participation without penalty

   2. There should be no implicit or explicit pressure from managers or peers

   3. Participation must not be linked to performance evaluations, promotions, bonuses, or access to benefits. 

The standards explicitly frame participation as voluntary and anonymous – the way you communicate and design your process must make that a lived reality. 

Ensuring voluntary participation and anonymity in practice

This guide outlines practical steps you can take to meet the requirements of JEDI 1.1.4 for voluntary and anonymous participation in order to collect high-quality data within a framework of trust and psychological safety. 

1. Design for voluntary participation

An “optional” checkbox alone is not enough. To be truly voluntary, your whole process should minimize direct and indirect pressure. 

• Implement "Prefer Not to Say" and non-participation options 

  • Include “Prefer Not to Say” (or equivalent) for each sensitive question

  • Allow workers to skip specific questions or the entire survey/discussion without any explanation

  • Ensure workers can submit a survey even if they leave identity questions blank or select “Prefer Not to Say”

• Use clear, transparent communication

 In your invitations and reminders:

• Clearly state that participation is optional

• Explicitly confirm that anonymity is protected and workers can choose not to disclose social identities

• Explain what you are measuring, why, and how the data will be used (e.g. to inform JEDI2 actions) 

• Affirm that non-participation will have zero impact on employment, compensation, performance, or advancement 

• Decouple incentives from participation

  • Avoid tying individual or team-level rewards to completion rates (e.g. vouchers for completing the survey, or targets for managers to hit 80% participation)

  • If you use general incentives (e.g. a company-wide celebration), ensure they are not framed in a way that pressures specific teams or individuals to participate

• Build trust over time

 Psychological safety is built over multiple cycles.

• Expect lower participation in early years, especially on sensitive questions

• Document and share the JEDI2 actions you take as a result of the data, and link them clearly to what people told you

• Continue to collect data at least annually and show how both the insights and the process are improving over time.

2. Establish the infrastructure and rules that protect anonymity 

Anonymity is a technical and procedural reality, not just a promise in an email. You must prevent “deductive identification” - working out who someone is from a combination of traits or from small-group reporting. The standards explicitly warn against accidentally de‑anonymizing worker feedback.

• Use appropriate tools and platforms

 Use tools that:

• Do not automatically associate responses with company accounts, emails, or IP addresses (e.g. Google or Microsoft Forms)

• Allow anonymous mode or separation between login (for access) and response data (for analysis)

• Enable export of disaggregated but de-identified data into secure analysis environments. 

• Apply a "Rule of 5" (or similar threshold)

Introduce a minimum reporting threshold (commonly 5-10 respondents) for any breakdown you share internally. For example: 

• If there is only one woman in a department, you should not publish “Department X - women’s satisfaction score” as this could identify her.

• Instead, aggregate into a larger group (e.g. “all technical staff” or “all staff in country Y”). 

• Limit access to raw data

If you cannot rely entirely on external anonymised tools: 

• Assign a single role (e.g. a JEDI lead, HR specialist, or external consultant) to handle raw data

• This person cleans and prepares anonymised, disaggregated reports and dashboards for internal use

• Direct managers, executives and other colleagues see only aggregated results, never line-by-line responses. 

Once reports are finalised, store raw data securely or destroy it in line with your data retention policy, in keeping with broader expectations about safe handling of disaggregated JEDI data. 

Case Study 1: Small companies

Company profile: A boutique architecture firm with 40 workers.

Goal: To meet JEDI 1.1.1 and 1.1.2 with a qualitative discussion and baseline survey.

Challenge: With only 40 workers, deductive identification is a serious risk. For example, if the firm were to report "100% of non-binary staff feel unsupported" and there is only one non-binary employee, anonymity would be compromised.

Strategy: The boutique architecture firm took the following steps:

1. The discussion (JEDI 1.1.2): The firm held a 90-minute "JEDI Visioning" session during work hours. To keep participation voluntary, workers could opt-out and work on other tasks instead, without having to justify their choice. 

2. Anonymized note-taking: Instead of using digital notetaking tools, an external facilitator was engaged to take notes and prepare a thematic report. Instead of using names or identifiable quotes, the facilitator grouped feedback into high‑level themes, such as “concerns about adequacy of parental leave,” avoiding direct attributions that could reveal individuals in small teams.

3. The data (JEDI 1.1.3 - planning ahead): Although the firm is below the 50‑worker threshold for JEDI 1.1.3, it decided to pilot a simple anonymous survey to prepare for future growth and to gather baseline data. It used a free, anonymous survey tool that did not require company logins. Workers were asked about:

• Gender identity (with the drop-down options being “Man”, ‘Woman”, “Non-binary”, and  “Prefer Not to Say”)

• Perceptions of workplace culture and belonging

• Experiences of fairness in workload and recognition

4. Applying the "Rule of Five": Analysis showed that only three workers identified as “Non-binary”. Because this was under 5 people, the firm did not include this data in the reporting. In order to prevent identification by deduction, the firm only shared the overall % of men and women, without disclosing the specific number of men and women that the % corresponded to when reporting on “Gender Identity”.

Result: The firm identified caregiver support as a key barrier and used this insight to design a JEDI 2 action: a flexible core hours policy supporting different care arrangements. Because employees saw a concrete change that clearly linked to the discussion and survey, participation in the following year’s survey increased by 25 percentage points.

Case Study 2: Large companies

Company Profile: A global manufacturing company with 700 employees across 8 sites.

Goal: To meet JEDI 1.1.3 with a full quantitative disaggregation and analyse the data across five worker-related measurements.

Challenge: The company’s HR system  already records “sex at birth” and “date of birth” for payroll, but:

• This information does not necessarily reflect how workers currently identify, and

• Using it without clear, voluntary consent for JEDI purposes would not align with JEDI 1.1.4.

Strategy: The company took the following steps:

1. Employee Self-ID Campaign: The company launched an Employee Self‑Identification (Self‑ID) campaign:

• Workers were invited to log into a secure, external questionnaire to review or update their gender identity.

• Communications clarified that participation was voluntary and that workers could choose “Prefer Not to Say.”

• The purpose was described as equity auditing and learning, not individual monitoring, performance management, or disciplinary action.

2. Data capture and handling (JEDI 1.1.3): 

   1. Responses were exported into a secure, encrypted environment managed by a single JEDI Lead.

   2. Direct identifiers (names, employee numbers, email addresses) were removed before analysis.

   3. The JEDI Lead created disaggregated but anonymised datasets for analysis and reporting.

3. Disaggregating across five worker‑related measurements data: With 585 people participating, the company had enough data to disaggregate gender identity across five worker‑related measurements, such as

• Hiring rates: comparing gender ratio of applicants to final hires at company and site level

• Promotion time: analyzing average time to promotion for different gender identities

• Retention rates: identifying that women in manufacturing roles left 15% faster than men

• Compensation levels: conducting a basic raw pay gap analysis by gender

• Training: assessing gender representation in the “Fast Track to Leadership” programme

All reporting respected minimum‑group thresholds to avoid accidental identification in smaller sites or departments.

Result: The company discovered that while hiring was relatively balanced, women were significantly under‑represented among Floor Managers and in the leadership pipeline. It used these insights to design a JEDI 2 action plan that included:

• Sharing the disaggregated findings and methodology transparently with workers

• Publishing a time‑bound action plan addressing promotion and progression barriers

• Rolling out bias‑interruption training for HR Business Partners, managers, and promotion committee members

The company committed to repeating the Self‑ID campaign annually to monitor progress over time, in line with the standards’ emphasis on continuous improvement across impact topics.