Audience
- Builders: Understand how insights from real-world stakeholders inform ethical system design.
- Funders: Gain confidence in the empirical grounding of the framework presented.
- Partners: Recognize how diverse perspectives contribute to shaping ethical data practices.
Overview
This playbook is grounded in qualitative research conducted within the Epiverse ecosystem. The methodology was designed to capture how ethical principles are understood, interpreted, and applied across different roles and contexts within a complex data environment.
Rather than evaluating compliance against predefined standards, the approach focuses on how ethics operates in practice. It examines how responsibility is distributed, how decisions are made, and how principles are translated into real-world actions.
Ethics in data systems is not only about rules—it is about how people interpret, negotiate, and apply those rules in practice.”
Project perspective
The study is based on semi-structured interviews with 18 participants across the Epiverse ecosystem. This method was selected to balance structure and flexibility: ensuring consistency across interviews while allowing participants to share detailed insights from their own experience.
The interview format enabled participants to:
- reflect on how ethical principles are applied in practice
- provide concrete examples from their work
- identify challenges, trade-offs, and uncertainties
- suggest areas for improvement
This approach ensured that the findings are grounded in real-world practice rather than abstract assumptions.
Participant Profiles and Stakeholder Groups
Participants represented a diverse range of roles within the Epiverse ecosystem, including:
- data scientists and developers
- program managers and capacity-building leads
- institutional and governance actors
- fellows and community practitioners
To support comparative analysis, participants were organized into three stakeholder groups:
- Group 1: Institutional and governance actors
- Group 2: Technical and development contributors
- Group 3: Community-level practitioners and fellows
This grouping enabled the analysis to capture differences in perspective across levels of responsibility, expertise, and proximity to implementation.
Why multiple stakeholder perspectives matter
- Ethical responsibility in data systems is distributed across actors, not centralized.
- Different roles interpret principles in different ways depending on context.
- Tensions emerge between values and implementation across levels.
- Understanding these differences is essential to building coherent systems.
Data was collected through open-ended, principle-based questions aligned with the six ethical principles. Each interview explored how participants understood and applied these principles within their specific roles.
Participants were encouraged to:
- describe real scenarios and use cases
- reflect on risks and ethical dilemmas
- explain how decisions are made in practice
- identify gaps between intention and implementation
This ensured that the data captured both conceptual understanding and operational experience.
Analytical Approach
The analysis followed a comparative and thematic approach, focusing on identifying patterns across stakeholder groups.
Three key dimensions guided the analysis:
- Interpretation: how principles are understood
- Operationalization: how principles are applied
- Tensions: where conflicts or trade-offs emerge
Rather than analyzing each principle in isolation, the study explored how they interact within a broader system. This allowed for the reconstruction of the six principles as an integrated ethical architecture.
From Empirical Insights to Framework
The transition from research findings to the framework presented in this playbook was guided by three core insights:
- Ethical responsibility is distributed, not centralized.
- Principles operate as interconnected layers, not standalone categories.
- Tensions between principles are structural and persistent, not anomalies.
These insights informed the layered model introduced in Chapter 1 and provide the foundation for the analysis in subsequent chapters.
Limitations
As a qualitative study, this research reflects the perspectives of participants within the Epiverse ecosystem and may not capture all possible viewpoints.
In addition:
- findings are context-dependent and may evolve over time
- results are not statistically generalizable
- implementation may vary across institutional and regional settings
However, the strength of this approach lies in its ability to provide deep, context-rich insights into how ethics operate in real-world systems.
