Financial sustainability, ethical governance, and defining “data” in the context of data cooperatives

This is the sixth post for the Community Data Hubs Documentation series. This series will document the thought and conversation trajectories within the process of creating the building blocks of our Community Data Hubs model and OEDP’s broader data stewardship work. The first of these blogs will document the progress of the Community Data Hubs Advisory Group, which is working alongside OEDP to tackle conceptual questions related to the model, including social and technical infrastructures, stewardship, and community data.

This post documents the fourth meeting of the Community Data Hubs Advisory Group on June 28, 2023. For this meeting, we asked members of the advisory group to read two brief accounts (this one and this one) of a data cooperative in Zurich, Switzerland. “Data producers” donate their mobility data to the co-op; through a democratic process, the co-op then decides to share or sell access to an anonymized version of that data to interested government agencies, universities, or private firms. Considering the virtues and pitfalls of this data cooperative served as a jumping off point for refining our emerging model of a Community Data Hub.

Our conversation centered on several major questions:

1. How to assure financial sustainability while adhering to ethical and democratic governance structures?
2. What does “data” really mean — in this and other contexts?
3. What are the differing timelines of researchers and community members, and how can partnerships endure in spite of these differences?
4. How can we make room for flexibility and reflection within our data stewardship models?

How to assure financial sustainability while adhering to ethical and democratic governance structures?

Advisory group members identified a fundamental tension that all data stewards must negotiate: the imperatives of ethical data governance can be out of step with the process of producing a profitable data “commodity.” Group members found various components of the co-op’s governance mechanisms inspiring — especially the concept of the elected ethics council — though they wondered about the reproducibility and financial sustainability of such a model. If selling access to democratically governed, anonymized data constitutes the primary income for the co-op, how often would the democratic process yield a dataset with an interested buyer? To achieve financial stability, group members indicated that cooperatives in the US might seek out additional sources of funding from government or private foundations, especially if they seek to compensate co-op members, as they hope to in Zurich.

One group member noted that the cultural context is important for understanding the data co-op’s applicability in other settings: Switzerland’s economy has a long history with cooperatives, making it a socio-economic norm. Swiss data cooperatives can build off of this baseline familiarity with the day-to-day operation of cooperatives, whereas a similar program in the US might need to educate its members both on data stewardship and cooperative governance.

Another group member worried about the demands this model of governance placed on co-op members, and wondered how challenging it might be to fill the elected administrative roles. This surfaced another tension: data projects have to create institutional procedures that allow space for slow, messy democratic debate, without making unsustainable demands on people’s time.

What does “data” really mean — in this and other contexts?

Group members commended the data cooperative for both its clear parameters regarding the data it planned to accumulate and its explicit articulation of the audiences or “data consumers” to whom that data might be relevant. Group members noted, though, that many projects often have (much) less clarity about both the relevant data and its final audience; that the cooperative model might be strained by the complexity these ambiguities introduce. Data, as group members noted, can take many forms. Qualitative data like oral histories — ”grandma’s stories” — present an array of challenges, not least of which being that community members sometimes resent the use of the term “data” to refer to these highly personal kinds of information. What would democratic stewardship look like if oral histories were the data in question?

Group members indicated that environmental projects need to consider, and consistently revisit, the question of what kind of data they wish to collect and why. Group members offered two examples of different models here: 1) an intra-community inquiry focused on sense-making or a process of internal self-discovery, and 2) an outward-facing effort to document well-known phenomenon for community members to an outside, perhaps skeptical, audience (e.g.,policymakers). What data looks like, and thus the appropriate data governance procedures, may be quite different in each case.

What are the differing timelines of researchers and community members, and how can their partnerships endure in spite of these differences?

Group members identified several tensions that tend to emerge between researchers and community members in the context of environmental research. One especially important tension is temporality: different incentives structures shape the timelines and behavior of researchers and community members.

One group member noted that researchers must respond to an incentive to “publish or perish,” an emphasis on speed that contradicts the “slowness” required for democratic and reflective practices within cooperative data stewardship models. Another group member noted that the tension is also often reversed: whereas community members are sometimes making urgent demands to ameliorate an injustice of some kind, it often takes years for academics to slowly accumulate data and move through the peer review process before publishing an authoritative account.

One group member suggested that the best way to navigate these tensions around temporality is transparency. This, obviously, does not resolve the problem of the differing timelines or incentives at play between communities and researchers, but clear communication can create an atmosphere of understanding that allows partnerships to negotiate their differences and endure despite them.

How can we make room for flexibility and reflection within our data stewardship models?

Group members consistently emphasized the importance of flexibility in data stewardship. The process of doing research is transformative, and so data governance must respond to the emergent nature of the research process. As we learn, we may wish to accumulate new data, or recruit a new audience.

One group member suggested CDHs should have built-in systems for generating reflection and prompting strategic reorientation when necessary. This might look like structured, regular meetings that provide space to consider broader questions. Or, as another group member suggested, perhaps a CDH might become a known, welcoming space that invites participation, not unlike a library.

Another group member noted that data sharing is key to building flexibility across time. If data is publicly available, later researchers or community members can make use of it, even if their ambitions differ significantly from that of the original researchers. One group member noted, for instance, that they found oral histories performed by past researchers in their community extremely useful, even though the researchers asked questions that did not pertain to their inquiry. Open data creates more opportunities for flexibility, including unanticipated inquiries down the road.

Insights and questions to revisit during the CDH Model Co-Design Process:

1. Ethical governance must exist alongside financial sustainability. The CDH model cannot recommend governance structures that will either overburden participants or conflict with sound financial stewardship.
2. The CDH model could guide potential partners through a deliberative process where they articulate their goals, the associated data they wish to collect, and its proposed audience. Sound recommendations regarding data governance procedures depend upon communities and researchers coming to clear agreements regarding these questions at the onset and in an iterative manner.
3. Transparency must be a central value among the partners working through a CDH, such that various parties understand the reasons and incentive structures for whatever differences may emerge between them over the course of their work. How can we incentivize transparency within collaborative environmental research projects?
4. CDHs will need built-in systems for prompting reflection and encouraging flexibility. As the research process generates new information, and as community goals shift, so too must data governance procedures. How can we balance building enduring, consistent data infrastructures with this imperative to change and adapt alongside shifting circumstances?