Compared to those in other sectors, civil society’s data initiatives:
- Are often multi-team collaborative efforts with a complexity of incentives and power-dynamics
- These are each different groups: the “user”, the “data subject”, the “stakeholder”.
- The purpose or role of data in the project may be fundamentally different
- Data sources must be treated as imperfect representations of reality
- Have different funding and sustainability models than we see in the private and public sectors.
Having acknowledged that we should think carefully about the use of data in civil society, we might ask, well, is this just a matter of directing our collective focus? Can we simply do what we have been doing elsewhere, now with civil society organizations? Can we not utilize, with a few adjustments, those tools, skills, and methods that we created for such initiatives in other problem-spaces?
I’m talking about the technical infrastructures, the agile methodologies and the coding practices. The product mindset, the user-centered design, the way we structure our teams. The funding ideas will be different, that’s clear, and some legal frameworks, but that feels not so relevant to the data problems themselves.
In fact, there are really important differences between a civil society data initiatives and those at a for-profit organization, or one run by the government. And these differences require that we evaluate carefully any direct application of familiar tools or methods, and that, in fact, we may need to build new approaches entirely.
And this understanding – that civil society is different, that the problems it tries to solve are different, that the values and principles we need to carry in its work are different, and that together we have to really identify these differences and plan for them – this undertanding allows us to see the whole sector with new perspective, one that is not just wary of pitfalls but is also full of fresh ideas and opportunities.
This understanding allows us to see the whole sector with new perspective, one that is not just wary of pitfalls, but also full of fresh ideas and opportunities and a renewed promise.
I have described a few of the differences below, with examples, to illustrate how these differences can impact the design, the team composition, the sustainability, and the impact of the tool or product.
Civil society data initiatives are often multi-team collaborative efforts with a complexity of incentives and power-dynamics.
One US county that we worked with was aware that understanding recidivism requires bringing together information from many organizations. An individual’s interactions with the justice system over time may include the Sheriff’s office, the courts, prison, the probation office, community based organizations (CBOs) and more. Whether or not they recidivate, they may encounter many of these multiple times.
These are all organizations that are part of the same system, the justice system, and they frequently interact with one another. They even have common goals – such as to reduce recidivism. So in a broad sense, there is a shared incentive to bring their data together So in a broad sense, there is a shared incentive to bring their data together.
And yet, there are good reasons why they may hesitate from cooperating. These organizations often have conflicting incentives as well. A public defender can be described as both coordinating with and at odds with the district attorney. A community based organization may have information about a program participant that they would not want to share with their probation officer.
The county justice system as a whole may benefit from better understanding how individuals traverse the justice system, how to improve interventions, to see how recidivism rates have changed over time and what programs lead to improved outcomes. Pooling their siloed data may help make better decisions about these programs, and might help identify problems in the system that could be addressed.
But who would get all that data? How will it be used, and who will decide how to use it? How can we ensure that such a system would be used for the benefit of the community, and not against it?
- Unequal power dynamics
- Alignining on objectives
The user, the data principal and the stakeholders
Consider these definitions:
- The user is the individual interacting with the tool or product, for whom a user interface has been created.
- The data principal or subject is the persons or entities that are described by the data, or who are directly affected by the decisions and actions enabled by the tool.
- The stakeholders are many, but include the funder and other vested organizations.
For a data product at, say, NetFlix, we might say that: I am the user, since I select and watch the movie; I am the data principal, since data is collected about my preferences and those of other users; and I am the funder, since I pay the monthly subscription to the NetFlix company.
In the recidivism example above, these entities are not the same. Although this happens in all sectors, the nature of work in civil society makes this common or default case. And for these organizations, for instance NGOs working with refugees or journalists investigating corrpution, this dynamic leads to threat models, governance challenges, and design patterns that are more important and challenging.
“User-centered design” and other approaches in a product mindset must be applied carefully or rebuilt entirely in such contexts.
Rethinking the role of data
The purpose or role of data in the project may be fundamentally different
Data, truth and reality
Data sources must be treated as imperfect representations of reality
Funding, sustainability, and the project lifecycle
Have different funding and sustainability models than we see in the private and public sectors.
Together, these kinds of differences present opportunities and challenges for the use of data that are unique to the sector, not often seen in profit-motivated corporations, private ventures, or government offices. These present good reason to reconsider appropriating approaches and tools from other contexts, and to assume that they will be just as successful here.Instead, civil society must address these needs by building internal capacity for data initiatives.
Nodes on the graph represent articles; try clicking on them. Experimental. May not work on mobile.