How should data stewards support collective and participatory data governance practice?

The GovLab and new non-profit The Datatank have launched a conversation about the job specification for a re-imagined data stewardship role. They argue that ‘Data Stewards’ are a much-needed role in both public and private organisations, operating to maximise re-use of data in the public interest, and are inviting feedback on a revised Data Steward job specification.

In this post I look at the feedback that a collective and participatory data governance frame might offer.

The draft job spec covers a number of areas from ‘audit, assessment and governance’, through to ‘internal coordination and data operations’. It also includes a heading on ‘Partnership and community engagement’, where it suggests, amongst other things, that a data steward should:

• “Establish and operate a policy of transparency and active engagement to foster trust between the organization and data-producing customers and communities
• Establish procedures and engagement models in order to create and sustain an organization’s social license for data re-use
• Keep partners and beneficiaries informed and up-to-date on insights generated from data initiatives

And

• Help establish a “social license” for data re-use”

This hints towards a role for data stewards in supporting more participatory data governance, but does not make it particularly explicit, and it stops short of bringing collective impacts of data into view.

The Open Data Institute will shortly be releasing a definition of ‘Responsible Data Stewardship’, that may help here, bringing to the fore the idea that any work with data should have in mind ‘structural inequalities’ that affect how it acts in the world.

Applying a Connected by Data lens to the question of what a data steward should do suggests also the following shifts of emphasis:

• Establish and operate a policy of transparency, active engagement and participatory governance to foster trust between the organisation and customers and communities contributing to, or affected by, relevant data
• Establish procedures and engagement models in order to understand, develop and sustain an organisation’s social license for data re-use
• Keep partners and beneficiaries informed and up-to-date on insights generated from data initiatives, and keep the organisation up-to-date on community views around data re-use and impacts

These edits might be small, but they drive at a couple of key points:

1. Stewards should be part of the process of making sure affected communities have a voice in data governance. It is important to move beyond a focus on ‘data subjects’ alone, and to consider wider impacts of any data collection or use.
2. Any work to develop ‘social licence for’ (i.e. the public legitimacy of) data re-use should start not from assuming it can be created, but by understanding existing social attitudes and views, and starting from engagement with these - including recognising that this might involve adapting organisational plans. There is a risk that framing a role as ‘creating social license’ slants it towards PR campaigning, rather than participatory engagement.
3. The stewardship role functions as an interface between the organisation and external stakeholders, and so has a key role in supporting two-way communication, rather than just organisational broadcast updates.

Expanding the skill set

The edits above might also suggest some additional ‘stakeholder engagement’ skills that it is desirable for a data steward to have or develop, moving beyond the ‘human centred design’ and ‘design thinking’ skills mentioned in the current brief, to also think about consultation and participatory engagement practices.

In our evaluation of Justice Data Matters we explored the importance of understanding different forms of public engagement and deliberation, noting that while some digital-centred organisations are used to reaching for ‘user research’ type of methods, these are generally only able to answer the constrained question: “How can a system, process or platform be designed to work better for target users?”, but they leave aside other important questions such as “What do key stakeholders think about these plans?” or “To what extent are the public supportive or concerned about these ideas?”.

Just as a data steward may be involved in formulating or assessing the questions for which data is needed, an excellent data steward may also be able to formulate and assess engagement questions to ensure all affected stakeholders are inputting into data governance decisions, and to select the appropriate engagement methods that can answer those questions.

In our practice work in the year ahead we’ll be thinking more about the kinds of resources that can support data stewards, and other data governance professionals, to make these kinds of choices.