In our latest blog, Head of Development and Impact Helen Miller-Bakewell, and Director General for Regulation Ed Humpherson discuss the state and sociotechnical elements of data sharing and linkage across government…
We published our report on data sharing and linkage in July. The report highlighted that data sharing and linkage in government is at a crossroads. Not all the routes from the crossroads lead to positive outcomes. The report set out 16 recommendations to ensure that the positive path – data sharing for the public good – is that one that is taken.
In highlighting this position, we were not taking a purely technical approach. We were not looking just at the technical and legal underpinnings of data sharing and linkage and assessing progress against them. We were not just looking at data standards and architecture.
For sure, technical and legal issues do appear in the report, and getting them right is a key enabler.
But the crossroads analogy comes from the broader perspective the report brings to data sharing and linkage. In the report, we consider four possible ‘future scenarios’ for data sharing and linkage, set five years from now. They are not predictions but stylised versions of possible futures, which help to bring out the impact on public good of acting on (or not acting on) the current barriers that exist to data sharing and linkage.
Because we wanted to look at data sharing and linkage not just in a technical way but in its broader social context, each scenario explicitly considers public understanding and buy-in to the benefits of data being shared and linked. And our future scenarios feature personas – people who work with data inside and outside government – and how the different scenarios impact on them and the things they are interested in.
In doing so, we consciously adopted a sociotechnical approach. It has helped us to demonstrate that public engagement and social licence are intrinsically linked to the ability for linked data to serve the public good. Without public support, data are less likely to fulfil their potential.
This term – sociotechnical – first came up in our work through the Ada Lovelace Institute. When we published our report on exam algorithms, Carly Kind of the ALI tweeted that she liked OSR’s ‘sociotechnical’ approach. At that time, it was not something we incorporated into our work.
However, being generally curious by nature, in OSR we started to wonder what lay behind the idea of behind sociotechnical.
We learnt that the sociotechnical denotes an interest in how technology, and technical issues more broadly, both influence and then in turn are influenced by how people behave, respond and think about them. I picture it as a sort of dance between the social – the human and community aspect of something – and the technical – which is the abstraction of something into a measurement or calculation structure of some kind.
The more we looked into this, the more we realise that we have always worked in a sociotechnical way. We care about how people use statistics – that is what confers value. We care about the trustworthiness of the organisational processes, which is about human issues like the role of the head of profession. And we care about quality, not in an abstract computational sense, but whether the statistics provide users with a good estimate of what they are aiming to measure.
For the data sharing and linkage work, we established a sociotechnical panel of advisors – people who we’ve met or worked with who have a good handle in this interface of the technical and the social. They were They were Nick Pearce, Brian Rappert, Rachel Coldicutt, Jessica Davies and Sabina Leonelli.
The panel was invaluable. They challenged the way we thought about public engagement. Our early thinking was too fixed. We talked of “engaging with the public”, as if there was a single public and a single best way to track public approval. The panel reminded us that the public is not homogenous and counselled us to advocate for public engagement that is targeted, based on the intended outcomes of projects and the demographics of the population that are most likely to be affected. And they helped expand our thinking on the connection between public willingness to share data and their trust in different types of organisation. The more subtle focus on the ever-evolving, context-dependent social licence for data sharing in the report came from the advice of this panel.
Finally, it is not just social context beyond government that can impact data sharing and linkage. Culture and people within government can be key determinants of progress. During our review we heard that, at every step of the pathway to share and link data, the people involved are instrumental to determining whether projects succeed or fail. We heard examples of departmental barriers becoming unblocked when new people arrive showing how many can be overcome simply with a new motivation, knowledge or skill.
Reflecting on the report, we set out to write about data sharing and linkage – a potentially dry and technical topic – and it ended up being more about people than technical architecture. And, in fact, we find this is a common pattern across almost all our work.
The final report, then, is an important diagnosis of the state data sharing and linkage in government. It is also the fullest expression yet of the sociotechnical angle to our work.