An encouragement for the participatory design of data governance policies

How data is managed has both technical and societal consequences. Data-based technologies determine which content people see, who gains access to services and who does not. Data governance policies set out how the data on which these technologies are based may be handled. In doing so, they substantially influence how fairly or unfairly digital spaces can be designed. Participatory design of these policies is one way of developing rules for handling that data together with the people who are represented in it or affected by its use. This concerns questions such as: which data is collected – and which is explicitly not? Who has access to the data, and under what conditions? Or: when is data deleted? Such questions are relevant in almost every project involving personal data. In the planning of mobility services, for example, it is relevant whether and how movement data is collected, how this data is subsequently anonymised and pooled, and whether third parties or the public can gain access to it – and for what purpose.

In the course of digitalisation, data is often regarded as a new commodity, whereas for research data there are established processes for handling personal data. Institutional review boards, for example, decide on data access or review research designs while involving a range of perspectives.^[1][2] The participatory design of policy documents that answer questions such as those set out above is still not widespread. Medical research – which deals with particularly sensitive data – is, however, a step ahead in this area too. It offers a view of how the interests of those affected can be taken into account in such processes. At the same time, the case examples given below indicate that participatory design of data governance can also increase a project's functionality. This may be because projects are better accepted by users, by those affected and by the public, because projects become more secure, or because the insights gained allow data to be used more effectively.

For a national data collection on emergency medical care in the United Kingdom, for example, principles for a data access protocol were developed in workshops with representatives of the public and with patients. The protocol governs access to the data collection and, as one outcome of the workshop, provides for a Data Trust Committee that advises on specific applications for data access. Among other things, the protocol that was developed sets out the composition of the committee and determines how members are to be recruited. Even though the body has no vote, it is established within the project as an advisory voice of the public.^[3]

A more far-reaching example can be found in a tuberculosis monitoring project in Canada. Here, legally binding data governance agreements were developed with Indigenous communities in a long and highly detailed process in order to safeguard their interests and their sovereignty. Among other things, this made it possible to align data availability and data ownership more closely with the communities' needs. Beyond this, the project is not to generate any insights on the basis of the data that do not also result in insights for the communities themselves. In addition, it also became possible to jointly investigate research questions relevant to the communities. This also improves the project's underlying data, since monitoring data was no longer merely collected passively and one-sidedly, but actively returned to the communities – enabling them to make better sense of local outbreaks and, in addition, to derive their own measures.^[4]

On the one hand, both examples show that participation processes are often deployed where long-entrenched power asymmetries exist; on the other, they also show how participation can be designed in such a way that it counteracts precisely these asymmetries. For participation to become effective, it is essential to state clearly what can be jointly decided, and to make visible how the results of participatory processes feed into projects.^[5][6] For participatory processes with a diffuse problem orientation are often costly for those responsible for the project and frustrating, or even misleading, for participants.^[7][8] Participatory policy design creates clear points of reference here: the needs of those affected are negotiated in relation to concrete topics such as data access, data collection, aggregation or retention periods, and are moreover translated into rules. The needs of those affected are at the forefront here, but at the same time there is an opportunity to improve a project's acceptance, security and functionality as well.

In fields oriented towards the common good in particular – from the planning of public transport through health services to urban and housing development – the experience gained from participatory policy design should be used, adapted and developed further. For every participatory process for which it cannot be clearly stated what difference it will make to the project risks losing its effect. And every project that does not seek productive exchange with those it affects forgoes the chance of a better outcome and greater acceptance.