The edited book ‘Good Data’ is a mammoth twenty-chapter collection of new writing on digital data and its consequences. We will be unpacking individual chapters over the next few weeks. To kick-off our reading, here are some initial notes and thoughts arising from the introductory chapter by Monique Mann and colleagues.
The prospect of ‘Good Data’ prompts some obvious questions – notably how data might be deemed ‘Good’ or ‘Bad’. Sensing this, Mann quickly acknowledges that ‘Good’ is a subjective descriptor. For example, from a technical viewpoint it might be argued that good data should be usable, fit for purpose, easily accessible and revisable. From an ethical viewpoint, these issues might extend into the ‘respectful’ collection of data – that is, collected with respect for human dignity and human rights, as well as the rights of the natural world. These issues might also extend into the social benefits of what can be done with this data – for example, data that allows people to “form useful social capital” (p.12).
Mann and colleagues are upfront about their own western philosophical perspective that good data work implies a concern with increased wellbeing, sustainability, fairness and justice. Conversely, then, ‘Bad Data’ can be seen as anything that lacks these qualities. As Mann continues, ‘bad’ data practices therefore:
“… include the mass gathering of data about individuals, in opaque, unethical and at times illegal ways, and the increased use of that data in unaccountable and discriminatory forms of algorithmic decision-making” (p.8).
These distinctions offer a helpful starting-point for judging the nature of the digital data that is being generated and/or being made use of within a school. They also point toward particular qualities that we might wish to see developed in new forms of data. For example, how might people who are having data collected about them have more of a say in their own visions for how this data is then used? How might “data discourses from underrepresented, disenfranchised and disempowered voices” (p.10) be prioritised?
As such questions suggest, the ‘Good Data’ approach chimes with concurrent concerns over ‘data ethics’, ‘data rights’ and ‘data justice’. Unlike some facets of critical data studies, however, the emphasis here is place less on critique and more on optimistic explorations of possible better uses of data. This is work that is less interested in detailing the current data ‘harms’, as it is in striving for positive alternatives and a hopeful vision of the datafied future. As Mann reasons,
“how could data and digital technologies be designed and used in ‘good ways’, for ‘good’ purposes? If digitisation and data are inevitabilities, then we have to (re)imagine the kind of digitised world and data we want to see rather than only offering a naysaying critique of the status quo” (p.9)
In elaborating on these foundations, some interesting suggestions and challenges are brought to the fore. We hope that these might prove useful for our own interventions into school data. For example:
- How might we work with students and staff co-produce design concepts of systems that improve personal control of data?
- How might data governance within a school be conducted along communal, cooperative lines?
- How might students, teachers and parents engage in the collection and interpretation of ‘alternative’ data-sets “to impose political pressure for social ends” (p.12)?
- How might the the “strong performative qualities” (p.18) of data visualisations offer a medium for students, teachers and parents to proactively shape the environments in which they live in?
- How might we support and/or agitate students, teachers and parents to protest against ‘Bad Data’ practices?
Mann’s opening chapter also raises some interesting caveats and possible limitations to these ambitions that are discussed in later chapters of the book. First, is the observation that digital data is quickly evolving in ways that complicate and compromise any attempt to reappropriate it for public good. For example, Flintham and colleagues draw attention to the co-constructed nature of much of the digital data that we might find in schools. This is not so much ‘personal data’ as ‘inter-personal data’ – in other words, data relating to many different people and their interactions. Think, for example, of the data generated from a shared laptop or a room occupancy sensor. Such ‘group data’ is difficult to disaggregate and confer individual control over. Moreover, this data “is drawn from, and carries consequences for, the relationships between intimate groups” (p.19). We clearly need to tread carefully in working out alternate uses of these types of data.
Secondly, Mann reminds us of the compromises implicit in reshaping data practices along lines of social justice, fairness and so on. These are not ‘win/win’ changes. Instead, adjustments that might be ‘good’ for some people who are hitherto marginalised in established data processes, might then not be good for others. For example, moves toward increased sharing and openness of datasets will lead to losses in terms of data protection or data privacy. Similarly, adopting particular definitions of ‘fairness’ (for example re-designing a predictive model in order to reduce the discriminatory impact of that model towards minority groups) is likely to come with a ‘trade-off’ of reduced predictive accuracy.
All told, this collection of ‘Good Data’ chapters offer an interesting take on issues of datafication. At its heart is an interest in subverting the promise to ‘Do No Harm’ that one often sees in computational and data science (most infamously the Google maxim of ‘Don’t Be Evil). Instead, this book explores the maxim to ‘Do Good with Data’. This is undoubtedly a complex ambition. Amidst the realities of any school setting, this book is most likely to provide us with a series of exhortative aims and provocative inspiration rather than easily achievable objectives. Nevertheless, the intention of ‘Doing Good’ is a much better starting-point than most others!
notes from: Monique Mann, S. Kate Devitt And Angela Daly (2019) What Is (In) Good Data? In Angela Daly, Kate Devitt And Monique Mann (eds) ‘Good Data’ Amsterdam, Theory on Demand (pp.8-23)