What Kinds of Data Literacies Do We Need in a Datafied Society?
Dr Lyndsay Grant, School of Education
The increasing datafication of everyday life is well documented. From state security agencies to Silicon Valley social media developers, applications of big data and artificial intelligence are hailed as bringing greater precision and efficiency to complex domains of social life. More recently, the potential of these new ways of ordering social life to create new risks and harms is also becoming more apparent, with a ‘techlash’ against increasing datafication seen in concerns over ‘fake news’, discriminatory algorithms, privacy and security breaches, and surveillance concerns.
The recent Data Power Conference highlighted various analyses of the limitations, risks and harms of a datafied society. But how can we – as citizens, researchers and educators – respond to this situation? In several discussions during the conference, educating citizens to be more ‘data literate’ was raised as a possible response.
These calls for data literacy echo the calls for digital literacy that were prominent in early 2000s as the scale and reach of the Web grew, which in turn drew heavily on earlier calls for media literacy from the mid-late 20th Century in response to the new mass media cultures emerging around TV and film.
What can this ‘history of literacies’ tell us about how the current calls for data literacy might work out? Is data literacy the answer we seek? What would it take for it to make a difference? We can explore some of the challenges and lessons from these previous literacies in three broad areas.
Making responsible personal choices
First, some models of data literacy education focus teaching the skills and knowledge that individuals need to make informed choices about data sharing. This echoes approaches to literacy education that focus on developing the skills of decoding and encoding texts. The idea behind this approach is that many people are operating with an information deficit around data sharing, and that more informed individuals will be able choose to opt out of data sharing to protect their privacy and security. As anyone who has ever sighed and clicked ‘Accept Cookies’ will recognise, informing yourself about how each different organisation intends to use your data is a significant task in itself. Dima Yarovinsky’s ‘I agree’ art installation brilliantly illustrates this, with print-outs of the terms and conditions of popular social media sites running to several metres long. At over 12,000 words, Snapchat’s Terms and Conditions are estimated to take over an hour to read in full. But, perhaps more importantly, framing this as an issue of personal informed choice overlooks the sometimes prohibitive social and economic costs of opting of out of data sharing. For example, social media can be a lifeline for marginalised people to access social support and health insurance companies offer discounts if you are willing to share your fitness data. Choosing to ‘opt out’ of data sharing may only be a realistic choice for people who are already privileged enough to do so. Even more significantly, framing data literacy as a matter of an individual’s knowledge, skills and choice fails to offer a route for understanding and critiquing the wider implications of a datafied society or for collective political responses such as regulation in the interests of the common good.
Schools responding from within a data-based system
Second, we might look at how calls for a kind of critical, empowering and transformative ‘digital literacy’ became translated into training pupils in basic computing skills and checking the authority of sources to identify ‘bias’ and authenticity. Within a constrained school curriculum, there were limited opportunities to develop the creative, critical, cultural and political aspects of digital literacy, in which learners might question not only the messages conveyed by digital media, but also their own cultural embeddedness and participation in digital networks. In such a situation, data literacy could easily be translated into a focus on developing the technical skills of interpreting numbers and statistics, and valuable though this is, it will not be sufficient to allow pupils to identify and critique the social effects of datafication. Moreover, when data literacy becomes about the ability to competently use numbers, teachers are also placed in a position in which they are not asked to critique data, but to increase their use of data in their own professional practice. We also have to consider how far it might be possible to critically question processes of the uses and effects of data from within a thoroughly datafied education system.
Weaponisation of critical media literacy
Third, explicitly critical approaches to literacy education have attempted to foreground the importance of questioning, resisting and using media discourses to resist oppression and further emancipatory projects. Emerging from an anti-authoritarian approach that aimed to empower people to resist state and mass media power and propaganda, this scepticism towards authority as a source of truth has, however, recently been appropriated by right-wing and populist politics. Claims of ‘fake news’ – initially appearing as a call for fact-checking online media – are now weaponised to the extent that almost any claim is easily dismissed as motivated, biased, or simply untrue. This is the era of post-truth politics, in which shared standards for what counts as truth no longer hold, and the emotional impact and reach of messages are more important than their content. The problem of responding to these trends with fact-checking or further calls for critical media literacy, as danah boyd argues, is that it can be seen as an establishment attempt to re-assert control over what counts as the ‘truth’ in a context where people have been encouraged to be sceptical of all such epistemological authoritarianism. Similar patterns can be seen with data. Following the divergence of the 2015 UK General Election from opinion polling data, data-based claims are now easily dismissed as either biased or simply irrelevant. Yet calling on ‘data literacy’ to re-establish the authority of hard data over what counts as the real facts overlooks the need to interrogate how that data was produced and mobilised to create particular meanings in the first place.
Where do we go from here?
So, given the past histories of critical, digital and media literacies, has the idea of data literacy failed from the beginning? Or can we conceive of a kind of critical data literacy that doesn’t reinstate the objectivity of numbers as an ultimate authority on truth, but which also does not caricature data as yet another weapon in the post-truth armoury? Can we develop a data literacy that not only equips individuals to be skilful users and informed consumers of data, but in which we also investigate how a datafied society is being constructed, and develop collective responses? Can we imagine ways of teaching new kinds of data literacy from within datafied institutions such as schools, working from our position within the data machine?
Some possibilities for ways forward might be found by drawing on approaches and insights from the fields of New Literacies Studies alongside posthuman and performative theoretical approaches. Emerging from an anthropological approach, New Literacy Studies sees the reading and writing of texts as a cultural and social practice, always emerging from and embedded within particular groups and communities. Expanding from an initial focus on everyday reading and writing practices, the New Literacies Studies developed to encompass the literacy practices associated with participation in digital spaces including gaming and social media. The emphasis on literacy as a fundamentally social practice may help us think more about data literacy not as the responsibility of individuals to acquire and deploy a specific technical set of skills, but as emerging within new kinds of social spaces. By situating literacy in specific sociocultural contexts, it also opens the way to thinking about how plural data literacies work to shape and constrain the uses and effects of data in different contexts.
Alongside the New Literacies Studies, posthuman, new materialist, performative theoretical approaches offer potentially productive directions for thinking about data literacies. These see knowledge and being as emerging in the relation of associations between a diverse array of human and non-human entities and practices. From this point of view, data practices do not just more or less accurately reflect the world, but are part of the dynamic and ongoing production of the world itself, allowing us to ask how processes of datafication bring new social and material worlds into being. These approaches, and related methodologies from the related field of Science and Technology Studies suggest a direction for data literacies in which we do not just ask whether data messages are truthful, but to explore how they come to create new facts, and how they come to matter in the world. They also allow us to think about what data literacies might look like from a our starting point within the data machine, accounting for the parts we ourselves play in processes of datafication.
This is not a new manifesto for one particular approach to data literacies, but offered as a way towards a critical, but also productive and creative approach towards data literacies that may move beyond some of the challenges encountered with previous developments around critical, digital and media literacies. These approaches might open the way towards a kind of data literacy that does not just ask whether or not we can ‘do’ data in terms of personal skills and choices, but creatively explores what data ‘does’, how it works, and how we are part of it.