The Datafication of Primary and Early Years Education

Playing with Numbers

Alice Bradbury, Guy Roberts-Holmes

• Publisher: Routledge
• Year: 2017
• Print ISBN: 9781138242159
• eBook ISBN: 9781315279039
• Topic: Education
• Subtopic: Education General

The Datafication of Primary and Early Years Education explores and critically analyses the growing dominance of data in schools and early childhood education settings. Recognising the shift in practice and priorities towards the production and analysis of attainment data that are compared locally, nationally and internationally, this important book explores the role and impact of digital data in the 'data-obsessed' school. Through insightful case studies the book critiques policy priorities which facilitate and demand the use of attainment data, within a neoliberal education system which is already heavily focused on assessment and accountability. Using an approach influenced by policy sociology and post-foundational frameworks, the book considers how data are productive of data-driven teacher and child subjectivities. The text explores how data have become an important part of making teachers' work visible within systems which are both disciplinary and controlling, while often reducing the complexity of children's learning to single numbers.

Key ideas covered include:

• The impact of data on the individual teacher and their pedagogical practice, particularly in play-based early years classrooms
• The problems of collecting data through assessment of young children
• How schools respond to increased pressure to produce the 'right' data – or how they 'play with numbers'
• How data affect children and teachers' identities
• International governance and data comparison, including international comparison of young children's attainment
• Private sector involvement in data processing and analysis

The Datafication of Primary and Early Years Education offers a unique insight into the links between data, policy and practice and is a crucial read for all interested in the ways in which data are affecting teachers, practitioners and children.

1
Introduction

The state, data and primary education

Introduction: why research the use of data in schools?

This book begins with two simple questions: how important are data in primary schools and early years settings? And what is the impact of data on these educational sites? The two research studies we discuss here explored assessment policy and practices in early years settings, and one specific policy – Baseline Assessment – which was an assessment for children aged four and five. Through both of these projects, we identified the increasing importance of data in the every-day lives of those working in primary schools and early years settings – a phenomenon we (and others) call ‘datafication’ (Roberts-Holmes 2015; Williamson 2016b; Lingard, Martino and Rezai-Rashti 2013). By this we mean a change both in the classroom, where data collection drives pedagogy and dominates workloads, and in school management, where inspection is organised around a school’s data. In this latter vein, the basis for Baseline Assessment itself is a reliance on numerical data as a method of judging schools. We also use ‘datafication’ to describe the changes in subjectivity wrought by the current obsession with data, particularly the changing role and status of teachers in the data-driven school and the way in which children come to be ‘made up’ through data.¹ Thus we explore, in all senses, what Ball has called ‘the tyranny of numbers’ (2015). Our subtitle ‘playing with numbers’ serves as a reminder that we discuss largely play-based contexts where numerical data are collected, and that numbers are something that can be manipulated.

This work is driven by the research data we collected in England, but also reflects a wider concern with the use of data in education from academics and practitioners internationally (Selwyn 2016b; Lawn 2013; Ozga et al. 2011).² Increasingly, issues relating to data have a huge impact on education practices, yet this remains an under-researched area. Thus, we respond here to Selwyn’s call to ‘make visible the flow and circulation of data and begin to understand the ways in which data are then integrated back into everyday education practices’ (2015, 76). This interest in data in education in turn reflects a wider interest in the impact of ‘big data’ in society and culture more broadly (Bowker and Star 1999; Kitchin 2014; Eynon 2013). Much of this work claims that digital technology is central to the functioning and values of society: as Beer comments, ‘Questions of data, metrics, analytics and number in the cultural sphere are not marginal, rather they are central to contemporary cultural formations’ (2013, 11). This change is based on the technological innovation of the advent of computing power; data are now cheaply produced, collected and analysed in contexts where in the past they were expensive, leading to claims of an ‘industrial revolution in data’ (Hellerstein cited in Manovich 2012). This has resulted in overwhelming volumes of digital data: ‘the production of data is increasingly becoming a deluge; a wide, deep torrent of timely, varied, resolute and relational data that are relatively low in cost and, outside of business, increasingly open and accessible’ (Kitchin 2014, xv). This volume of data is largely due to increased computing power, which has meant ‘we no longer have to choose between data size and depth’ (Manovich 2012, 3); thus the age of either in-depth studies or mass surveys is over. These developments have resulted in new fields of academic study focused on the impact of data as a productive force:

This wave of interest in the constitutive power of data, software and code is increasing as it becomes clear that data are, in a number of ways, central to the make-up of contemporary social formations of different types. Despite this excellent emergent body of literature, we are still only at the foothills in our critical analysis of the role of data in culture and society.

(Beer 2015, 3)

As in many fields, the educational research world has only recently begun to consider the impact of data, and in many cases this has been limited to consideration of governance, known as ‘governing through data’ (Ozga et al. 2011). In this text, we aim to redress this imbalance by adding a significant contribution to the limited research focused on school and classroom practices (Finn 2016; Pratt 2016).

The context for this study is the primary education system (age 4–11) and early years settings (age 2–4) in England; education in other areas of the United Kingdom is governed by devolved policies. This is pertinent as England is an extreme example which demonstrates many concerns voiced in the international literature; the education system is described as ‘the most “advanced” in Europe in terms of data production and use’ (Ozga 2009, 149; see also Silliman 2015). Since this was written, the reliance on data has only increased with policy changes, so that the issue has attracted media attention with headlines such as ‘Teachers “worn out by demand for data”’ (Metro 2017). An article by a school inspector in the Times Educational Supplement commented that the guidance for Ofsted inspectors on analysing data for primary schools was 50 pages long.

Think about it: how have we reached a situation where the data associated with primary assessment need 50 pages of explanation to professionals involved in inspection? What does this say about our current data-obsessed assessment system? It says that it’s far too complicated; far too dependent on numerical data which have a spurious air of precision, reliability and validity; far too impenetrable; far too far from the judgments that teachers need to make about real children’s progress towards greater, genuine understanding. The guidance gives a glimpse into a parallel universe far too removed from classroom reality.

(Richards 2016)

Thus, our work provides a case study – or cautionary tale – of what happens when an education system becomes data-obsessed, and loses sight of the complexities of the identities and learning taking place in an early years or school setting. This data obsession is an essential part of the current functioning of a neoliberal education system driven by values of competition and comparison (Ball 2012a; Ball 2013a), so, for us, studying data is an important part of studying the operation of policy and how these values are both accepted and contested. As Selwyn comments, ‘the discourses, practices and objects of digital data offer a direct “way in” to many of the struggles and conflicts that now characterise contemporary education’ (2015, 79). These struggles include social justice concerns: who is defined as what by data, and how can datafication work to solidify, reveal or challenge disparities between different groups of children?

In this introductory chapter, we consider the politics of data, how data and datafication are defined, and then attempt to map out the existing research on data and education internationally. We end the chapter with a description of the relevant policy context in England, and an outline of the structure of the book. This sets the scene for our overall argument, which has these main strands:

• datafication is productive – of particular data-driven subjectivities, including new roles and hierarchies, and reproductive of some inequalities;
• datafication is reductive – reducing complexity of learning to single numbers and defining quality through proportions;
• datafication results in increased visibility of performance and has thus become an important part of performativity;
• the attraction and danger of datafication both reside in the permanence of beliefs about accuracy and therefore the usefulness of tracking and prediction.

The politics of data

Data are³ political; they reinforce arguments, ‘prove’ effectiveness and demonstrate the success or failure of policy. At a local level, data determine the course of children’s educational trajectories and teachers’ careers, and, in England, a school’s rating by the Office for Standards in Education (Ofsted). Data are fundamental to the relationships between the state and schools (Ozga et al. 2011; Ozga 2016). The use of data must be seen within the political context of an international education system which is driven by neoliberal values of managerialism and accountability (Ball 2015; Lingard, Martino and Rezai-Rashti 2013; Apple 2006). The collection of data facilitates accountability at greater levels of precision and fosters increasingly reliance on numerical comparisons as the basis of assessments of quality; an ‘audit society’ (Power 2013). This makes it very attractive to those who wish to engender market-driven values of competition into the education system: data are the ‘ideal means of bringing market values and free market mechanisms into otherwise closed public education settings’ (Selwyn 2016a, 92). Nowhere is the competition dimension of data in education more prominent than the PISA tests (Programme for International Student Assessment), the Organisation for Economic Cooperation and Development’s international comparison of attainment at secondary school. The PISA results rank countries by attainment in the international tests, driving national policies, encouraging ‘policy borrowing’ from successful systems and, in the case of lower-ranking countries, generating ‘PISA shocks’ (Grek 2009; Sellar and Lingard 2013), which provoke public debate. In 2017 a new international comparative assessment from the OECD began to be piloted for younger children – the International Early Learning Study for four- and five-year-old pupils – which has been described as a ‘pre-school PISA’ (Moss, Dahlberg, Grieshaber et al. 2016). We return to this recent development in later chapters. On a national level, statutory assessment systems such as the ‘Sats’ tests in England, NAPLAN⁴ in Australia, state-wide standardised tests after the No Child Left Behind Act in the United States and the SIMCE⁵ in Chile provide examples of the use of data to provide parents with information to choose schools, within a marketised system. In these cases, data become powerful indicators of the ‘quality’ of schools and teachers, with performance made visible.

Data in education are integral to a particular political understanding of what matters in education, based on neoliberal values of competition. Related to this is the great attraction of the apparent precision of data, a positivist discourse, within the world of education which is seen by many as ‘messy’, ad hoc or unregulated – what Biesta calls the ‘pseudo-security of numbers’ (2017, 317). This is part of a ‘broader deference to statistical reasoning which permeates our understanding of social practices’ (Hardy and Boyle 2011, 214). As we discuss in later chapters, the allure of reduction can be seen as relating to the uncertainties of what has been described as the ‘late neoliberal’ era (McGimpsey 2017).

At an institutional level, data use in schools also has to be seen in the context of discourses of ‘digital revolution’ and the concordant disruption to established educational practice. From the 1990s on, the promise of technology to transform education became an established trope, but ‘it should be clear to all but the most zealous technophile that the much heralded technological transformation of schools and schooling has yet to take place’ (Selwyn 2010, 5). This tale of ‘high tech hope and digital disappointment’ (ibid.) has not prevented new waves of hype about the power of technology to improve education, most recently through online learning, and the continued interest of private companies in the educational market. When considering the use of data in schools, we need to acknowledge the complex interests involved in reproducing this discourse about the power of data and other digital technology to change education; as Selwyn argues, ‘we need to recognize the corporate, commercial and economically-driven nature of much of the prevailing talk about disruption and deinstitutionalization’ (2016a, 21). The scale of money described through this discourse is vast: he quotes a 2013 McKinsey Global Institute report which argued that efficient use of data could add 1 trillion dollars to the value of education each year through improved effectiveness (Selwyn 2016a, 92). Education data are big business; as we discuss in more detail in later chapters,⁶ the private businesses tasked with providing Baseline Assessment received a share of between £3.5 and £4.5 million (Heavey 2016).

Defining ‘data’

In a world where discourses of ‘big data’ and the ‘data revolution’ regularly circulate, the term ‘data’ has become one which is commonly used without much thought to its exact definition. As Pratt notes, in schools the term ‘data’ has been reduced to mean only numbers: ‘the very meaning of the term has been commandeered to ensure only that which is enumerated counts; as if data could only be numbers and not more qualitative descriptions of the children’s world’ (2016, 897). We include in our definition non-numerical forms of data (written observations, photographs, colour coding), though the majority of the data discussed do take numerical form. In the study of data beyond education, they are described as: ‘the raw material produced by abstracting the world into categories, measures and other representational forms – numbers, characters, symbols, images, sounds, electromagnetic waves, bits – that constitute the building blocks from which information and knowledge are created’ (Kitchin 2014, 1). Kitchin argues that data are: ‘epistemological units, made to have a representational form that enables epistemological work’ (2014, 19); the knowledge gained from data is contingent on how and why the data were collected in the first place. The simple scientific framing of data as benign, technical and neutral is mistaken; there are always influences in how data are collected, stored, analysed and used. As Gitelman and Jackson put it, data are always ‘cooked’; there is no such thing as ‘raw data’, untouched by human influence (2013, 5 cited in Kitchin 2014, 20; see also Bowker 2005). This is obviously true in education, where the most common form of data collected, assessment data, are inevitably affected by the mode of assessment, as well as many other factors. There is also no neutral method of analysing data: however neutral the processes of code and algorithms may appear, they are always imbued with some underlying values about what matters in education:

The work of policymakers, education leaders and educators, the choices of parents, and the behaviour and progress of learners alike are all being sculpted or governed by technologies that are instructed by the code and algorithms written by technical experts according to particular discourses about what education is or should be.

(Williamson 2016b, 4)

This view of the technology as part of the social context – a ‘sociotechnical system’ – is drawn from a range of fields including science and technology studies, software studies, geography, philosophy and sociology (Williamson 2016b). This perspective draws attention to the original context in which the technologies were produced, and then ‘fold back to re-shape the contexts in which they originated’ (Williamson 2016b, 6). In this book we conceptualise data in this broader sense as a record of something which also does work itself; data are not a purely technical form of information, but a socially created set of information or knowledge which also have influence on practices and subjectivities.

Defining ‘datafication’

The use of data in education is not a new phenomenon: schools have always kept records of attendance, attainment and other practices such as punishments (Selwyn 2016a; Goldstein and Moss 2014). However, the advent of digital technology and societal expectations around the promise of this new technology to improve education have led to data having an increased prominence in schools and other educational settings – on an ‘an unprecedented scale’ (Selwyn 2016a, 81). Ozga et al. describe ‘the incessant production of data to monitor performance in education’ (2011, 1).

We use the term datafication to describe this broad phenomenon of increased prominence, as it offers a shorthand for a complex process. In earlier work we have defined datafication in terms of increased significance, visibility and constant governance through dataveillance, and as being what happens when people or systems are ‘subjected to the demands of data production’ (Roberts-Holmes and Bradbury 2016). Ozga describes the use of data as a policy instrument as growing in ‘strength, speed and scope’ (2009, 150). Others have described the increased use of data in education using the ‘three Vs’ (Laney 2001 cited in Selwyn 2016a, 81): volume, describing increased amounts of data produced; variety, in terms of types of data and their sources; and velocity, of both the production and processing of data. This is a useful categorisation which helps us to move forward from descriptions of ‘more data’. It is helpful to think about how data are produced from multiple sources and have multiple forms in classrooms, how they are transferred from one site to another and with whose permission, who has the power to alter them, and of course who controls how they are processed and delivered back to serve some purpose. For us, changes in volume, variety and velocity are all component parts of datafication. We are interested in the different forms and purposes of data, such as the distinctions between ‘compliance data’ produced to fulfil a commitment (usually policy-based) and ‘useful data’ which aid learning (Selwyn, Henderson and Chao 2015). Thus this work builds on Bradbury’s previous explorations of assessment processes in early years and their complex relationships with concepts of ‘teacher knowledge’ (Bradbury 2013c). We also develop our discussions on the changing status of data within schools (Bradbury and Roberts-Holmes 2016a) and the related ‘artefacts’ of its collection (Souto-Otero and Beneito-Montagut 2016).

However, in this book we also use the term datafication to examine the impact of data, particularly on subjectivities, and so the process described is not simply a change to what is done and how, but also a change to who people are, or who they are expected to be. As Gitelman and Jackson note, data ‘need to be understood as framed and framing’ ...

Table of contents

• Cover
• Title
• Copyright
• Dedication
• CONTENTS
• List of illustrations
• Preface
• Acknowledgements
• List of abbreviations
• 1 Introduction: the state, data and primary education
• 2 Theorising and researching data in education
• 3 Datafication in the classroom: the production of data-driven subjectivities
• 4 Data, reductionism and the problems of assessing young children
• 5 Schools’ responses to datafication and the visibility of performance
• 6 International datafication and business opportunities
• 7 Datafied learner identities: towards an understanding of the future of data in education
• References
• Index