Education Governance and Social Theory

Interdisciplinary Approaches to Research

Andrew Wilkins, Antonio Olmedo, Andrew Wilkins, Antonio Olmedo

• Publisher: Bloomsbury Academic
• Year: 2018
• Print ISBN: 9781350159723
• eBook ISBN: 9781350040076
• Edition: 1
• Topic: Education
• Subtopic: Education Theory & Practice

The study of 'education governance' is a significant area of research in the twenty-first century concerned with the changing organisation of education systems, relations and processes against the background of wider political and economic developments occurring nationally and globally. In Education Governance and Social Theory these important issues are critically examined through a range of innovative theoretical perspectives and methodological approaches to assist in guiding those interested in better understanding and engaging with education governance as an object of critical inquiry and a tool or method of research. With contributions from an international line-up of academics, the book judiciously combines theory and methodologies with case study material taken from diverse geo-political settings to help frame and enrich our understanding of education governance. This is a theoretically and empirically rich resource for those who wish to research education governance and its multifarious operations, conditions and effects, but are not sure how to do so. It will therefore appeal to readers who have a strong interest in the practical application of social theory to making sense of the complex changes underway in education across the globe.

Part One

Data Regimes

1

Digitizing Education Governance: Pearson, Real-Time Data Analytics, Visualization and Machine Intelligence

Ben Williamson

Digital technologies have become significant non-human actors in education governance. This chapter examines emerging techniques of digital governance in education. The global education company Pearson has become a key actor in contemporary education governance (Hogan et al. 2015), with aspirations to become a ‘digital-first’ company using data-processing techniques to govern educational institutions and individuals (Williamson 2016). Pearson’s ambitions involve generating massive databases of educational data, producing data visualizations and utilizing real-time data analytics, machine learning algorithms and artificial intelligence to monitor, measure, make calculations and produce objective facts and knowledge about education. Its strategic business objective since 2016 has been to create a ‘global digital learning platform’ utilizing cloud services, data analytics and machine learning across all its products and services (Ismael 2016).

Pearson exemplifies how digital data, real-time analytics and machine intelligence are being recruited to the task of enacting education governance. Moreover, Pearson’s activities in education are characteristic of an emerging ‘data politics’ (Ruppert et al. 2017), whereby power has been distributed to non-human systems and to those human actors able to translate what digital data are ‘saying’ for public audiences, political agents and states (Davies 2017). Through both its technological innovations and its narration of the impacts of digital data in education, Pearson is seeking to not only speed up conventional techniques of governance, such as data-driven performance measurement, but also circumvent conventional modes of policymaking by making educational processes of improvement and reform into real-time, automated tasks performed with non-human computational systems rather than administrative tasks enacted via bureaucratic organs of state. In so doing, it is seeking ‘data monopoly’ over the production and narration of educational data. This chapter provides a case study of Pearson’s efforts to digitize education governance, focusing on its data visualization, machine learning and artificial intelligence developments, and theorizes its activities as a form of ‘algorithmic governmentality’ (Rouvroy and Berns 2013) that enmeshes educational institutions, teachers and students in the computational logics of algorithmic data analysis. Methodologically, the chapter draws on ‘policy network analysis’ (Gulson et al. 2017) and ‘software studies’ (Kitchin and Dodge 2011) approaches to analyse Pearson’s influence network and its digital product development.

Governance and governing

The term ‘governance’ signifies two distinctive but related ways of analysing the circuits and functions of power in contemporary states, and has been taken up as a key concept in understandings of contemporary educational policy processes. Firstly, governance signifies a structural shift from centralized state government control to distributed and interactive networks of actors and experts working together across sectors on policy problems and policymaking (Ozga et al. 2011). Contemporary governments are increasingly seeking to ‘decentralize’ powers away from central bureaucratic agencies and ‘devolve’ processes of ‘state monopoly’, in some cases with new ‘private monopolies’ of business networks supplementing or supplanting the formal authority of government (Wilkins 2017). The second definition of governance is inspired by Michel Foucault’s influential conceptualization of techniques of governing. In the term ‘governmentality’ Foucault (2007) captured the historically grounded techniques, calculations, analyses and procedures employed by specific authorities and political powers for directing human behaviour, as captured in the phrases ‘conduct of conduct’ or ‘acting upon action’. Importantly, for authorities to govern, studies of governmentality insist, it is essential to possess knowledge of whatever they wish to govern in order to then administer and intervene in the lives of individuals or activate and manage populations (Rose 1999).

The task of studying governance methodologically involves tracing the networks of authorities and experts, which together administer society and the state, and examining the specific ideas and techniques they enact to know, analyse and guide human behaviour to achieve their objectives. Education governance, then, can be understood as the increasingly devolved ways in which educational policies are influenced and generated – through networks and relations between state, civil society and private sector actors – and refers to the manifold techniques being developed through these relations to establish new practices, routines, technologies and discourses within the institutions of state education. In order to examine the specific issue of education governance, education researchers therefore address and ‘follow’ how ‘policy actors, discourses, conceptions, connections, agendas, resources, and solutions of governance’ increasingly move across sectors and spaces, and how they are exerting ‘significant impact on the formulation and reformulation of teaching and learning, assessment and the curriculum, and the general directions and conceptualization of education policy and governance’ (Ball 2016: 1–2).

In order to articulate how Pearson is involved in both a structural shift to ‘de-governmentalized’ governance (Gulson et al. 2017) and designing techniques to govern conduct within education, the focus for the chapter is on the ‘policy networks’ (Ball 2016) that Pearson occupies, and on the ‘policy instruments’ (Lascoumes and le Gales 2007) it has produced to intervene into, manage and activate the actions of policymakers, teachers and students to achieve desired outcomes. Methodologically, network analysis has been conducted on Pearson to identify its key actors, events, connections and partnerships, and then to follow some of its activities, its products and its technical developments over a period from 2011 to 2017. Network analysis involves mapping actors and their connections, and then ‘studying the chains, circuits, networks, and webs in and through which policy and its associated discourses and ideologies are made mobile and mutable’ (McCann and Ward 2012: 43). As such, policy networks produce what McCann and Ward (2012: 43) describe as ‘policy assemblages’, using the term ‘ “assemblage” in a descriptive sense to encourage both an attention to the composite and relational character of policies . . . and also to the various social practices that gather, or draw together, diverse elements . . . into relatively stable and coherent “things” ’. A ‘software studies’ approach is also employed to examine the specific digital policy instruments that Pearson and its partners have produced as part of these assemblages. This involves tracing the evolution and contextual unfolding of ideas, decisions, constraints, actions and actors that shape software projects, in order to ‘excavate ‘the social lives of ideas into code’ (Kitchin and Dodge 2011: 255). The examples of Pearson’s policy instruments reveal how specific ideas, assumptions, values and intentions structure the software products they produce or form partnerships to promote. Taken together, policy network analysis and software studies methodologies allow us to trace the organizational webs that produce composite policy assemblages, and then to excavate the social practices, values and expert knowledges that have informed the development of the digital ‘things’ these policy networks are seeking to embed into educational settings and practices.

Digitizing governance

In order to approach and understand Pearson as a key actor of digital education governance, it is important to locate their activities in the broader context of shifts in the ways that governance is structured and practised. One key shift is the temporal acceleration and spatial distribution of policy processes, termed ‘fast policy’ (Peck and Theodore 2015). Peck and Theodore (2015) argue that modern policymaking may still be focused on centres of political authority, but is also distributed to sprawling networks of human and non-human actors. As such, policy is increasingly accomplished through connected webs of consultants, think tanks, research institutes, guru performers as well as websites, blogs, social media and other non-human technologies and material objects.

The non-human actors of fast policy can be approached as ‘policy instruments’. Policy instruments are defined as any kind of device, method, tool or technique designed to put a particular policy into practice (Lascoumes and le Gales 2007). Importantly, policy instruments are not value-neutral. Because they are designed in particular settings, they carry values and worldviews that then may partly shape policies. Policy instruments constitute ‘a condensed form of knowledge about social control and ways of exercising it’ and ‘are not neutral devices: they produce specific effects . . . which structure public policy according to their own logic’ (Lascoumes and Gales 2007: 3). With digital policy instruments – software programs designed to operationalize key policy ideas – particular values and ways of approaching problems and solutions can therefore be understood to be coded-in to their functioning. Fast policy is partly the accomplishment of digital policy instruments, with technologies of measurement, ranking and comparison creating new continuities and flows that can overcome physical distance in an increasingly interconnected and accelerating digital world (Lewis and Hogan 2016).

The adoption of digital policy instruments to accomplish particular policy objectives is part of a wider and ongoing transformation in the organization of the state termed ‘digital governance’ (Dunleavy and Margetts 2015). In the era of social media, automation and big data, governments, organizations and individuals alike leave digital traces of everything they do. Digital governance describes the use of these digital data to generate insights to inform future policy development and political intervention. Digital education governance is part of this turn to the use of digital data to generate the knowledge required to govern the state. Techniques of digital governance can already be seen in the expansion of highly complex technical data infrastructures required for the collection, storage, analysis and dissemination of test data at national and international scales (Sellar 2015). The test data collected from schools are themselves the products of chains of decisions about what should be collected, and how it should be processed and reported, with these decisions made not just by actors within the formal education system but by various testing companies, software firms, consultancies, university consortia and philanthropic foundations. As a consequence, educational data are ‘the products of complex assemblages of technology, people and policies that stretch across and beyond the boundaries of our formal education system’ (Anagnostopoulos et al. 2013: 2).

The existing infrastructure of test-based performance measurement, however, is beginning to evolve to include algorithmic data analytics in which huge quantities of continuously collected ‘big data’, real-time analysis and automated feedback – and the technical and statistical experts that handle it – are to play a crucial role (Hartong 2016; Mayer-Schonberger and Cukier 2014). The role of computer algorithms to make sense of massive quantities of digital data is ushering in a new condition of ‘algorithmic governmentality’ whereby individuals can be ‘known’ from digital data traces of their activities and then acted upon on the basis of that knowledge. Rouvroy and Berns (2013: 7) have detailed how algorithmic governmentality is enacted through the collection and automated storage of vast ‘data warehouses’ from various sources. These data can then be subjected to ‘data mining’, or ‘the automated processing of these big data to identify subtle correlations between them’, leading to ‘action on behaviours’, or the application of this knowledge to infer ‘probabilistic predictions’ that might be used to ‘anticipate individual behaviours’ (Rouvroy and Berns 2013: 7–8).

The use of statistical knowledge to govern populations has a long genealogy stretching back to nineteenth-century censuses, surveys, accounting and other bureaucratic practices of state management and control (Foucault 2007). Statistical knowledge of the population was a key source of modern governmental power, enabling ‘a machinery of government to operate from centres that calculate’ (Rose 1999: 213). Algorithmic governmentality registers a shift from population data to fine-grained individual data and its use for purposes of precise behaviour management. While digital governance conceptualizes the changing practices of the state – and the technical institutions that increasingly co-constitute state power – as digital data become available to conduct a constant audit of the population (Ruppert et al. 2017), algorithmic governmentality registers in a more Foucauldian sense how algorithms that process digital data may be used to intervene in and govern people’s lives (Rieder and Simon 2016). This algorithmic governmentality has been described by Cheney-Lippold (2011: 167) as a ‘soft biopolitics’ whereby algorithmic sorting of users’ data may be used ‘to determine the new conditions of possibilities of users’ lives’. Digital education governance, then, signifies how policy and governance are becoming more accelerated and distributed to include non-human technical instruments and networked systems that collect, process and communicate data. Education might therefore be measured through digital data to allow state governments and other organizations to know and intervene in education at scales from the classroom to the nation state, while new and emerging forms of algorithmic analysis of data may ena...

Table of contents

• Cover
• Half Title
• Series Information
• Title Page
• Contents
• Series Editor’s Foreword
• Notes on Contributors
• List of Abbreviations
• Foreword
• Introduction: Conceptualizing Education Governance: Framings, Perspectives and Theories
• Part One  Data Regimes
• Part Two  Evaluation Regimes
• Part Three  Knowledge Regimes
• Part Four  Institutional Regimes
• Index
• Copyright Page