Magic Numbers? Education Policy, "Big Data," and a Critical Race Theory of Statistics

• In Event: Advancing Critical Race Theory in Educational Quantitative Methods

Abstract

1. Objectives
The paper exposes and critiques the mythology surrounding the power of numbers in education policy. On both sides of the Atlantic administrations have argued that accountability through statistics would lead to improved social justice and millions have been spent developing ‘Big Data’ applications in the belief that numbers contain solutions to key economic and social problems (Kalil 2012; RCUK 2015). The paper offers an empirical critique of these positions and sets out principles for statistical critical race scholarship in the future.

2. Theoretical framework
The paper builds on and extends the nascent movement to combine critical statistical analyses with understandings informed by Critical Race Theory (CRT), e.g. Covarrubias and Velez (2013) and Gillborn (2010).

3. Methods and techniques of inquiry
We adopt a mixed method approach, combining quantitative and qualitative data, shaped by CRT policy analysis.

4. Data sources and evidence
Our quantitative analyses draw on the National Pupil Database (which records the attainment of all students attending state-maintained schools in England and includes variables on poverty, school type and ethnic origin) and the Longitudinal Study of Young People in Education (with around 15,000 respondents) which gathers additional data on education and employment experiences. Critical policy discourse analyses will be conducted using data on government debates, policy texts, and media coverage of relevant issues, especially ‘Big Data’, drawing on official sources (e.g. Hansard) and dedicated search engines such as LexisNexis.

5. Results and substantiated conclusions
The paper debunks the myth that numbers have special qualities that (a) reveal truths inaccessible to human analysts; and (b) drive forward equitable policy reform by exposing current injustices.

Our quantitative analyses demonstrate that far from revealing unrecognized race inequities, statistics have been used to downplay the significance of institutional racism while constructing an image of White racial victims. Our analyses of political and media texts show how this myth of White under-achievement has been recycled year-after-year and remains impervious to critique regardless of the data sources that are mobilized.

The analyses reveal that, far from exercising an objective and automatic influence on policy, statistics are effectively subject to the ‘rules of racial standing’ (Bell 1992) such that, under normal circumstances, judgements of their quality and significance are a direct function of their position in supporting or challenging White racial privilege.

6. Scientific and scholarly significance of the study
The paper provides the empirical basis to generate a set of principles that might usefully act as a check on the unthinking race bias that frequently accompanies the use of statistics in educational research, policy and practice. These ‘QuantCrit’ principles draw attention to the socially constructed, partial and contested nature of statistical data and argue for its explicit use in the service of social justice goals.

Authors

• David Gillborn, The University of Birmingham

• Paul Warmington, The University of Birmingham

• Sean Demack, Sheffield Hallam University