Mind the Gap: Educational Inequalities during Covid-19
School closures have affected roughly 10m children and young people across the UK, interrupting their learning and placing considerably more responsibility for educational activities on the home environment than ever before. This is expected to slow the progress of a whole generation of students and to widen gender, socio-economic, and ethnic inequalities (EEF, 2020). Addressing widening inequalities is essential to avoid undesirable long-term consequences, including negative labour market and health outcomes leading to lower economic productivity, increased health care costs, and reduced social mobility.
To be able to mitigate the disadvantages experienced by students in a timely manner, we need to understand the way school closures have differentially affected children. Students due to take GCSEs and A-levels this summer have been the first to feel the effects, with exams replaced by teacher assessments and algorithms. This raises concerns about the possible presence of biases in assessment outcomes, which might have a differential impact by students’ gender, socio-economic background and ethnicity.
For year groups not currently due to take exams, at this early stage we can best assess inequalities by comparing differences in the home learning environment, comprising investments of time and resources provided by schools, parents, and the students themselves. We will analyse changes in all learning inputs during and after the school closures and combine this with detailed information about students’ background. This will provide an accurate picture of the development of gaps in education which can be used by policy makers and practitioners to help alleviate the expected inequalities.
The aim of the research project is to understand whether school closures due to the Covid-19 pandemic result in widening educational inequalities, and if so, which are the groups most affected. Early analysis of new data collected during the pandemic (Andrew et al., 2020; Bayrakdar and Guveli, 2020; Benzeval et al., 2020 Green, 2020; Lucas et al., 2020) alert to the presence of large differences in learning inputs, with state schools providing significantly less regular access to on-line classes than private schools, parents spending more time helping boys rather than girls, and wide variation in the hours students themselves spend on homework by free school meal entitlement. Somewhat surprisingly though, inequalities in educational investments do not always follow a socio-economic gradient. For example, the availability of computers or parents’ help with home schooling does not vary much by educational background (Benzeval et al. 2020). Low educated parents may have more time available to dedicate to home schooling as they are more likely to work fewer hours as a result of the pandemic (Adams-Prassl et al., 2020). This suggests that we need a more comprehensive analysis of family circumstances to understand how educational inequalities have developed during the period of school closures.
In this project we plan to go beyond the current state of knowledge by:
1) considering detailed information on students’ background and circumstances to get a nuanced picture of the developing inequalities. Our focus is on gender, ethnicity and socio-economic background (including parent’s income, education, employment status, housing circumstances, household composition and health). We will study how background factors combine to create disadvantage. For example, we will investigate the presence of an income or education gradient in parental inputs and child own study time and whether parents’ hours of work constraints mitigate or accentuate this differential (Ermisch and Francesconi, 2013);
2) taking a comprehensive view of educational inputs, comprising inputs provided by schools (such as live lessons), parents (such as help with homework or paid-for tutoring) and the students themselves (such as time spent studying) and studying how they interact with each other. Only by assessing the total inputs received by each student are we able to assess the overall impact of school closures, as the inputs combine to produce educational outcomes (Rabe, 2019; Greaves et al., 2020). For example, we will ask whether parents compensate for the absence of school inputs, and whether this compensatory behaviour is different across different groups of the population.
3) adopting a longitudinal perspective and looking at changes in learning inputs experienced by students. Large gaps in parental time spent with children by socio-economic status and income have been widely documented in the education literature (Fiorini and Keane, 2014; Carneiro and Ginja 2016, Del Bono et al., 2016). Here we focus on changes in parental education investments to understand whether existing inequalities have increased due to and in the course of the pandemic;
4) providing a first assessment of inequalities in school outcomes. We will study attainment gaps in current teacher-assessed GCSE and A-level results and investigate whether, compared to exam-based grades in previous years, the attainment gap has widened between students of different gender, ethnicity and free school meal eligibility, and how this varies by teachers’ characteristics (Wyness and Murphy, 2020; Wyness, 2017; Campbell, 2015). This analysis will allow us to assess the short-term inequalities that may already be visible in student outcomes.
Our first contribution will be to analyse changes in learning inputs during the period of the Covid-19 pandemic. To do so, we will use monthly data from the Understanding Society Covid-19 survey.
A second contribution will be to assess changes in learning inputs before and after the Covid-19 pandemic, and how these changes differ across students from different backgrounds. For this analysis we will exploit Wave 11 of Understanding Society.
Our third contribution will be the analysis of educational outcomes for the students completing their GCSEs and A-levels in the current academic year 2019/20. We will use National Pupil Database GCSE and A-level attainment data linked to school characteristics (e.g. state/private schools, composition of the student body, prior attainment, Ofsted grade) as well as the School Workforce Census which includes teacher characteristics such as gender, ethnicity and experience. We will also use UCAS data to examine the impact of A-level grading on higher education participation, by background.