The Effectiveness and Equity of Grammar Schools in England

Abstract

In England, secondary education is mainly comprehensive in nature. However, a small group of grammar schools retain ability-based selection at the age of 11. Students in these 163 schools obtain high grades at Key Stage 4 (KS4), aged 16. Some commentators and policy-makers believe that these high grades are an indication that grammar schools are therefore good schools, and suggest increasing the number of grammar school places. They also claim that grammar schools provide opportunities for economically poorer but able students. At the same time, other commentators believe that grammar schools are elitist and socially divisive, actually providing further opportunities mostly for the already privileged. They urge successive governments to abolish the remaining grammar schools. The issue has become a particular ‘hot topic’ in the last two elections in England.

This study looks at the impact of the presence of grammar schools in some local authorities (LAs), and considers whether the expansion or removal of grammar schools would lead to a more effective and equitable education system. This new study uses the National Pupil Database (NPD), the Higher Education Statistics Agency (HESA) data on university attendance, and a local dataset providing the results of the grammar schools’ selection test in one LA for one year. The study looks at access to grammar schools, and its relationship with pupil’s prior attainment, geographical location, and family background. It considers whether the opportunity to attend grammar schools, and outcomes of attending grammar schools, are distributed equally among different social groups. The effectiveness of grammar schools is evaluated based on pupil’s KS4 attainment and their Higher Education (HE) participation patterns. The analysis is based on standard regression models, controlling for pre-existing differences between pupil groups, and uses the innovative design of regression discontinuity (RDD) which is more robust in making causal inferences. The study also looks at the influence of the presence of grammar schools on the overall performance level in the local area, looking at the effectiveness of selective LAs and non-selective LAs in a holistic way. Lastly, the link between family background and post-18 destination in both types of LA are compared. Based on access to and outcome of grammar schools, the analysis demonstrates the possible impacts of the selective system on the redistribution of educational opportunities. This study is in a position to help decide whether the policy of grammar school expansion would work as claimed by the government, or whether the removal of grammar schools would make any difference.

Overall, once prior attainment is accounted for, the study found a small advantage in KS4 attainment associated with grammar school attendance, only for borderline pupils, rather than the highest performers. The study found no advantage from grammar school attendance for HE participation. Selective LAs have no better results than non-selective LAs. And pupils in selective LAs who failed to attend grammar schools gained worse results than equivalent ones in comprehensive LAs. Therefore, the claim that increasing grammar school places will help raise national academic standard is unrealistic, and not founded in existing grammar school outcomes.

In addition, attainment at an early age differs considerably between social groups. Selection at an early age means that the opportunity to attend grammar school is heavily unbalanced between these groups. The layered early attainment compounded by the imperfect selection process means that the small advantage that may be associated with attending grammar schools for borderline entrants and the cost of not attending grammar schools is not equally spread between social groups. The mechanism of selection and differentiation thus amplifies the attainment gap between pupils from high and low socioeconomic groups. Therefore, the selective system fails to comply with the demands of equity, whether merit-based or needs-based, and whether taking pupil’s prior attainment into account or not. Moreover, separating pupils into different schools at an early-age creates segregated school compositions, which ultimately endangers the integration and long-term development of the society. To conclude, this study does not find any substantial benefit associated with early-age academic selection in England, and the expansion of grammar schools is unlikely to raise national performance standard, or to provide more opportunities for the poor.

The study also includes the first attempt at using the RDD to evaluate the current effectiveness of grammar schools in England. However, the incomplete data available on the results of the selection test means the application of the RDD is more a feasibility trial than a definitive test intended to settle the debate on the effectiveness of grammar schools. Conducting this design with national data on grammar school selection would create the most powerful evidence so far. The study thus advocates that the government should make the responsible decision to disclose grammar school selection data for the purposes of research. Promoting an effective and equitable education system is important, and the public deserve to know more. If the results are confirmed with full data on the results of the selection test, then perhaps the next step would be to phases the 163 grammar schools out altogether.