Funding, School Specialization, and Test Scores: An Evaluation of the Specialist Schools Policy Using Matching Models

Abstract

We evaluate the effect on test scores of a UK education reform which has increased funding of schools and encouraged their specialisation in particular subject areas, enhancing pupil choice and competition between schools. Using several data sets, we apply cross-sectional and difference-indifferences matching models, to confront issues of the choice of an appropriate control group and different forms of selection bias. We demonstrate a statistically significant causal effect of the specialist schools policy on test score outcomes. The duration of specialisation matters, and we consistently find that the longer a school has been specialist the larger is the impact on test scores. We finally disentangle the funding effect from a specialisation effect, and the latter occurs yielding relatively large improvements in test scores in particular subjects.

The Effects of the Specialist Schools Education Policy on School and Post‐School Outcomes in England

“…Figure shows that specialist schools have tended to outperform non‐specialist schools throughout the history of the policy, and Bradley et al . () present evidence to suggest that specialist schools are increasingly likely to have average test scores in the highest quintile of test scores, which is strongly suggestive of non‐random assignment of certain types of school into the specialist schools initiative. A closely related source of bias is the non‐random selection of pupils into specialist schools ( pupil selection bias ), insofar as unobservably more able pupils are ‘cream‐skimmed’ by ‘good’ (specialist) schools.…”

“…There is a growing literature on the evaluation of the specialist schools policy on test score outcomes at school which we review in some detail in Bradley et al . (). Many of the early studies are aggregate, or school‐level, studies that use cross‐sectional data and tend to show quite large effects of the specialist schools policy on test scores, increasing the probability of pupils achieving 5 or more General Certificate of Secondary Education (GCSE) grades A*–C by approximately 5 percentage points (Gorard, ; Jesson, ; Jesson and Crossley, ; Office for Standards in Education, ).…”

“…A closely related source of bias is the non‐random selection of pupils into specialist schools ( pupil selection bias ), insofar as unobservably more able pupils are ‘cream‐skimmed’ by ‘good’ (specialist) schools. Specialist schools are, for instance, more likely to have a higher percentage of pupils from richer social backgrounds and are more likely to have a lower percentage of ethnic minority pupils (Bradley et al ., ). It is difficult to disentangle these two sources of bias with available data; nevertheless, in trying to measure the impact of the specialist schools policy on educational outcomes from compulsory schooling, it is important to try and do so.…”

The joint evaluation of multiple educational policies: the case of specialist schools and Excellence in Cities policies in Britain

“…Benton et al (2003), again using multi-level modelling techniques on cross-sectional pupil-level data, find that the specialist schools policy raised GCSE grades, or points, by 1.1, whereas Levacic and Jenkins (2004) found a very similar effect (1.4 GCSE points). Bradley, Migali, and Taylor (2012) use DID matching methods and find a modest causal effect on pupil test scores of between 0.4 and 0.9 of a GCSE point.…”

“…It is also possible that specialist schools cream-skimmed the best pupils so ensuring high average test scores (generating a pupil selection bias). Bradley, Migali, and Taylor (2012) provide some evidence for the existence of these selection effects which could generate a positive bias in the estimated effect of the specialist schools policy. Clearly, the estimated effect would not be causal if attempts are not made to mitigate these sources of bias.…”