Abstract Alternative provision (AP), in particular Pupil Referral Units (PRUs), have been criticised as a forgotten part of the education system, side‐lined and stigmatised as somewhere only the very worst behaved pupils go. In response to this criticism, PRUs have now been academised to become AP Academies and new AP schools have been set up—AP Free Schools. A sample of 5 years of AP pupils sitting their GCSE examinations from 2016/17 to 2020/21 ( N = 15,019) was used to compare the academic attainment of AP Free School pupils and AP Academy pupils. AP Free School pupils achieved 13.26% more capped GCSE points than AP Academy pupils, which increased to a difference of 18.65% when controlling for selection bias, suggesting that not only do AP Free Schools outperform AP Academies academically, but they do so with more disadvantaged pupils. However, this difference is equal to one grade in one GCSE subject and the national average score for ‘Attainment 8’ is 7.3 times higher than the average capped GCSE points of AP Free School pupils, such that pupils may fare better academically if they remain in mainstream education. The controlled disadvantageous pupil characteristics of being a looked after child ( β = −1.67), being eligible for free school meals ( β = −1.42), having special needs ( β = −1.26) and being of an ethnic minority ( β = −1.23) were found to be stronger predictors of capped GCSE points than AP type ( β = 1.04) and are characteristics more likely to be found in pupils who are in AP as a result of exclusion. The superior academic performance of AP Free Schools over AP Academies is not sufficient to overcome the disadvantages faced by pupils excluded into AP. Permanently excluding should be a last resort and mainstream schools should work in partnership with AP Free Schools so that the relative strengths of both types of schools are leveraged in the interest of pupils at‐risk of permanent exclusion.
Hills et al. (Sat,) studied this question.