Social mobility and selection do not mix

Effective whole-population analysis is much easier now we are in the world of big data and the National Pupil Database. However, even though the problems of sampling are avoided, the move from raw data, through information to understanding to action, requires careful thought. It is easy, for example, to confuse correlation with causation, or to get the causal factors the wrong way round. All research should be read with a sceptical eye, and with an eye to alternative explanations.

What, then, does the report show? The key finding is that, as with health, there is a clear link between poor education outcomes and low income. So far, so well known, but the richness of the National Pupil Database enables fine-grained analysis. While in primary schools many children on free school meals do as well as more affluent pupils, at secondary the large majority of pupils on free school meals make significantly less progress than their peers, even when they did well in primary school. There are clear questions here for policy makers, practitioners and researchers: what is behind this failure to maintain progress? What about school funding? What can secondary schools do to overcome this lack of progress? The report hypothesises that out-of-school effects, such as parental support and the home learning environment, are important.

Another finding is that there is a positive urban effect, especially in London. My hypothesis would be that this is linked to school funding and the City Challenges set up by the last Labour government. Also, the progress gap is smaller in schools with very few pupils on free school meals, and with very many such pupils. The report also hypothesises about the Pupil Premium and peer group effects.

The most important finding for education policy relates to school type, where the report is inconsistent, stating: "Focusing on differences between schools is… a distraction", before adding: "Low-income pupils are less likely to make good progress in schools where they are grouped by ability from an early age". This inconsistency may have been expected from commission members Alan Milburn and Baroness Gillian Shephard, former politicians. However, the evidence seems clear - across the country, selection disadvantages low-income pupils.

Many observers have made convincing theoretical arguments against selection at age 11. This report goes further, and provides explicit evidence of what is actually happening in our schools - selection reduces the chances of low-income pupils succeeding at secondary school. I fear Prime Minister Theresa May's obsession with grammar schools is proving antithetical to her stated aim of promoting social mobility. We need evidence-based policy.

John Freeman is a former director of children's services and now a freelance consultant