Gabriel Sahlgren: Seek and thou shall find - exposing the confirmation bias of British free school critics

Gabriel H. Sahlgren was a Visiting Research Fellow with the Institute of Economic Affairs in the summer of 2010 and is the author of the paper Schooling for Money: Swedish Education Reform and the Role of the Profit Motive. He recently graduated from Cambridge University with Starred First Class Honours in Politics.

Various commentators aiming to undermine the government’s free school programme have argued that recent research in Sweden suggests a decline in standards and increased segregation. It is interesting that critics have suddenly started to cite such research given that the overwhelming majority of studies display that the Swedish free school system has led to improvements. It is a bit like a punter claiming to be a master at betting on the horses when, after having already lost thousands of pounds, he finally wins back a tenner on a lucky nag.

So what does the evidence say? Regarding educational achievement and Swedish school competition, the academic research only covers grades given on standardised tests and the grade point average. This research displays unequivocally positive effects from free schools.

The paradox, then, is why Swedish pupils generally perform worse in international comparisons today compared to before the school choice reform. The paradox is resolved when it is realised that the argument that school competition has depreciated achievement in international tests is based on mere correlation. Achievement in those tests has fallen during the 1990s/2000s, which was the same period during which Sweden carried out education reforms. Critics of free schools simply ignore all other changes that occurred at the same time – which altered the way education was carried out in much more significant ways than the school choice reform. A key example that has been highlighted recently is the policy influence of educationalists, who have been allowed to implement pedagogical experiments without proper academic grounding. A majority of Sweden’s leading educationalists consider tests and other forms of measuring student achievement, as well as ‘teacher-centred education’, harmful – which have been reflected in education reforms. Interestingly, 65% of Swedish education professors have not been cited in any acknowledged publication outside of Sweden in the past 10 years.

This should be contrasted with Finland – a high-performer in international surveys – where teachers and final exams remain key. Rather than dumbing down teacher education to adapt to lower student quality, as Sweden has done, in Finland all teachers are recruited from the top 20% of upper-secondary school students; only about 10% of applicants are accepted to become teachers. Yet, this key difference is never voiced by British critics of the Swedish system.

Given that all academic evidence displays positive effects from competition from free schools according to standardised tests and the grade point average, it is highly likely that the school choice reform in fact has made the fall in performance in international tests less severe. This interpretation is further reinforced by an OECD paper, which provides micro-econometric, cross-national evidence utilising PISA mathematics and science scores from 265,000 pupils in 37 countries, finding that larger shares of privately managed schools increase individual pupils’ performances. Further displaying the importance of tests/teachers, too, the paper finds that external exit exams and monitoring of classes, aiming at holding teachers responsible for students’ achievements, predict higher PISA scores.

What about segregation? It is true that Swedish socio-economic and ethnic school segregation has increased over the past 20 years, and, again, British critics have blamed school competition. But the evidence is not conclusive: some papers find that this increased segregation is mainly due to increased housing segregation although new studies suggest that competition is a more important contributor to the phenomenon.

Firstly, there are no obvious lessons for the UK here as its education system is already very segregated. But the more important observation is that, although this increased segregation inflates variation in educational attainment between schools, it does not inflate variation amongst students. In other words, there is no “cream-skimming”/”sink-school” problem. Despite increased variation between schools, the better-off or particular ethnic groups do not do better as a result (and the worse-off and other ethnic groups do not do worse). Just because poorer students are more concentrated in the same schools does not mean that they perform worse because of it. Furthermore, one of the advantages of competition is that, rather than having to move to switch schools (a privilege of the rich), the poor have the right to exit any school if they so wish! It is also notable that the Swedish decline in international tests has affected low, middle and high income students basically equally.

Having dealt with the selective citing of evidence by British critics of free schools, one conclusion is in order: there were so many changes in the 1990s that it is extremely hard to evaluate which ones are responsible for Sweden’s fall in international comparisons and, since no one has directly analysed the impact of Swedish school competition on such scores, it is more due to ideology than facts when critics blame school choice. If anything, the evidence suggests that school competition prevented scores from falling even further.

Three things, however, are very clear:

1. Variation in educational achievements between individual pupils has not increased even if variation between schools has; the system has not become more unequal in terms of pupil achievement.
2. School competition has increased domestic educational achievement, parental/teacher satisfaction and teacher salaries. There is no doubt about this conclusion.
3. Profit-making schools appear to benefit the less-well-off the most – while little suggests that competition would have increased significantly without such schools.

There are clear implications for the coalition government. Its free school policy must be robustly defended. However, by not admitting profit-making schools, it is doing the poor a great disservice and puts the entire reform at risk of utter failure.