Since August I have been living in Bradford in Yorkshire. In the 1870s the local MP for Bradford was a Liberal, William Edward Forster. Forster has a special place in the history of social policy in Britain because he was responsible for introducing the 1870 Education Act. The 1870 Education Act is to many the reform that started the welfare state because it introduced the idea that the government had a responsibility to see that no child was denied an education because of their family circumstances. The prevailing view at the time was that politics and politicians should interfere as little as possible in social problems. The 1870 Act therefore marked a significant change in the role played by the government in regulating social affairs.
Education, and particularly the role of education in promoting social mobility, has continued to feature prominently in the policies of the modern Liberal party. Indeed the most significant contribution that the Liberals have made to the present coalition government has been the introduction of the pupil premium which in 2012-2013 will provide additional funding of £625 for each child from a disadvantaged background in a schools intake. The additional funding for children from disadvantaged backgrounds is being provided to schools in order to try to close the gap in achievement that exists between children from disadvantaged and more affluent families. The gap in achievement between children from disadvantaged and more affluent families is considerable. In 2011, 34.6 percent of children eligible for free school meals received 5 or more GCSE’s at grade A-C compared to 62.0 percent of all other pupils. Low educational qualifications increases the risk of a range of poor economic and social outcomes as adults and interventions that raise the achievement of children from disadvantaged backgrounds can be seen as having the character of an investment rather than being a direct cost to the government.
The level of government funding that a school receives has little relationship to where a school is placed in national league tables of pupil performace. Policy makers have therefore concluded that in order to raise the achievement of disadvantaged children we need a better understanding of the effectiveness of different approaches, rather than simply higher spending. In order to help schools decide how to spend the pupil premium, the Department of Education established the Education Endowment Foundation (EEF) with a grant of £110 million in 2010. The approach of the EEF to evaluating the effectiveness of interventions has two aspects. Firstly, the EEF funds projects aimed at improving the attainment of disadvantaged children and many of the funded projects have used randomised designs in which children are assigned at random to treatment and control groups. Secondly, the EEF produce a Teaching and Learning Toolkit which attempts to synthesise the evidence on the effectiveness of different types of educational interventions. The effectiveness of each type of interventions is evaluated using meta-analysis in which the results from individual published studies are pooled into an overall estimate.
In meta-analysis, a common problem is that different studies assess the effectiveness of an intervention using different measures (eg. different achievement tests) which makes it difficult to compare results across studies. In order to allow comparison of results from different studies, in a meta-analysis the published data from each study are first used to calculate an effect size or standardised treatment effect which expresses the effect of the intervention in units of standard deviations. The figure below illustrates how an effect size is calculated. In this case, the children in control group have a mean score of 100 on an achievement test while the children in intervention group have a mean score of 112. The standard deviation of the test scores in each group is 15 so the difference in achievement in the two groups is equal to 0.8 standard deviation units. Although this might not look like a large effect, researchers generally consider effects of this size to have large overall impacts and nearly 80 percent of the children in the control group have lower scores than the average child in the treatment group.
The EEF Toolkit concludes that two of the most effective approaches to raising achievement are providing effective feedback on performance to children and encouraging children to reflect on their own learning strategies termed thinking skills. The EEF reaches its conclusions concerning the effectiveness of thinking skills using the results of a meta-analysis published in 2005 which conveniently includes an Appendix giving the summary statistics used in the analysis. In total the meta-analysis used results from 29 studies conducted in a range of countries. The effect size and confidence interval for the effect on children’s cognitive outcomes are plotted below for the seven studies conducted in Britain. Although the confidence intervals of the individual studies overlap there is significant variation in the effect sizes which range from 0.09 to 1.25. The studies are all relatively small, however, with the largest study only involving 188 children. The summary overall effect (0.52; 95% CI 0.29 – 0.76) is calculated as a weighted average of the individual study effects where the weights the precision with which the effect is estimated in each study. An effect of this size is roughly equivalent to the advance that the average child would make in a period of around 6 months. While this is encouraging I wouldn’t say the current evidence is exactly strong, and it will be interesting to see the results produced by the projects funded by the EEF.
Whether interventions targeted on individual children will be sufficient to improve the achievement of the poorest children is something about which I would be far less certain. As we have looked at previously, the relationships between low educational attainment and poverty have a durable character. Although it is clearly necessary to focus interventions on individual children it would seem unlikely that interventions at the child level alone can solve problems of low educational attainment, particularly for the poorest children.