The role of school governor requires a wide range of sophisticated skills, among them understanding their school's performance and evaluating policy and evidence for interventions. However, with limited statistical skills on governing boards, is there a way in which statisticians could help interpret the often large amounts of data that is presented to them?
This was the key question asked in this session chaired by Harvey Goldstein, who is renowned for his statistical work in educational assessment and school league tables.
Session organiser Tom King of Newcastle University kicked-off proceedings by suggesting that there is potential for the RSS to work with Ofsted regarding its data output. At the moment there are two key sources of data measuring schools' progress; the publicly available Data Dashboard and RAISEonline, which is not a public document but is available to school governors. Tom pointed out that on the Data Dashboard, context (such as ethnicity) is only included after data on value-added, attainment and attendance. He also pointed out that the RAISEonline document runs to some 70 pages with little assistance to prioritise. While some governors will have a good understanding of how to interpret the data, the report is not structured to be helpful and governors may not be picking out the most useful information. In this aspect, he said, the RSS should look to engage.
Nancy Wilkinson, who leads the Wellcome Trust's work in school governance, stressed that governors are volunteers, often not from scientific backgrounds; therefore they might not have the right skills to identify issues in data sets such as RAISEonline in order to hold headteachers to account. Statisticians, she said, would make excellent governors and she encourages them to volunteer. It’s important, for example, that governors understand variations in the data, such as cohort changes, that can affect results.
Neil Sheldon, the Royal Statistical Society’s honorary officer for education, highlighted the types of statistical concepts that many leaders - and providers of data - lack knowledge of. These include the use of the word ‘significance’, confidence intervals, regression to the mean, difference in estimation for individuals and groups. Variability in general, he said, is often misunderstood. The RSS education committee sees this as an area that RSS fellows can do more to influence.
Ros Sutherland of Bristol University has seen the issue from both perspectives - as an educationalist and governor. It is eye-opening, she said, to see how extensively data are used in schools and the extent to which it is scrutinised by school leadership. Sometimes, however, this scrutiny is insensitive to variation and instability in the population. While in her school, focusing on this data had clearly improved results, it was now encouraging rote learning in some lessons which was leading to problems for those moving from GCSE to A level.
Much of the discussion that followed focused on how the RSS might have a role to play in educating school governors and improving understanding of education data. The RSS’s successful journalism training programme might be expanded in this area, perhaps in collaboration with Ofsted and others, where inspectors and school leaders could benefit from training.
RSS representatives present at the meeting felt that there was scope for fellows to engage in this issue and perhaps form a working group. Potential actions could be to seek funds to report on this issue, perhaps in collaboration with others, and proactively engage with educational organisations such as Ofsted and others.