Opinion

“If I were education secretary I would…”

I would publish more data and set fewer targets. I would make it easier for research organisations to access the treasure trove of information on the national pupil database. And I would fund education statisticians to run workshops around the country for interested teachers and parents who want to understand how they can use this data.

I would do all this for two reasons: first, focusing narrowly on one target — or “small data” — has been shown to have damaging unintended consequences and, second, greater access to “big data” has the potential to transform teaching and learning.

Collecting, collating and publishing accurate and reliable statistics on schools is one of the most vital jobs of government — and a job few other institutions have the power or capacity to do. Although it may feel like the government already collects vast amounts of data from schools, a lot of what it collects is, in fact, designed to create targets.

Collecting, collating and publishing accurate statistics on schools is vital

For example, it is easy to find out how many pupils in certain schools got five A*–C at GCSE including English and maths, but the government does not collect statistics on how many pupils got A*–A. It is also hard to see breakdowns by individual subject area. Why not publish more data, but leave it up to individuals and non-governmental organisations to decide what’s important? When the government sets the targets itself, unintended consequences abound. Relentless focus on its preferred five A*–C model has led to schools chasing that target to the exclusion of all else. Schools who excel in other areas often find the data that would prove this is not collected or published.

By contrast, when external organisations, such as Education Datalab, the Open Public Services Network and the Education Endowment Foundation, analyse data, they often come up with insightful and unexpected analyses. Why not make their job easier by publishing even more data? This would also make it harder for schools to game any one particular measure, as attempts to focus exclusively on one measure would be shown up in others.

This would be particularly valuable at primary level. Currently primaries are going to be judged on the numbers of pupils achieving a certain score in their end of key stage 2 tests. It is easy to predict what impact this will have: schools will focus on those pupils at or around that arbitrary cut-off mark. Instead, why not just publish as much information about the results of the test as possible, and let schools and other organisations decide what matters most?

Daisy Christodoulou is author of Seven Myths About Education (Routledge, 2014) @daisychristo

Latest education roles from

Systems and Reporting Manager

Systems and Reporting Manager

Richmond and Hillcroft Adult & Community College

Funding and Planning Manager

Funding and Planning Manager

Richmond and Hillcroft Adult & Community College

Head of Management Information Services and Funding

Head of Management Information Services and Funding

Richmond and Hillcroft Adult & Community College

Associate Assistant Principal: English, Maths or Science

Associate Assistant Principal: English, Maths or Science

Harris Federation

Teacher of Science

Teacher of Science

Harris Academy Clapham

Pastoral Manager

Pastoral Manager

Harris Academy Clapham

Your thoughts

Leave a Reply

Your email address will not be published. Required fields are marked *

One comment

  1. Absolutely, 100% agree, very achievable too, Statistical First Release is archaic and hard to use, RAISE also out of date and not queryable. Queryable Web portal would be easy and relatively cheap with current tools like sequal server.
    100% agree a broad set of measures would dilute gaming.
    Governors and those really interested in schools always ask for a broad range of measures.
    Unfortunatly DfE have mooted cutting data services not expanding them and the new parliment will see cuts. Plus DfE hae scapped all ambitious data projects. There appears no reason small feasablily pilots could not be outsourcesd.
    To me this is silly as a broader data portal would be an easy win and relatively cheap.