School value added scores are unreliable
The value added score published annually for all secondary schools in England is an unreliable indicator of a school’s performance, and should not be used to compare one school with another, according to Professor Jim Taylor of Lancaster University.
The research, carried out in the Department of Economics, Lancaster University Management School, shows that a school’s value added score is only partly determined by factors within the school’s control.
The school performance tables, first published in 1992, originally used the proportion of pupils obtaining five or more A to C grades in the GCSE exams as a measure of each school’s performance. This has no value as a performance indicator since a school’s GCSE results are determined primarily by the ability and family background of its pupils. Some schools are more able to attract pupils with high levels of ability or who have parents who are more supportive of their child’s education.
Since 2002, the School Performance Tables have included a value added score, which is designed to measure the amount of knowledge gained by pupils during their secondary school years. By taking the educational attainment of its pupils at the end of their primary education into account, the DfES had hoped that the new value added measure would be a useful indicator of each school’s performance. This is not the case.
The new research, which is based on DfES data for all state-maintained secondary schools in England, challenges the Government’s view that the value added score is a valuable performance indicator.
It shows schools with a high proportion of pupils from poorer families, for example, are more likely to have a lower value added score compared to schools with a high proportion of pupils from better off families.
Many schools with low value added scores may not therefore be performing badly. Some schools with low value added scores are in fact performing well, given the social and economic background of their pupils.
Other factors affecting a school’s value added score include its gender mix. Girls-only schools have a substantially higher value added score than either boys-only or co-educational schools. There is also evidence that having a subject specialism is associated with a higher value added score, but this may be because schools with a high value added score are more likely to acquire specialist status. Not surprisingly, a school’s authorised absence rate also has a detrimental effect on its value added score.
A school’s value added score therefore depends as much on the personal characteristics and family background of its pupils as on its own efforts to add value to its pupils. These factors need to be incorporated into the value added measure if the school’s own contribution to the value added to its pupils is to be accurately measured.
Professor Taylor said: “We need a performance indicator which actually measures the performance of the school and not the performance of its pupils and their parents. We need a far more accurate way of measuring performance than is currently available.
“A low value added score may simply indicate that a school has a high proportion of disadvantaged pupils. Many schools are being wrongly labelled as ‘poor’, with the consequent harmful effects on pupil and teacher morale.”
Although the published value added scores may not be useful for measuring the performance of schools, it is nevertheless valuable to measure and record the exam performance of pupils at each Key Stage. According to Professor Taylor, the measurement and statistical analysis of each pupil’s academic progress is absolutely essential if appropriate action is to be taken to improve educational outcomes. This is especially crucial for those social groups who are consistently underperforming relative to their potential.
[A paper describing this research is to be published in the Oxford Bulletin of Economics and Statistics. A copy of the Working Paper is available on request from Professor Jim Taylor, Department of Economics, Management School, Lancaster University, Lancaster. Email: firstname.lastname@example.org. Tel: Work: 01524-594227; Home: 01524-770644]