Making a meal of tests;Letter
Research has shown that the most appropriate way of comparing schools' performance is in terms of the progress made by pupils, taking account of prior attainment. Performance indicators based on attainment at school-level (league tables) are inherently unreliable and misleading.
This problem is even greater in the case of 5-14 attainment which is a curious mixture of teachers' subjective judgments and confirmatory national tests. Assessment and testing in primary schools is not intended for comparisons between schools and is therefore not carried out on consistent and comparable basis.
FME is not a consistent measure of schools' characteristics and does not provide an appropriate measure to take account of differences in school intake. My research shows only a weak correlation between FME and baseline attainment in reading and mathematics in primary schools (-0.53).
FME is associated with local economic conditions including unemployment and deprivation. Significant random differences in FME between secondary schools show it is an inconsistent measure for contextualising school comparisons.
Statistical performance indicators should not be regarded as conclusive evidence of schools' effectiveness, but should be considered in the light of other evidence. Elizabeth Maginnis was surprised at the "startling variations" in the data on Edinburgh schools. If the statistical evidence contradicts other evidence of school performance this should alert her to the need to question the validity of the statistics.
Linda Croxford Centre for Educational Sociology University of Edinburgh