It is widely assumed that some schools improve more rapidly than others. However, unlike the well‐established finding that schools differ in their effectiveness, evidence that schools improve at different rates is sparse. Using data on pupils’ examination results and prior achievements from five cohorts of pupils passing through over 30 schools in one LEA, the study considers three questions. First, the extent to which some schools improve more rapidly than others. Second, whether certain ‘types’ of school are more likely to improve. And third, whether there were any ‘strategies’ which substantial numbers of schools employed to bring about improvement. A multilevel model was employed to generate estimates of schools’ ‘effectiveness’ and ‘improvement over time’. The analysis showed that around 1 in 7 schools ‘improved rapidly’ over the 5‐year period. Both ‘less effective’ and ‘more effective’ schools’ improved. However, differences in schools’ effectiveness remained substantial when compared with the extent of improvement; it would take several years, on this evidence, for a ‘less effective’ school to move into the pack and as long again for it to pull ahead. One reason why some schools were improving more rapidly than others could be attributed to the fact that they had increased the average number of exam subjects their pupils were entered for above the rate of increase across the whole sample. Other factors not explored in this study are also likely to have been influential. [ABSTRACT FROM PUBLISHER]