Abstract
This article reports the results of analysing data from 2,583 children aged around 5 years in their 1st year of schools in the city of Edinburgh. The best predictor of attainment at the end of the year was attainment at the start of the year but a number of other variables, such as entitlement to free school meals and special needs, also had significant explanatory power. A multilevel analysis applied to the data found that class effects were more important than school effects. The policy implications of the findings for school improvement and monitoring are discussed.