This paper presents results of a study of the effectiveness of schools
in teaching reading. Some 2,500 pupils in primary schools were given
a standard test at ages six and eight, and their progress compared ove
r the period. A number of other school-level and pupil-level variables
were collected and sieved to see which had a relation with reading pr
ogress. After allowing for the measures found to be effective in the p
reliminary sieving process using multilevel modelling, the measure of
school effectiveness thus formed was compared with raw mean scores for
schools, and substantial differences were found between the two ranki
ngs. The method used gives a fairer comparison of effectiveness betwee
n schools than does comparison of the raw mean results, and may be a p
articularly attractive proposition since it uses data already availabl
e from administrative sources.