Accounting for previous performance of students by means of growth curves analyses to estimate the size, stability and consistency of school effects

The current study explored the size, stability, and consistency of school effects, using two different effectiveness indicators, being the achievements of students at the end of primary school (gross school effects) and the growth in achievement across three years of schooling (value added effects). The empirical data are from a Dutch data set including the scores of 25,269 students on Reading comprehension, Spelling, and Mathematics tests, taken in grade 4, 5 and 6 among three cohorts of students in 319 primary schools. Multivariate multilevel growth curve modelling was applied in which the measurements (level 1) were nested within students (level 2), which were nested in school-cohorts (level 3), which were nested in schools (level 4). The results showed that 1) for students’ growth of achievement the relative proportion of variance at the school level seemed larger compared to achievement at the end of primary school; 2) the total variance in growth was substantially smaller compared to variance in achievement at the end of primary school; 3) school effects for growth are less stable across different cohorts of students than school effects established at a particular point in time; and 4) school effects for growth over time are less consistent across multiple subject domains than school effects as indicated by students’ achievement at a particular moment. Differences with findings from previous research are discussed.