This study aims to determine effective school and class processes by analysing data of the European Science Foundation project in which 1,927 students were followed during fourth grade. There was an assessment of scientific literacy at the beginning and at the end of the school year, enabling us to estimate learning gains and to investigate value-added differences between schools and classes in Flanders. Using variables from student, teacher and headmaster questionnaires in multilevel regression-analyses, we answer the question how these differences in effectiveness and equity can be explained. Our focus is threefold. (1) We estimate the effect of socio-economic and ethnic group composition on learning gains. (2) We determine the extent to which schools and teachers succeed in narrowing the achievement gap between high and low-risk students (in terms of SES and ethnicity). (3) We test the advantages of the Dynamic Model of Effectiveness and compare our results with the results of SiBO, a longitudinal study in primary education in Flanders. SiBO revealed no effects on learning gains of neither the students’ at-risk status nor the group composition, indicating that the achievement gap between high and low-risk students arises at an early stage and does not grow during primary education.