Summary.Following the work of Luchins, a test suitable for 11‐year‐olds was designed to explore rigidity in problem solving among pupils in two junior schools. One of these is based on a child‐centred, progressive regime while the other believes in a mainly adult‐directed, traditional form of organisation. The total sample consisted of 211 children. Intelligence quotients and the results of two arithmetic tests were available for every child.No over‐all difference in rigidity was found between the two schools. However, within each school, children in the lowest stream and pupils of more limited intelligence showed a significantly higher degree of rigidity. Moreover, in the progressive school, the less able children showed greater flexibility in problem solving than the comparable groups in the school using traditional teaching methods. Thus, it seems, that intelligence and educational approach have a differential effect on rigidity, rather than an “all or none” influence as has been suggest