The Singapore bar-model has recently arrived in many Spanish classrooms, where it has been identified as an effective modelling tool for problem-solving in primary schools. This model introduces a pictorial approach prior to the abstract symbolic resolution. Schools usually change over to this scheme by introducing it in grade 1 (6/7 years old) and progressing yearly with the children. In this study we perform a pre-post analysis of 71 fifth graders. An experimental group of 16 children were introduced to the Singapore bar-model, while the control group, the remaining 55 children, solved the exact same set of problems using their usual symbolic-only approach. Our results show statistically significant improvements in the performance of the experimental and the control groups in the cognitive domains of "knowing" and "applying" as well as in the content domain of geometry.