The purpose of the experiment reported was to investigate the effect of a mathematics intervention on counting strategies and levels of performance of young children with special educational needs. For six months, 32 children participated in the intervention. They were compared with 26 children in a non-intervention control group. Observation of four types of tasks took place at two, four and six months after the intervention had begun. Results suggested that the particular instructional programme used with the experimental group was no better than the teaching programmes used by teachers in the control group. These results may, however, be explained by ceiling effects. Implications for future research are discussed.