Four different groups of young subjects (Mental Retardation, Learning Difficulties and two Control groups) were tested in a particular type of dual-task paradigm. The primary task was a speeded left-right manual response to stimulus position (R1). The unspeeded secondary task consisted of reporting verbally whether the two letters that composed each stimulus were the same or different (R2). Each participant was tested in two experimental conditions. In condition A he had only to respond to the stimulus position (R1), whereas in condition B he had to respond to the stimulus location and than to say aloud if the two letters composing the stimulus were the same or different (R1 and R2). It was assumed that the slower RTs to the spatial discrimination in condition B respect to condition A are a measure of the presence of the Central Executive to permit queueing the two different responses. The results suggest that the temporal cost due to this process is higher in the Mental Retard and in Learning Difficulties groups and lower in the Control groups. We conclude that the two experimental groups have a deficit in the strategic control processes.