Reviews of the dyslexia literature often seem to suggest that children with dyslexia perform at a lower level on almost any task. Richards et al. (Dyslexia 2002; 8: 1-8) note the importance of being able to demonstrate dissociations between tasks. However, increasingly elegant experiments, in which dissociations are found, almost inevitably find that the performance of children with dyslexia is lower as tasks become more difficult! By looking for deficits in dyslexia, could we be barking up the wrong tree? A methodological approach for circumventing this potential problem is discussed.