It is sometimes argued that implementation of an overall similarity classification is less effortful than implemen- tation of a single-dimension classification. One piece of evidence taken to be in support of this argument is that highly impulsive individuals appear to be more likely to sort on the basis of overall similarity than individuals with low impulsivity (Ward, 1983); presumably, higher impulsivity results in lower effort. In the current arti- cle, we identify some limitations in Ward’s procedure and, using a more standard measure of impulsivity and a less ambiguous measure of overall similarity classifica- tion, re-investigate the relationship between impulsivity and overall similarity classification. Using a match-to- standard procedure, the current experiment finds that overall similarity classification is less prevalent in highly impulsive individuals. The implications of this result, which is opposite to that reported by Ward (1983), are discussed.