Categories, Controls and Ceilings

In the target paper, I outlined several methodological issues associated with attempts to document category specific deficits; and a potential solution based around certain minimal criteria. The main argument being that an accurate interpretation of patient performance requires a comparison with a normal control group and that the group performs below ceiling. Neither of these requirements is new or one would imagine, especially contentious. It is therefore surprising that a review of the category specific literature reveals no single study that meets these criteria (Laws, in this issue). Moreover, this has to be viewed in the context that the commonly used analyses (within-patient χ2 and betweensubject comparisons with controls at ceiling) produce false positive, false negative and even paradoxical deficits. [opening paragraph]