A new test of strategy application was designed to be relatively free of the constraints that limit the standard neuropsychological assessment of supervisory abilities. The validity of the test was assessed in 3 samples of participants with varying degrees of supervisory deficits and frontal systems dysfunction: focal frontal lesions, traumatic brain injury (TBI), and normal aging. Inefficient strategy application varied systematically across the 3 groups and was not due to extraneous factors such as forgetting the test instructions. Previous case studies have emphasized strategy application deficits in the face of normal neuropsychological test performance. In this study, it was shown that strategically impaired participants from a consecutive series can include those both with and without deficient neuropsychological test performance. When neuropsychological impairment was present, it was greatest on executive functioning tasks. Among participants with nonstrategic performance, there was evidence for a dissociation of knowledge from action. This finding was not specific to focal frontal lesions. A number of supervisory processes contributing to strategy application were identified. Exploratory analyses indicated differential effects of lesion location on these processes, especially inferior medial frontal and right hemisphere lesions. Overall, the results supported the use of unstructured tasks in the assessment of supervisory abilities. (JINS, 1998, 4, 247–264.)