Although high levels of reliability are emphasized in the construction of many measures of psychological traits, tests that are intended to measure patient change following psychotherapy need to emphasize sensitivity to change as a central and primary property. This study proposes 2 criteria for evaluating the degree to which an item on a test is sensitive to change: (a) that an item changes in the theoretically proposed direction following an intervention and (b) that the change measured on an item is significantly greater in treated than in untreated individuals. Outcome Questionnaire (Lambert et al., 1996) items were subjected to item analysis by examining change rates in 284 untreated control participants and in 1,176 individuals undergoing psychotherapy. Results analyzed through multilevel or hierarchical linear modeling suggest the majority of items on this frequently used measure of psychotherapy outcome meet both criteria. Implications for test development and future research are discussed.