Objective To assess the relationship between state anxiety and performance on neuropsychological tests in older adults. Methods Nine hundred fifty-five community-dwelling individuals without dementia age 66 and over were evaluated at home by a psychologist. State anxiety was measured by the State-Trait Anxiety Inventory Y. Cognitive assessment included general cognitive functioning (Mini-Mental State Examination), verbal fluency (Isaacs Set Test), short-term visual memory (Benton's Visual Retention Test), speed of information processing/visuomotor coordination (Digit Symbol Coding), conceptual knowledge (Similarities), episodic memory (Verbal Paired Associates), and working memory (Digit Span forward/backward). Covariates included age, education, sex, depressive symptoms (Center for Epidemiologic Studies–Depression Scale), subjective health, subjective cognitive complaint, chronic diseases, functional abilities in basic and instrumental activities of daily living, and use of medication. Results Adjustments for confounders substantially modified the relationship between state anxiety and cognitive performance. Multivariate analyses revealed positive effects of mild and moderate state anxiety for verbal fluency and general cognitive functioning, respectively. High and moderate anxiety also had beneficial influence on short-term visual memory performance in participants with low education level and on the speed of information/visuomotor coordination processing in participants using medications. Conclusions These results suggest that when confounders are taken into account, state anxiety in older adults is not necessarily deleterious for cognitive performance and has no appreciable negative effect on many cognitive domains or can even be beneficial. Relationships between state anxiety and cognitive performances are complex because they are influenced by many factors and differ according to anxiety severity and cognitive domains.< Réduire