Purpose: To identify stroke patients showing a match between an objectively meaningful improvement in upper-limb motor capacity and subjectively meaningful improvement in upper-limb performance after stroke, as well as determinants of matching and non-matching. Methods: Changes in upper-limb capacity in 39 stroke patients were measured over 17 weeks using the Action Research Arm Test (ARAT). Subjective change was measured with the Stroke Impact Scale Hand domain (SIS-Hand) and the Motor Activity Log (MAL). Patients showing a match between objective and subjective improvements (matchers) and patients showing a mismatch (non-matchers) were identified. Subsequently, determinants for the matchers were modeled using a logistic regression analysis. Results: A significant association was found between improvements on the ARAT and MAL (p=0.011), but not for improvements on the ARAT and SIS-Hand. For the ARAT and SIS-Hand level of education and mood significantly improved the model. The probability of a match between objective and subjective outcomes in lower-educated patients with a positive mood was estimated to be 0.83, whereas in higher-educated patients with a negative mood it was estimated to be 0.07. Conclusions: Mismatches between objective and subjective outcomes are relevant for practitioners evaluating rehabilitation goals, and for defining outcome variables in future upper limb rehabilitation trials.Implications for RehabilitationIn upper limb rehabilitation after stroke, a match between capacity and self-perception outcome measures is not self-evident.Regarding these potential mismatches, practitioners and patients should consider carefully which goals to set, and how to evaluate upper limb rehabilitation after stroke; preferably with capacity measures and self-perception measures combined. © 2013 Informa UK Ltd. All rights reserved.