Background and Purpose: Regarding people with stroke, simple outcome measures in clinical settings capable of representing the actual use of the upper extremity (UE) would be useful to rehabilitation professionals for the purposes of goal setting. This study seeks to describe the relative levels of paretic UE use, investigate the association between manual dexterity and task-related UE use, and to establish the manual dexterity cutoff points that correspond to relative levels of paretic UE use. Methods: Forty-six adults with chronic hemiparesis participated in this cross-sectional study. Behavioral mapping was employed to ascertain the actual amount of UE use by the identifying the unimanual and bimanual activities performed in the participants' homes within a 4-hour period. Participants were classified into 4 levels of paretic UE integration into activities considering the data from the behavioral mapping (activity, hand function, and type of grasp). The Box and Block Test (BBT) and the Nine Hole Peg Test (NHPT) were used to evaluate dexterity. The Spearman test was used to evaluate the correlations. In analyzing the receiver operating characteristic curve, we applied the Youden index to determine the cutoff points. Results: Participants with full/almost full (n = 11), partial (n = 12), and limited (n = 12) integration of the paretic UE into activities and with little/no use (n = 11) were identified. Unimanual and total paretic UE activities were found to have a high correlation with the BBT scores. The boundaries between the integration levels were between full/almost full and partial integration, BBT greater than 30 blocks or NHPT of 41 seconds and less; between partial and limited, BBT greater than 16 blocks; and between limited and little/no use, BBT greater than 3 blocks. Both tests show good accuracy (≥0.81).