BACKGROUND Various outcome measures are used for the assessment of balance and mobility in patients with stroke. The purpose of the present study was to examine test-retest reliability, construct validity, and responsiveness of the Timed Up and Go Test (TUG), Berg Balance Scale (BBS), and Dynamic Gait Index (DGI) for measuring balance in patients with chronic stroke. METHODS Fifty-six patients (39 male and 17 female) with chronic stroke participated in this study. A senior physical therapist assessed the test-retest reliability and validity of three scales, including the DGI, TUG, and BBS over two testing sessions. In addition, the third assessment of each scale was taken at the time of discharge to determine the responsiveness of the three outcome measures. RESULTS The reliability of the TUG (intraclass correlation coefficient [ICC2,1] = 0.98), DGI (ICC2,1 = 0.98) and BBS (ICC2,1 = 0.99) were excellent. The standard error of measurement (SEM) of the TUG, DGI, and BBS were 1.16, 0.71, and 0.98, respectively. The minimal detectable change (MDC) of the TUG, DGI, and BBS were 3.2, 1.9, and 2.7, respectively. There was a significant correlation found between the DGI and BBS (first reading [r] = 0.75; second reading [r] = 0.77), TUG and BBS (first reading [r] = -.52; second reading [r] = -.53), and the TUG and DGI (first reading [r] = 0.45; second reading [r] = 0.48), respectively. CONCLUSIONS The test-retest reliability of the TUG, BBS, and DGI was excellent. The DGI demonstrated slightly better responsiveness than TUG and BBS. However, the small sample size of this study limits the validity of the results.