The Berg Balance Scale (BBS) was designed to help determine change in functional standing balance over time. The purpose of this paper was to estimate the minimum detectable change score (MDC) using the standard error of measure (SEM), thereby providing a means to decide if genuine change had occurred. Calculation of the agreement regarding the presence of change as determined by the MDC and clinicians' perceptions was performed to give an indication of the validity of this criterion value. Forty-eight subjects who were receiving inpatient rehabilitation after stroke were assessed on consecutive days by two raters using the BBS. The MDC analysis suggests that a change of +/- 6 BBS points is necessary to be 90% confident of genuine change. Only 25/45 subjects showed agreement between the statistically derived presence of change and clinicians' perceptions of change. The lack of agreement may relate to the validity of the SEM/MDC methodology to determine the criterion BBS value, the heterogeneity of the subjects, or the use of clinician gestalt impressions of change.