As the pressures on hospitals mount to keep costs down while maintaining the quality of care, it becomes increasingly important to document the outcome of inpatient hospital course. A new scale has been developed for the concurrent evaluation of function during hospitalization after total hip arthroplasty. The initial development and face validity of this scale have been previously reported. The purpose of this study was to report the interobserver reliability of the scale. Each patient was evaluated to determine the level of assistance (assisted or unassisted) required to perform 4 milestones of functional recovery: transferring, ambulating using a walker, ambulating using crutches, and stair negotiation. Patients were judged on the performance of each function and were classified as "unassisted" if they did not require the presence of another person to perform the function. Twelve patients were evaluated simultaneously by 2 physical therapists who remained blinded to each other's evaluation. Agreement was assessed using the kappa coefficient of concordance ([kappa]) for binary data. The following guidelines have been suggested for the strength of [kappa]: poor: 0, slight: 0 to .20, fair: 0.21 to 0.40, moderate: 0.41 to 0.60, substantial: 0.61 to 0.80 and almost perfect: 0.81 to 1.00. There were 86 pairs of observations for all the functions, with 31, 32.17, and 16 pairs for transferring, using a walker, using crutches, and climbing stairs, respectively. Agreement was "almost perfect" overall (k = 0.86) and for transferring (k = 0.93). Agreement was "substantial" for using a walker (k = 0.82), using crutches (k = 0.75), and climbing stairs (k = 0.75). This data establishes the interobserver reliability of a scale for functional evaluation immediately after total hip arthroplasty. Reliability testing is necessary to standardize outcome tools.