BACKGROUND: Three-dimensional motion analysis has provided important information about functional outcome for patients who have undergone total knee replacement, yet many comparisons to control populations are inconclusive. Some of these findings may be explained by variability in patients' repeated performance, however this has not been previously reported. Therefore the objective of this study was to determine the variability of repeated performance in patients who have undergone total knee replacement during walking and selected activities of daily living using three-dimensional motion analysis. METHODS: The variability of kinematic and kinetic parameters over six repetitions of walking at comfortable and fast speeds, lunging, stair ascent and descent, squatting and standing up from sitting were measured using motion analysis. Variability was calculated using both the co-efficient of variation, and the group average of the range of data within each subject. FINDINGS: During level walking the within-subject difference was less than 3.2 degrees or 0.9 Nm/kg and co-efficients of variation ranged from 2% to 29%. When patients performed other activities of daily living the within-subject difference was up to 8.2 degrees or 1.3 Nm/kg and co-efficients of variation were between 0.5% and 50.5%. INTERPRETATION: Although patients walked and performed the activities with reasonable consistency, the variability of their performance appeared to be related to the size of the group mean for each parameter. Comparison of patient groups on parameters with small mean values requires larger sample sizes to prevent intra-subject variability from obscuring differences between groups.