Objectives: To compare clinical knee arthrometer tests with theoretical force–displacement curves as a potential source of test error. Design: A single auditor reviewed 280 KT-2000 knee arthrometer tests conducted by six physical therapists. Setting: A University Sport Medicine Center. Participants: Six physical therapists with at least 3 years experience tested 280 individuals with suspected or confirmed ACL injury referred for arthrometer evaluation of knee laxity. Main outcome measures: An independent test auditor, using a priori rules, evaluated computer-acquired measurements of knee ligament laxity. The auditor judged the test quality as technically correct, indefinite or incorrect based on the theoretical force–displacement curve. A Chi-square analysis was used to assess the statistical significance of the frequency distribution of test quality. Results: The highest percentage of technically correct tests for a single physical therapist was 71% while the lowest was 21%. There was a statistically significant difference between testers in the percentage of technically correct tests. Conclusions: Diagnosis of ACL injury and outcomes following reconstructive surgery have been evaluated with quantitative measurement of sagittal plane knee laxity. The technical quality of arthrometer tests may be a potential source of error in clinical evaluation and research studies. In this study, physical therapists demonstrated discrepancies in the frequency of valid tests when compared to theoretical force–displacement curves.