Background The Active Movement Scale is a frequently used outcome measure for children with obstetric brachial plexus palsy (OBPP). Clinicians observe upper limb movements while the child is playing and quantify them on an 8 point scale. This scale has acceptable reliability however it is not known whether it accurately depicts the movements observed. In this study, therapist-rated Active Movement Scale grades were compared with objectively-quantified range of elbow flexion and extension and shoulder abduction and flexion in children with OBPP. These movements were chosen as they primarily assess the C5, C6 and C7 nerve roots, the most frequently involved in OBPP. Objective quantification of elbow and shoulder movements was undertaken by two-dimensional motion analysis, using the v-scope. Methods Young children diagnosed with OBPP were recruited from the Royal Children's Hospital (Melbourne, Australia) Brachial Plexus registry. They participated in one measurement session where an experienced paediatric physiotherapist facilitated maximal elbow flexion and extension, shoulder abduction and extension through play, and quantified them on the Active Movement Scale. Two-dimensional motion analysis captured the same movements in degrees, which were then converted into Active Movement Score grades using normative reference data. The agreement between the objectively-quantified and therapist-rated grades was determined using percentage agreement and Kappa statistics. Results Thirty children with OBPP participated in the study. All were able to perform elbow and shoulder movements against gravity. Active Movement Score grades ranged from 5 to 7. Two-dimensional motion analysis revealed that full range of movement at the elbow and shoulder was rarely achieved. There was moderate percentage agreement between the objectively-quantified and therapist-rated methods of movement assessment however the therapist frequently over-estimated the range of movement, particularly at the elbow. When adjusted for chance, agreement was equal to chance. Conclusion Visual estimates of elbow and shoulder movement in children with OBPP may not provide true estimates of motion. Future work is required to develop accurate, clinically-acceptable methods of quantifying upper limb active movements. Since few children attained full range of motion, elbow and shoulder movement should be monitored and maintained over time to reduce disability later in life.