A reliable assessment of clinical performance requires multiple cases, as performance varies between cases depending on previous experiences and knowledge of the case. However, behavioural attributes, including communication and teamwork, may be expected to be less dependent on specific case knowledge and thus be more stable across cases. This has implications for training and assessment design. In this study we measured the psychometric properties of assessment of behaviour in simulated anaesthetic emergencies. Twenty anaesthesia trainees were rated in three simulated emergencies by four assessors. The psychometric properties of scores for behaviour were determined and were compared with scores for medical management and overall performance. We found that scores for behaviour were less dependent on the specific clinical context than the scores for overall performance and medical management, implying transferability of crisis management behaviours between cases.