BACKGROUND Shared mental models are crucial for constructing mutual understanding of the patient's condition during a clinical handover. Yet, scant research, if any, has empirically explored mental models of the parties involved in a clinical handover. OBJECTIVES This study aimed to examine the similarities among mental models of incoming and outgoing nurses, and to test their accuracy by comparing them with mental models of expert nurses. DESIGN A cross-sectional study, exploring nurses' mental models via the concept mapping technique. PARTICIPANTS 40 clinical handovers. DATA COLLECTION Data were collected via concept mapping of the incoming, outgoing, and expert nurses' mental models (total of 120 concept maps). Similarity and accuracy for concepts and associations indexes were calculated to compare the different maps. RESULTS About one fifth of the concepts emerged in both outgoing and incoming nurses' concept maps (concept similarity=23%±10.6). Concept accuracy indexes were 35%±18.8 for incoming and 62%±19.6 for outgoing nurses' maps. Although incoming nurses absorbed fewer number of concepts and associations (23% and 12%, respectively), they partially closed the gap (35% and 22%, respectively) relative to expert nurses' maps. The correlations between concept similarities, and incoming as well as outgoing nurses' concept accuracy, were significant (r=0.43, p<0.01; r=0.68 p<0.01, respectively). Finally, in 90% of the maps, outgoing nurses added information concerning the processes enacted during the shift, beyond the expert nurses' gold standard. DISCUSSION AND CONCLUSIONS Two seemingly contradicting processes in the handover were identified. "Information loss", captured by the low similarity indexes among the mental models of incoming and outgoing nurses; and "information restoration", based on accuracy measures indexes among the mental models of the incoming nurses. Based on mental model theory, we propose possible explanations for these processes and derive implications for how to improve a clinical handover.