Sarah Atkins, Celia Roberts, Kamila Hawthorne, Trisha Greenhalgh

Assessment of consulting skills using simulated patients is widespread in medical education. Most research into such assessment is sited in a statistical paradigm that focuses on psychometric properties or replicability of such tests. Equally important, but less researched, is the question of how far consultations with simulated patients reflect real clinical encounters - for which sociolinguistics, defined as the study of language in its socio-cultural context, provides a helpful analytic lens. In this debate article, we draw on a detailed empirical study of assessed role-plays, involving sociolinguistic analysis of talk in OSCE interactions. We consider critically the evidence for the simulated consultation (a) as a proxy for the real; (b) as performance; (c) as a context for assessing talk; and (d) as potentially disadvantaging candidates trained overseas. Talk is always a performance in context, especially in professional situations (such as the consultation) and institutional ones (the assessment of professional skills and competence). Candidates who can handle the social and linguistic complexities of the artificial context of assessed role-plays score highly - yet what is being assessed is not real professional communication, but the ability to voice a credible appearance of such communication. Fidelity may not be the primary objective of simulation for medical training, where it enables the practising of skills. However the linguistic problems and differences that arise from interacting in artificial settings are of considerable importance in assessment, where we must be sure that the exam construct adequately embodies the skills expected for real-life practice. The reproducibility of assessed simulations should not be confused with their validity. Sociolinguistic analysis of simulations in various professional contexts has identified evidence for the gap between real interactions and assessed role-plays. The contextual conditions of the simulated consultation both expect and reward a particular interactional style. Whilst simulation undoubtedly has a place in formative learning for professional communication, the simulated consultation may distort assessment of professional communication These sociolinguistic findings contribute to the on-going critique of simulations in high-stakes assessments and indicate that further research, which steps outside psychometric approaches, is necessary.