Objectives: This study aimed to determine the impact of giving junior medical students control over the level of emotion expressed by a simulated patient (SP) in a teaching session designed to prepare students to handle emotions when interviewing real patients on placements. Methods: Year 1 medical students at Keele University School of Medicine were allowed to set the degree of emotion to be displayed by the SP in their first 'emotional interview'. This innovation was evaluated by mixed methods in two consecutive academic years as part of an action research project, along with other developments in a new communications skills curriculum. Questionnaires were completed after the first and second iterations by students, tutors and SPs. Sixteen students also participated in evaluative focus group discussions at the end of Year 1. Results: Most students found the 'emotion-setting switch' helpful, both when interviewing the SP and when observing. Student-interviewers were helped by the perception that they had control over the difficulty of the task. Student-observers found it helpful to see the different levels of emotion and to think about how they might empathise with patients. By contrast, some students found the 'control switch' unnecessary or even unhelpful. These students felt that challenge was good for them and preferred not to be given the option of reducing it. Discussion: The emotional level control was a useful innovation for most students and may potentially be used in any first encounter with challenging simulation. We suggest that it addresses innate needs for competence and autonomy. The insights gained enable us to suggest ways of building the element of choice into such sessions. The disadvantages of choice highlighted by some students should be surmountable by tutor 'scaffolding' of the learning for both student-interviewers and student-observers.