The article discusses psychological issues of the role repertoire of university students in a situation of danger. Topicality of the problem is related to the wide range of threats a group may face, and its low scientific knowledge. The purpose of the research presented in the article is to study the nature of the distribution of roles in the group when solving security problems. The main research hypothesis is the assumption that the need to overcome danger may contribute to the repertory and functional specialisation of roles in the group caused by the fact and type of danger. Achieving this goal required an experiment. When organising it, ethical restrictions related to the inadmissibility of purposeful placement of a person in a situation of threat to life and health were taken into account. This determined the use of role-playing capabilities as the basis of the experiment. 10 student groups (10≤n≤15) with a high level of group cohesion according to the Carl Emil Seashore’s index were involved in the experiment. Each group was offered three dangerous situations, differing in the “field” and “effect” of danger, and its participants were required to discuss, play and reflect the repertoire and functional content of the necessary roles. Experts were engaged to analyse the empirical data. According to the results of the study, two types of roles demanded by a dangerous situation were identified – those aimed at overcoming the threat (initiator, expert, developer, coordinator, controller, commentator) and those aimed at supporting group members (optimist, nihilist, conformist, dogmatist). It is established that danger by its very fact specialises the functions of group roles. Each type of hazard increases the significance of certain roles. Regardless of the type of danger, the group is most in demand for the roles of optimist, conformist and developer. The study is promising to continue on other sample compositions and arsenals of dangerous situations.