A. Marcus (Ed.): DUXU/HCII 2013, Part III, LNCS 8014, pp. 362–371, 2013. © Springer-Verlag Berlin Heidelberg 2013 Sense of Presence in a VR-Based Study on Behavioral Compliance with Warnings Emília Duarte 1 , Francisco Rebelo 2,4 , Luís Teixeira 2,4 , Elisângela Vilar 2,4 , Júlia Teles 3,4 , and Paulo Noriega 2,4 1 Unidcom, IADE – Creative University. Av. D. Carlos I, 4, 1200-649 Lisbon, Portugal emilia.duarte@iade.pt 2 Ergonomics Laboratory, Faculdade de Motricidade Humana, Universidade Técnica de Lisboa {frebelo,lteixeira,elivilar,pnoriega}@fmh.utl.pt 3 Mathematics Unit, Faculdade de Motricidade Humana, Universidade Técnica de Lisboa jteles@fmh.utl.pt 4 CIPER – Interdisciplinary Center for the Study of Human Performance, Universidade Técnica de Lisboa, Estrada da Costa, 1499-002 Cruz-Quebrada, Dafundo, Portugal Abstract. Recent researches suggest that Virtual Reality (VR) is amongst the best tools for examining behavioral compliance with warnings, therefore over- coming some ethical and methodological constrains that have been limiting this type of research. Yet, such evaluation using VR requires both usable and engag- ing virtual environments (VEs). This study examines the sense of presence ex- perienced by the participants after having been immersed in a VE designed for evaluating the effect of sign type (static vs. dynamic) on compliance. The VR simulation tested here allowed participants to perform a realistic work-related task and an emergency egress, during which they were supposed to interact with warnings and exit signs. A neutral condition (i.e., no/minimal signs) was used as a control condition. Subjective and objective data were gathered from two sources, respectively, i.e., a post-hoc questionnaire administered to the partici- pants, and a video analysis of the participants’ interaction behavior during the VR simulation. Results reveal high levels of presence across the three experi- mental conditions. Keywords: Virtual Reality, Presence, Behavioral Compliance, Warnings. 1 Introduction Behavioral compliance with warnings generally requires that people take some sort of action. Thus, evaluating compliance involves observing what people do, i.e., if indi- viduals carry out the warning-directed behavior. Although behavior is considered one of the most important measures of warning effectiveness, it is usually quite difficult to conduct behavioral tests [1]. The reasons include, among others: ethical and safety concerns (research participants cannot be intentionally exposed to real hazards), rarity and unpredictability of the hazardous events, difficulty of creating scenarios that mimic real-life situations which appear to be risky yet are safe, and the costs in terms