Many studies have shown the ability of media--television, movies, and virtual reality (VR) experiences--to elicit emotions. Nevertheless, it is still unclear how the different factors involved--user related and medium related--play a role in producing an emotional response during a VR experience. We investigate this issue, analyzing the role played by the cultural and technological backgrounds of the users in the emotional responses to VR. Specifically, we use the "core affect" model of emotions developed by Russell (2003) to explore how these factors influence the way in which participants experience virtual worlds. Our sample includes 20 Mexican participants: 8 living in El Tepeyac, a small rural and isolated Mexican village characterized by a very primitive culture, and 12 high civilized inhabitants of Mexico City. The "Green Valley," a noninteractive, relaxing immersive environment showing a mountain landscape around a calm lake, was used to induce relaxation in the two groups during an ambulatory surgical operation. To investigate the effects of VR on the relaxation process, we measured participants' physiological (heart rate) and emotional (VAS-A) responses before, during, and after the operation. The results show that VR significantly modified the core affect (reduced arousal) in all participants but that the final emotional response produced by this change was influenced by the attribution process: the civilized inhabitants of Mexico City, who were able to attribute the reduced arousal to the VR experience, reported a significant reduction in the self-reported level of anxiety, while people from El Tepeyac showed a reduction in their physiological reactions but not in their perceived anxiety.