Thirty-five adult participants tested the hypothesis that the mere belief in having a social interaction with someone improves learning and understanding. Participants studied a passage on the mechanism that causes a fever. They then entered a virtual reality environment with an embodied agent on the other side of a table. The participant read scripted questions relevant to the fever passage, and the agent gave scripted responses. In the avatar condition, participants heard that the virtual

more »

... esentation was controlled by a person whom they had just met. In the agent condition, participants heard that the virtual representation was computer controlled. The avatar condition exhibited better learning at posttest, even though all interactions within VR were held constant. Skin conductance indicated that the avatar condition caused more arousal and higher arousal was correlated with learning on a problem-by-problem basis. Further results suggest that the learning effect was not due to social belief per se, but rather in the belief of taking a socially relevant action.