Symbolic generalization refers to the transfer of learned responses from one situation to another because the two are seen as symbolically related. In this study, the researchers tested the role of symbolic generalization in near-misses using a simulated slot machine task. They trained 91 participants on the relations between seven images (e.g., cherry, bell, lemon) as being same or opposite. One image, cherry, always depicted win during training. As expected, participants then rated win outcomes with images that they had learned to be same as cherry to be closer to a win than win outcomes with images that they had learned to be opposite. They rated near-misses as closer to a win than losses.
