J. Levy’s research while affiliated with Columbia University and other places

In the Tilted Room Illusion (TRI) Ss judging verticality use conflicting visual (V) and nonvisual (NV) cues. Consistent individual differences occur and sex differences develop with age. The hypothesis that learning contributes to these differences was tested by attempting to find a training situation which altered the TRI. In the first experiment, five groups of 14 Ss were trained in different judgments. Three groups showed a reduction in the TRI compared to a group of 14 Ss who received no training.In a second preliminary experiment 4 Ss were intensively trained, each in a different task. The TRI was reduced only when the NV cues were positively reinforced in a V-NV cue conflict situation. Intensive training which reinforced either modality, in situations where there was no intermodal discordance, had no effect.