Discrimination of geometric angles by adult humans

Men and women learned to discriminate between two different size angles presented to them as objects within a real-world task. During Experiment 1, participants in group 50 were trained to choose a 50° angle and participants in group 75 were trained to choose a 75° angle. During testing, both groups were provided with a choice between their training angle and one of a set of test angles that was either smaller or larger than the training angle. Results showed a generalized pattern of responding, with group 50 showing increased responding to test angles smaller than 50° and group 75 showing increased responding to test angles larger than 75°. Further analysis of the response patterns revealed that participants in group 50 showed evidence of absolute learning, whereas participants in group 75 showed evidence of relational learning. During Experiment 2, a third group of participants (group 25) trained to choose a smaller angle (25°) was included in addition to group 50 and group 75. Participants were trained with three angles present and tested with just two, one being their training angle and the other being one of a set of novel test angles. Similar to the participants from Experiment 1, group 75 showed evidence of relational learning. Group 50, for which no relational rule could be applied during training, showed an absolute learning pattern with no response shift to test angles smaller or larger than their training angle. Group 25 showed evidence of absolute responding that was more pronounced than that found for the smallest training angle during Experiment 1. These findings suggest differential learning of geometric angles based on amplitude with smaller angles perceived as more distinct and thus more resistant to broader generalization than larger angles. Implications of these results are that certain geometric properties may be subject to different learning processes based on the specific magnitude of that property.