Quality of learning of severely retarded adolescents

A modified discrimination-reversal task identical to that used by Rumbaugh (1971) was employed to assess the quality of learning of severely retarded adolescents. Subjects initially were trained to criterion on a series of multi-dimensional object-discrimination problems (A+B-). Subsequent to one reversal trial on each problem (A-B+), one of three postreversal conditions was introduced: (a) a standard discrimination-reversal condition (A-B+), (b) a condition that substituted a novel stimulus for A (B+C-), or (c) a condition that replaced B with a novel stimulus (A-C+). Subjects' performance on the modified discrimination-reversal task was the same level for all three conditions. In comparison to earlier findings with nonhuman primates, results were interpreted as indicative of mediational learning processes.