Analysis of the Partial Reinforcement Extinction Effect in Humans Using Absolute and Relative Comparisons of Schedules

The partial reinforcement extinction effect (PREE) was analyzed in humans using a quasi-multiple schedule of reinforcement with two discrete trial tasks. The first task was a videogame analog of a shuttle box. Using a joystick, subjects were required to move a cue in one of four directions for reinforcement. The second was a simple concept-formation task consisting of two binary dimensions; responding to one of the four alternatives was reinforced. All subjects were trained on the two tasks on a randomly alternating basis. Two groups were trained on the tasks using either continuous or partial reinforcement schedules for both tasks. An additional two groups received both schedules of reinforcement with continuous reinforcement on one task and partial reinforcement on the other task. Groups exposed to only a single schedule of reinforcement displayed a conventional PREE, with the group receiving only partial reinforcement showing a greater resistance to extinction than the group receiving only continuous reinforcement. In contrast, subjects exposed to both schedules of reinforcement during acquisition did not evidence a conventional PREE; rather their behavior was equally persistent on both tasks and schedules. The data are compared with results of nonhuman within-subjects experiments that indicate a transfer of persistence effects across schedules. Finally, the status of the PREE, as an empirical generalization, is reviewed.