Date of Graduation


Document Type


Degree Type



Eberly College of Arts and Sciences



Committee Chair

John Crosbie.


On fixed-interval (FI) schedules of reinforcement, humans and other animals often have different response patterns. For example, the pattern typically associated with animals is the scallop: a pause followed by a gradual acceleration in responding. Humans generally show terminal-minimum (i.e., pause until the end of the interval, then make 1 or 2 responses) or constant-rate patterns. There have been numerous investigations of the variables that control such different response patterns. The most common variables in those studies are instructions, verbal ability, response cost, concurrent tasks, and food and points as reinforcers. The purpose of the present experiments was to investigate a variable that has received little experimental attention in this area: effects of an immediately consumable reinforcer, timeout from avoidance. In Experiment 1, adult human subjects responded on a concurrent schedule with avoidance of point-loss in one component, and timeout from avoidance in the other. Timeout was scheduled on FI with inter-reinforcement intervals ranging from 30 s to 480 s. Functions relating response rate and post-reinforcement pause to schedule value were similar to those reported for animals, but running-rate functions were not. Subjects showed terminal-minimum response patterns in most intervals. In Experiment 2, there were three conditions: (1) FI 240 s for point reinforcers, (2) concurrent FI 240 s for points and avoidance, and (3) FI 240-s timeout from avoidance. Response rates were highest and post-reinforcement pauses were shortest in conditions with point-gain only. Response rate was lowest and post-reinforcement pause was longest in timeout-from-avoidance conditions. There were few intervals with scalloped response patterns. The most terminal-minimum intervals occurred in timeout-from-avoidance. There are three main implications of the present results: (1) timeout from avoidance was a reinforcer which produced differential responding across FI values, but it was not more valuable than points as reinfocers in the present procedure; (2) scalloped response patterns on FI may be primarily due to averaging data over intervals; and (3) the best way to compare data across studies is to use strict stability criteria and quantitative measures that minimize data averaging.