Date of Graduation


Document Type


Degree Type



Eberly College of Arts and Sciences



Committee Chair

Michael Perone.


The present experiments were designed to evaluate accounts of how behavior is allocated on concurrent variable-interval (VI) VI schedules. The generalized matching law assumes that the proportion of behavior allocated to the two schedules will match the proportion of reinforcers obtained on the two schedules (Baum, 1974). An alternate view holds that under ideal conditions, behavior should "fix" almost exclusively at the richer schedule. The leaner schedule is only visited, or "sampled", when a reinforcer becomes available on that schedule (Houston & McNamara, 1981). Attempts were made to create optimal conditions for producing an ideal fix-and-sample response pattern. Pigeons were exposed to a series of concurrent VI VI schedules with and without signals for the availability of reinforcement. In Experiment 1, reinforcers at the lean alternative were signaled, which produced a robust tendency to fix at the rich schedule. In Experiment 2, reinforcers at the rich schedule were signaled, and the allocation of behavior shifted in favor of the lean schedule. No evidence of an ideal fix-and-sample pattern was found in either experiment, but a less extreme form of the pattern was observed in both experiments, regardless of the presence or absence of signals. Results were consistent with other research indicating that less extreme fix-and-sample patterns may be a fundamental characteristic of behavior engender by concurrent schedules (e.g., Aparicio & Baum, 2006). Although the present data were well described by the generalized matching equation, they were better described by a model that captures the way in which reinforcers are obtained by fix-and-sample response patterns (MacDonall, 2005).