Ezra G. Hall

Date of Graduation


Document Type


Degree Type



Eberly College of Arts and Sciences



Committee Chair

Kennon A Lattal

Committee Co-Chair

Daniel E Hursh

Committee Member

Michael Perone

Committee Member

Miranda N Reed

Committee Member

Claire C St Peter


Timeout punishment is among the most commonly reported disciplinary procedures (Barkin, Scheindlin, Ip, Richardson, & Finch, 2007). Despite the frequent use of timeout, little basic research has systematically examined different schedule effects of timeout from positive reinforcement. Using pigeons as subjects, the current series of experiments arranged variable schedules of timeout from positive reinforcement within a multiple-schedule arrangement where 20-s timeouts were response-dependent, response-independent, and delayed. Experiment 1 used a within-subject yoking procedure to compare schedules of variable-ratio (VR) and yoked-interval (YI) timeouts. Experiment 2 arranged separate parametric analyses of variable-interval (VI) and VR schedules of timeout. Within-session, yoked-control components delivered response-independent timeouts according to the same temporal distribution as in the preceding response-dependent timeout components in an attempt to isolate a direct response-decreasing effect of timeout presentations from indirect reductions in reinforcement rate. In Experiment 3, delays to timeout were studied using the same yoked-control procedure as in Experiment 2. Experiment 4 was designed to address confounds in the control conditions that were arranged for reduced timeout rate during delays in Experiment 3. The primary findings were: 1) response-dependent VR 2 and VR 3 timeout resulted in the most response reduction and the highest timeout rates across Experiments 1, 2, 3, and 4, 2) schedules of VI timeout reduced responding relative to baselines for the most frequent mean schedule values in Experiments 1 and 2, 3) response rate increases occurred during the introductions of delays to timeout in Experiment 3 and were partially attributed to the introduction of the delays in Experiment 4, and 4) response rates in the response-dependent timeout components were not always lower than their corresponding response-independent timeout components.