Association for Behavior Analysis International

The Association for Behavior Analysis International® (ABAI) is a nonprofit membership organization with the mission to contribute to the well-being of society by developing, enhancing, and supporting the growth and vitality of the science of behavior analysis through research, education, and practice.


44th Annual Convention; San Diego, CA; 2018

Event Details

Previous Page


Symposium #412
Extinction: The Devil is in the Details
Monday, May 28, 2018
9:00 AM–9:50 AM
Marriott Marquis, San Diego Ballroom A
Area: EAB; Domain: Basic Research
Chair: Brian R. Katz (West Virginia University)
Abstract: Contingencies of reinforcement and punishment, like apparatuses, “sometimes break down” (Skinner, 1956/1961, p. 109). Such breakdowns may occur through experimenter error or degradation, transformation, relaxation, or removal of the contingencies. Although much of the experimental analysis of behavior has been concerned with how intact contingencies maintain behavior, effects of the aforementioned changes also are of interest. These changes have been studied in various ways, but one of the most widely studied, and perhaps best understood, is extinction. Comparatively little research, however, has investigated the influence of parameters of extinction procedures on the overall course of behavior change and terminal patterns of behavior. Thus, the three talks in this symposium will address effects of different procedural variations on responding during extinction. These will include comparisons of how conventional and unconventional extinction influence changes in response rate, how response variability changes as a function of the discriminability of extinction, and how aspects of the extinction procedure influence the presence of the extinction burst.
Instruction Level: Basic
Keyword(s): Extinction, Extinction burst, Response variability, Unconventional Extinction
An Analysis of the Extinction-Induced Response Burst
BRIAN R. KATZ (West Virginia University), Kennon Andy Lattal (West Virginia University)
Abstract: Although the extinction burst often is cited as one of the most frequently reported generative effects of extinction, few controlled investigations into the nature of the effect have been conducted. Thus, the purpose of the present study was to determine whether the extinction burst could reliably be produced. To that end, seven pigeons were exposed to a procedure involving repeated cycles of 5-session blocks of baseline and 8-session blocks of extinction. This sequence of sessions was then repeated at least five times. Depending on the condition, baseline sessions consisted of either a fixed-ratio (FR) or variable-ratio (VR) schedule, and transitions to extinction were conducted either between or within sessions. Preliminary analyses suggest that the extinction burst is a transitory phenomenon. Analyses of whole-session response rates during extinction have failed to demonstrate an extinction burst. Conversely, restricting analysis to solely the first minute or thirty seconds of an extinction session has revealed a burst, although this finding has not been consistent. Furthermore, burst-like increases in responding were more frequently observed following within-session transitions to extinction. These data suggest that aspects of the extinction procedure influence whether or not a burst-like increase in responding is observed at the onset of extinction.

Repetition and Variability of Operant Behavior in Extinction After Continuous Reinforcement

IVER H. IVERSEN (University of North Florida)

Understanding response variability in extinction is relevant for implementation of effective shaping techniques. Rats were pressing a very soft lever that could be moved down 2 cm. Lever movement was recorded in detail. Food pellet delivery was accompanied by a "beep" sound. After acquisition of lever pressing on continuous reinforcement (CRF), two conditions of extinction compared the effect of no sound from the feeder with the effect of retaining the beep sound. For the first 20 responses in extinction, the condition with no beep decreased inter response times (IRT) compared to CRF (extinction burst) whereas extinction with beep retained IRTs from CRF. In extinction, beep sounds were initially followed by entry to the food tray, which did not occur without the beep. Response topography with beep was more similar to topography under CRF. Without the beep, response topography varied more in extinction. In both types of extinction, response duration increased in extinction (slightly more for the no-beep condition). Data will be presented for six rats. The experiment shows the importance of trivial procedural details for response changes in early extinction.

Higher Response Rates Are More Resistant to Change When the Response-Reinforcer Dependency Is Removed
JAMES E. COOK (University of Mississippi Medical Center), Kennon Andy Lattal (West Virginia University)
Abstract: When reinforcement rates are equated, behavior that is maintained at lower response rates is more resistant to change when disrupted by adding additional reinforcers or conventional extinction (i.e., terminating reinforcer deliveries), but whether response rates affect resistance to change when only the response-reinforcer relation is removed (i.e., time schedules, unconventional extinction) is unclear. The responding of 20 rats was maintained on identical variable-interval 30-s schedules for 10 sessions. The 7 rats with the highest and lowest average response rates over the 6 most recent sessions in baseline were respectively placed into the High-Response-Rate (HRR) and Low-Response-Rate (LRR) groups. The schedule was then changed to a variable-time (VT) 30-s schedule for 10 sessions. HRR rats showed greater resistance to change under the VT schedule than the LRR rats. A direct correlation was shown for all rats between baseline response rates and resistance to change under VT schedules in contrast to previous experiments showing an inverse relation between baseline response rates and resistance to change. These data demonstrate that response rates should be considered when evaluating resistance to change when responding is disrupted and that response rates may have different effects depending on the nature of the disrupter.



Back to Top
Modifed by Eddie Soh