Behavioral momentum theory states that resistance to change of operant behavior is governed by the Pavlovian stimulus-reinforcer relation in a given discriminative stimulus situation. That is, higher reinforcer rates in the presence of a discriminative stimulus result in a stronger stimulus-reinforcer relation and, thereby, greater resistance to change. Within the momentum-based quantitative framework of resistance to change, the construct relating persistence to pre-disruption reinforcer rates is termed “behavioral mass.” All research on which momentum theory is based has examined resistance to change following prolonged exposure to stable reinforcer rates in multiple schedules of reinforcement. Thus, at present little is known about the time frame over which behavioral mass accumulates or the manner by which newly experienced stimulus reinforcer relations are incorporated into mass when these rates change. The experiments described in this dissertation aimed to clarify these facets of the construct. Chapters 1 and 2 provide a detailed overview behavioral momentum theory and resistance to change. Topics discussed include quantitative models of resistance to change, clinical implications of resistance-to-change research, some notable limitations of behavioral momentum theory, and extensions of the theory to account for diverse behavioral outcomes. A recently published study is presented in Chapter 3 that aimed to determine how resistance to change and behavioral mass of pigeons’ key pecking adapts in the face of stimulus-reinforcer relations that change across time during baseline. Results suggest that resistance to change is a function of recently experienced stimulus-reinforcer relations and that behavioral mass depends most heavily on these recent experiences. The experiment described in Chapter 4 extended the findings reported in Chapter 3 by examining whether behavioral mass changes during operant extinction. Pre-exposure to extinction in an alternative multiple-schedule component decreased resistance to extinction of target-component key pecking relative to conditions without pre-exposure to extinction. Between-condition differences in extinction were well accounted for quantitatively by either variation in behavioral mass or changes in the magnitude of factors that are assumed to disrupt responding during extinction. Chapter 5 offers an integrative discussion of this research and emphasizes theoretical implications, practical applications, and areas for future research.

Behavioral momentum theory (BMT) provides a theoretical framework for studying the persistence of behavior when challenged. The typical experimental arrangement to study persistence involves reinforcing a behavior according to a multiple schedules design. Unique schedules of reinforcement are programmed to each component. When steady-state responding occurs, the schedules of reinforcement are disrupted by a challenge condition (e.g., extinction, distraction, or prefeeding). The multiple schedules component that maintains the greatest level of responding during disruption is described as being more persistent. Basic research has shown that rate of reinforcement is a reliable predictor of persistence. The multiple schedules component associated with the higher rate of reinforcement persists longer than the multiple schedules component associated with the lower rate of reinforcement during disruption. Applied researchers have recently begun translating BMT to problems of social significance. The success of these initial translations suggests that relations between other dimensions of reinforcement and persistence should be studied. The current two-experiment study investigated the effect of quality of reinforcement on the persistence of task completion. Three participants with a history of engaging in problem behavior to escape from demands participated in Experiment I. After showing the conditions under which participants would and would not allocate away from a work task to engage with a preferred item, a baseline measure of task completion was obtained. Task completion was then reinforced with attention or tangibles within a multiple schedules design. Orange tokens signaled access to tangible reinforcement and yellow tokens signaled access to attention reinforcement. After steady-state responding occurred, preference for attention and tangibles was assessed within a concurrent schedules design. Extinction was then implemented to disrupt task completion within each component of the multiple schedules design. Results showed modest differences in the persistence of task completion with task completion in the multiple schedules component associated with the delivery of the more preferred reinforcer persisting longest. The modest differences in persistence were smaller than what has previously been shown in the literature. Thus, a follow-up experiment was conducted to evaluate the effect of reinforcer potency on the persistence of task completion. Three participants with a history of engaging in problem behavior to escape from demands participated in Experiment II. After identifying relatively more and less preferred stimuli with a multiple stimulus without replacement (MSWO) preference assessment, a unit price analysis was conducted to evaluate the potency of these two items within a concurrent schedules design. Task completion was then reinforced with the more and less potent reinforcers according to a multiple schedules design. After showing steady-state responding, task completion was disrupted by extinction. Results clearly showed greater persistence of task completion under the component associated with the delivery of the more potent reinforcer for two of three participants. Results from both experiments are discussed in terms of their conceptual and applied implications.

Behavioral momentum theory provides a conceptual framework for the study of the recurrence of previously extinguished operant behavior. Commonly referred to as treatment relapse, this is the failure to maintain treatment gains (e.g., reductions in challenging behavior) when there is a change in conditions under which these gains were achieved. One treatment relapse paradigm previously examined in basic and applied research is reinstatement. Reinstatement of challenging behavior has been shown to occur when functional reinforcers are delivered on a fixed-time schedule following extinction of challenging behavior. Although examinations appropriate behavior have applied value, analyses of reinstatement have been conducted almost exclusively with challenging behavior. During the current study, a reinstatement methodology was applied to communicative responses with five children diagnosed with developmental disabilities who exhibited comorbid communication deficits, as well as challenging behavior maintained by positive reinforcement. In the first phase of the reinstatement evaluation, each child received functional communication training (FCT) within a positive reinforcement context within a two-component multiple schedule design with each schedule paired with a distinct communicative response. After achieving steady-state responding in the first phase, in which all participants were independently emitting both communicative responses, all appropriate communication was placed on extinction in the second phase. Extinction continued until rates of appropriate communication were at or near zero. In the third phase, positive reinforcement was delivered and the recurrence of appropriate communication was evaluated. For two of five participants, communicative responding recurred following the fixed-time delivery of the functional reinforcer, indicating a successful demonstration of reinstatement. For three of five participants, communicative responding recurred prior to the delivery of fixed-time reinforcement, indicating that an alternative recurrence phenomenon likely occurred. These results suggest that reinstatement methodologies can be applied to cases in which FCT treatment failures have occurred to efficiently restore clinical gains for some participants. Implications for clinical practice and future directions of this line of research are discussed.

Interventions based on a token economy effectively reduce problematic behavior. Yet, treatment gains deteriorate once an intervention is discontinued. It is important to better understand the persistence of behavior maintained by token reinforcement in simple experimental procedures. A Pavlovian association with primary reinforcement is said to endow neutral stimuli (e.g., coins, poker chips, lights, signs, stickers, etc.) with their own function to strengthen behavior as conditioned reinforcers. Behavioral momentum theory suggests that resistance to change under conditions of disruption is the appropriate measure of response strength. However, some animal studies have suggested that conditioned reinforcement may not affect resistance to change of a response. Here, a novel token reinforcement procedure was developed to investigate the resistance to change of responding maintained by token reinforcement. Pigeons responded on a key to produce tokens displayed on a touchscreen monitor in two signaled token-production components. Tokens accumulated over the two production components prior to a common exchange component where pecks to the tokens on the touchscreen produced food reinforcement. Resistance to change of responding maintained by different rates of token reinforcement was assessed by disrupting baseline token-production responding with presession feeding. Token reinforcement rates had inconsistent effects on baseline token-production response rates. However, small effects of token reinforcement rate on resistance to change were found. Results provide weak support for a response-strengthening account of conditioned reinforcement and insightful directions for future studies of token reinforcement in related procedures.

Interventions based on a token economy effectively reduce problematic behavior. Yet, treatment gains deteriorate once an intervention is discontinued. It is important to better understand the persistence of behavior maintained by token reinforcement in simple experimental procedures. A Pavlovian association with primary reinforcement is said to endow neutral stimuli (e.g., coins, poker chips, lights, signs, stickers, etc.) with their own function to strengthen behavior as conditioned reinforcers. Behavioral momentum theory suggests that resistance to change under conditions of disruption is the appropriate measure of response strength. However, some animal studies have suggested that conditioned reinforcement may not affect resistance to change of a response. Here, a novel token reinforcement procedure was developed to investigate the resistance to change of responding maintained by token reinforcement. Pigeons responded on a key to produce tokens displayed on a touchscreen monitor in two signaled token-production components. Tokens accumulated over the two production components prior to a common exchange component where pecks to the tokens on the touchscreen produced food reinforcement. Resistance to change of responding maintained by different rates of token reinforcement was assessed by disrupting baseline token-production responding with presession feeding. Token reinforcement rates had inconsistent effects on baseline token-production response rates. However, small effects of token reinforcement rate on resistance to change were found. Results provide weak support for a response-strengthening account of conditioned reinforcement and insightful directions for future studies of token reinforcement in related procedures.

Several variables have been shown to influence resistance to change including rate, magnitude, and delay to reinforcement (Nevin, 1974). One variable that has not been studied with humans concerns the evaluation of resistance to change with unconditioned and conditioned reinforcers. In Experiment 1 (Resistance to Extinction Assessment), 5 participants' behaviors were reinforced during a baseline phase on a mult VI 30 s VI 30 s schedule with either a conditioned (i.e., token) or unconditioned reinforcer (i.e., food). Following equal reinforcement rates across components, extinction was introduced as a disruptor. All participants showed greater resistance to extinction in the component associated with conditioned reinforcers than unconditioned reinforcers. In Experiment 2 (Varied Distractors Assessment), 4 participants experienced a baseline phase the same as Experiment 1 (i.e., mult VI 30 s VI 30 s). Each participant was then exposed to prefeeding and distraction as disruptors in separate analyses. Results showed that behaviors were more resistant to distraction with conditioned than unconditioned reinforcers, similar to Experiment 1. However, when prefeeding disrupted responding, greater resistance to change was observed with unconditioned reinforcers than conditioned reinforcers. Implications of the results are discussed.

Preference for one stimulus context over another and resistance to disruption within those contexts are a function of the conditions of reinforcement arranged within those contexts. According to behavioral momentum theory, these measures are converging expressions of the concept of response strength. Most studies have found that preference in concurrent chains and resistance to change are greater in contexts presenting higher rates or larger magnitudes of reinforcement. The present series of experiments attempted to extend behavioral momentum theory by examining whether differences in reinforcer type affect relative response strength with rats lever pressing for different types of food. In Experiment 1 of Chapter 2, several nonuniform disrupter types were examined that provided free access to a food type that was the same as one reinforcer type. Responding decreased more in the context presenting the same type of reinforcer as the disrupter, suggesting that many traditional disrupters (e.g., satiation) are inappropriate for examining how reinforcer type impacts response strength. Therefore, extinction was used throughout the remainder of the experiments to more uniformly disrupt responding across contexts. In Experiment 2 of Chapter 2, resistance to extinction was assessed when food pellets and a sucrose solution maintained responding across contexts. Moreover, relative reinforcer type was manipulated by changing the sucrose concentration across conditions. Relative response rates were systematically affected by changing sucrose concentration, but relative resistance to extinction was not. In Experiment 3 of Chapter 2, qualitative difference between reinforcers was enhanced and preference also was assessed to provide a converging measure of response strength. Preference and relative response rates were systematically affected, but relative resistance to extinction again was not. Finally, in Chapter 3, relative reinforcer rate and type were manipulated while assessing preference and resistance to extinction using the matching law. Preference, but not resistance to extinction, consistently was affected by changes in reinforcer rate and type. Systematic deviations in sensitivity and bias, however, suggested that different reinforcer types interacted with reinforcer rate. Overall, these findings suggest that the overall context of reinforcement, including interactions between different reinforcer types, should be considered when assessing preference and relative resistance to change.

Behavioral momentum theory (BMT) has become a prominent method of studying the effects of reinforcement on operant behavior. BMT represents a departure from the Skinnerian tradition in that it identifies the strength of responding with its resistance to change. Like in many other operant research paradigms, however, responses are considered to be momentary phenomena and so little attention has been paid to non-rate dimensions of responding. The current study takes up the question of whether or not the degree of effort defining a discriminated operant class has any meaningful effect on its resistance to change. Using a force transducer, rats responded on a two-component multiple VI 60-s VI 60-s schedule where each component was correlated with a different force requirement. Resistance to change was tested through prefeeding and extinction. Proportional declines in response rate were equal across components during all disruption tests. Differentiated response classes remained intact throughout. The negative result suggests several future research directions.