Association for Behavior Analysis International

Individualized assessment and treatment is a hallmark of applied behavior analysis. First, we conducted a functional analysis (Iwata et al., 1994/1982) of severe problem behavior displayed by a female with autism. Then, a subsequent analysis was added to evaluate whether response effort (i.e., ambulation) associated with tasks in the escape condition would influence the results. Finally, the effects of differential reinforcement of alternative behavior (i.e., task completion) using edible and leisure reinforcers were compared to an escape condition in a preliminary treatment evaluation. The functional analysis indicated that problem behavior only occurred in the escape condition if tasks involved ambulation, and the treatment package appeared promising to reduce escape-maintained problem behavior for this individual. Implications for research and clinical practice will be discussed.

Matching (Herrnstein, 1961) has been demonstrated with appetitive and aversive stimuli, as well as when appetitive and aversive stimuli are simultaneously presented (Farley & Fantino, 1978). Interestingly, in contexts where a single response produces both reinforcement and punishment, some research has demonstrated that a punisher subtracted more value than a reinforcer added (Rasmussen & Newland, 2008). This purported asymmetry in the effects of reinforcement and punishment was assessed in one individual with autism spectrum disorder, a population for whom the effects of simultaneous reinforcement and punishment has not been evaluated. To do so, we established tokens as reinforcers and evaluated the effects of a progressive token gain/loss schedule where losses gradually became denser to identify a schedule at which the individual would not respond. We then compared responding at the identified token gain/loss schedule where the participant stopped responding to a schedule in which an equal density of reinforcement was available for gain without the loss contingency. This was done to demonstrate that the loss contingency was directly responsible for the cessation of responding. Responding persisted when an equal density of reinforcement was available for gain (when the loss contingency was removed). Results are consistent with some previous findings suggesting that the punisher subtracted more value than the reinforcer added.

Communication is a prominent variable in studies of social dilemmas which affects the selection and maintenance of cooperation between individuals who interact. The Prisoner's Dilemma, an imaginary situation employed in Game Theory, is the most famous game in investigations of cooperation. The Behavioral Analysis of Culture uses such games as experimental tasks, due to the possibility of reproduction of social phenomena in laboratory. This study analyzed the interaction and communication of two groups with four participants each in an Iterated Prisoner Dilemma. Different opportunity costs and benefits to cooperate were manipulated. The results replicated some findings of the literature, that is communication facilitated cooperation between individuals. However, communication may not a sufficient condition for the emergence of cooperation, especially in situations where opportunity costs to cooperate and benefits to compete are high. Moreover, correspondence between verbal agreements and choices made by the participants might be a necessary condition for the selection and maintenance of cooperative behaviors.

Inequity aversion may be defined as refusal reinforcement when there is an unequal distribution of reinforcements. The literature on inequity aversion shows that humans experience both, aversion to disadvantageous inequity (DI) and advantageous inequity (AI). However, few researches have reported how DI and AI aversion are learned or can be modified by individual experience in multiple conditions. The present study investigates DI and AI through an ABA reversal design, each phase consisting of sixteen trials in a computer game. Participants picked between blue or green card. The combination of blue-blue choices resulted in unequal distribution of reinforcement. AI and DI reinforcers were delivered by computer software, according to each experimental phase. In Group DI (n=10), participants were exposed to DI (phase1), AI (phase 2), and DI (phase three), and Group AI (n=10) to AI, DI, and AI. The results showed that 5 participants present aversion to DI and only 2 to AI. The second phase programmed interaction modulated the aversion to inequity for 8 and 2 participants in the DI and AI aversion, respectively. Confounding variables, such as cultural differences may have influenced aversion to AI and should be addressed in future research.

Following the treatment of problem behavior, resurgence may occur when a response-reinforcer relation is discontinued. Implementing differential reinforcement of alternative behavior can address resurgence of the target problem behavior by making it less valuable compared to a more desirable, alternative behavior. Previous research has shown that concurrent schedules of reinforcement can be used to treat problem behavior and its resurgence (Peterson et al., 2012). In the present study, a computer task involving three phases (1, 2, and 3) with different schedules of reinforcement was in effect. Phase 1 consisted of points delivered contingent on making the low quality target response. Phase 2 consisted of points available for all three response options, where five points were delivered contingent on the low quality target response, ten points were available contingent on the medium quality response, and 15 points were delivered contingent on the high quality alternative response. In Phase 3, all response options were on extinction. The goal of the study was to investigate whether target responding would increase briefly at the start of the Phase 3.

Searching for variables that affect the rate of temporal discounting is important to understand better how delay discounting (DD) phenomena works. We want to know if the daily exposure to a DD experienced task could reduce the number of impulsive choices in the task. Six students of twenty years on average from a Colombian University have done a procedure for sixteen sessions. In each session, six blocks of ten concurrent-chains choice trials were presented. The participants could choose between the smaller-sooner outcome and a larger-delayed outcome, two and six candy balls respectively. In each block, the delay of larger outcome varied from 0.1, 5, 10, 20, 40 and 80 seconds. Every 4 sessions the condition changed, the order of presentation of delays of larger outcome varied in each condition (ascending, descending, random and re-ascending). We found two things. First, participants made exclusive preference choices by block after some sessions. Second, the exposure reduces the discounting rate in three participants with a degree of confidence of 0,05. A post-experiment interview shows that the perceived time of the experiment is lower than elapsed time. A version of the experiment with time of a video game as a outcome is well underway.

Conditioned inhibition can be trained in different ways. In differential inhibition training a cue X is presented by itself, interspersed with another cue Y that is paired with the outcome. In Pavlovian inhibition training X and Y are presented together without the outcome, intermixed with trials where Y is presented with the outcome. In both cases, X becomes an inhibitor. In this experiment, participants were presented a computer task to compare the effects of both types of training on the prediction of food (an image of mexican tacos), as well as the effects of using the word (“Nothing”) or a blank screen on trials where the outcome was not shown (No image of mexican tacos). We ran both summation and retardation tests, to compare which training cue becomes a better conditioned inhibitor. The results are that the differential inhibition procedure was more effective than the Pavlovian procedure on training inhibition. Also, the summation and retardation tests showed that the blank screen was more effective than the word “Nothing”. The results are discussed with the comparator theory of associative learning.

Renewal is a process related to craving in humans, and effort is been made to build a human model. Van Gucht et al. (2008) used chocolates (Unconditioned Stimulus, US) and a tray (Conditioned Stimulus, CS) to investigate the effects of context on chocolate craving. They ran an experiment where chocolate craving was acquired–extinct–put on test. AAA group ran the three phases on the same room light condition, while ABA group ran it with the lights on–off–on. Visual Analogic Scales (VAS) gave the US-Expectancy and US-Craving they would receive/eat chocolate. We suggest the use of physiological data to enhance the validity of their model. This pilot study used two women allocated each one in one group. The same procedure was kept, while sensors were attached to the participant (electromyography; skin conductance – SC; electrocardiography and temperature). The data show ABA acquisitions and extinction on both US-Expectancy and US-Craving, but no renewal. AAA doesn’t show learning. ABA SC follow her VASs acquisition and extinction patter, and it also shown renewal. Other physiological measures doesn’t show consistency. AAA shown no patter on all her physiological measures. Data discussion compares VAS and SC renewal for ABA, and argues about methodological improvements.

Post-reinforcement pauses are a common behavioral measure utilized within the behavior analytic literature to evaluate the effect of different slot machine outcomes on gambling. Previous research has reported differences in post-reinforcement pauses demonstrated by gamblers as a function of slot machine outcomes (Barton et al., 2017). Most of the studies incorporating post-reinforcement pauses as the primary dependent measure have been conducted utilizing computerized simulations of slot machines, with limited research on post-reinforcement pauses being conducted in a naturalistic gambling setting. The current study evaluated the effect of three different slot machine outcomes, wins, losses disguised as wins (LDW), and losses, on post-reinforcement pause in a Midwestern casino. Post-reinforcement pauses were collected for each of the three different of outcomes across 18 participants, which were compared to determine if a difference existed between the outcomes. Preliminary data suggest that the post-reinforcement pause following a win (M = 3.492) is longer than that of an LDW (M = 1.688) or a loss (M = 0.833), which supports the existing gambling research.

Previous studies have suggested that compared to contingent delivery of reinforcers by others, self-selection of the same stimuli may increase reinforcer effectiveness. To evaluate preference for choice versus no-choice, we replicated a study by Thompson, Fisher, and Contrucci (1998) using the Portable Operant Research and Teaching Lab (PORTL). First, typically-developing adults learned a choice/no-choice discrimination via shaping using playing cards as initial links within a concurrent-operant arrangement. Next, we evaluated preference for choice vs. no-choice when (a) selection of choice and no-choice initial links produced equal, continuous rates of reinforcement (FR1) and (b) selection of choice was placed on a progressive ratio schedule (PR). During the FR schedule, all participants allocated more selections to the choice link. When the choice link was placed on a PR schedule, all participants shifted selections to the no-choice link. Unlike results from Thompson et al., initial evaluations of choice preference under equal conditions did not predict whether the preference would maintain when the conditions were no longer reinforced on the same schedule. Ultimately, this study allowed for further development of previous applied work on choice and preference.

The present study is a parametric analysis of the effects of percentile and progressive schedules of reinforcement on the accuracy and rate at which individual youngsters dribble a basketball. A percentile schedule of reinforcement provides a systematic and arithmetic procedure for response shaping. The equation for percentile reinforcement is k = (m+1) (1-w), where m is the distribution of observations, w is the probability of reinforcement, and k is the rank the current response must exceed to deliver a reinforcement. A progressive schedule of reinforcement, also used in response shaping, utilizes ongoing assessment and calculation of the incremental effects of differential reinforcement alone. Performance data are displayed on the Standard Celeration Chart. Changes in the accuracy and rates of dribbling over time are portrayed as celeration values, which pertain to response differentiation in the w value of the percentile schedule and the values of the progressive schedule over time.

Timeout from positive reinforcement is widely used as a punishment procedure however, there are still many unanswered questions about the controlling variables that influence the punishing effects of timeout. Two reinforcement variables that influence punishment effects include the schedule of reinforcement and the availability of an unpunished alternative. The present study is a systematic replication of Dardano and Sauerbrunn (1964), using TO as a punishment procedure instead of electric shock. The purpose of replicating Dardano and Sauerbrunn was to investigate the effects of TO under two variables: schedule of reinforcement and availability of a non-punished alternative. Investigating TO under these two variables provided information on both the TO and time-in environment. In the current study, pigeons responded under a Findley-switching-key concurrent schedule with equal schedules of positive reinforcement. Response-contingent TO was added to one of the alternatives. If TO is aversive, it was expected that pigeons would switch to the other alternative even though the time-in environment progressively became leaner. TOs from positive reinforcement necessarily result in a decrease in the overall reinforcement rate in the environment; that is, TO presentation is confounded with a decrease in overall reinforcement. Shifts in preference to the nonpunished alternative may result from the decrease in reinforcement rate, not from the punishing effects of TO per se. Therefore, in the current study, a progressive-interval (PI) schedule was used on both alternatives. Using a PI schedule, allowed overall reinforcement rates to be equated across the alternatives.