Association for Behavior Analysis International

Resurgence is the recurrence of previous reinforced behavior by the worsening of current alternative behavior. Jarmolowicz and Lattal (2014) showed that delayed alternative reinforcement could serve as a worsening condition to produce resurgence, but it is unclear if the recurrence obtained was function of the delay or decrease of reinforcement rate. The aim of this study was exploring resurgence under delayed reinforcement without changing reinforcement rate. First, chain-pull was reinforced as target response. Second, lever-press was reinforced as alternative response while target response contacts extinction. Lastly, delays in alternative reinforcement were progressively increasing while the value of the reinforcement schedule were diminished, equaling reinforcement rate. Preliminary results showed that lower target response and decreasing alternative rates were maintained with most of delays. Resurgence was found until the longest delay were introduced for two of three rats but where lower than level exhibit under extinction. Findings suggest that introducing delays with more richer schedules of reinforcement rate could late resurgence of target behavior.

The contribution of different response requirements with or without a deadline to the occurrence of procrastinating behavior was evaluated. Specifically undergraduate students clicked on either of three buttons, presented on the screen of a smartphone, to earn points according to the following contingency. For 6 participants, a variable ratio of 10, 20 or 30 pressings for a point was programmed for a left, central, or right button, respectively. All participants had 6 trials of 90 seconds each to complete whatever of the three requirements. The main dependent variable was the temporal distribution of responses to each option in 9 10-s subintervals of each trial. As was expected, the frequency of responses to the button with the highest response requirement increased gradually as each trial elapsed for both groups. According to procrastination theory, as the response effort or requirement of a task increases the delay to initiating and completing it increases.

Previous research has found that rich-to-lean transitions may function aversively for both humans and non-humans. This study's purpose is to insert a neutral period between reinforcer delivery and the start of a new transition to decrease pausing; time-out periods occurred after half or all the food reinforcers. This time-out period may assist in mitigating the aversiveness of transitioning from more to less favorable conditions by diminishing the interaction between the past and upcoming reinforcer magnitudes. Pigeons’ pecking was reinforced under a multiple FR FR schedule in which completion of half of the ratios was reinforced with small (lean) and half was reinforced with large (rich) amounts of access to grain. For the two-color condition, different key light colors signaled only the upcoming magnitude; for the four-color condition, different key light colors were associated with each of the four transition types. All pigeons experienced both conditions. Overall, pigeons paused on average 2 to 10 times as long during rich-to-lean transitions than all other transition types. Different timeout durations were examined in a reversal design. Stimulus conditions were switched for each group allowing for the analysis of the relative contributions of the past and upcoming magnitudes on pausing.

The present study aimed to evaluate the occurrence of recombinative generalization when following instructions (action-object), using a teaching matrix and video-modeling. The actions “circle”, “cross out”, “underline” and “make a triangle over” were used as stimuli and the Greek letters zeta, sigma, gamma and lambda, organized in sub-matrices, were used as objects. A design of multiple probes between sub-matrices was used. The teaching task consisted of presenting visual stimuli (Greek letters) in the four corners of the computer screen simultaneously with the presentation of the target instruction. The child should perform the action (using the mouse) in relation to the corresponding object/letter. If correct, a potentially reinforcing consequence was presented; and if incorrect response, a video model was presented (in which the model/researcher performed the action in relation to the Greek letter corresponding to the dictated instruction), followed by a new opportunity for the participant to respond. The assessment task employed the same configuration, but without presenting different consequences. Data collection was carried out remotely. The four children with ASD aged between 6 and 9 years followed the instructions taught and showed recombinative generalization. The data suggest that teaching following instructions via video modeling was effective and the use of the teaching matrix contributed to promoting recombinative generalization, despite intra- and inter-participant and inter-sub-matrices variability.

Delay discounting describes how an outcome loses value as the delay to the outcome increases. We measured discounting of different types of talk-therapy in standard discounting tasks and cross-commodity discounting tasks. In a standard discounting task, the reinforcer on one choice alternative is the same reinforcer on the other choice alternative. In cross-commodity discounting tasks, the reinforcers on each alternative are qualitatively different. We compared the hyperbolic discounting model (Mazur, 1987) to a modified hyperbolic model to determine the highest quality model for the standard and cross-commodity discounting task. Model quality was compared using two popular information criteria, Akaike information criterion and Bayesian information criterion. For data from the standard discounting task, the information criteria behave as expected and select the simpler model. In one case for the cross-commodity discounting, the information criteria are strict enough to prefer the simpler model even though the more complex model is visually, noticeably better. Conceptual, theoretical, and practical considerations will be discussed.

Altruism has been shown to decrease as a function of increasing social distance. This relation is captured by social discounting which outlines a quantitative decline between reward value and increasing social distance. In the social discounting questionnaire, participants are prompted to imagine a list of the 100 people closest to them in order to determine reward values to socially distant persons. Similarly, in the reciprocal discounting questionnaire, participants are required to imagine the same list but instead report how a socially distant person would value rewards to themselves (the participant). Previous research has found that participants tend to report more altruism towards others than others are in return (towards the participant). But in the discounting questionnaires, the questionnaires usually do not specify the name of the socially distant person; rather, a vague indication (i.e., ‘Person 1’) is used. Perhaps if specific names were employed, altruism and generosity from others would be more similar. The present study evaluates the effect of specific naming on social and reciprocal discounting similarity.

Effects of delay of reinforcement is one of the most studied topics in the behavioral sciences. The present work aimed to test predictions of the “Mathematical Principles of Reinforcement” (MPR: Killeen, 1994), concerning the effects of delays on performances under variable interval (VI) schedules. 28 university students participated in the study. Participants were divided equally into two groups: Delay Group and Immediate Group. In a one-session experiment, participants of both groups responded under five VI values: 1 s, 6 s, 20 s, 35 s, and 60 s. In the Delay Group, reinforcers for participants’ responding occurred after a 1 s delay. In the Immediate Group, reinforcers were immediately delivered to participants’ responses. A two-way ANOVA revealed a main effect of VI schedules but not of delay. There were no interaction effects. MPR fits revealed an r2 ≈ 0.80 for the averaged data of both groups, but fits for the individual data prevented a conclusion concerning the predictions of the model. It is suggested that longer sessions, using stability criteria, can answer the question with greater reliability.

The generalized matching equation (GME) is a quantitative model of a given organism’s response allocation across concurrent schedules of reinforcement. Though some studies have found evidence of matching behavior among human subjects, others have observed marked departures. Given that non-human response allocation more commonly adheres to the GME in the published research, this has led some to question the GME’s generality vis-à-vis human subjects. One critique revolves around the presence of ordinal, schedule-correlated stimuli in several studies that have found positive evidence of matching in human subjects, a feature that is largely absent in non-human matching research. Therefore, the present pilot study exposed adult human participants to concurrent schedules of reinforcement accompanied by ordinal, schedule-correlated stimuli. Presses on two buttons displayed on a laptop screen were reinforced with points putatively exchangeable for money on variable interval schedules of reinforcement. Participants were exposed to three concurrent schedules of reinforcement for five-minute intervals, and response allocation was subsequently subject to generalized matching analysis. As of this writing, most subjects have exhibited indifference to the schedule parameters and idiosyncratic descriptions of the contingencies in subsequent interviews.

Social distance describes how close an individual feels to someone else. There have been issues measuring it, as physical distance is measured in ratio units while social distance is measured on an ordinal scale. Safin and Rachlin (2020) proposed a method to measure it with a ratio scale. However, they asserted that their method produced high variability at the lowest positions in the social distance scale, affecting its precision. We developed the drawing-a-line method [DLM] to measure social distance with a ratio scale more accurately. We asked 76 participants to place people in positions 1, 2, 20, 50, and 100 of the social distance scale and to draw a line representing how close they feel to each of them. We extracted the line’s length and used it as the value to construct the discount function. Results show minimal variability in the distance assigned to people placed at low positions. We also found that the hyperbolic pattern of discounting remained, even when ordinal social distance was replaced directly. We discuss the results in terms of how the DLM stands against the traditional method and Safin and Rachlin’s one and conclude that it measures social distance with a ratio scale effectively.