The Use of Contingency-Based Procedures to Promote Tolerance of Schedule Thinning for Hypothesized Multiply Maintained Problem Behavior

Abstract

Contingency-based procedures have been found to be effective in facilitating delays to reinforcement for escape-maintained behavior within research and clinical contexts. Few studies have evaluated the use of these procedures for multiply maintained problem behavior. A contingency-based procedure was conducted with two participants to evaluate the effectiveness this procedure had on increasing tolerance to schedule thinning for problem behavior that was hypothesized to be multiply controlled. Results suggested that the procedure was effective at facilitating tolerance for a delay for both participants after switching to a lower effort task.

Functional communication training (FCT) is an important intervention used to reduce problem behavior and increase communication with individuals with disabilities (Carr & Durand, 1985). This intervention enables individuals to emit a functional communicative response (FCR) that produces the same consequences that were provided contingent on problem behavior. Once the desired outcomes of the intervention have been observed, schedule thinning is often conducted so that previous responses or mands reinforced on a continuous schedule can be thinned to a more practical schedule similar to that seen in the natural environment. It is crucial to implement interventions that address both schedule thinning and the negative effects that might occur once a delay has been introduced, such as reduced rates of the FCR and the resurgence of problem behavior. Previous research has evaluated several procedures used to thin the schedule of reinforcement in FCT and the effects these procedures have on behavior. For example, researchers have used time-based procedures to incorporate a delay between the communicative response and reinforcement (e.g., Fisher et al., 2000). Although time-based procedures have helped increase the delay to reinforcement without increasing the occurrence of problem behavior, such procedures are not always effective. In a study conducted by Sidener et al. (2006), the researchers implemented a time-based procedure that included a signaled delay component to reduce the number of mands emitted during FCT and to facilitate tolerance of a reinforcement delay. Their results indicated that the use of a signal proved ineffective at maintaining mands at the terminal criterion for the delay-to-reinforcement condition.

Researchers have also evaluated the use of other procedures such as multiple schedules, mixed schedules, and denial tolerance to facilitate schedule thinning (Hanley et al., 2014; Sidener et al., 2006). One procedure that has been evaluated during schedule thinning is the use of a contingency-based procedure. After emitting the FCR in a contingency-based procedure, the individual is instructed to engage in another activity before accessing the reinforcer. The negative effects that can occur when using a time-based procedure are often reduced when using this type of procedure, producing an effective thinning technique.

In a study conducted by Ghaemmaghami et al. (2016), the researchers evaluated the effects of using a response contingency during delays to reinforcement. The researchers were interested in comparing the effects of contingency-based and time-based tolerance training on facilitating tolerance to delayed reinforcement. During the contingency-based progressive delay procedure (CBPD), the therapists arranged the specific evocative situations for each participant, and reinforcement was provided contingent on completion of the response requirement. Results indicated that problem behavior decreased to zero or near-zero levels and participant compliance and engagement with alternative activities increased, indicating that the CBPD procedure was more effective in building the participant's tolerance to delayed reinforcement. Previous literature has also evaluated the effects of chained procedures on multiply maintained challenging behavior (Falcomata et al., 2013). However, few studies have evaluated the use of a contingency-based procedure exclusively with participants whose problem behavior is hypothesized to be multiply maintained by both escape and tangible functions. Therefore, the purpose of the current study was to extend previous research by evaluating the effectiveness of a contingency-based procedure in facilitating tolerance to an increasing delay for behavior hypothesized to be maintained by access to escape and tangible items.

Method

Participants and Settings

The participants in this study were two males, Santiago and Ari, ages 7 and 6 years, respectively, both diagnosed with autism spectrum disorder. Santiago was initially referred to the study based on parental reports that indicated that he engaged in aggression, screaming, and crying behaviors when preferred items were removed. Additionally, he would engage in noncompliance, specifically when asked to sit down. In the first several clinic sessions, we were unable to replicate the contexts during which Santiago engaged in aggression, screaming, or crying; however, we were able to observe instances of noncompliance. Thus, the decision was made to target compliance with the instruction to sit down. Ari was referred to the study based on teacher reports and direct observations that indicated that he engaged in high rates of aggression, self-injurious behavior (SIB), and screaming when preferred items were removed or when transitioning to work tasks. The researchers identified work tasks that resulted in the highest amount of problem behavior for Ari but that were also a part of his individualized education plan (IEP). Placing Velcro sticks through a small slit in a closed container resulted in the highest levels of problem behavior, so this task was targeted. Sessions were conducted at a local university-based behavior analysis clinic for Santiago and in a self-contained special education classroom at a local elementary school for Ari; sessions were 5 min long, and session materials consisted of preferred items, work tasks, and data collection devices.

Response Measurement and Interobserver Agreement

Researchers collected data on each participant’s behavior during the study. For both participants, data were collected on independent mands, which were defined as any unprompted vocal utterance of the FCR for Santiago (i.e., “I want [item] back, please.”) or any unprompted handing of an picture card that had an iPad image for Ari. Compliance, which was defined as complying with an instruction within 10 s of its presentation, was also scored. For this measure, Santiago additionally had to sit for a predetermined interval duration to meet his definition of compliance. Compliance was then expressed as a percentage by dividing the number of instructions that the participants complied with by the total number of instructions and multiplying by 100. For Ari, aggression was defined as any attempt at hitting, scratching, or biting directed toward another person. SIB was defined as any attempt at hitting his head with a closed fist or hitting his head against an object (e.g., desk, wall, fidget toys). Screaming was defined as any nonfunctional vocalization with a negative affect, separated from a previous occurrence by a minimum of 2 s. For this study, negative affect was operationally defined as observable physical characteristics that signaled distress that accompanied the screaming behavior. These characteristics included the occurrence of visible tears, audible groans and whines, and facial signs of discomfort/distress such as frowning. Data collected on Ari’s problem behavior were then expressed as a rate of total problem behavior. During demand fading for Santiago, data were collected on the total number of demands placed and compliance with each demand.

Data were collected on a handheld device (phone, iPad, iPod) using the Countee data collection application (Peić & Hernández, 2016). The criterion for moving from baseline to treatment was based on visual analysis of the data, and the mastery criterion for each treatment phase was three consecutive sessions at 100% for compliance and/or zero or near-zero levels of problem behavior. A second data collector independently collected data on all target behaviors for at least 30% of all sessions. To calculate interobserver agreement (IOA), each session was divided into 10-s intervals. An agreement was scored when both observers scored the same response within the interval. Agreement percentages were calculated by dividing the number of agreements by the number of intervals and multiplying by 100. For Santiago, mean IOA was 98.3% (range 90%–100%) for the total number of mands and 92.8% (range 70%–100%) for compliance. For Ari, mean IOA was 99% (range 93.3%–100%) for aggression, 95.2% (range 85%–100%) for screaming, 97.3% (range 86.1%–100%) for SIB, 97.5% (range 73%–100%) for compliance, and 96.5% for the total number of mands (range 84%–100%).

Procedure

Preassessment

Prior to the start of the assessment, a free operant (FO) preference assessment (Roane et al., 1998) was conducted with each participant. For Santiago, animal figurines were initially identified as highly preferred; additional FOs were conducted to identify alternatives to use during subsequent sessions, as the reinforcing value of the animals decreased over time. For Ari, we identified the iPad as highly preferred. Afterward, FCT was conducted with both participants to teach the FCR prior to the start of the assessment.

Functional Assessment

For Ari, based on the context in which the intervention was conducted, we implemented the use of a functional behavior assessment (FBA). The use of this assessment was based on the school’s standard procedure regarding functional assessments, thus creating a situation where we had to adapt to using alternative methods instead of conducting a traditional functional analysis (FA). Results of the FBA identified escape and access to tangible items as the two hypothesized functions of Ari’s behavior. For Santiago, an open-ended functional assessment interview (Hanley, 2012) was conducted in lieu of a traditional FA. Results of the interview identified escape and access to tangible items as the two hypothesized functions of Santiago’s behavior.

Baseline

During baseline, participants were provided access to their preferred items for 1 min before the start of the session. Afterward, the researcher removed the items, and the session began. For Santiago, the researcher placed a demand (i.e., “sit down”) once the items were removed; this was done to obtain a baseline duration of sitting. No consequences were provided for any independent mands or problem behavior.

Treatment

Treatment sessions were similar to baseline sessions except that when the FCR was emitted, the researcher stated, “Nice job asking for [item], but before you get it back, you need to ___________.” Santiago was instructed to sit down for a predetermined amount of time based on his baseline levels, and Ari was instructed to put Velcro sticks through a slit in a closed container. If the participant complied with the instruction, the researcher provided praise and access to the item for 30 s before starting the next trial. If the participant did not comply within 10 s, the researcher used a least-to-most prompting procedure to prompt the compliance response; prompted responses resulted in reinforcement. If the participant engaged in problem behavior, then the researcher continued to withhold the item until problem behavior ceased. The researcher then started a new trial and waited until the FCR was emitted; prompts were provided every 30 s if the participant did not engage in mands for the reinforcer. The response requirement increased after three consecutive sessions with no problem behavior and/or 100% compliance.

Demand Fading

For Santiago, a demand-fading phase was added to teach motor demands, which were deemed to be a lower effort task. Our decision to incorporate a less effortful task was based on visual analysis of the data that depicted high levels of variability during treatment, with the data moving in a countertherapeutic trend. Additionally, previous research on demand-fading procedres have found these procedures to be effective in the case of negative reinforcement (Hagopian et al., 2011) which supported our decision. This change would allow the participant to contact the contingency more frequently and build momentum to comply with the more effortful task. Additionally, motor demands were part of the participant’s repertoire and were deemed to be instructions that he would have minimal issues complying with. Once training was complete, treatment sessions were conducted as previously mentioned. When mastery of one motor demand was achieved, the therapist increased the response requirement during subsequent phases by one until sitting was reintroduced. For Ari, his initial response requirement consisted of placing five Velcro sticks inside the container. After observing a steady increase in problem behavior during the first half of his treatment sessions, we decided to decrease the response requirement to one Velcro stick to increase Ari’s opportunities to contact reinforcement. The experimental design used in this study was an A-B-C-B-A-B design for Santiago and an A-B-A-B design for Ari.

Results

Figure 1 displays the percentage of compliance and rate of mands for Santiago. During baseline sessions, compliance was at zero while manding occurred at low levels. During treatment sessions, Santiago was initially prompted to sit in a chair for 5 s. Here, compliance varied across sessions, though a gradual decrease was observed. Thus, we switched to a lower effort task in the next phase. During the demand-fading phase, compliance increased to 100%, resulting in an increase of the response requirement during subsequent phases. After the demand-fading phase, the therapist reinstated the higher effort task. During this condition, the mastery criterion was eventually achieved.

Fig. 1.

Percentage of Compliance (Primary y-Axis) and Rate of Mands (Secondary y-Axis) for Santiago. Note. The dashed lines indicate changes in the response requirement during each treatment phase. BL = baseline; MD = motor demand

During the reversal back to baseline, both compliance and manding reduced to zero occurrences in the final two sessions. A return to treatment resulted in Santiago meeting the mastery criterion. The sitting requirement was increased, and the criterion was reached in the subsequent 10-s and 20-s phases. At the 40-s phase, a decrease in compliance prompted a return back to the 20-s phase; during this phase, we observed a concomitant increase in compliance, although due to time constraints, we were unable to obtain more data for Santiago during this phase.

Figure 2 displays the percentage of compliance and the rates of total problem behavior and mands for Ari. During baseline, variability was initially observed, but problem behavior steadily increased after Session 5. Because of the increasing trend in problem behavior after introducing the treatment phase, we decreased the work requirement after Session 15. Subsequently, problem behavior reduced to zero and compliance increased to 100%. A return to baseline produced similar levels of responding compared to the previous baseline phase. In the subsequent treatment phase, problem behavior decreased to zero and compliance reached 100%. At this point, his teacher decided to focus on other IEP goals, so we were unable to continue with our intervention.

Fig. 2.

Percentage of Compliance (Primary y-Axis) and Rate of Total Problem Behavior and Rate of Mands (Secondary y-Axis) for Ari. Note. The dashed line during the first treatment phase indicates a change in the response requirement. BL = baseline

Discussion

The results of this study indicate that implementing a contingency-based intervention was successful in facilitating tolerance to delayed reinforcement for behavior that was hypothesized to be multiply maintained once the work/response effort was reduced. The decision to use a less effortful task for both participants was made to increase the number of opportunities to contact reinforcement, which would allow the participants to establish a reinforcement history where compliance led to access to the reinforcer and would help maintain compliance as task effort increased.

It should be noted that there are several limitations to these findings. First, the inclusion of the demand-fading phase for Santiago provides inconclusive results about whether this phase was effective in building tolerance toward a higher effort task because there were no baseline data to demonstrate experimental control between phases. Additionally, the intervention was unsuccessful in achieving the mastery criterion past 20 s. It is possible that the establishing operation for tangible items was not potent enough to compete with the aversiveness of sitting down. This could suggest that escape from demands was more reinforcing than accessing tangibles. Because a traditional FA was not conducted, it is unknown whether the function of Santiago’s behavior was equally maintained by both escape and access to tangible items, although the results of his open-ended functional assessment interview indicated those were the two possible functions for his behavior. For Ari, the intervention was successful in reducing problem behavior and increasing compliance after lowering the task’s response requirement, a response pattern that was also initially observed with Santiago. Unfortunately, additional data were needed to see if these levels would maintain as we gradually increased the response requirement. In spite of these limitations, we were able to demonstrate the feasibility of implementing this type of procedure within different contexts and when researchers are unable to replicate the same controlling variables within contrived, clinical settings. Regarding the targeted tasks, alternative tasks may have been more effective during the schedule-thinning procedure, but we chose to incorporate tasks that were essential for each participant’s academic performance.

Using alternative tasks to help facilitate delays might be unsuccessful if the initial requirement is too effortful. Future research should evaluate the effects of using lower effort tasks from the onset of the intervention versus higher effort tasks if an individual’s behavior is, or hypothesized to be, multiply maintained, as was done by Hanley et al. (2014). Future research should also evaluate the effects of the order in which chained responses are introduced. For example, previous research has required individuals to complete a work task before they can request a break (Wacker et al., 2011). Research has also evaluated having the individual emit the FCR followed by an instruction, as was done in the current study. To determine if the order of the response chain influences the effectiveness of the intervention, future research should conduct a comparative analysis evaluating the procedural order of these interventions.

Funding

Not applicable.

Declarations

Conflict of interest

Karla A. Zabala declares that she has no conflict of interest. Kara L. Wunderlich declares that she has no conflict of interest.

Ethical approval

All procedures performed in the study involving human participants were in accordance with the ethical standards of the University of Georgia and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed consent

Informed consent was obtained from all individual participants included in the study.