Designing Research
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/357057067
A Parametric Single-Case Analysis and Social Validation of the High-
Probability Request Sequence
Article in Journal of Positive Behavior Interventions · December 2021
DOI: 10.1177/10983007211062610
CITATIONS
3 READS
128
3 authors, including:
Alissa N Baida
1 PUBLICATION 3 CITATIONS
SEE PROFILE
Joshua Jessel
Brock University
90 PUBLICATIONS 1,722 CITATIONS
SEE PROFILE
All content following this page was uploaded by Alissa N Baida on 21 June 2024.
The user has requested enhancement of the downloaded file.
https://doi.org/10.1177/10983007211062610
Journal of Positive Behavior Interventions 1 –12 © Hammill Institute on Disabilities 2021 Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/10983007211062610 jpbi.sagepub.com
Empirical Research
The Centers for Disease Control and Prevention (2018) currently estimates that one in 59 children have autism spectrum disorder (ASD). In addition to more prevalent symptomology (e.g., deficits in language and appropriate social interactions), many children who are diagnosed with ASD tend to exhibit noncompliance with adult instructions (Kaat & Lecavalier, 2013). It is often reported that many children with ASD tend to refuse to initiate or complete tasks that they are asked to do within a specific time frame. This can be a problem when the incompletion of a task can interrupt classroom activities or interfere with the learning of daily life skills in the home. That is, noncompliance can potentially impact important learning opportunities and stunt developmental growth (Cipani, 1998).
The severity of noncompliance as a form of problem behavior can range from the minimal dismissal of an adult instruction to aggressive or self-injurious behavior (Davis et al., 1992) and tend to occur in socially relevant contexts such as academic settings (Lee et al., 2004). Although the topography of noncompliance may vary within and across individuals, the function tends to be relatively consistent (Ardoin et al., 1999; Mace & Belfiore, 1990). For example, Mace and Belfiore (1990) conducted an evaluation of escape-maintained noncompliance in the form of stereotypy that occurred in the presence of low-probability (low-p) instructions identified as household tasks. When the partici- pant was instructed to complete household tasks (e.g., put- ting dirty plates in the sink, hanging up coats) she would
often engage in stereotypy of excessive touching of objects and people. The researchers developed a behavioral inter- vention that involved providing a series of high-probability (high-p) instructions as an antecedent procedure before delivering the low-p instruction. High-p requests are instructions that the participant has a history of cooperating with and were used to build behavioral momentum in over- all compliance. Mace and Belfiore successfully improved compliance with the household tasks when preceding those low-p instructions with the high-p request sequence.
Since the introduction of the high-p request sequence, there have been many replications of the procedures often supporting its efficacy. Cowan et al. (2017) conducted a meta-analysis of single-subject research studies that inves- tigated the high-p request sequence on compliance and on task performance for students who have ASD. The review included 16 studies and a total of 40 students between the ages of 3 and 13 years. The researchers suggested that there was an average of 80% improvement for compliance when high-p instructions were implemented prior to the low-p
1062610 PBIXXX10.1177/10983007211062610Journal of Positive Behavior InterventionsBaida et al. research-article2021
1Queens College, City University of New York, Queens, USA
Corresponding Author: Joshua Jessel, Queens College, City University of New York, 65-30 Kissena Boulevard, Queens, NY 11367, USA. Email: Joshua.Jessel@qc.cuny.edu
Action Editor: Joshua Harrower
A Parametric Single-Case Analysis and Social Validation of the High-Probability Request Sequence
Alissa N. Baida, MA1, Sharon Azizi, MA1, and Joshua Jessel, PhD1
Abstract Noncompliance with adult instruction is a common problem exhibited by individuals diagnosed with intellectual and developmental disabilities. The high-probability (high-p) request sequence was designed to increase compliance with low-probability (low-p) instructions by rapidly presenting high-p instructions immediately prior to the targeted low-p instruction. This study evaluated the use of three different levels of the high-p request sequence (i.e., one instruction, three instructions, and six instructions) to increase the compliance of five children who were diagnosed with autism spectrum disorder (ASD). Results indicated that all three levels of the high-p request sequence were often successful in increasing compliance with low-p instructions; however, when given the opportunity to choose, participants and caregivers (i.e., mothers and therapists) tended to prefer the high-p request sequence with three instructions.
Keywords autism, compliance, high-probability request sequence, parametric analysis, social validity
2 Journal of Positive Behavior Interventions 00(0)
instruction, across all studies reviewed. Interestingly, the number of high-p instructions within the sequence seems to be a standardized procedure, with 11 out of 16 studies implementing the three high-p request sequence prior to the low-p instruction.
For example, Riviere et al. (2011) studied the effects of a high-p request sequence to increase compliance with two boys diagnosed with ASD during medical examinations (e.g., getting teeth cleaning done, cutting toenails, and checking their throats). Three sets of low-p instructions were implemented with each participant. When the mother conducted the three high-p request sequence, the two par- ticipants’ compliance with those medically related instruc- tions increased. In addition, medical professionals were trained in the high-p request sequence and were able to suc- cessfully implement the procedure in the medical offices when providing general care.
In another example of the standard, three high-p instruc- tions, Pitts and Dymond (2012) studied the effects of the high-p request sequence with and without programmed reinforcement contingent on the completion of each high-p instruction. That is, as an antecedent-based strategy, the high-p request sequence does not necessarily require the use of supplemental reinforcers identified during a formal preference assessment and often includes general praise fol- lowing compliance with high-p instructions. The authors found that the high-p request sequence, using three high-p instructions, was successful in increasing compliance; how- ever, the sequence including programmed reinforcement was more effective.
Although using three high-p instructions in the sequence has become commonplace, its dissemination among researchers and clinicians may be more influenced by tradi- tion rather than empirical support. That is, there is a dearth of research supporting the efficacy of one sequence over another. Davis et al. (1992) extended the high-p request sequence by including a range of three to five high-p instructions randomly introduced into the sequence to increase compliance among two boys diagnosed with devel- opmental disabilities. Thus, the introduction of the low-p instruction became less predictable. Not only did this varied high-p request sequence increase compliance but the authors also found that the effects generalized to low-p instructions introduced with other adults who did not imple- ment the treatment. The results suggest that high-p sequences can be effective when the number of instructions are varied; however, because there was no systematic com- parison between different sequences, it is difficult to deter- mine relative efficacy.
One of the only studies to compare the utility of different high-p request sequences conducted a parametric analysis of three different high-p and low-p ratios (Ertel et al., 2019). Ertel et al. (2019) compared treatments with one, three, and five high-p instructions for every low-p instruction to
increase compliance of three children diagnosed with ASD. Interestingly, all sequences were successful in increasing compliance for two of the three participants, with the most effective ratio including five high-p instructions. Ertel et al. suggested that the improvements obtained with greater ratios may have been indicative of increases in behavioral momentum with the additional opportunities to contact reinforcement before the low-p was introduced. That is, if the efficacy of the high-p request sequence was influenced by behavioral momentum, then more compliance should be associated with larger sequences.
We conducted this study as an extension to Ertel et al. (2019) by implementing a parametric analysis of three dif- ferent ratios of high-p sequences (i.e., 1:1, 3:1, and 6:1). In addition to the measure of efficacy in improving compli- ance among the five participants, we also evaluated the acceptability and preference of the high-p request sequences using two different social validity measures. After the treat- ment comparison, the participants were asked to select the condition they wished to experience, and the caregivers were asked to complete a follow-up questionnaire. The pur- pose of this study was, therefore, to evaluate the relative efficacy of, and preference for, high-p instructions of vary- ing sequences.
Method
Participants
This study included five participants diagnosed with devel- opmental disabilities. Sessions for all participants were conducted in their respective homes. Max was a 7-year-old Caucasian boy who was diagnosed with ASD and attention- deficit/hyperactivity disorder (ADHD). Max attended a special education classroom in a local private school and received 15 hr a week of in-home, Applied Behavior Analysis (ABA) and speech services. During his ABA ser- vices, Max experienced programs targeting his social skills, play skills, and verbal behavior, including reading. Max was verbal and capable of understanding basic task instruc- tions. In addition, Max was able to make requests and use full sentences. However, Max exhibited difficulties with understanding complex language. Max was compliant with many basic instructions such as “Touch your head,” “Sit down,” “High five,” and “Stand up.” Max experienced dif- ficulties when instructed to relinquish preferred items, such as a tablet or phone. When Max was noncompliant, he would say “No,” ignore the instruction, or yell.
The second participant was Sam, who was a 2-year-old Caucasian boy diagnosed with ASD. He attended an early intervention program at a center for children who have ASD for 10 hr a week and received an additional 10 hr of in- home ABA and speech services. Sam’s ABA services were focused on improving his receptive/expressive language,
Baida et al. 3
activities of daily living skills, and play skills that included appropriate socialization and turn taking. Sam was mini- mally verbal but was capable of understanding basic task instructions and requests. He could also make basic requests by pointing and occasionally labeling. The participant was compliant with some simple activities and instructions. Sam exhibited noncompliance with many tasks, specifically during transitions. Often, he engaged in noncompliance when transitioning from preferred to non-preferred tasks. When he was noncompliant, he would yell, cry, say “No,” and throw himself on the floor.
Jacob was a 5-year-old, Caucasian boy diagnosed with ASD. Jacob attended a special education classroom at the local public school and received 20 hr a week of ABA home services. Similar to Sam, the ABA therapists targeted recep- tive/expressive language and play skills. In addition, Jacob also experienced academic programming targeting his read- ing, writing, and math. Jacob was verbal and capable of understanding basic task instructions and requests and he could make basic requests. He was compliant with some activities and instructions, (e.g., “Tell me what sound you hear,” “Clap your hands,” “Touch your nose,” and “Touch your head”). However, Jacob often engaged in noncompli- ance when relinquishing preferred items. When Jacob was noncompliant, he would yell, cry, say “No work,” push/ throw the item on to the floor, scratch, push over the table, or leave the table and go underneath. Parents reported that many of the problem behaviors that he engaged in while noncompliant often placed himself or others in danger.
Dana was a 7-year-old Caucasian girl diagnosed with ASD and ADHD. Dana was receiving 16 hr a week of ABA services and experienced programming targeting academic skills (reading, writing, and math), social skills, and school readiness skills (e.g., compliance with teacher instruction). Dana also received some services for occupational therapy. Dana was verbal and was capable of reading full sentences as well as solving simple math equations (addition and sub- traction). Dana was also capable of making clear requests using full sentences. Dana was compliant to any requests regarding math but struggled with compliance to requests involving reading. When engaging in noncompliant behav- ior, Dana would scream “No,” start making noises (i.e., “Me! Me! Me!”), or ignore the given instruction. Dana attended online public school in a special education class- room during the completion of this study.
Finally, Deb was an 8-year-old African American girl with an ASD diagnosis, who was provided 20 hr of ABA services a week. During that time, ABA therapists targeted activities of daily living skills, receptive/expressive lan- guage, play skills, academic skills (reading), and commu- nity safety skills. In addition to ABA services, Deb also received services for speech therapy, occupational therapy, and physical therapy. Deb communicated verbally and understood basic task requests. She also was capable of
making verbal requests. She was cooperative with basic instructions (e.g., “Touch your head,” “Touch your shoul- ders,” and “Turn around”). She was capable of doing physi- cal therapy activities with guidance when counting the number of times she was engaging in the given instruction (i.e., 20 sit-ups) but often engaged in noncompliance during this time with her physical therapist. Deb’s physical thera- pist reported difficulty completing sessions, with Deb often screaming, whining, stomping her feet, and/or ignoring the instructions.
The first and second authors served as the experimenters conducting sessions. Both experimenters were graduate stu- dents completing a master’s degree in ABA. In addition, each experimenter had around 2 years of experience directly working with children diagnosed with ASD, providing one- on-one, in-home ABA therapy.
Setting and Materials
We conducted sessions in a room of the participants’ respec- tive homes. The rooms measured around 10 m × 10 m and typically included a desk with two chairs. The material used for this experiment was a data sheet or tablet with the data sheet on it, to collect data on compliance. Many preferred items such as tablets, phones, music toys, pretend play objects, and cars were used with participants who exhibited noncompliance with relinquishing those items (Max, Sam, and Jacob). Dana exhibited problem behavior during reading tasks and so cue cards with different words were included in her sessions. Deb did not require materials specific to the low-p instructions because they involved imitating gross motor movements. The experimenter used a timer to mea- sure the duration of the trials. In addition, the experimenter used a green card, yellow card, blue card, and a red card. Each card was visible throughout the session to signal the specific condition. That is, each condition was correlated with a specific color. The color–condition pairing was ran- domly determined for each participant to reduce the emer- gence of color preferences. The use of the color-coded cards are described further in the “Concurrent Chains Analysis” section.
Dependent Measures
The dependent variable was compliance with low-p instruc- tions. Compliance was defined as engaging in the response that corresponded to the delivered low-p instruction within 10 s of when the instruction was given and completing the instruction. To be considered compliance, the participants needed to complete the instruction without engaging in any problem behavior. Noncompliance was defined as not com- pleting the specified instruction within 10 s or exhibiting any disruptive behaviors, such as yelling, crying, saying “No,” ignoring the instruction, throwing items, or hurting
4 Journal of Positive Behavior Interventions 00(0)
themselves or others. No additional prompts were provided for opportunities to comply with the low-p instruction. The dependent measure was calculated as a percentage by divid- ing the number of compliant trials by the total number of five trials and multiplying that number by 100%. Each ses- sion was broken up into five trials. Data were not collected on compliance with high-p instructions.
To measure the participants’ preferences for one of the three ratios of high-p request sequences, a concurrent chains analysis was conducted. The dependent measure for the concurrent chains analyses was the selection of one of four cue cards, each correlated with a different ratio of the high- p request sequence. Choosing a card was defined as the act of touching one card from the array. Each instance of choos- ing a card was recorded and represented as cumulative selections or a total count. See the “Concurrent Chains Analysis” section for more information on directly evaluat- ing preference of treatment procedures.
Social Validity Questionnaire. A social validity questionnaire was given to the participant’s therapist(s) and/or mother by the experimenter following participation in this study. Ses- sions were conducted in the presence of the therapists or mothers when possible to provide direct observation of the treatment procedures. All therapists and mothers were able to observe at least one session from each condition. How- ever, prior to completing the social validity questionnaire, the experimenter described the procedures and provided open access to any videos of sessions that were available. The experimenter also provided time to answer any ques- tions regarding the procedures the therapists or mothers may have had.
The questionnaire contained six questions and was answered using a 7-point Likert-type scale with 1 indicating the lowest score and 7 indicating the highest score. The questions asked whether the caregivers saw improvement in their child or student’s compliance. It also asked whether the caregivers were satisfied with the improvements, whether they would use the interventions in the future, and which treatment level they preferred, if any. It asked whether the treatments were acceptable to the therapist and whether they thought that increasing the student’s compli- ance was important to them. There was also a question that asked the therapists whether they were likely to suggest this intervention to others who work with the child or to other individuals who have worked, or work, with a child who struggles with noncompliance. The social validity question- naire was completed by each participant’s mother (n = 5) and any therapists working directly with the participant (n = 9). The results of the social validity questionnaire are summarized as means across questions separated by group of mothers and therapists.
Interobserver Agreement and Treatment Fidelity
The experimenter served as the primary data collector for the participants. A secondary observer collected data independently during at least 30% of the sessions in each condition for all participants. The secondary observer was present to observe live sessions. Agreements were defined as both observers either scoring “compliance” or “noncom- pliance” with the low-p instruction during a trial. Because the dependent variable was a binary measure (i.e., compli- ance or noncompliance), disagreements were defined as one observer recording the trial as the opposite of the other. The observers calculated the percentage occurrence of agree- ments by dividing the number of trials that the two observ- ers agreed by the total number of trials. The agreement coefficient for Max, Jacob, Sam, Deb, and Dana was 100%, 98% (range = 88%–100%), 98% (range = 90%–100%), 98% (range = 80%–100%), and 100%, respectively.
We also calculated treatment fidelity for 40% of sessions with all participants. The secondary observer recorded whether the experimenter gave all instructions in the correct order and followed all the procedures properly using a treat- ment fidelity checklist (see Supplemental Material). For example, the secondary observer recorded whether the experimenter provided praise following compliance or whether the experimenter implemented the proper number of high-p instructions followed by the low-p instruction. The number of occasions that the experimenter conducted the procedures appropriately was divided by the total pos- sible number of treatment components that were imple- mented within a session and multiplied by 100%. During all sessions observed, there was 100% treatment integrity.
Experimental Design
We used a combination of a nonconcurrent multiple base- line design across participants and an alternating treatments design. Functional control was demonstrated during the multiple baseline design when greater levels of compliance with the low-p instruction was observed following the stag- gered introduction of the treatment comparison. Functional control between each high-p treatment condition was dem- onstrated during the rapid alternation between the condi- tions of the alternating treatment design. In addition, three participants (Jacob, Dana, and Deb) included a reversal. The multiple baseline design, supplemented by the rever- sals, provided evidence for the effects of treatment in com- parison with a condition in which any treatment was absent (i.e., baseline). The alternating treatments design allowed the experimenters to examine the relative efficacy between the three ratios of high-p request sequences during the direct parametric comparison. Thus, differentiated outcomes
Baida et al. 5
obtained during the alternating treatments design would be indicative of each parameter of the high-p request sequence being evaluated (i.e., 1:1, 3:1, and 6:1) having individual- ized effects on compliance.
Procedures
Pretreatment Assessment of Low-p and High-p Instructions. The low-p instruction was selected based on caregiver- reported difficulty in the home environment and need for behavioral intervention. To ensure that there was a low probability of compliance with the target instruction, the experimenter conducted a brief assessment of 10 trials presenting the caregiver-informed, low-p instruction. The experimenter provided praise for any compliance and ignored any noncompliance during this time. The low-p instruction was included in further treatment evaluation if the participant complied in two or fewer trials. Greater compliance (i.e., three or more trials) would have resulted in a return to open-ended questioning with the caregiver or reevaluation of their participation in the study; however, this did not occur. The low-p instruction for three partici- pants (Sam, Max, and Jacob) was asking them to relinquish access to a highly preferred item. Dana was given four- word sentences to read as her low-p instruction (e.g., the grass is green, the pig is pink, and the bike is red). The low-p instruction for Deb was doing sit-ups, a gross motor activity suggested by her physical therapist.
The high-p instructions were all basic requests that the participant had a history of compliance with and had mas- tered in previous programming. The brief, 10-trial assess- ment was conducted for each high-p instruction and only those instructions in which the participant complied with in all 10 trials were included in further treatment evaluation. A minimum of five different high-p instructions were identi- fied for use with each participant. The high-p instructions for most participants (Sam, Max, Jacob, and Deb) included simple gross motor movements (e.g., “Touch your nose,” “Touch your head,” “Give me a high five,” “Stand up,” “Sit down,” and “Touch the table”). The high-p instructions for Dana included completing one-digit math problems (e.g., 1 + 4, 5 + 2, and 3 + 3).
Baseline. The experimenter placed a colored cue card on the table and presented a verbal prompt (e.g., “You are cur- rently in the blue condition”), which signaled that the base- line condition would be taking place. The experimenter began a baseline trial by presenting the low-p instruction (e.g., “Give back the [preferred item],” “Read this word,” and “Do a sit-up”). The experimenter provided verbal and physical (e.g., tickles) praise for compliance with the low-p instruction during the baseline sessions. If the participant did not comply after 10 s, the experimenter physically prompted the response with light guidance and provided a
brief corrective statement (e.g., “This is how you give back the [preferred item],” “This is how you do a sit-up”). For example, the experimenter would lightly press on Dan’s back to help guide him in the sit-up. Because the low-p response of reading a sentence could not be physically prompted for Dana, the experimenter provided the correc- tive answer before moving on to the next trial (e.g., “This says [target sentence]”). Trials were conducted as they natu- rally occurred (e.g., when the participant needed to relin- quish items during early intervention programming, when the participant was scheduled to complete reading home- work) or during specific services (e.g., gross motor move- ments during physical therapy).
High-p Treatment Comparison. Similar to baseline, the ses- sion began with the colored cue card corresponding to the specific condition being placed on the table. The cue card remained in view throughout the session. During the high-p treatment comparison, the experimenter verbally presented one, three, or six high-p instructions, depending on the spe- cific condition the participant was experiencing. Immedi- ately following the sequence of high-p instruction(s), the participant was presented with the low-p instruction. If the participant complied with any of the instructions, high-p or low-p, they received praise and or tickles from the experi- menter (i.e., no additional reinforcers were included in any high-p request sequence). Similar to baseline, the partici- pant was physically prompted to complete the low-p instruction following noncompliance. For the high-p (1) condition, the experimenter presented the participant with one instruction immediately before presenting the low-p instruction. The high-p (3) condition consisted of three high-p instructions prior to the low-p instruction and the high-p (6) condition consisted of six high-p instructions before the low-p instruction. When presented in a sequence, instructions were provided in rapid succession, such that the next instruction was provided within 5 s of compliance with the previous instruction. Trials were discontinued if the par- ticipant did not comply with a high-p instruction. The experimenter did not collect data on noncompliance with the high-p instruction; however, it was a rare occurrence and was reportedly observed less than 5 times across all participants. The order of the conditions was quasi-random- ized, with each condition having to be experienced before it could be repeated.
Concurrent Chains Analysis. A concurrent chains analysis was conducted following the completion of the high-p treat- ment comparison and was used to determine which treat- ment conditions were preferred by the participant (Hanley, 2010). The baseline from the high-p treatment comparison was used as the control condition in the concurrent chains analysis. In other words, if the participant selected the color cue card associated with the control condition, a session
6 Journal of Positive Behavior Interventions 00(0)
was initiated in which only low-p instructions were deliv- ered (i.e., the absence of any high-p sequence).
The experimenter placed the assigned color cue cards on the table in front of the participant. The participant was pre- sented with all four color cue cards for each trial simultane- ously. This included one color that represented the control condition, that is, one that represented the high-p (1) condi- tion, one that represented the high-p (3) condition, and one that represented the high-p (6) condition. The experimenter told the participant to “Pick one.” The experimenter changed the order in which she placed the cue cards on the table for each trial. During each trial, the participant chose one cue card, all other cue cards were removed, and they then expe- rienced the corresponding condition associated with the selected cue card. For example, if the participant chose the red card, three high-p instructions were presented prior to the low-p instruction (i.e., the participant experienced one session of the high-p [3] condition from the treatment com- parison). The array of all four cue cards were then replaced onto the table and the process was repeated. The experi- menter conducted the trials of the concurrent chains analy- sis until the participant consistently picked the same condition. If the participant did not choose a card, the experimenter verbally prompted them to pick a card and the trial was not counted.
For two participants (Dana and Max), the array was restricted in a separate phase by removing the most consis- tently selected option. For example, if the participant selected
the high-p (3) condition 4 times in a row, the high-p (3) condi- tion would have been determined to be the most preferred choice and the color cue card associated with that condition would have been removed from further arrays. The therapist would have placed three cue cards on the table (control, high-p [1], and high-p [6]) and would have allowed the participant to choose from those available options. The purpose of conduct- ing the restricted array was to determine a potential hierarchy of preference for conditions (Jessel & Ingvarsson, 2017). A hierarchy cannot be established when the most preferred con- dition is always available in the array because the participant is likely to select that option every time.
Results
Figure 1 displays the results of the treatment comparison for two participants (Max and Sam). During baseline, Max (top panel) did not comply with any of the low-p instructions. When high-p instructions were introduced, there was an immediate increase in compliance across all three treatment conditions; however, compliance during the high-p (6) con- dition was low and eventually decreased until Max was no longer complying with any low-p instructions (M = 8%, range = 0%–20%). Moderate levels of compliance were observed during the high-p (1) condition (M = 44%, range = 20%–60%), with the highest levels of compliance observed during the high-p (3) condition (M = 96%, range = 80%–100%).
Sam (see Figure 1; bottom panel) exhibited a decreasing trend and overall low level of compliance during baseline (M = 20%, range = 0%–40%). When high-p instructions were introduced, there was an increase in compliance, with differentiation observed across high-p conditions. Compliance during the high-p (6) condition was the lowest (M = 63%, range = 60%–80%), whereas moderate levels of compliance was observed during the high-p (1) condition (M = 77%, range = 60%–100%). The highest levels of compliance was observed during the high-p (3) condition (M = 97%, range = 80%–100%).
The results of the three participants (Jacob, Dana, and Deb) who experienced a reversal are presented in Figure 2. A reversal was conducted for these participants because all high-p sequences tended to improve compliance and differ- entiated outcomes between the treatments was not initially observed. Jacob (top panel) rarely complied with the low- probability instruction (M = 12%, range = 0%–20%) dur- ing the baseline. High and undifferentiated percentages of compliance were obtained in the high-p (1) condition (M = 84%, range = 80%–100%), high-p (3) condition (M = 90%, range = 80%–100%), and high-p (6) condition (M = 80%, range = 60%–100%). When the baseline was reintroduced, compliance returned to low levels (M = 13%, range = 0%– 20%). The most preferred condition (i.e., three high-p
0
20
40
60
80
100
High-p (1) High-p (3) High-p (6)
BL High-p Comparison
Max
5 10 15 20 25 30 35
0
20
40
60
80
100
Sessions
Sam
C om pl ia nc e (% )
Figure 1. The Results of the High-p Comparison for Max and Sam. Note. High-p = high-probability; BL = baseline.
Baida et al. 7
instructions) was then implemented alone and compliance again increased to elevated levels (M = 93.3%, range = 80%–100%).
Dana (see Figure 2; middle panel) did not comply with any low-p instructions during the baseline. Once the high-p instruction sequences were introduced, there was an imme- diate increase in compliance across all three treatment con- ditions. Although relatively indifferent, the highest levels of compliance were observed during the high-p (1) condition (M = 97%, range = 80%–100%), followed by the high-p (6) condition (M = 91%, range = 60%–100%) and high-p (3) condition (M = 86%, range = 60%–100%). After returning back to the baseline condition, compliance with the low-p instructions remained high and comparable to the previous treatment conditions (M = 87%, range = 60%– 100%). Therefore, a return to treatment was deemed unnecessary.
Deb (see Figure 2; bottom panel) initially demonstrated low levels of compliance during the baseline condition (M = 22%, range = 0%–60%). Compliance increased after implementing the high-p treatment conditions; however, the
level of improvement varied with both the high-p (1) condi- tion (M = 80%) and high-p (3) condition (M = 83%, range = 80%–100%) resulting in higher levels of compliance in comparison with the high-p (6) condition (M = 56%, range = 40%–60%). Following the return to baseline, compliance remained somewhat elevated but below overall treatment effects (M = 56%, range = 40%–60%). Therefore, the most preferred treatment, high-p (3), was reintroduced and com- pliance again increased (M = 80%, range = 60%–100%).
The results of the high-p treatment comparison for all participants is summarized in Figure 3. Overall, the high-p request sequence with the most compliance included three high-p instructions (M = 90%, range = 83%–97%). This was followed by the high-p request sequence including a single high-p instruction (M = 76%, range = 44%–97%). Finally, the high-p request sequence with the least amount of compliance included six high-p instructions (M = 60%, range = 8%–91%).
The results of the concurrent chains analyses is presented in Figure 4. Three participants displayed a preference for the high-p (3) condition (Deb, Sam, and Jacob), whereas the preferences for two participants varied (Dana and Max). Dana displayed a preference for the high-p (6) condition. The array was then restricted to determine a potential hier- archy of preference and Dana began to select the high-p (3) condition. Once the full array was reintroduced, Dana returned to selecting the high-p (6) condition. Max pre- ferred the high-p (1) condition and, once the array was restricted, removing the high-p (1) option, he began select- ing the high-p (3) condition. This indicated a secondary preference for the high-p (3) condition for both participants
0
20
40
60
80
100
Jacob
BL High-P Comparison HP-3BL
High-p (1) High-p (3) High-p (6)
0
20
40
60
80
100
C om pl ia nc e (% )
Dana
5 10 15 20 25 30 35
0
20
40
60
80
100
Sessions
Deb
Figure 2. The Results of the High-p Comparison for Jacob, Dana, and Deb. Note. High-p = high-probability. BL = baseline.
Hi gh -p (1 )
Hi gh -p (3 )
Hi gh -p (6 )
0
20
40
60
80
100
C om pl ia nc e (% )
BL High-p Comparison
Figure 3. Summary Results of the High-p Treatment Comparison Across Participants. Note. Bars represent the mean. Each symbol represents an individual participant. High-p = high-probability; BL = baseline.
8 Journal of Positive Behavior Interventions 00(0)
who did not initially select the high-p (3) condition. In sum- mary, the high-p (3) condition was the primary preference for three of the five participants and was the secondary pref- erence for the remaining two participants who experienced the restricted array.
Once the participants completed the concurrent chains analysis, the mothers and therapists were given the social validity questionnaire to complete (see Figure 5). Overall,
the mothers and therapists were highly satisfied with the improvement in noncompliance and found treating non- compliance to be very important. The reported likelihood of using the high-p procedures was somewhat lower for both the mothers and therapists; however, ratings were still posi- tive. Finally, the mothers and therapists reported that they were very likely to recommend the high-p procedures and found the overall treatment to be highly acceptable.
0
1
2
3
4
5
C um ul at iv e S el ec ti on s
Sam
2 4 6 8 10
0
1
2
3
4
5
Sessions
Jacob
5 10 15 20
0
2
4
6
8
10
Sessions
Max
0
2
4
6
8
10
Dana
Full Array Restricted Array
Full Array
0
1
2
3
4
5
Deb
High-p (1) High-p (3) High-p (6)
Control
Figure 4. The Results of the Concurrent Chains Analyses. Note. High-p = high-probability.
Baida et al. 9
In addition, the mothers and therapists were given the opportunity to identify which procedures they preferred (see Figure 6). The majority of individuals selected the high-p (3) condition, with the second highest selected con- dition being high-p (1). Neither the mothers nor the thera- pists selected the high-p (6) condition or the baseline condition.
Discussion
The high-p request sequence was effective in increasing all three participants’ compliance. The high-p request sequence with three instructions was the most effective condition for two out of the five participants and was just as effective for the remaining participants. Furthermore, the high-p (3) con- dition was overall most preferred across participants, care- givers, and therapists. This study provides empirical support for the standardized use of the high-p sequence with three instructions. However, it is important to point out that indi- vidual differences were observed and that the high-p (1) condition was the second most effective procedure when differentiated outcomes were obtained.
A potential conclusion regarding the partial effective- ness of the high-p (1) sequence is that this condition may not have built up enough momentum to compete with the high-p (3) sequence. Behavioral momentum is an analogy that explains the effects of resistance to change in the pres- ence of reinforcement. In other words, the more reinforce- ment that is provided in the context of compliance, the more likely the child will comply with the low-p instruction (Nevin, 1996). Therefore, the single high-p instruction with minimal access to reinforcement before the presentation of the low-p instruction could have contributed to less momen- tum, and consequently, less compliance. That being said, we cannot entirely attribute these findings to behavioral momentum alone because the theory would suggest that compliance would correspondingly improve as the sequence is extended with greater amounts of high-p instructions.
1 2 3 4 5 6 7
Ratings
A bb re vi at ed Q ue st io ns
Therapist (n = 9) Mother (n = 5)
Lowest Highest
Satisfied with improvements in noncompliance
Importance of compliance
Likelihood of using high-p
Likelihood of recommending
high-p
Acceptability of high-p
Figure 5. Results of the Social Validity Questionnaire. Note. Bars represent mean values. High-p = high-probability.
0
2
4
6
8
Treatment Condition
T ot al Se le ct io ns
Mother (n = 5) Participant (n = 5)
Therapist (n = 9)
High-p (1) High-p (3) High-p (6)
Figure 6. Preference for High-p Conditions Across Participants, Mothers, and Therapists. Note. Participant data representative of the results of the concurrent chains analyses. High-p = high-probability.
10 Journal of Positive Behavior Interventions 00(0)
We found that the high-p (6) condition was the least effective at increasing compliance for most of the partici- pants. This suggests that momentum in the context of com- pliance has its limits and increasing the high-p instructions will not necessarily always produce an increase in compli- ance. The reduction in compliance creates a sort of bimodal curve, implicating a potential “goldilocks zone” that bal- ances the momentum from reinforcement presented after the high-p instruction and the effort of complying with the high-p instruction. That is, providing too few high-p instructions (e.g., high-p 1) may limit access to the rein- forcement crucial to build enough momentum to evoke compliance with the low-p instruction. On the contrary, with too many high-p instructions (e.g., high-p 6) the child may experience overexertion with the repeated require- ment to complete specified tasks. In fact, it seems that the reinforcement for the high-p instructions may largely be contributing to the benefits of the high-p request sequence, rather than the instructions themselves (Bullock & Normand, 2006; Zuluaga & Normand, 2008).
For example, Zuluaga and Normand (2008) evaluated the effects of using the high-p request sequence with and without reinforcement to improve the compliance of two boys diagnosed with developmental disabilities. The authors found that compliance only increased when there was a reinforcement contingency in place for the high-p instructions. Furthermore, Bullock and Normand (2006) found that the high-p instruction itself was not necessary and a sequence of response-independent presentations of reinforcers (i.e., noncontingent reinforcement) could improve compliance. For example, the researchers pre- sented preferred edibles every 10 s without high-p instruc- tions prior to the low-p instruction, regardless of the participant’s behavior, and found improvements in compli- ance comparable to that of the high-p request sequence. Therefore, the antecedent introduction of reinforcers, and not the high-p instructions, may be far more integral to the effectiveness of the high-p request sequence.
Although response-independent reinforcers can improve behavioral momentum (Nevin et al., 1990), it may be less likely to be considered socially acceptable when the high-p request sequence is found to be equally effective. For exam- ple, a caregiver may be less inclined to agree to implement a treatment that involves providing highly preferred items before the child has even been asked to complete any tasks. It may be considered more socially acceptable to have stu- dents comply with instructions before providing reinforcers to ensure that students continue to, in some way, “earn their rewards.” It is difficult to draw any conclusions regarding comparative acceptability of response-independent rein- forcement strategies and the high-p request sequence because only variations of the high-p request sequence were presently conducted. We are, therefore, limited to interpret- ing a parametric range of response-contingent behavioral
expectations (i.e., compliance with one, three, or six high-p instructions). Interestingly, when the caregivers were all informally asked prior to this study if one condition should include nine high-p instructions, all adults found the use of nine instructions to be excessive, and the condition with the most high-p instructions was agreed upon to include six. Thus, from a practical and socially relevant standpoint, if the caregivers are considering presenting reinforcers con- tingent on some behavioral requirement in a high-p request sequence, the sequence should be capped to reduce the effort of the (a) participant in complying with those instruc- tions, and (b) the caregiver implementing the treatment.
Normand et al. (2010) presented one of the few failures of the high-p (3) procedure with a typically developing pre- schooler who exhibited noncompliance when instructed to clean up preferred toys. Interestingly, the authors found that not only did compliance with the low-p remain low but also the participant began to exhibit noncompliance in the pres- ence of the high-p instructions as well when signals were presented, indicating the eventual presentation of the low-p instruction. Therefore, it may be best, in some cases, not to rely on a set sequence of one, three, six, or other high-p instructions that can become predictive of the low-p instruc- tion and instead use varied sequences that are unpredictable to the participant (cf. Davis et al., 1992). Future researchers may want to consider evaluating the influence of predict- ability on compliance before establishing the high-p (3) sequence as the standard. In fact, unpredictable reinforce- ment is often preferred (e.g., Daly, 1985, 1989) and has been suggested to be a potentially effective treatment com- ponent (Jessel et al., 2016).
A limitation of this study is the continued necessity of the high-p request sequence to maintain treatment effects. We did observe some maintenance of compliance in two of the three participants who experienced the return to base- line. However, the levels of compliance were comparable to treatment effects in only one participant (Dana). This is important to note because the high-p procedures should be a temporary treatment and eventually faded out. In other words, caregivers should not have to continuously present high-p instructions, and the expectations are likely going to be a return to the typical arrangement of low-p instructions in the classroom or home. Ardoin et al. (1999) conducted a study beginning with the high-p request sequence as an ini- tial treatment component and faded out the instructions to transfer stimulus control to the low-p instructions. The researchers systematically decreased the number of high-p instructions before the low-p instruction, based on the par- ticipants’ performance. The high-p instructions were suc- cessfully faded out for two out of three of the participants, with these positive outcomes maintaining in follow-up probes conducted 3 weeks later. Future research may want to consider including fading in further studies of the high-p request sequence to elucidate the entire process, from initial
Baida et al. 11
treatments requiring some form of high-p request sequence to the eventual goal of eliminating supports to improve compliance.
While fading the high-p instructions could potentially help return the child to the more natural home and school preparations, other strategies beyond the breadth of this study could be incorporated as well to improve the ecologi- cal validity of the high-p request sequence. For example, future researchers may want to conduct similar evaluations in the home with caregiver implementers or in the school with teachers. That is, program effectiveness and reports of social validity may be impacted by the inclusion of natural agents of change in the intended setting. Broad applicabil- ity is also limited by the fact that this study only included five participants. Future research conducted in more eco- logically relevant settings should consider including a larger sample of participants to improve interpretations of external validity. Furthermore, the high-p request sequence was, similar to other studies, conducted over a relatively brief amount of time and it is unknown whether treatment fatigue to these antecedent procedures without additional supports would maintain compliance over longer periods. Longitudinal studies could help to identify the mainte- nance of the effects of the high-p request sequence if care- givers or teachers were to incorporate the procedures into regular programming for months or years.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Ethical Approval
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee 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.
ORCID iD
Joshua Jessel https://orcid.org/0000-0002-1649-2834
Supplemental Material
Supplementary material for this article is available on the Journal of Positive Behavior Interventions website at http://jpbi.sagepub. com.
References
Ardoin, S. P., Martens, B. K., & Wolfe, L. A. (1999). Using high probability instruction sequences with fading to increase student compliance during transitions. Journal of Applied Behavior Analysis, 32(3), 339–351. https://doi.org/10.1901/ jaba.1999.32-339
Bullock, C., & Normand, M. P. (2006). The effects of a high- probability instruction sequence and response-independent reinforcer delivery on child compliance. Journal of Applied Behavior Analysis, 39(4), 495–499. https://doi.org/10.1901/ jaba.2006.115-05
Centers for Disease Control and Prevention. (2018). Prevalence of autism spectrum disorder among children aged 8 years— Autism and developmental disabilities monitoring network. Surveillance Summaries, 67(6), 1–23.
Cipani, E. (1998). Three behavior functions of classroom non- compliance: Diagnostic and treatment implications. Journal of Applied Behavior Analysis, 13(2), 66–72. https://doi. org/10.1177/108835769801300201
Cowan, R. J., Abel, L., & Candel, L. (2017). A meta-analysis of single-subject research on behavioral momentum to enhance success in students with autism. Journal of Autism and Developmental Disorders, 47(5), 1464–1477. https://doi. org/10.1007/s10803-017-3076-6
Daly, H. B. (1985). Observing response acquisition: Preference for unpredictable appetitive rewards obtained under conditions predicted by DMOD. Journal of Experimental Psychology, 11(2), 294–316. https://doi.org/10.1037/0097-7403.11.2.294
Daly, H. B. (1989). Preference for unpredictable food rewards occurs with high proportion of reinforced trials or alcohol if rewards are not delayed. Journal of Experimental Psychology, 15(1), 3–13. https://doi.org/10.1037/0097-7403.15.1.3
Davis, C. A., Brady, M. P., Williams, R. E., & Hamilton, R. (1992). Effects of high probability requests on the acquisition and generalization of responses to requests in young children with behavior disorders. Journal of Applied Behavior Analysis, 25(4), 905–916. https://doi.org/10.1901/jaba.1992.25-905
Ertel, H., Wilder, D. A., Hodges, A., & Hurtado, L. (2019). The effect of various high-probability to low-probability instruc- tion ratios during the use of the high-probability instructional sequence. Behavior Modification, 43(5), 639–655. https://doi. org/10.1177/0145445518782396
Hanley, G. P. (2010). Toward effective and preferred programming: A case for the objective measurement of social validity with recipients of behavior-change programs. Behavior Analysis in Practice, 3(1), 13–21. https://doi.org/10.1007/BF03391754
Jessel, J., Hanley, G. P., & Ghaemmaghami, M. (2016). A transla- tional evaluation of transitions. Journal of Applied Behavior Analysis, 49(2), 359–376. https://doi.org/10.1002/jaba.283
Jessel, J., & Ingvarsson, E. T. (2017). Using rich-to-lean transitions following errors during discrete-trial instruction. European Journal of Behavior Analysis, 18(2), 291–306. https://doi.org /10.1080/15021149.2017.1404396
12 Journal of Positive Behavior Interventions 00(0)
Kaat, A. J., & Lecavalier, L. (2013). Disruptive behavior disor- ders in children and adolescents with autism spectrum disor- ders: A review of the prevalence, presentation, and treatment. Research in Autism Spectrum Disorders, 7(12), 1579–1594. https://doi.org/10.1016/j.rasd.2013.08.012
Lee, D. L., Belfiore, P. J., Scheeler, M. C., Hua, Y., & Smith, R. (2004). Behavioral momentum in academics: Using embedded high-p sequences to increase academic productiv- ity. Psychology in the Schools, 41(7), 789–801. https://doi. org/10.1002/pits.20014
Mace, F. C., & Belfiore, P. (1990). Behavioral momentum in the treatment of escape motivated stereotypy. Journal of Applied Behavior Analysis, 23(4), 507–514. https://doi.org/10.1901/ jaba.1998.21-123
Nevin, J. A. (1996). The momentum of compliance. Journal of Applied Behavior Analysis, 29(4), 535–547. https://doi. org/10.1901/jaba.1996.29-535
Nevin, J. A., Tota, M. E., Torquato, R. D., & Shull, R. L. (1990). Alternative reinforcement increases resistance to change: Pavlovian or operant contingencies? Journal of the Experimental Analysis of Behavior, 53(3), 359–379. https:// doi.org/10.1901/jeab.1990.53-350
Normand, M. P., Kestner, K., & Jessel, J. (2010). An analysis of stimuli that influence compliance during the high-probability instruction sequence. Journal of Applied Behavior Analysis, 43(4), 735–738. https://doi.org/10.1002/bin.1474
Pitts, L., & Dymond, S. (2012). Increasing compliance of chil- dren with autism: Effects of programmed reinforcement for high-probability requests and varied inter-instruction inter- vals. Research in Autism Spectrum Disorders, 6(1), 135–143. https://doi.org/10.1016/j.rasd.2011.03.013
Riviere, V., Becquet, M., Peltret, E., Facon, B., & Darcheville, J. C. (2011). Increasing compliance with medical examina- tions requests directed to children with autism: Effects of a high probability request procedure. Journal of Applied Behavior Analysis, 44(1), 193–197. https://doi.org/10.1901/ jaba.2011.44-193
Zuluaga, C. A., & Normand, M. P. (2008). An evaluation of the high-probability instruction sequence with and with- out programmed reinforcement for compliance with high- probability instructions. Journal of Applied Behavior Analysis, 41(3), 453–457. https://doi.org/10.1901/jaba. 2008.41-453
View publication stats