Abstract
The use of restrictive and punishment procedures has been debated by professionals in the field of behavior analysis since the field's early days, being joined by disability rights advocates, particularly those in the neurodiversity movement, in modern decades. Despite arguments to eliminate or reduce the use of these procedures, no current data exists to provide an estimate of their frequency of use. The current study surveyed 534 behavior analysts on the frequency of their use of restrictive and punishment procedures with learners on their caseload at the time of the study. Results include descriptive summaries of the use of multiple procedures, as well as differences in use across several personal and professional factors. Implications for future directions in research and practice are discussed.
Punishment and other aversive procedures have been subjects of evaluation and debate within the behavior analytic field for several decades, yet consensus on their ethics and implementation limitations continues to elude the discipline (Axelrod, 1987; Snell, 1987; Wilkenfeld & McCarthy, 2020). Notably, disability rights advocates and proponents of the neurodiversity movement have voiced opposition to behavior analytic techniques, citing concerns over historically employed technologies perceived as abusive, particularly those rooted in the behavioral principle of punishment (Ne’eman, 2010; Veneziano & Shea, 2023). Despite this longstanding debate, there remains a notable gap in the literature regarding the current usage of these procedures. To address this gap, the present article presents the findings of a comprehensive survey aimed at investigating the contemporary practices and attitudes surrounding the utilization of punishment and restrictive procedures within behavior analysis. By shedding light on the current landscape, we aim to provide valuable insights for researchers, practitioners, and advocates navigating the complexities of ethical decision-making and intervention strategies in behavior analysis.
While Applied Behavior Analysis (ABA) has demonstrated effectiveness in promoting positive behavior change (Baer et al., 1968), it remains a subject of ongoing debate within the disability and neurodiversity communities. Although many practitioners aim to align ABA practices with the values of client-centered care, tensions persist regarding historical and current uses of behavior analytic technologies, particularly concerning restrictive and punishment-based procedures. Critics, including disability rights advocates and proponents of the neurodiversity movement, have raised concerns about the ethical application of ABA and its potential for harm (Ne’eman, 2010; Veneziano & Shea, 2023). These tensions highlight the need for continued dialogue, self-reflection within the field, and a commitment to practices that honor the autonomy, dignity, and lived experiences of individuals receiving behavior analytic services.
A note on terminology
It is essential to clarify the usage of terms like “punishment” and “restrictive procedures” within the context of behavior analysis in the current article, as they are derived directly from The Ethics Code for Behavior Analysts (ECBA; BACB, 2020). We have chosen to use these terms in alignment with ethical guidelines. While these terms are commonly used, it is important to note that there is no official definition of restrictive procedures, and punishment specifically refers to any consequence that effectively reduces the future probability of a behavior occurring (Cooper et al., 2020). This distinction is crucial because, in behavior analysis, the effectiveness of a consequence in reducing the likelihood of a behavior's recurrence determines whether it qualifies as punishment. If a consequence does not result in a decrease in the future occurrence of a behavior, it cannot be considered punishment according to the principles of behavior analysis. Reprimanding is typically considered a form of punishment. For example, an analyst or caregiver might deliver a reprimand believing it will reduce the future probability of the behavior occurring. However, sometimes the client or child enjoys the attention from the reprimand, which can actually increase the future probability of the behavior. Therefore, behavior analysts must carefully assess and analyze the effects of consequences to ensure ethical and effective intervention practices.
Ethical Guidelines of Using Restrictive and Punishment-Based Behaviors
The ECBA (BACB, 2020) addresses restrictive and punishment-based procedures (RPBPs) in multiple standards. Standard 2.15 states “They [cis: behavior analysts] recommend and implement restrictive or punishment-based procedures only after demonstrating that desired results have not been obtained using less restrictive means, or when it is determined by an existing intervention team that the risk of harm to the learner outweighs the risk associated with the behavior-change intervention.” Furthermore, this section states that behavior analysts only implement RPBPs after following all required review processes, and continually monitor the effectiveness of RPBPs throughout treatment. In standard 4.06, the ECBA states that when supervising and providing training, behavior analysts should focus on positive reinforcement procedures. In standard 2.14, the ECBA advises practitioners to limit the use of RPBPs and emphasize positive reinforcement procedures in both practice and training. Further, the ECBA is clear that when underlying medical conditions are present or suspected, or when otherwise indicated, behavior analysts engage in consultation with and/or referral to other professionals and the client to better address the needs of the learner (BACB, 2020). We conceptualized the behavior decelerators discussed in this article under the term “RPBP” based on the language in the ECBA; however, the ECBA does not define RPBPs within the Standards or glossary of the document.
Defining RPBPs
Defining restrictive procedures within behavior analysis poses inherent challenges due to the subjective nature of their perception, varying among individuals (Cooper et al., 2020). To address this complexity, our study adopts the term “restrictive and punishment-based procedures” to encompass a spectrum of techniques, including punishment and extinction methods. While some interventions, like physical restraint, may be widely viewed as highly restrictive, others, such as attention extinction procedures, might impose different subjective constraints depending on the individual's perspective (Cooper et al., 2020). RPBPs cover a broad range of approaches intended to decelerate behavior, including positive punishment involving the introduction of an aversive stimulus following behavior, negative punishment entailing the removal of a desirable stimulus, and extinction procedures involving the removal or blocking of access to desired consequences (Cooper et al., 2020).
Several specific procedures fall under the umbrella of RPBPs, including response interruption and redirection, overcorrection, timeout, and response blocking (Pokorski & Barton, 2021). It is essential to distinguish between merely describing a procedure and confirming its function through observed behavioral changes, particularly when delineating between punishment and extinction interventions (Cooper et al., 2020). This nuanced understanding is crucial for ensuring that the ethical implications and efficacy of RPBPs are adequately evaluated and applied within behavior analytic practice, thereby promoting responsible and effective intervention strategies. It is also important to note that many of the strategies under this umbrella are employed by nonbehavior analysts, such as healthcare workers, teachers, and parents. Notably, physical restraint is often researched in acute healthcare settings (Minnick et al., 2007).
RPBPs and Applied Behavior Analysis
The controversy surrounding the use of restrictive procedures in behavior analysis has deep historical roots, dating back to the late 1990s with the introduction of Positive Behavior Intervention and Support (PBIS) and the emergence of Self-Injurious Behavior Inhibiting System (SIBIS). During this period, debates over the use of RPBPs sparked intense discussions within the field, driven by concerns over their ethical implications and potential negative effects. Notably, behavior analysts actively advocated against the use of restrictive procedures during this time, marking a significant shift in the field's perspective. Carr et al. (2002) note that these early debates paved the way for the development and promotion of alternative strategies, such as PBIS, which prioritize proactive, preventive, and positive approaches to behavior change. These efforts underscored a transition towards person-centered, positive interventions aimed at supporting behavior change while minimizing the use of restrictive procedures.
As a result of this advocacy, current behavior analysis considers it essential to contextualize the use of RPBPs within the framework of addressing behaviors that significantly interfere with an individual's wellbeing and participation in meaningful activities. Behavior analysts aim to enhance clients’ overall quality of life by targeting behaviors that pose risks to their wellbeing or impede their development and inclusion. Thus, while RPBPs may be employed, their use should be guided by the overarching goal of promoting clients’ active involvement and improving their overall functioning in various contexts.
Alternatives to RPBPs
Arguments for the use of less-restrictive procedures can be seen across the behavior analytic literature and in the ECBA (BACB, 2020; Snell, 1987). These procedures include assent-based practice (redacted for review, 2023; Breaux & Smith, 2023; Morris et al., 2021), reinforcement strategies (e.g., differential reinforcement), and antecedent strategies. Antecedent strategies are strategies that manipulate a stimulus that is contingency independent, ultimately leading to an improved learning environment. The primary purpose of antecedent strategies is to prevent behavior and/or to teach necessary skills in a nonrestrictive way (Snell, 1987), while differential reinforcement aims to increase the probability of behaviors (Cooper et al., 2020). Recent research found that antecedent manipulations and differential reinforcement were more effective than extinction or punishment procedures at reducing problem behavior (Chazin et al., 2022; MacNaul & Neely; 2018; Trump et al., 2020), indicating not only do these procedures result in a less-intrusive behavior program, they may also be a more effective alternative for behavior change.
Despite the need for less-restrictive interventions being recommended, particularly for severe behaviors, several decades ago (Snell, 1987), current research and practice continues to implement punishment procedures without the use of alternative procedures (Lydon et al., 2015; Pokorski & Barton, 2021). Some trends indicate a move forward for the field, such as a decrease in the number of published studies using only punishment procedures over time (Lydon et al., 2015). Even with a decrease in the use of only punishment procedures, Lydon et al. (2015) observed about 30% of published behavior analytic articles in the 2010s reported using punishment alone. Further, they reported a significant lack of the collection and reporting of social validity data (Lydon et al., 2015).
Neurodiversity and Behavior Analysis
As described in previous sections, behavior analysts have often been advocates for progress in the field related to the use of RPBPs. Another group that has provided a voice for change are disability advocates, particularly those associated with the neurodiversity movement. The neurodiversity paradigm is a set of beliefs that, at their core, are founded on the assumption that autism and other neurodiversity experiences are caused by biological factors leading to natural human variation, which alone does not lead to a disorder. Rather, the individual is labeled as disordered based on society's nonacceptance of diversity and lack of accommodations and supports for individuals with differences (Kapp, 2013). This set of beliefs is aligned with the social model of disability. The neurodiversity movement, then, is an advocacy movement made up of neurodivergent people and allies to change how society views neurodiversity, and indeed, change society to be more accommodating for neurodivergent people. Within the neurodiversity movement, autism is not a disorder, but rather an evolutionary variation in human functioning that comes with its own strengths and barriers, based on the current norms in society, that should be celebrated rather than prevented or treated (Kapp, 2013). This perspective is echoed by behavior analysis, which posits that behavior is a result of environmental contingencies and an individual's learning history. As such, behavior analysts attempt to alter environmental variables to provide the necessary supports for individuals who experience a mismatch in their environment and support needs.
Current Study
The current study aimed to summarize practicing behavior analysts’ rate of use of RPBPs and some alternatives to RPBPs. Our research questions were: (1) What is the current usage of multiple RPBPs, antecedent procedures, and consultation and referral by behavior analysts, defined as the frequency of learners on their current caseload with whom they use the procedures? and (2) What demographic and practice variables may be related to the usage of these procedures? The answers to these questions will provide a rough baseline estimate of the use of certain procedures, as well as some potential predictors of use of these procedures, which may guide the field in recommendations for training, research, and practice.
Methods
Survey Development and Content
A survey was developed to collect data on practicing BACB certificants’ current usage of a variety of applied behavior analytic practices (i.e., with their current caseload). The survey included demographic items, as well as information about respondents’ current caseloads and number of learners with whom they currently use several procedures. Inclusion criteria for participation was holding a Board Certified Behavior Analyst (BCBA), Board Certified Assistant Behavior Analyst (BcaBA), or BCBA-D (Doctoral level BCBA) credential. Before the survey began the participants affirmed they were an active certificant. Certificant numbers were not gathered, due to the survey being anonymous.
Demographic items were selected based on (a) their ability to describe the sample in relation to external validity and (b) their utility in evaluating characteristics and experiences that may be related to use of procedures. Participants’ professional experiences, such as primary client ages and disabilities (aggregate/general, rather than individual), work setting, and years as a BCBA and Registered Behavior Technician (RBT) prior to BCBA were collected to describe the sample and as potential predictors of procedure use. It was anticipated that there may be some characteristics of respondents’ professional setting and experiences that may influence use of procedures. For example, a person working in a residential setting may be more likely to utilize more RPBPs than a person whose primary setting is in an early intervention day program. Personal experiences with autism and other disabilities were also queried, including whether respondents had a diagnosis of autism or another disability, whether they had children with autism, and whether they had other family members with autism. Finally, participants were asked about their familiarity with the neurodiversity movement and their agreement with statements about different models of disability. Given current controversies in the field of behavior analysis, we anticipated these variables may predict differential use of behavior analytic practices.
Several, but not an exhaustive list of, RPBPs were included in the survey, as well as items on general antecedent procedures as a precursor to implementing consequence procedures, and the extent to which practitioners consulted or referred to other professionals when an underlying condition may be present. We would like to reiterate that each procedure was selected as a result of being defined as a punishment procedure (Cooper et al., 2020) and/or having some source indicating its potential restrictiveness. Inclusion does not indicate that all learners would find the procedure to be restrictive or that the procedure would consistently result in punishment (e.g., reduce the future probability of the behavior). Instead, inclusion is based on existing literature; however, the final determinant of a restrictive procedure is the recipient of the procedure. Further, we recognize that each included procedure may be perceived along a range of restrictiveness or aversiveness.
The RPBPs selected included escape extinction, sensory extinction, attention extinction, overcorrection, contingent exercise, exclusionary time-out, physical restraint, and response blocking. Cooper et al.’s (2020) definition of restrictive procedures was used as the primary source to develop this survey. Each included RPBP is defined in Table 1. Response blocking, contingent exercise, and overcorrection were included in the survey and are often considered positive punishment procedures, as they include the addition of a stimulus with the intention of, and research support documenting their success at, reducing the probability of the targeted behavior (Cooper et al., 2020). Escape extinction, which eliminates the ability of a person to escape or avoid certain stimulus conditions contingent on a behavior, was included as withholding of reinforcement (e.g., escape/avoidance) may be perceived as painful or uncomfortable to the individual experiencing extinction (Chazin et al., 2022). Additional procedures were included in the survey due to literature suggesting they may be restrictive to recipients: attention extinction (Delahooke, 2020; Quenell & Allison, 2007), sensory extinction (Rincover & Devany, 1982; Rincover et al., 1979), physical restraint, and exclusionary time-out (Gast & Nelson, 1977). Reprimands and response cost procedures were not included due to a preliminary literature search finding no information on the restrictiveness or aversiveness of the procedures. Questions about the use of alternative procedures were broad and included questions on how many learners they implemented antecedent interventions for a period prior to implementing RPBPs, and for how many learners they referred or consulted with other professionals when an underlying condition was present or suspected.
Definitions and Examples of RPBPs.
Content Review and Ethical Approval
Following development of the first draft of the survey, the survey was distributed for content validity and general feedback to three current practitioners of ABA and a disability justice advocate. In the current study, we took a proactive approach to ensure a comprehensive and inclusive perspective by involving a disability justice researcher in the development process. Recognizing the importance of incorporating diverse viewpoints and expertise, we collaborated closely with the disability justice researcher to provide insights into the ethical considerations and implications of our study's focus on behavior analysis and restrictive procedures.
To recruit content reviewers, the first author posted in a Facebook group (Applied Behavior Analysis) that content reviewers for a survey looking at behavior analysts’ usage of RPBPs in practice was needed. Anyone who responded to the post and provided information about their qualifications (i.e., current certificant) was sent the survey. The disability justice advocate is a colleague of the first author and was recruited specifically for their experiential knowledge. The directions for providing feedback were open-ended, directing reviewers simply to review the survey and provide any content or wording changes they felt appropriate. All feedback from these experts was reviewed and revisions were made to incorporate this feedback in the final survey items. Following these experts’ review, the primary revisions involved revising items to obtain continuous quantitative data, including examples in some items, and some language/vocabulary changes.
Final Survey
The final survey collected ratio data for current usage of procedures and nominal data for percentage of historical use of procedures. Given the noncomparability of the historical and current usage data, only current usage results for participants reporting their current caseload are reported here. The survey took an average of 10 to 15 min to complete. The full survey response dataset can be found at https://osf.io//(redacted for review). McDonald's ω, a measure of internal consistency of responding when data within a variable may not be normally distributed, indicated adequate internal consistency of the survey items reported in the current study (ω = 0.799).
Survey Distribution
Prior to recruitment, ethical approval was sought from the Institutional Review Board (IRB) of the host university of the first author. The process involved a thorough review of the study's methodology, objectives, and potential impact on participants by the committee, ensuring compliance with ethical principles and regulations governing research involving human subjects. Following IRB approval, survey participants were recruited through social media sites (Facebook, Twitter, and Instagram), including groups for BCBAs. A second recruitment strategy was through email listserv. The listserv was created through the first author's BCBA continuing education (CE) business. The social media post and email contained the study title, purpose of the study, the parts of the study the survey would include, inclusion criteria, and a link to the Qualtrics survey platform (Qualtrics, Provo Utah, 2020). The survey was open for 4 months following distribution. There was no compensation for participation.
The CE business utilized for recruitment focused on providing content in various subject areas, including compassionate care, trauma-informed practice, school-based ABA, and functional behavior assessment. While the first author contributed some of the CE content, a substantial portion was delivered by guest speakers, including the second and third authors. The recruitment process involved sending emails through an opt-in email listserv, which was initially populated by individuals who had purchased courses from the business’ website. Over a span of 4 months, four email invitations were sent to this listserv. In addition to email outreach, graduate students were also involved in recruitment efforts. During the 4-month recruitment period, the authors disseminated multiple posts across five distinct social media groups: “Applied Behavior Analysis,” “Learn Listen Lead,” “BCBA Clinical Questions,” “BCBA CEU and Supervision Support,” and “BCBA Share.” This multiplatform approach was employed to enhance the reach and accessibility of the survey to a diverse audience within the behavior analysis community.
Participants
Respondents included 534 ABA practitioners. Analyses included only those participants for whom current caseload data was either reported or could be extrapolated from the response (N = 481). Not all respondents answered all questions, so reported percentages may not reflect the total sample. See Table 2 for n and missing data for each demographic question, as well as additional sample demographic information.
Personal and Professional Demographic Characteristics of Sample.
Note: Client/learner population: ABA= Applied Behavior Analysis; AUT = autism; BCBA= Board Certified Behavior Analyst; DD = developmental delay; ID = intellectual disability; RBT= Registered Behavior Technician.
Of the respondents who identified their gender, 419 (87.8%) identified as female, 45 (9.4%) identified as male, and 12 (2.5%) identified as nonbinary. Participants selected age ranges, and most participants were 30 years or older (77.3%), with 18.1% 26 to 30 years old, 4.4% 22 to 26 years old, and 0.2% 18 to 22 years old. Most respondents were from the United States (87.5%) and White (84.6%). Most participants had a master's degree (92%), and about half completed their ABA coursework online, about one-third in person/on campus, and the remainder completed hybrid coursework.
Participants were asked with which populations they worked and were instructed to select all that applied. Participants reported working with only autistic individuals (n = 120, 24.9%), only individuals with intellectual disabilities (n = 4, < 1%), only individuals with developmental disabilities (n = 5, 1%), individuals with autism and intellectual disabilities (n = 20, 4.2%), individuals with autism and developmental disabilities (n = 250, 52%), individuals with intellectual and developmental disabilities (n = 2, < 1%), and “other.” Respondents also indicated in which settings they worked. Most respondents indicated a combination of school, clinical, medical, home-based, and/or residential facility work settings. Schools were the most common work settings (n = 92, 19.1%), followed by ABA clinics (n = 87, 18.1%), then home-based settings (n = 82, 17%).
Participants were asked about their personal experiences with autism and other disabilities. About 8% of the sample reported having autism, with about half of those respondents reporting a formal diagnosis and half reporting no formal diagnosis (e.g., self-diagnosis). An additional ∼20% reported having a diagnosis other than autism. Almost 9% of the sample reported having at least one child with autism, with an additional 5% suspecting one of their children had autism. Of those reporting a child with autism, 57.14% report their child(ren) receive ABA services. Nearly 30% of the sample reported having one or more other family members with autism. Of these, 33 reported multiple family members with autism. The highest frequencies for those with one family member with autism were observed for cousin (n = 35), niece/nephew (n = 20), sibling (n = 14), and spouse (n = 7). Frequency for other family member categories can be found in Table 2.
Participants were asked about their familiarity with the neurodiversity movement and belief alignment with the medical model of disability for autism. A large portion of the sample reported being familiar with the neurodiversity movement (n = 321, 66.7%), while 21% reported having heard of the neurodiversity movement “only in passing” (n = 101). When asked to rate their agreement for the statement (labeled as medical model in Table 1), “Autism is a disease that needs to be cured,” 87% of the sample disagreed or strongly disagreed, while only ∼5% agreed or strongly agreed. When asked to rate their agreement of the statement (labeled as neurodiversity paradigm in Table 1), “Autism is a component of human diversity that does not need to be cured,” 71% of the sample agreed or strongly agreed, while ∼12% disagreed or strongly disagreed. The term “cure” was used in both items as it is consistent with discussion of ableism and the medical model of disability in the literature and social discourse (Shyman, 2016), although we recognize it represents extreme ends of both models of disability.
Data Analyses
Prior to data analyses, the data file was cleaned to enable statistical analyses. Specifically, for the open-response questions on current usage of technologies, it was ensured a single numerical response was entered. Several participants responded with written numbers, percentages, ranges, or with a N/A response. For cells where one of the previous responses was evident, the response was modified to reflect a numerical value (e.g., participant response of “one” was modified to “1”). Percentages were transformed into numerical values based on the respondent's entry for “current caseload” and rounded down to the nearest integer, ranges were replaced with the lowest number in the range to avoid overestimation, and N/A responses were treated as missing data. Written responses where a numerical value could not be inferred were treated as missing data (e.g., “some”).
For the current caseload cells, three values had to be transformed. Each respondent had entered a range rather than a single numerical value. These rages were 7 to 10, 10 to 20, and 5 to 7, and the minimum value for each was used in data analyses. For the remaining cells on procedure use, 55 respondents entered data that needed to be transformed, for a total of 194 cells (3.41% of cells were transformed in the data file, including current caseload cells). The most common transformation was converting “all,” “every client,” and “100%” or similar responses to the value of that respondent's reported current caseload (24 participants). The next most common was converting “none,” “no client,” and “0%” or similar response to a value of 0 (19 respondents). Twelve participants reported percentages of clients, which were converted to values based on the participant's reported caseload. The remaining transformations included removing words or symbols from an otherwise numerical response (4 participants), adding values together when participants reported multiple values with a narrative that indicated an additive approach (2 participants), and using the minimum value from a reported range (2 participants).
All continuous data were summarized with descriptive statistics (mean, median, standard deviation, and missing data). Percentage of use is also reported, which was calculated by dividing the average reported use for each RPBP by the average total caseload. Categorical data were summarized with the number and percentage of respondents selecting each response option. In addition to summarizing the data collected from participants, three nonparametric analysis of variances (ANOVAs), specifically Kruskal–Wallis ANOVAs, were conducted to examine differences between respondent demographic characteristics and use of procedures. Reported values include omnibus Kruskal-Wallis (X2) and pairwise post hoc comparisons (W). The Bonferroni correction was applied, accounting for three comparisons, resulting in an alpha of .016. Effect sizes (epsilon-squared) were reported for the omnibus Kruskal–Wallis findings in the tables. Nonparametric statistics were used as data violated the assumption of normality via positive skew (Nwobi & Akanno, 2021). Analyses were conducted using jamovi, an open-access statistical software (The Jamovi Project, 2024).
Results
Use of Procedures
Respondents reported an average current caseload of 16.74 clients (Mdn = 10; range: 1–200). Three RPBPs were reported at noticeably higher rates than the others: attention extinction, response blocking, and escape extinction. Attention extinction was the most frequently reported RPBP, used with an average of 35.89% of clients on respondents’ caseloads (M = 6.11, Mdn = 2, range: 0–100). Response blocking was the next most frequently reported RPBP, which was used with an average of 19.70% of the clients on respondents’ caseloads (M = 3.35, Mdn = 1, range: 0–100). Escape extinction was the third highest reported RPBP, used with an average of 17.63% of clients on the respondents’ caseloads (M = 2.98, Mdn = 0, range: 0–100).
The remaining RPBPs were reported with fewer than 6% of clients on respondents’ caseloads. Of these, physical restraint was the most reported, used with an average of 5.89% of clients on respondents’ caseloads (M = 1.01, Mdn = 0, range: 0–61). This was followed by overcorrection (5.71%, M = 0.97, Mdn = 0, range: 0–75), exclusionary time-out (4.12%, M = 0.70, Mdn = 0, range: 0–61), contingent exercise (3.66%, M = 0.63, Mdn = 0, range: 0–80), and sensory extinction (1.69%, M = 0.29, Mdn = 0, range: 0–20).
Alternative procedures were reported at higher rates. Respondents reported using antecedent strategies with 51.98% of the clients on their caseloads (M = 9.26, Mdn = 4, range: 0–183). Respondents reported using consultation and referral with 52.99% of clients on their caseloads (M = 9.21, Mdn = 5, range: 0–100). Descriptive summaries, as well as number of missing data points, for aggregated usage of all technologies can be found in Table 3.
Reported Usage of Aversive Technology With Current Clients.
Note: Mean and mode represent the number of current clients with whom respondents reported using each procedure.
Potentially Related Variables
There are several demographic variables and personal experiences with disability and neurodiversity that may be related to use, or lack thereof, of behavioral technologies. Significant Kruskal–Wallis ANOVA comparisons are reported below, and all comparisons by variable can be found, reported as ε2 effect sizes, in Table 4. Several participant characteristics demonstrated significant class imbalances, where some categories had fewer than five participants. As such, several potentially predictive variables were not used to evaluate differences in use of procedures. Those with less variability in groups were utilized as predictors in the analyses reported below.
Effect Size Comparisons of Procedure Use Across Respondent Characteristics.
Note: All values are ε2 effect sizes; * = p < .016; ** = p < .001; df = 4 for all respondent characteristics. RBT= Registered Behavior Technician.
Respondent Characteristics
Small differences were observed between how many years a respondent had worked as an RBT prior to becoming a BCBA and use of physical restraint (X2(4) = 13.68, p = .008, ε2 = .029), where those who had never worked as an RBT reported higher levels of physical restraint than those with 3 to 5 years of RBT experience (W = −4.37, p = .017).
Personal Experiences With Disability
Some participants reported personal diagnoses of non-ASD (non Autism Spectrum Disorders) disabilities, where small differences were observed in use of escape extinction (X2(3) = 9.74, p = .021, ε2 = .021) and attention extinction (X2(3) = 8.68, p = .034, ε2 = .019) procedures when using an alpha of .05. Following Bonferroni correction, these comparisons fall above the .016 threshold, indicating no significance.
Neurodiversity and Other Perspectives
Statistically significant differences were observed between neurodiversity familiarity and several procedures. For the use of escape extinction, a moderate effect was observed across categories of familiarity (X2(3) = 41.66, p < .001, ε2 = .088), where those who reported they were familiar with the neurodiversity movement reported less use of escape extinction than those who were not familiar (W = 8.02, p < .001) and those who had only heard of it in passing (W = 4.98, p = .002). Those who reported not being familiar reported higher levels of escape extinction than those who had only heard of the movement in passing (W = −4.02, p = .023).
For the use of attention extinction, a moderate effect was observed across categories of neurodiversity familiarity (X2(3) = 29.63, p < .001, ε2 = .063), where those who were familiar with the movement reported lower use of attention extinction than those who had not heard of it (W = 6.30, p < .001) and those who had only heard of it in passing (W = 4.71, p = .005). For the use of sensory extinction, a small effect was observed across categories of neurodiversity familiarity (X2(3) = 20.84, p < .001, ε2 = .044), where those who were familiar with the movement reported lower use of sensory extinction than those who had not heard of it (W = 5.93, p < .001).
For the use of overcorrection, a small effect was observed across categories of neurodiversity familiarity (X2(3) = 27.51, p < .001, ε2 = .058), where those who were familiar with the movement reported lower use of overcorrection than those who had not heard of it (W = 6.73, p < .001) and those who had only heard of it in passing (W = 4.31, p = .012). For the use of contingent exercise, a small effect was observed across categories of neurodiversity familiarity (X2(3) = 15.82, p = .001, ε2 = .034), where those who were familiar with the movement reported lower use of contingent exercise than those who had not heard of it (W = 4.63, p = .006) and those who had only heard of it in passing (W = 4.61, p = .006). For the use of exclusionary time-out, a small effect was observed across categories of neurodiversity familiarity (X2(3) = 15.09, p = .002, ε2 = .032), where those who were familiar with the movement reported lower use of exclusionary time-out than those who had not heard of it (W = 5.37, p < .001), and those who had only heard of it in passing reported lower use than those who had not heard of it (W = −3.69, p = .045). For the use of response blocking, a moderate effect was observed across categories of neurodiversity familiarity (X2(3) = 28.64, p < .001, ε2 = .061), where those who were familiar with the movement reported lower use of response blocking than those who had not heard of it (W = 5.59, p < .001) and those who had only heard of it in passing (W = 4.97, p = .003). Effect sizes for all comparisons can be found in Table 4.
Discussion
The current study measured the use of select behavioral technologies currently used by behavior analysts in practice. Several of these procedures can be classified as RPBPs, indicating a need for special care and ethical considerations, such as the use of antecedent and positive reinforcement-based interventions prior to their use. This survey was intended as a preliminary look at the current rate of usage of behavior analytic procedures. In other words, these findings should serve as a starting point for further research and discussion, and recommendations are theoretical in nature rather than fully supported by these findings.
Respondents were 534 BCBAs (481 of whom were included in analyses) who completed an online survey about their use of behavioral technologies with their current caseload. Results indicate that several RPBPs were reported as being infrequently used by practitioners, with practitioners reporting using each RPBP with fewer than an average of 40% of clients on their caseloads and many procedures with fewer than 10% of clients. Further, respondents reported using antecedent interventions and consultation and referral each with about half of their caseloads. Several participant characteristics were related to differences in the use of these procedures. In general, BCBAs who had never worked as an RBT were more likely to report using physical restraint procedures. This could indicate the benefits of experience as an RBT prior to becoming a BCBA, particularly in minimizing the use of physical restraint procedures. Participants who were familiar with the neurodiversity movement were the least likely to report using many of the RPBPs.
Given the current climate of the discussions around ABA therapy (Veneziano & Shea, 2023), these results provide an important estimate of procedural use that can be used as a starting point for understanding this issue. Confirming these results through multiple sources of data, as well as recognizing the limitations of these data is important to fully understanding the issue. As the debate on the use of RPBPs continues in the field, it is important to make decisions based on data as well as input from the client receiving these treatments, including the rate of use of these procedures and the social validity of behavior analytic practices to clients and consumers. This study contributes to our understanding of the frequency of the use of several behavior analytic practices; however, it is not without its limitations.
Limitations
First, this sample was a convenience sample, recruited online via multiple social media professional pages. While over 500 participants responded to the survey, we are not able to report the response rate given the way surveys were distributed. Additionally, the sample is heavily skewed toward individuals who are White, female, and aged 30 years or older. While 83.8% participants in our sample self-identified as White, demographic data from the BACB indicate only 69.16% of BCBAs identify as White (BACB, 2020). Demographic data from the BACB also indicate that 86.73% of BCBAs identify as female, which is similar to our sample (87.5%). Although not reported in ranges that match the current survey procedures, the BACB reports 99.12% of BCBAs are over the age of 24, while 95.43% of participants in the current sample reported being 26 years of age or older. Together, these data indicate the current sample has a larger proportion of White respondents than would be representative of BACB certificant data; however, it appears to be representative in terms of gender and age. Additionally, the topic of our study may have disproportionately attracted individuals who were more familiar with and open to the principles of neurodiversity and autistic acceptance based on the study title and some recruitment strategies (e.g., utilizing Facebook groups with this focus). Given these limitations, additional work may be necessary to validate these results with a sample more globally representative of behavior analysts.
Another limitation relates to the influence of various contextual factors that may affect the appropriateness of RPBPs. These factors, such as the specific behavioral challenges presented by clients, input from the clients, environmental conditions, and available resources, can significantly impact the use and effectiveness of RPBPs. The survey did not collect detailed information on these contextual factors, making it challenging to provide a nuanced analysis of why behavior analysts may choose to utilize RPBPs. While the survey asked about function-based interventions (e.g., extinction procedures), it did not gather information on participants’ clients’ behavioral functions or the severity of behaviors. Therefore, results cannot be interpreted in terms of the suitability of procedures respondents reported using with clients’ and may have skewed reports of using some procedures. For example, a client engaging in behavior with an attention function would not be prescribed escape extinction as an intervention, nor should a student exhibiting a minor, nondangerous behavior be physically restrained. Gathering data on function, behavior severity and the type of intervention would have strengthened the survey and our interpretation of results.
Similarly, information was not collected on whether procedures were implemented alone or as part of multicomponent intervention plans, limiting our ability to state whether the clients for whom participants reported implementing antecedent interventions and consultation and referral were the same as those with whom they reported using RPBPs. The survey also asked about a closed list of strategies. While these strategies were based on current literature, it is possible several RPBPs were left out of the survey. An alternative strategy could be to allow respondents to report add-in procedures on which RPBPs they use with clients.
A final limitation is that this research relies on self-report data of clinical practice. The problems with self-report survey data are well documented (Fisher, 1993). Data may be skewed toward respondents presenting themselves in a positive way. Seeing questions about neurodiversity on the survey also may have skewed the results. Our research questions may be more accurately answered through an archival data review of behavior intervention plans (BIPs) at organizations providing behavior analytic therapy. Archival data collection (when confidential information is involved) is labor intensive research and requires careful consideration for the maintenance of client privacy. The authors did not have the resources at this time to engage in this type of data collection, though we believe this would greatly supplement our findings.
Implications and Future Research
While several notable findings were observed in this study, replication studies, either survey or archival, would improve upon and validate our findings. Additionally, the findings of this survey, combined with the neurodiversity movement, have implications for the future directions of the field of behavior analytic therapy, particularly as it relates to social validity. Future research should examine additional predictors of RPBPs use, as well as client and family perspectives on the use of these practices.
While relatively low rates of use of RPBPs were reported, some procedures were more frequently reported than others (e.g., attention extinction, response blocking, and escape extinction). Given their higher frequency of use, care settings may wish to examine, and potentially modify, policies, and procedures, including how they determine the restrictiveness of procedures and select a hierarchy of procedures with less-restrictive procedures implemented first. Treatment for any client needs to be individualized, and there will likely be situations where many of the RPBPs surveyed and discussed here are appropriate and represent the least-restrictive and most socially valid procedure for a given client.
The intriguing potential relationship between neurodiversity awareness and acceptance and the selection of interventions warrants further investigation. While our study has offered valuable insights by potentially indicating a correlation between these variables, a future experimental study holds promise in uncovering a functional relationship. Such research endeavors could employ methodologies to explore the causal links between neurodiversity awareness, acceptance, and intervention selection. By manipulating variables in a controlled setting, we can gain a deeper understanding of how these factors interact and influence behavior analysts’ choices in real-world practice. This approach not only expands our theoretical understanding but also holds practical implications for enhancing the effectiveness and inclusivity of interventions for individuals across the neurodivergent spectrum.
Some training and continuing education recommendations can be extrapolated from these data. Respondents in the current study who were familiar with the neurodiversity movement reported lower use of most RPBPs when compared to those unfamiliar or familiar only in passing, indicating it may be beneficial for training and continuing education programs to include information on the neurodiversity movement and disability rights. It may be beneficial to include activities that allow trainees to interact and become familiar with Autistics and the neurodivergent community, as descriptive data indicated less reported use of RPBPs with closer experiences to people with autism (e.g., self, child, and family member).
Another training and continuing education recommendation is to emphasize the importance of utilizing antecedent and reinforcement-based procedures, social validity measurement and consideration, and the use of consultation and referral when underlying conditions may be present, consistent with ethical standards of the field. Despite ethical guidance that less-restrictive procedures should be used prior to RPBPs, our data align with previous findings (Lydon et al., 2015; Pokorski & Barton, 2021), in that participants reported using antecedent strategies for a period of time prior to more-restrictive procedures with only about half of the clients on their caseloads. Finally, practitioners may want to self-assess their process for selecting procedures, evaluate whether this process is aligned with ethical standards and indicates the use of least restrictive and socially valid procedures, and seek continuing education where necessary.
Conclusion
Results of the current study indicated RPBPs are currently used at relatively low rates when compared to the use of prior antecedent interventions and consultation and referral with underlying conditions by behavior analysts. Some variables predicted use of RPBPs: years of experience as an RBT and familiarity with the neurodiversity movement. Several additional variables evidenced lower use of RPBPs and more use of antecedent intervention and consultation and referral when examining the descriptive results, such as having autism, having a child with autism, and having a family member with autism; however, these observations are based only on descriptive summaries, given class imbalances across categories within these variables. Familiarity with the neurodiversity movement was a negative predictor of use of several RPBPs, with several small-to-moderate differences in procedure use between those who were familiar with the neurodiversity movement, those who had only heard of it in passing, and those who were not familiar with the movement.
Proponents of the neurodiversity movement and some within the field (Mathur et al., 2024; Veneziano & Shea, 2023) have presented criticisms of the practice of behavior analysis. Rather than doubling down on keeping the science and implementation of behavior change as is because it is effective, behavior analysts should hear the concerns of clients, clients, and their families, and make positive changes to implement sound behavior change practices in a way that promotes respect for, and autonomy of, clients. While behavior analysts are the experts on behavioral assessment and intervention, they should not discount the perspectives of those they serve.
Footnotes
Declaration of Conflicting Interests
The first author acknowledges a potential conflict of interest related to the implications of this study. Specifically, the first author stands to benefit financially from actions that may be taken in response to this article. Two-thirds of the implications discussed in the article pertain to additional training and continuing education in this field, which aligns with the services offered through the CEU platform (
). The first author receives financial compensation from owning and developing content for the CEU platform, which was utilized for participant recruitment in this study. The platform provides Continuing Education Unit (CEU) opportunities for behavior analysts, particularly focusing on neurodiversity.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Ethical Approval
Ethical approval was granted by the host university. All participants included in this report provided informed consent.