Response Allocation of Board-Certified Behavior Analysts toward Categories of Evidence-Based Practice

Abstract

Board certified behavior analysts (BCBAs) are often called upon to recommend treatments while working with autistic individuals. As practitioners of the science of human behavior, behavior analysts must make recommendations supported by scientific evidence. However, at times, individual practitioners may inadvertently recommend interventions that are not evidence-based. This study sought to examine if the severity level of the present symptoms of autism impacted the recommendations made by behavior analysts. A survey of 122 BCBAs gathered information about how they allocated resources toward interventions across three categories: evidence-based, emergent, and nonevidence-based. The results indicate that up to 62% of BCBAs allocated resources toward nonevidence-based or emergent practices and that these resource allocations were affected by the autism severity of hypothetical client presentations. There were statistically significant differences between allocations to resources between individuals with the lowest symptom severity and those maximally affected for both evidence-based practice (p < 0.0009) and nonevidence practice (p < 0.0011).

Autism spectrum disorder (ASD) is a complex neurological and developmental condition affecting approximately 1 out of every 36 children in the United States Maenner et al., 2023). Individuals diagnosed with ASD1 often present with persistent challenges related to social communication, restricted interests, and repetitive behavior patterns (Centers for Disease Control & Prevention [CDC], 2024). Although ASD is considered a lifelong disorder, the provision of evidence-based practices (EBP) has been shown to positively affect consumers (Kasari & Smith, 2016; Odom et al., 2010; Weiss et al., 2008). For example, interventions based on applied behavior analysis (ABA) are effective in treating symptoms related to ASD (Dillenburger & Keenan, 2009; Peters-Scheffer et al., 2011).

Determining which treatments qualify as EBP can be challenging due to the variety of methods professionals and organizations use when assessing and conceptualizing the available evidence (Campbell et al., 2021; Reichow et al., 2011). Selecting EBP can be particularly challenging for those serving individuals with ASD due to the prevalence of ineffective, costly, and, at times, hazardous interventions marketed to this population. The proliferation of such interventions has added complexity to the decision-making process for individuals seeking appropriate and beneficial treatments for ASD (Foxx, 2008). Attempts have been made by various professional groups (e.g., Association for Science in Autism Treatment, n.d.) and state organizations (e.g., New York State Department of Health Bureau of Early Intervention, 2017) to disseminate information via guidelines related to best practices for ASD. For example, the National Standards Project (NSP) published a report in 2015 on the current state of treatment practices for individuals diagnosed with ASD between the ages of 0 and 22. Of the 45 interventions assessed, 14 were considered established, 18 were considered emerging, and 13 were considered unestablished. Although these attempts have had some success in guiding treatment selection, the efficacy of dissemination of this information alone does not appear to be adequate (Stephenson et al., 2012). Therefore, practitioners working with individuals with ASD must be able to assess the evidence available for a variety of interventions before using or recommending their use to consumers.

Although complementary and alternative interventions are prevalent in treating individuals with ASD, board certified behavior analysts (BCBAs) are ethically required per the Behavior Analyst Certification Board (BACB) to select and utilize only EPBs. For example, the Ethics Code for Behavior Analysts (BACB, 2020) states that BCBAs must “provide services that are conceptually consistent with behavioral principles, based on scientific evidence, and designed to maximize desired outcomes'' (BACB, 2020, 2.01). In addition to remaining current in their knowledge of best practices in the field of ABA, the BACB Ethics Code requires behavior analysts to “remain knowledgeable and current about interventions (including pseudoscience) that may exist in their practice areas and pose a risk of harm to clients” (BACB, 2020, p. 4). However, some BCBAs continue to recommend the use of nonevidence-based practice, citing (1) recommendations by others; (2) online marketing; (3) a priori beliefs in treatment effectiveness; (4) popularity of the treatment; or (5) monetary gain as common reasons for recommendations of nonevidence-based practices (Marshall et al., 2023; Schreck et al., 2016; Schreck & Mazur, 2008).

It has also been found that BCBAs, similar to parents of children diagnosed with ASD, tend to overestimate the evidence available for non-ABA-based interventions (Campbell et al., 2021; Miller et al., 2012). Surveys of parents of children with ASD indicate that on average seven to nine interventions are employed at a time, with intervention selection being most sensitive to the age of the child at the time of intervention, the type of diagnosis, and the severity of presentation (Goin-Kochel et al., 2007; Green et al., 2006). Speech-language therapy is often reported to be the most common intervention, followed by visual schedules, sensory integration therapy, and finally, ABA. In addition, the use of complementary and alternative medicine is commonly used with individuals with ASD. In one study, 27% of parents reported that their child was following a specialized diet (e.g., gluten-free, casein-free, yeast-free), and 43% reported that their child was taking vitamin supplements (Green et al., 2006).

Although prior research has shown a variety of reasons why behavior analysts may fail to exclusively endorse evidence-based practices, no study to date has examined the role of client needs as it relates to the endorsement of non-EBP intervention. The level of support required by a client may be a factor potentially contributing to endorsement drift, as varying levels of severity may impact the behavior analyst's decision-making process. For example, when a behavior analyst is familiar with a client's unique strengths and needs, likely due to repeated interactions with individuals with similar profiles, the potential effectiveness of interventions becomes more evident. On the other hand, a drift away from evidence-based interventions may be more likely when the needed support level appears novel or extreme. A first step towards investigating the function of such behavior could be to explore parametrically if such a drift would remain constant or change across levels of reported disability. The current study sought to extend research on BCBA treatment selection related to therapy services for individuals with ASD by examining the relationship between the severity of client presentation and BCBA treatment allocation.

Methods

Participants and Setting

The authors distributed a survey online to BCBAs from across the United States through social media platforms such as LinkedIn and Facebook and via email to Association of Professional Behavior Analysts (APBA) members. From this sample, 356 respondents initiated the survey, and 122 respondents completed the survey for a completion rate of 34%. The attrition rate may be related to the deliberate vagueness of the descriptions provided for the severity scale, as indicated by some written comments submitted to the primary investigator or the length of time required to complete the survey.

Of the participants who completed the survey, 12% had been certified as a BCBA for 2 years or fewer, 51% for 3 to 10 years, and 37% for more than 10 years. Participant age ranges were varied, and the primary population group with whom participants worked was children aged 4–12 (76%). Most participants reported working within their current population for over 5 years (85%). For complete demographic information, see Table 1.

Table 1.

Participant Demographic Information

	Sample (Total N = 122)
Participant Characteristic	N	%
Years Certified
< 2	15	12
3–10	62	51
>10	45	37
Age Range
24–29	10	8
30–35	26	21
36–41	34	28
42–47	20	16
48–53	10	8
54-59	13	11
60	9	7
Primary Population Working With
Infants and Toddlers (age 0–3)	6	5
Children (age 4–12)	93	76
Adolescents (age 13–21)	14	11
Adults (age > 21)	9	7
Years Working with Primary Population
0–5 years	18	15
between 5–10 years	26	21
between 10–15 years	32	26
> 15 years	46	38

Survey Construction

The survey was distributed via the online survey platform Qualtrics. Consent was obtained at the beginning of the survey, where participants were provided with a description of the project's purpose. Participants who consented to the survey were presented with the initial screening question regarding their certification as a BCBA. Participants who selected no, indicating that they were not board certified, were excluded from the survey. Participants who selected yes were taken to the next screen, where they provided basic demographic information about their time certificated, age range, primary population with whom they worked, and time working with this population. The four demographic questions were included to identify potential variables that may have affected responses. A response was required for each demographic question to continue with the survey. No identifying data were gathered, thus protecting the anonymity of all respondents. Each question was presented in a multiple-choice format. The next screen provided brief instructions about the survey.

The instruction page included a summary of the survey, which detailed that each page of the survey would present the diagnostic criteria of autism from the Diagnostic and Statistical Manual of Mental Disorders (5^th ed, text rev.; DSM-V, American Psychiatric Association, 2022) and the severity of autism paired with evidence-based, emerging, or non-evidence-based practices. Participants were provided a scale of severity for autism ranging from 1 to 7, in which a severity level of 1 was described as “likely the person is indistinguishable from their peers.” In contrast, a severity level of 7 was described as “symptoms are pervasive and interfere with the individual's daily living.” The severity level was presented in bold blue font to draw the participants’ attention to this variable. In addition, the severity level corresponded with hover text, which provided the deliberately vague definition of severity as indicated above.

Each intervention also corresponded with hover text that provided a short, neutral definition of the therapy (see Table 2). Categorical assignments of evidence-based, emerging, or nonevidence-based practices were determined based on the most recent iteration of the NSP (National Autism Center, 2015). Practices that were not listed in the NSP were deemed to be nonevidence-based for this study, although there may be other bodies of research that support these practices. One example is acceptance and commitment therapy (ACT; Dixon & Hayes, 2022; Dixon et al., 2020),2 which is not listed in the NSP for treating individuals with autism and, as such, was placed in the nonevidence-based practice category. Evidence-based practices selected from the NSP included medical interventions, speech therapy, ABA, cognitive behavioral therapy (CBT), parent-mediated interventions, and social stories.3 Music therapy was included and categorized under the NSP as an emergent therapy for children; however, it should be noted that it is listed as unestablished for adults, and survey respondents were not provided the age of the fictitious individual they were considering. The inclusion of music therapy was a small sample of the category of emergent intervention listed on the NSP, which also included a cognitive behavioral intervention package, modeling, and sensory integration package, all similar to interventions already captured in the survey. Although this is categorized differently from unestablished, it should be noted that emergent interventions are also not established and still lack a sufficient evidence base to be considered evidence-based practice. Nonevidence-based (or not included in the NSP categorical designations) were ACT, occupational therapy, physical therapy, nutritional therapy, facilitated communication, dolphin therapy, art therapy, pet therapy, sensory integration therapy, chelation, craniosacral therapy, auditory integration training, or "no therapy recommended." To increase the likelihood that participants would complete the entire survey, not all the interventions listed in the NSP were included. The authors chose to include 20 of the listed interventions, selecting those believed to be the most familiar to most participants, based upon the authors' own experiences.

Table 2.

Treatment Options and Definitions

Treatment	Definition	Research Category*
Medical treatment	Treatment advised/prescribed by a medical practitioner.	Established
Speech-language Therapy	Treatment prescribed by a licensed Speech Therapist.	Established
Applied Behavior Analysis (ABA) Therapy	Therapy based on the science of learning and behavior.	Established
Cognitive Behavior Therapy (CBT)	A type of therapy in which negative patterns of thought about the self and the world are challenged in order to alter unwanted behavior patterns.	Established
Parent Mediated Therapy	Therapeutic intervention in which parents are trained by professionals to implement specific therapies to their own child.	Established
Social Stories/Social Thinking	Individualized instruction based on stories or other curricular components designed to address social situations that a person with autism may encounter.	Established
Music Therapy	Therapy provided by a certified music therapist in which music is purposefully utilized to support development, health, and well-being.	Emerging
Acceptance and Commitment Therapy (ACT)	A psychological treatment that uses acceptance and mindfulness strategies along with commitment and behavior-change strategies to increase psychological flexibility.	Unestablished or not listed in the NSP2 as a treatment for autism
Occupational Therapy	Treatment prescribed by a licensed Occupational Therapist.	Unestablished or not listed in the NSP2 as a treatment for autism
Physical Therapy	Treatment prescribed by a licensed Physical Therapist.	Unestablished or not listed in the NSP2 as a treatment for autism
Nutritional Therapy	Treatment based on nutrition.	Unestablished or not listed in the NSP2 as a treatment for autism
Facilitated Communication	Treatment to support non-vocal learners in which a facilitator provides physical support to the individual to type on a keyboard or other device.	Unestablished or not listed in the NSP2 as a treatment for autism
Dolphin-Assisted Therapy	A sensory therapy in which the individual is provided an aquatic experience with dolphins.	Unestablished or not listed in the NSP2 as a treatment for autism
Art Therapy	A form of therapy which encourages self-expression through painting, drawing, or other forms of art.	Unestablished or not listed in the NSP2 as a treatment for autism
Pet- Facilitated Therapy	Treatment utilizing dogs or other animals to improve the physical and/or mental health of patients.	Unestablished or not listed in the NSP2 as a treatment for autism
Sensory Integration Therapy	Treatment prescribed in which individuals are exposed to sensory stimulation in a structured way to improve the organization of sensory information from the environment.	Unestablished or not listed in the NSP2 as a treatment for autism
Chelation	A method utilized to remove certain heavy metals from the bloodstream.	Unestablished or not listed in the NSP2 as a treatment for autism
Craniosacral Therapy	A system of alternative medicine intended to relieve pain and tension by gentle manipulations of the skull regarded as harmonizing with a natural rhythm in the central nervous system.	Unestablished or not listed in the NSP2 as a treatment for autism
Auditory Integration Training	A type of sound therapy that aims to reduce sensitivity to sounds or problems with how sounds are processed.	Unestablished or not listed in the NSP2 as a treatment for autism
No therapy recommended	n/a	n/a

^*Research categorical designations were gathered from National Autism Center, 2015

Participants were instructed that they had 100 tokens to allocate to various interventions, that the tokens should be considered “the overall weight of (their) recommendation based on time, commitment, and resources,” and that they were able to allocate tokens in any way, including leaving some interventions blank or allocating all of their tokens to a single intervention. Participants manually advanced to the next screen, selecting “next,” which displayed the first question in the remaining seven questions. In each section of the survey, participants were asked to allocate their tokens across 20 different interventions with a reminder that they could allocate their tokens in any way they saw fit. In addition to selecting from the interventions assigned, they were also allowed to select “no therapy recommended.” The interventions were listed below the instructions and summary, with a box containing a zero on the right of intervention listed. Participants entered the number of tokens they would allocate toward each intervention, with the option to leave it set at zero. A total accumulated at the bottom of the page, beneath the token allocation box. This total was in red text until the total reached 100. If the total number of tokens exceeded 100, this text returned to red. If participants did not allocate the minimum number of 100 tokens or exceeded 100 tokens and attempted to advance to the next screen, the existing screen turned yellow with a red textual warning at the top asking them to total the choices to 100. Participants were only able to successfully advance to the next screen once their token allocation was exactly 100 tokens. Once advancing to the next screen, participants could not go back and make changes to their initial token allocations. Each subsequent screen was identical to the first one, varying only in the severity level of autism. At the end of the survey, participants were thanked for their participation.

Data Analysis

Mean and median token allocation were calculated for each intervention at each severity level (see Table 3 for means). Token allocations were aggregated by EBP categorical designation (see Table 4). Paired sample t-tests were conducted to compare mean differences between participant token allocation to severity levels one and seven across the three different categories of evidence-based practice (see Table 5 and 6). This was done to compare the difference in how BCBAs may allocate recommendations for interventions when examining individuals who are minimal to maximally impacted.

Table 3.

Mean Token Allocation per Intervention by Severity Level

Intervention	Means of Token Allocation by Severity Level (range)
	1	2	3	4	5	6	7
Medical treatment	3.16 (0, 100)	2.50 (0, 75)	2.75 (0, 50)	3.28 (0, 50)	5.08 (0, 100)	4.11 (0, 50)	4.33 (0, 50)
Speech-language Therapy	9.79 (0, 50)	11.97 (0, 50)	11.60 (0, 50)	12.05 (0, 50)	10.93 (0, 50)	9.82 (0, 50)	9.06 (0, 50)
Applied Behavior Analysis (ABA)	51.30 (0, 100)	56.96 (0, 100)	60.61 (0, 100)	63.34 (0, 100)	65.30 (0, 100)	67.18 (0, 100)	68.18 (0, 100)
Cognitive Behavior Therapy (CBT)	2.66 (0, 33)	1.39 (0, 20)	0.82 (0, 25)	0.25 (0, 15)	0.16 (0, 10)	0.08 (0, 5)	0.08 (0, 5)
Parent Mediated Intervention	6.97 (0, 100)	7.85 (0, 75)	6.76 (0, 80)	7.25 (0, 80)	5.80 (0, 80)	6.76 (0, 100)	5.44 (0, 80)
Social Stories/Social Thinking	3.80 (0, 50)	2.79 (0, 60)	1.50 (0, 30)	0.44 (0, 25)	0.11 (0, 10)	0.22 (0, 20)	0.25 (0, 20)
Music Therapy	0.17 (0, 10)	0.12 (0, 10)	0.20 (0, 10)	0.01 (0, 1)	0.11 (0, 10)	0.03 (0, 4)	0.02 (0, 2)
Acceptance and Commitment Therapy (ACT)	7.35 (0, 75)	5.39 (0, 90)	2.75 (0, 35)	1.93 (0, 50)	1.23 (0, 25)	0.70 (0, 25)	0.66 (0, 25)
Occupational Therapy (OT)	3.69 (0, 25)	4.46 (0, 34)	5.73 (0, 25)	5.38 (0, 25)	4.97 (0, 35)	4.47 (0, 40)	4.60 (0, 60)
Physical Therapy (PT)	0.53 (0, 25)	0.67 (0, 25)	1.43 (0, 25)	1.16 (0, 25)	1.02 (0, 25)	1.27 (0, 25)	1.20 (0, 25)
Nutritional Therapy	0.57 (0, 30)	0.41 (0, 25)	0.82 (0, 40)	0.62 (0, 25)	0.83 (0, 30)	1.16 (0, 30)	0.70 (0, 30)
Facilitated Communication	-	0.16 (0, 20)	0.16 (0, 20)	0.53 (0, 20)	0.61 (0, 25)	0.86 (0, 30)	0.57 (0, 25)
Dolphin-Assisted Therapy	-	-	-	-	-	-	-
Art Therapy	0.17 (0, 10)	0.12 (0, 10)	0.12 (0, 10)	0.01 (0, 1)	0.11 (0, 10)	0.03 (0, 4)	0.84 (0, 100)
Pet-Facilitated Therapy	0.04 (0, 5)	0.12 (0, 15)	-	0.08 (0, 10)	-	-	0.02 (0, 2)
Sensory Integration Therapy	0.79 (0, 25)	0.49 (0, 25)	1.02 (0, 80)	1.00 (0, 50)	1.56 (0, 80)	0.88 (0, 60)	1.14 (0, 90)
Chelation	-	-	-	-	-	-	0.61 (0, 75)
Craniosacral Therapy	-	-	-	-	-	-	-
Auditory Integration Training	0.11 (0, 10)	-	-	-	-	-	-
No therapy recommended	8.89 (0, 100)	4.58 (0, 100)	3.73 (0, 100)	2.66 (0, 100)	2.17 (0, 100)	2.42 (0, 100)	2.30 (0, 100)

Table 4.

Tokens Allocated to ABA Services by Severity Level Out of 100 Available Tokens

Severity Level	# Participants Who Allocated to ABA (N =122)	Percentage of Sample	M (range)	SD	Median
1	109	89	51.3 (0, 100)	30.42	50
2	113	93	56.96 (0, 100)	28.45	55
3	118	97	60.61 (0, 100)	26.68	60
4	118	97	63.34 (0, 100)	27.76	62.5
5	118	97	65.3 (0, 100)	27.81	70
6	117	96	67.18 (0, 100)	28.64	70.5
7	117	96	68.18 (0, 100)	29.43	75

Table 5.

Paired Sample T-Test Data

Comparison of Differences in Token Allocation from Level 1 to Level 7 Autism Severity
	N	Mean ± SD	t-test
Evidence-Based Practice	122	11.87 ± 31.95	t = 3.391, df = 121 p = 0.0009 *
Emerging Practice	122	-0.05 ±.3873	t =1.647, df = 121 p = 0.1022
Non-Evidence-Based Practice	122	-9.520 ± 31.47	t = 3.342, df = 121 p= .0011 *

^*Results are significant for p < .05

Table 6.

Percent of Participants Who Allocated 100% of Tokens to Various Arrangements of Interventions (Total N=122)

	Severity Level
	1	2	3	4	5	6	7
ABA	13	14	15	20	22	26	27
Evidence-based practices	38	35	37	46	48	56	58
Evidence-based practices and ACT	57	58	50	54	53	59	60
Evidence-based practices and OT	48	61	62	70	69	75	75
Evidence-based practices and PT	38	41	45	46	48	56	57
Evidence-based practice, ACT, and OT	75	81	75	78	76	78	78
Evidence-based practice, ACT, and PT	57	58	50	54	53	59	60
Evidence-based practice, OT, and PT	52	65	69	75	75	81	81
Evidence-based practice, ACT, OT, and PT	78	87	84	85	83	83	83
No intervention	6	2	2	2	>1	2	2

Results

Participants predominantly allocated tokens to established intervention practices as defined by the NSP. However, throughout the survey, 43% to 62% of participants allocated some tokens to both nonestablished practices and emerging procedures at various severity levels (see Table 7). It should be noted that only one intervention was deemed emergent (music therapy). The mean allocation of tokens toward EBP steadily increased from approximately 13 to 15 as severity increased. The median token allocation for ABA increased as severity increased, from 50 to 75, with high variability in responses. There was a steady increase in the mean and median allocation of tokens to ABA interventions based on increasing autism severity (see Table 4). As the severity level of autism increased, the percentage of participants allocating tokens to nonevidence-based practices steadily declined by 20%, resulting in less than half of the participants allocating tokens toward nonevidence-based practice at a severity level of 7. In comparison, over 60% had allocated in this manner at severity level 1.

Table 7.

Percentage of Participants Who Allocated Tokens to Each Category Across Severity Levels (Total N = 122)

	Severity Level
	1	2	3	4	5	6	7
EBP	93 (114)	97 (118)	98 (119)	98 (120)	99 (121)	98 (120)	97 (118)
Emerging	3 (4)	28 (34	2 (3)	1 (1)	2 (2)	1 (1)	1 (1)
Non-EBP	62 (76)	59 (72)	63 (77)	54 (66)	52 (64)	44 (54)	43 (52)

Number of participants are presented in parentheses

The paired sample t-test showed a significant difference between token allocation for level 7 and level 1 severity levels for evidence-based practice (t(121) = 3.391; p =.0009). The mean differences for token allocation toward evidence-based practices (M ± SD = 9.676 ± 31.52) significantly increased as severity increased. Results for nonevidence-based practice were also significant (t (121) =3.342; p =.0011). The mean difference for token allocation toward non-evidence-based practices (M ± SD = -9.520 ± 31.47) demonstrated a significant decrease in token allocation as severity increased. Results for emerging practice were not statistically significant.

At the mildest iteration of symptom severity, 76 of 122 BCBAs failed to allocate all their tokens toward evidence-based practices. This number decreased to 52 out of 122 BCBAs when the severity of the autism disability increased. These data suggest that some of the BCBAs in the current study might recommend practices that are not evidence-based in conjunction with those that are and that the degree to which they do so is related to the severity of a client’s disability. Most participants continued to recommend ABA services regardless of the severity level; however, they were also likely to recommend combining these treatments with other evidence-based treatments despite evidence suggesting eclectic treatment may not be as effective (Eikeseth et al., 2007; Howard et al., 2014).

Some unestablished interventions that may have been perceived to be evidence-based (i.e., ACT, OT, and PT) for individuals with autism were grouped. The allocation of tokens toward these groupings was combined with the total allocation toward all EBP categories. These groupings were driven by speculation that behavior analysts might attribute evidence-based status to these interventions based on their familiarity with and commonality in treating autistic individuals. In addition, it is noteworthy that ACT is commonly referenced in ABA journals. Thus, in a field that predominantly serves individuals with autism, it may be perceived that this intervention is an EBP matched with the autism diagnosis. The results of these groupings are displayed in Table 5. The grouping arrangements account for some of the technological drift observed in this study, indicating that participants may have viewed these three interventions as being in the EBP category. Although substantial research articles support these three practices, they were not included with the EBPs noted by the NSP for autism. They were thus counted as unestablished, the collective opinion of the authors notwithstanding. This may indicate that some additional clarification is necessary from organizations such as the NSP or that practitioners may need to focus more specifically on recommendations that are considered evidence-based for the population with whom they work. Although other allocations occurred, trends were not well-established, with minimal tokens allocated to any given intervention at any given severity level. In addition, some BCBAs advocated for no intervention at all seven presentations of severity level.

Discussion

In practice, BCBAs are often in the position of making recommendations for treatment to parents of autistic children, inclusive of recommendations for ABA services. Evidence-based practice and rigorous adherence to the scientific method are the foundations of behavior analysis (BACB, 2020; Baer et al., 1968, 1987; Bailey & Burch, 2022; Cooper et al., 2019; Skinner, 1957). Thus, behavioral scientists must practice within such parameters when evaluating recommendation options. Behavior analysts’ ability to discern which treatments have or have not been established as effective relies on frequent contact with research, which is often influenced by information available within their communities of professionals. However, some behavior analysts may find themselves isolated from behavior-analytic mentors familiar with pertinent research (Drahota et al., 2020) and may be influenced by input from professionals in other fields. Likewise, behavior analysts may have a preexisting history of exposure and possibly social reinforcement for implementing non-ABA practices. Any drift from empirically validated research may lead to a pattern of recommendations for interventions that are not evidence-based, wasting precious intervention time.

Considering that the expertise of behavior analysts consists wholly of the practice of ABA, it is prudent to discuss variables that may influence the selection of other treatments to be combined with ABA. Identifying the contingencies for selections beyond the scope of practice of behavior analysts may yield helpful information for practitioners and researchers alike. Such information may help inform graduate course development to train future behavior analysts on navigating available treatment options in favor of evidence-based interventions. Further, those providing behavior-analytic supervision may consider embedding experiences related to the practical application of the selection of evidence-based intervention in ways that address the contingencies in which drift may occur. Although this is certain to benefit novice practitioners entering the field, ongoing continuing education and mentorship in this area are also essential as behavior analysts continue in their career development.

Another notable variable for non-adherence to evidence-based treatment selection may be that behavior analysts have a history of reinforcement surrounding collaborative relationships with other practitioners (e.g., Campbell et al., 2021; Marshall et al., 2023; Schreck et al., 2016). There has been some criticism of ABA practitioners and their ability to collaborate with professionals in other fields (e.g., Taylor et al., 2018). In response to this criticism, behavior analysts may make concerted efforts to engage in collaborative relationships to the degree that they may make recommendations to combine treatments or recommend treatments that are less adherent to scientific principles to demonstrate behaviors aligned with collaboration. Further investigation may be warranted on how behavior analysts view their role in collaborating while adhering to the science of behavior by exploring the extent to which behavior analysts compromise regarding recommendations. In addition, the role of ongoing mentorship and support in addressing recommendations of nonevidence-based practice may need further exploration. Additional research may include developing tools and resources that support behavior analysts in selecting evidence-based practices, particularly related to collaboration with other professionals. Further, it may be worth exploring how frequently behavior analysts recommend nonevidence-based practices in their daily practice rather than simply not speaking up regarding an eclectic approach, particularly to be considered a better collaborator.

Ultimately, many participants in the current study allocated some portion of their tokens toward nonevidence-based practices across all severity levels. This may suggest that behavior analysts responding to this survey may not be aware of which treatments have sufficient evidence to support their use with autistic clients. One possible reason could be that research may be misunderstood or misinterpreted regarding the evidence available to support its use with varying populations. Some of the drift away from evidence-based practices may have arisen from allocation decisions favoring ACT under the assumption that it held the same evidence-based status as ABA, given its apparent association with ABA. More allocations toward ACT were observed at lower levels of autism severity, hinting at the potential need for more language-based treatments once these clients reached a certain level of cognitive complexity. This finding may support reconsidering the inclusion of ACT approaches under the umbrella term of behavior analysis by the NSP.

Although this may explain some of the drift from evidence-based practice, another possible explanation is how behavior analysts seek and gain information about intervention practices. Some behavior analysts gain information through communication and discussion with others, such as seeking treatment advice via social media (O’Leary et al., 2017) rather than through contact with the literature. Exploring how frequently behavior analysts contact sources that delineate evidence-based practices for the populations they serve may be worthwhile. Further, it could be useful to consider extending this current research by adding questions about how frequently respondents contact behavior analytic literature and in what capacities (e.g., as a graduate course instructor, part of journal clubs offered through their employer).

It should be noted that a small percentage of the participants did not recommend ABA services at various severity levels, which is the greatest concern in the results of this survey. This is consistent with some previous research, which indicates that some BCBAs do not support ABA as an intervention (e.g., Campbell et al., 2021), which poses a threat to the field of behavior analysis. Future studies should consider why some behavior analysts do not consistently support ABA and the factors that may deter their support of the field. Future studies could explore the arrangement and amounts of interventions, including balancing the interventions for evidence-based and nonevidence-based practices.

Some limitations of the present study include the relatively small sample size, high attrition rate, and diversity of experience within the sample of BCBAs. A larger sample population should be considered for replications and extensions to yield more robust results. The high attrition rate contributed to this small sample size and should be addressed in future studies in multiple ways, including the survey length. In addition, future studies could consider how different presentations across symptoms of autism may affect response allocation. Demographic information indicates an overinflated representation of more experienced behavior analysts, likely due to the recruitment method. Although participants were recruited from the APBA membership email, participants were also sampled from social media. Because social media sampling was utilized, this may have affected the participant pool. Future studies should include a more diverse experience base more representative of the practicing population. Another weakness of this sampling method is that it is impossible to determine an accurate response rate, as it is unknown how many individuals were exposed to the survey. As such, the completion rate was calculated, but future studies may wish to use methods to ensure it is possible to determine the response rate. Additional demographic information may also be useful to determine other possible variables that influence the response allocation of participants, such as dual certification or experience with populations other than autistic individuals.

Some additional limitations include some components of the survey construction. As the survey provided a sequential presentation of increasing intensity with no ability to return to previous screens, this may have affected participant response. For example, some participants may have changed their responses as they moved forward but could not change their previous response allocation. In addition, repeated presentations may have influenced participant responses. It should also be noted that the survey results are based on self-report and may not indicate how behavior analysts make actual recommendations in practice. Another factor that should be considered is that the survey did not provide respondents with information about the age of the client served. This would impact the categorical assignment of music therapy as it is an emerging practice for children, but for adults, it is unestablished. It may have affected how tokens were allocated by practitioners who served different populations. Further research may explore how self-report and actual behavior correspond with one another as it applies to the practice of treatment recommendations.

Another limitation of the survey construction was the option to select “no intervention.” Although it was not common, 12 respondents did allocate toward “no intervention” while simultaneously selecting other interventions in various EBP categories. It could be interpreted that these respondents were allocating a portion of their resources toward “no additional interventions.” However, it is impossible to interpret from the survey as it exists. Future extensions of this work may consider either removing this option or designing the survey so that if this option is selected, all 100 tokens are allocated to “no intervention.”

Another critical consideration in interpreting and assessing these data is the impact of the provided definitions for the chosen treatments, as listed in Table 2. The broad scope of these definitions may have influenced respondents' allocation of their tokens. For example, the classification of medical treatment as anything prescribed by a medical doctor may pose a limitation in this study, as some doctors have been known to prescribe unsafe and even dangerous practices such as chelation. This could influence token allocation, and future studies may wish to provide narrower definitions that provide contextual information on the recommended interventions. This would offer clarity to respondents in the evidence base of these recommendations.

This initial study examined how behavior analysts allocate their responses when recommending services for autistic individuals. Although there were limitations to this study, the preliminary results show promising future directions, such as particular areas of interest for the field, which include the role of behavior analysts in making recommendations in collaboration with other professionals, the scope of competence for behavior analysts, and motivational influences on technical drift. It is imperative that future research further explore how behavior analysts make decisions related to recommending interventions and how this may result in technical drift, as it is of vast importance that behavior analysts continue to support applied behavior analysis and are cautious about recommending other treatments that may be outside of their scope or may be unsupported by science.

Data Availability

Data associated with this article are available by request from the corresponding author upon reasonable request

Declarations

Conflicts of Interest

The authors have no relevant financial or nonfinancial interests to disclose.

Ethical Approval

The study was approved by the Endicott College Institutional Review Board (Project ID:1885307-1] ). The procedures used in this study adhere to the tenets of the Declaration of Helsinki.

Consent to Participate

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