Key Takeaways
- Dr. Robert K. Ross and Dr. Fernando Armendariz argue that center-based ABA delivery is driven by structural convenience, not by evidence that clinics are where behavior change should occur.
- Caregivers respond to their child’s behavior hundreds of times a day, whether or not anyone calls it therapy. Ross and Armendariz argue that clinics often treat their few hours as if they are the whole picture, while the interactions that most reliably shape behavior are happening around the clock at home.
- The field’s generalization problem, they contend, is a product of teaching skills in the wrong environment, not a failure of behavioral science itself.
- Both clinicians raise concerns about BCBA credentialing at scale, arguing that rapid growth has produced practitioners trained in procedures rather than underlying principles. Dr. Dennis Dixon, a clinical officer at one of the largest national providers, says the resulting workforce inexperience often leaves new behavior analysts hesitant to engage parents as partners.
- Dixon’s research and field experience point to clinic settings producing faster skill acquisition, but he agrees that gains are meaningless without generalization to the home, and calls a 100 percent center-based model a red flag.
- All three clinicians converge on the same path forward: parent training as the core intervention at home, clinic and community settings used where they fit, and the field’s focus shifted from process disputes to measurable outcomes.
Dr. Fernando Armendariz, PhD, BCBA-D, opens many of his conference talks with a parable. A man has lost his wedding ring and is searching for it beneath a street lamp when a passerby stops to help. After a few minutes of fruitless searching, the passerby asks where, exactly, the ring was dropped. The man gestures into the darkness well beyond the lamp’s reach. “Why search here, then?” the passerby asks. “Because there is light here,” the man replies.
Armendariz uses the story to describe how the ABA therapy industry has organized itself around center-based service delivery: not because the evidence points there, but because the clinic is where conditions are controlled and hours are billable. “They are doing the clinic because it is easy there,” he said. “You get nowhere, but it is easier.”
Armendariz is the founder and director of FABAS, Inc., a consulting and habilitation agency in Tucson, Arizona. His long-time friend and colleague, Dr. Robert K. Ross, Ed.D., BCBA-D, CPBA-AP, LABA, brings an equally long record of direct practice. Ross spent more than three decades as Chief Clinical Officer at Beacon ABA Services of Massachusetts and Connecticut and remains Clinical Advisor there.
Together, Armendariz and Ross represent a combined eight decades of direct clinical work. Their collective argument is not that ABA does not work, but rather that the version being delivered at scale today bears only a procedural resemblance to what the science prescribes. They are currently co-authoring a chapter revisiting the foundational 1968 paper “Some Current Dimensions of Applied Behavior Analysis,” by Baer, Wolf, and Risley, arguing that the field has drifted from nearly all seven of its defining dimensions.
That argument has not gone uncontested. Dr. Dennis Dixon, Ph.D., Chief Clinical Officer at 360 Behavioral Health and a co-author of one of the most-cited empirical comparisons of home-based and center-based ABA outcomes, sees the same overcorrection that Ross and Armendariz describe but locates the field’s real problem elsewhere. Speaking with Acuity Media Network at the 2026 Council of Autism Service Providers (CASP) conference in Las Vegas, Dixon said the binary itself is the issue. “I don’t understand folks that are only one or the other,” he said. “It has to be a blend.”
The Competing Behavior Program Running at Home Before Any Clinic Arrives
One of the central arguments Ross and Armendariz make is that caregivers shape their child’s behavior around the clock, whether or not anyone calls it therapy. Every mealtime, every transition, every moment of resistance involves a response from the adults in the child’s life, and those responses either reinforce or discourage the child’s behavior. “Clinics often make the mistake of assuming that the only ABA therapy their patient is receiving is at their program in the center,” Ross said. “The problem is that the child has an instructional program running all day, every day, and usually it is reinforcing the behavior we are not interested in.”
Ross and Armendariz believe that the reinforcing consequences that most reliably shape a child’s behavior do not occur in a therapy room. They occur at home, across hundreds of daily interactions, delivered by parents and adults who have never been trained in behavioral science. Clinic-based services account for a small fraction of the child’s total waking hours, and the interactions outside those hours often work against those inside them. The goal, Armendariz explained, is to train parents to respond to the hundreds of daily interactions they already have with their child in ways that build the behaviors the family actually wants. That shift in how caregivers respond, he argues, is the most important intervention available.
ABA Therapy Generalization: Why Clinics Create the Problem They Cannot Solve
Generalization, the transfer of a skill from one setting to others, is one of the most studied problems in behavior analysis. A 2022 scoping review published in Perspectives on Behavior Science found that reported rates of improved generalization outcomes ranged from 14 percent for emotional skills to 46 percent for language and communication, reflecting how inconsistently the field programs for transfer. Trefor Stokes and Donald Baer cataloged the problem in their seminal 1977 paper in the Journal of Applied Behavior Analysis, identifying nine strategies for actively programming generalization. The first, which they named “Train and Hope,” was not a recommendation. It was a label for the most common approach in the field: teach a skill in one setting and assume it will transfer to others without any deliberate programming to make that happen. Nearly fifty years later, it remains the most common approach.
Armendariz’s answer is not to program generalization more deliberately. It is to make the problem irrelevant. By teaching skills in the settings where they need to occur, with the caregivers who are present every day, he argues that transfer becomes automatic: the behavior was never removed from its natural context in the first place. When parents are trained to respond correctly, the intervention continues every waking hour. The clinic session, by contrast, ends.
Dixon, whose own research has compared rates of skill mastery across the two settings, frames the question the same way. The clinic, in his account, is an artificial environment that providers create because it lends itself to consistent, intensive treatment, but that intensity, he said, only matters if it generalizes back. As Dixon put it: “Doesn’t matter if you can master all these things in an artificial environment if you’re not generalizing it back. A lot of providers have gone way too far in the other direction.”
What Clinic-Based ABA Therapy Actually Delivers (and What It Does Not)
Ross and Armendariz are not calling for blanket elimination of center-based services. Armendariz draws a line between one-to-one drilling of decontextualized academic skills in a cubicle, which he considers unjustifiable, and a preschool-style group setting where children practice social and functional skills in an environment that approximates home. “Working on social skills and play skills in a kindergarten-like environment is perfectly okay and will result in strong outcomes,” he said.
The distinction Ross draws is between efficiency and effectiveness. Clinics are more efficient: more instructional trials per hour, faster skill acquisition, cleaner data. Studies exist to demonstrate this. “You can have an empirical demonstration of the efficiency of the clinic,” Ross said, “and use that as data to support doing more clinic-based services. But do you really get a better outcome?” Dixon’s 2017 retrospective study of 313 children with autism, published with colleagues then at the Center for Autism and Related Disorders, is one of the studies most often cited for that efficiency claim, finding that children mastered learning objectives at roughly twice the rate per hour in clinic settings as at home. The more common error since, Dixon said, has been overcorrection: a wave of providers shifting to exclusively center-based delivery, which he calls a mistake.
The skills acquired fastest in the clinic setting tend to be the ones easiest to teach in a controlled environment: shape discriminations, colors, and matching. Autism insurance mandates generally fund clinic-based ABA for behavior reduction and skill acquisition rather than academic instruction, which falls to schools through individualized education plans. The critique Ross and Armendariz raise is that discrete-trial work in practice often drifts toward decontextualized cognitive drilling, because that drilling produces clean data faster than the functional life skills do. The skills that determine adult independence (preparing food, using transit, managing money, sustaining a conversation) are rarely drilled.
A colleague of Ross’s who has spent over four decades working with adolescents and adults with autism, Peter Gerhardt, Ed.D., Executive Director of the EPIC School in Paramus, New Jersey, has long argued that children exit intensive ABA programs unable to take a bus or shop independently. “They can discriminate hexagons from pentagons,” Ross said. “They do not feed themselves. They do not toilet themselves.” The same structural logic that pushes clinics toward easy-to-document skills also shapes how behavioral health operators adopt technology: optimizing for what is measurable rather than what is meaningful. Dixon, in agreement with the underlying critique, adds a caveat: simply moving services into the home does not, by itself, ensure that what is being worked on is meaningful. The thought process Ross brings to clinical decisions matters more than the location, in his view. Setting is a neutral variable to be individualized and adjusted toward the best outcomes the child can reach.
A child who spends most therapeutic hours responding to adult prompts is being trained, inadvertently, to wait. “Initiation gets killed in that context,” Ross said. Children with autism already tend toward social initiation deficits as a feature of the diagnosis, and Armendariz argues the standard clinic model compounds that deficit directly. “They just sit there and wait for an adult to ask or tell them something. And if you do not initiate, you have no social interactions.”
Dixon’s field experience tracks with that critique even from the operator’s perspective. He said he has seen clinics where children “magically graduate as soon as they hit kindergarten age, not based on any clinical markers,” simply because the cookie-cutter model has aged them out. “Anytime I see a clinic that’s reporting out 100 percent center-based, to me, that’s a red flag,” he said. Treatment setting, in his view, has to be individualized client by client. Some clients need a fully home-based program; for others, home-based care is not feasible and a clinic-only model may be the right choice. What he wants to see across an organization or a single clinic is significant generalization to other settings, on the order of 20 percent of programming occurring outside the clinic. The choice is not binary between home and clinic. The work has to generalize to where the behavior matters: home, school, and community.
BCBA Credentialing at Scale: How Rapid Growth Traded Depth for Volume
Before the Behavior Analyst Certification Board (BACB) was founded in 1998, only six state-level credentialing programs operated in the United States, most of which were modeled on the Florida Behavior Analysis Certification Program, which had operated since 1985. Training was concentrated in a small number of university programs with deep ties to established clinical practice, and new practitioners typically spent years working under senior behavior analysts before being considered competent to practice independently. When the BACB issued its first national credentials, roughly 1,500 practitioners held certification worldwide. By October 2025, that figure had reached 317,699, driven largely by the proliferation of graduate programs designed primarily to move students through an exam. Hundreds of those programs now credential practitioners in as little as eighteen months of coursework.
Dixon has watched that growth from inside one of the largest national providers, and he locates the resulting clinical-quality concern in a place adjacent to where Ross and Armendariz do. With more than half of practicing BCBAs having earned their credential within the last five years, he said, a substantial share of the workforce is still building the confidence to engage parents as partners rather than positioning themselves as experts who arrive with the answers. “They have this fallacy that they’re supposed to be the expert that’s coming in to tell the parent how to do everything right,” Dixon said. “They instead need to be coming in as an expert in behavior analysis that then partners up with the parent as the expert in their child.” Many newer practitioners, he said, are taking on expectations of themselves that are not needed and then faking it, and that uncertainty is one reason caregiver coaching is often the first thing to drop out of a clinical session.
Private Equity in ABA: Where Operators and Critics Disagree
Meanwhile, the entry of private equity into the ABA market has compounded the problem in the view of Ross and Armendariz. Cookie-cutter programs delivered for maximum billable hours, with structural variables like staff availability and authorization windows driving treatment decisions instead of clinical need. “Whatever problems we were having,” Armendariz said, “with them coming in, they became even worse.” Payers have responded in kind: state Medicaid programs are increasingly capping hours and restructuring reimbursement in ways that bear little relationship to individual clinical need. The actors driving those state-level interventions, however, have not been exclusively PE-backed; recent enforcement actions and Medicaid reforms have followed conduct by clinician-owned and independently held providers as well, suggesting that the operational distortions Ross and Armendariz describe are not solely a function of ownership structure.
Dixon, who has worked with two PE sponsors over the course of his career (Blackstone in his prior role at the Center for Autism and Related Disorders, and DW Healthcare Partners at 360 Behavioral Health) offered a sharply different view. “They’re the first to be the champions of quality,” he said of his current and former sponsors. The reason, in his telling, is structural rather than altruistic: an early sponsor explained to him that the entire value of a healthcare company resides in its brand, and that a reputation for poor care collapses that value to zero. “You’re not going to get patients,” Dixon said. “The payers are certainly not going to work with you. You’re on a downward spiral as soon as you let quality slip.” He acknowledged the cookie-cutter operators Ross and Armendariz describe, but argued that the underlying drive to find a repeatable, scalable model is not unique to PE-backed providers, and that access in many parts of the country depends on it.
What Home-Based ABA Therapy and Parent Training Actually Look Like in Practice
Armendariz describes his assessment process as the “PBI” (Pocket Book Instrument), a deliberate joke at the expense of formal assessment culture. It consists of following a family through a child’s morning routine and setting goals based on what the child cannot yet do independently. “Do I think I am going to be teaching him to discriminate between a triangle and a hexagon? No way. I am going to be teaching him to get up by himself, to dress himself, to come and sit, to eat what everybody else eats.”
The parent training that follows communicates in concrete behavioral terms rather than theory. “My nutritionist does not teach me the chemistry of lettuce,” Armendariz said. “They tell me to eat greens at the beginning of the meal. They give me something to do, and I go do it.” The goal is a parent who can, in the moment, discriminate between behavior worth reinforcing and behavior worth ignoring.
Dr. Armendariz recalled watching a therapist run a flashcard vocabulary session with a child who had no functional verbal requesting behavior. At one point, the child reached toward a banana sitting on a counter just out of reach. The therapist noticed, got up, and pushed the banana within reach. The child took it. “That therapist had a perfect teaching opportunity right in front of them and walked straight past it,” Armendariz said. “By handing the banana over, they reinforced the opposite of what the whole session was supposed to build. The child did not learn to request. The child learned that reaching works.” Ross noted that the moment is not a one-off failure. It is, in his view, the predictable result of a training system that prepares practitioners to run procedures in controlled settings without giving them the behavioral science to recognize what is happening when a child reaches for a banana.
The Argument the Field Has Not Yet Had: Process or Outcomes
Where Ross, Armendariz, and Dixon converge most clearly is on what they think the next argument should be about. The location debate, in Dixon’s view, is a proxy for a question the field has been slow to answer with data. “Let’s talk about outcomes,” he said. “Let’s get to a point as a field where we don’t squabble about process as much as using an analysis of what process got to the best outcome.” He pointed out that while the Vineland Adaptive Behavior Scales may not be the best standalone tool for measuring treatment outcomes, it remains a reliable and valid window into the functional changes parents observe in daily life. “The increased use of objective autism assessments will allow us to conduct the retrospective analyses needed to understand which treatment processes lead to the best outcomes, and for whom,” he said. He also said he is unworried by the wave of OIG audits and payer scrutiny now reshaping the field, calling it a maturation step that should leave behind operators capable of documenting that their care produced value. Ross and Armendariz would arrive at the outcomes question from the other direction: their argument is that without a return to the home, the outcomes the field is being asked to demonstrate will continue to be the wrong ones.
Frequently Asked Questions
What is the difference between home-based ABA therapy and clinic-based ABA therapy?
Home-based ABA therapy delivers intervention in the child’s natural environment, working with the family members and caregivers who are present during the behaviors that matter: mealtimes, dressing, sibling interactions, community outings. Goals are built from a functional assessment of what the child cannot do independently in real daily routines. Clinic-based ABA therapy delivers intervention in a controlled, center-based setting, typically in one-to-one sessions with a therapist using structured materials and contrived reinforcers. Clinics are more efficient for data collection and billable hours, and existing research, including a 2017 retrospective study by Dixon and colleagues, has documented faster rates of skill mastery in clinic settings. Whether that efficiency translates into better outcomes depends on whether skills generalize back to the home, which Ross, Armendariz, and Dixon all identify as the bottom line. Ross and Armendariz argue that the clinic has become the default not because the science demands it but because it is operationally easier to staff, schedule, and bill. Dixon argues that a 100 percent center-based model is itself a red flag.
Why does ABA therapy have a generalization problem, and how is it addressed?
Generalization, the transfer of a learned skill to new settings, people, and situations, is one of the most studied problems in behavior analysis. It is not a flaw in ABA’s principles. It is a predictable consequence of teaching skills under controlled conditions that bear little resemblance to where the skill needs to occur. Stokes and Baer identified nine strategies for actively programming generalization in their 1977 paper in the Journal of Applied Behavior Analysis. The first, “Train and Hope,” describes the most common approach: teach the skill in one setting and assume transfer will follow without deliberate programming. Armendariz’s alternative is to bypass the problem entirely by teaching skills in the setting where they are needed, with the people who need to deliver them. When parents are trained to respond appropriately, generalization is automatic because the behavior was never isolated from its natural context in the first place. For a deeper look at how to evaluate ABA research claims, including whether outcome data reflects real-world generalization or controlled-setting performance, see Acuity’s conversation with Dr. Gina Green.
Why is parent training so important in ABA therapy for autism?
Parent training is important because parents are the most consistent, most available, and most powerful source of behavioral consequences in a child’s life. Every interaction between a parent and a child is a learning trial. Whether or not anyone designs those interactions, they are shaping behavior. When parents respond to crying by providing what the child wants, they reinforce crying. When parents ignore words and respond to reaching, they discourage language. Clinic-based ABA cannot compete with those contingencies because the clinic sees the child for a fraction of the day. Effective parent training teaches caregivers to respond correctly to the hundreds of behavioral interactions they have with their child every day, using simple behavioral principles communicated in concrete terms, not in the technical language of the science. When that training works, the intervention continues every waking hour. The outcome, as Armendariz has observed across decades of international consulting, is a family that can manage new problems as they arise without returning for a service authorization.
How has private equity investment changed ABA therapy quality?
Private equity entry into the ABA market has reshaped how care is delivered, and clinicians inside the field disagree on the net effect. Critics, including Ross and Armendariz, argue that PE-backed operators favor high-volume, center-based programs because they maximize billable hours per staff member and simplify scheduling, with treatment decisions driven by structural variables (staff availability, authorization windows, room capacity) rather than individual clinical need. The result, in their view, is intervention standardized around what is easy to document and bill, and a model that is financially self-reinforcing. Operators inside PE-backed providers, including Dixon at 360 Behavioral Health and previously at the Center for Autism and Related Disorders, push back on the demonization framing and argue that quality is in fact a sponsor priority because brand reputation drives company valuation. Dixon also argues that the drive to find a repeatable, scalable model is shared by founders and nonprofits, and that access in rural and mid-sized markets often depends on it. Both views are likely to continue coexisting as payers, regulators, and merger and acquisition activity across the behavioral health industry continue to reshape the market.
What are the seven dimensions of applied behavior analysis, and has ABA drifted from them?
Donald Baer, Montrose Wolf, and Todd Risley published “Some Current Dimensions of Applied Behavior Analysis” in 1968, establishing seven characteristics that define ABA as a discipline: applied (addressing socially significant behavior), behavioral (measuring actual behavior), analytic (demonstrating experimental control), technological (described precisely enough to be replicated), conceptually systematic (grounded in behavioral principles), effective (producing meaningful change), and capable of generalized outcomes. The seventh dimension, generality, requires that skills transfer beyond the treatment setting and maintain over time. Ross and Armendariz are co-authoring a chapter arguing the current delivery model most directly violates this dimension. Teaching skills in a controlled setting without any deliberate programming for their occurrence in the environments where they matter fails the dimension Baer, Wolf, and Risley considered essential for any intervention to be called truly applied.
What does a corrected ABA therapy model look like for families and providers?
A corrected model begins with a functional assessment of what the child cannot do independently in the environments where it matters: getting dressed, eating what the family eats, managing transitions, initiating interaction with a sibling. Goals follow from that observation, not from standardized curricula. The primary site of intervention is the home, with the people who are present every day. Parent training uses concrete, behavioral instructions rather than theory, teaching caregivers to respond correctly to the behaviors they value and to stop inadvertently reinforcing the ones they do not. When children need group social experience, a preschool-style environment that approximates home is acceptable. One-to-one drilling of decontextualized academic tasks in a cubicle is not. Telehealth parent-mediated models can be effective when they deliver real-time coaching on actual behavior rather than replicating a remote clinic session. Dixon’s view, shared in part by Ross and Armendariz, is that the clinic-versus-home debate is finally giving way to a more important one: whether the field can produce, measure, and stand behind the outcomes its families came for. Behavioral health technology adoption follows the same logic: tools that help practitioners deliver intervention in the natural environment and train caregivers effectively have clinical value. Tools that simply make clinic-based documentation faster do not change the underlying problem.
