Seven Dimensions Are Not Enough: Actively Disseminating Applied Behavior Analysis

Abstract

Within a popular seven-dimension framework, applied behavior analysis (ABA) interventions have historically been characterized by a focus on socially important behaviors. We propose that an intervention does not truly count as applied until it is being widely used and the world has improved in some meaningful way as a result. This perspective places active dissemination of ABA, with the goal of promoting adoption, on equal footing with other defining features, making it functionally ABA’s eighth dimension. According to this perspective, much of the work from ABA first 6 decades qualifies as pilot or developmental, with widespread adoption yet to be achieved. Successful dissemination will require some changes to ABA’s modus operandi.

Applied behavior analysis (ABA), or something like it, was first imagined in B. F. Skinner’s (1948) description of a fictional community organized around the wise and kind application of behavior principles. Walden Two exuded optimism that the laws of behavior could be harnessed to better the human condition, and together with Science and Human Behavior (Skinner, 1953), it fueled a revolutionary spirit that early behavior analysts trumpeted in rallying cries like, “Better living through behaviorism!” (Birnbrauer, 1978) and, “ABA can change the world!” All that was needed, it seemed, were the right behavioral technologies and, with guidance from Baer et al.’s (1968) iconic seven-dimension framework, ABA has now generated more promising interventions than can easily be counted (Heward et al., 2022). Yet somehow the anticipated behavioral revolution is still gestating. The promise of Walden Two was that once we had built and explained our better mousetrap the world would beat a path to our door to utilize it. We did our part, but the response was a resounding “meh” (e.g., Detrich, 2018; Dixon et al., 2018; Doughty et al., 2012; Bailey, 1991; Foxx, 1996; Skinner, 1987). As a result, the world remains as broken as ever.1

It Don’t Come Easy2

Why hasn’t ABA’s amazing behavioral technology been more widely adopted by those who would benefit from it? The answer, strangely enough, is that we behavior analysts, the world’s foremost experts in behavior, underestimated the behavioral complexities of connecting consumers to better mousetraps (Critchfield & Detrich, 2022; Detrich, 2018). Even in 1968, Baer and colleagues had conveyed some wariness about this particular can of worms. In a passing mention of societal redesign, they offered this tepid observation (our cynical translations are inserted in italics):

A society willing to consider a technology of its own behavior [the lightbulb must want to change!] apparently is likely [not guaranteed] to support that application when it deals with socially important behaviors. . . . Better applications, it is hoped [please please please], will lead to a better state of society. . . . [I]t is at least a fair presumption [we think, but aren’t sure] that behavioral applications, when effective, can sometimes lead to social approval and adoption [but expect a lot of failure]. (Baer et al., 1968, p. 91)

Baer and colleagues were right to be wary because, as it turns out, behavior analysts can build a whale of an intervention, but in most instances other people will decide whether or not to adopt and implement it (e.g., Baer et al., 1987; Skinner, 1987). Those people need some persuading.

If it makes behavior analysts feel better, we are not the only folks to have bruised our foreheads on barriers to adoption. History is bursting with fine solutions that were not readily embraced. One famous example is the discovery that scurvy, once the leading killer of seagoing sailors, is prevented and cured by consuming Vitamin-C-rich citrus (Griffin, 2013). Nearly 200 years passed between this discovery and the first fleet-wide application of “citrus therapy.” Keep that lag in mind because it vividly illustrates a central theme of this article: Adoption is something people do (behavior), and as every behavior analyst knows, behavior doesn’t just magically happen. Adoption is a dependent variable, and the trick is to find independent variables that influence it.

Dissemination encompasses all of the things we may do to influence adoption. Behavior analysts often have taken an “If you build it they will come” approach to dissemination by devising effective interventions but doing little promotion. Greenhalgh et al. (2008) called this approach passive dissemination, and it’s not really dissemination (see Dixon et al., 2018). True dissemination—active dissemination—requires making innovative solutions obvious and irresistible to adopters. It’s based on the lay concept of “good communication” and the Skinner (1957) notion of reciprocal influence between speaker and listener. Its Prime Directive: Seek out potential adopters, of which many kinds exist,3 and tailor persuasive messages to their needs. Active dissemination, it should be apparent, cannot be accomplished by talking to other behavior analysts, by talking to potential adopters as if they were behavior analysts, or by approaching different types of potential adopters in the same way all.

The Missing Eighth Dimension

Although the seven-dimension framework is a good guide to developing interventions, it is mute on how to get them adopted. This is lamentable because an intervention that is not adopted cannot make a difference in the world, and humbling because if ABA interventions don’t make a difference, well, why have ABA at all?4 It’s great to show something works by conducting small-N experiments and crafting well-controlled demonstration projects (things at which the ABA community excels). But let’s not delude ourselves into thinking that the world is a measurably better place for such efforts. In defining the applied dimension of ABA, Baer et al. (1968) emphasized studying socially important behavior that is defined by, “the interest which society shows in the problems being studied” (p. 92). That is not identical to the interest which society shows in the resulting interventions. When the study of socially important behavior does not attract attention, inspire adoption, or yield socially important outcomes, it might as well be basic research for all of the good it directly accomplishes. Hence we propose that, moving forward, active dissemination be considered ABA’s eighth defining dimension, with the original seven viewed as necessary but insufficient to create the socially important outcomes that are the true essence of application. Below we briefly discuss how a concern for dissemination changes the up-front work of creating and validating interventions, and how it defines the subsequent challenge of persuading adopters.

Creating and Validating Interventions

Focus on Socially Important Outcomes

Dissemination is about getting people to adopt a mousetrap that is more beneficial than current practices. In the language of the operant choice (Baum, 1974), a better mousetrap should be a relatively rich source of reinforcement that can shift control of adopter behavior away from some existing practice. Before dissemination is attempted, therefore, attention must be paid to verifying that the mousetrap is genuinely better, which involves at least two considerations. The first is to recognize that most effectiveness evidence in behavior analysis shows that an intervention works, but not that it works better. Single-case ABA research is usually designed to tell us whether an intervention is more effective than doing nothing. Less often do ABA studies compare the innovations of behavioral technology directly with practices that might be replaced. Treatment-comparison studies are not rare in ABA, but they usually compare different variations of behavioral interventions (e.g., Shabani & Lam, 2013). This leaves the defining feature of a better mousetrap rather unverified.

The second consideration is to ensure that the mousetrap transparently addresses a problem that is important to the adopter (Rogers, 2003). This can be a stumbling block for behavior analysts, because in ABA we tend to describe and measure things differently than potential adopters do. Behavior analysts take for granted that the discrete behaviors on which ABA focuses can aggregate to create outcomes that are valued in the everyday world, but that’s not evident to everyone. A research team (Strain et al., 2012) was reminded of this point when they designed an in-home program to teach children developmentally-appropriate skills, only to discover that parents were more concerned with being able to complete daily routines on time. Teaching children developmentally appropriate skills certainly can help them perform daily routines, so researcher and parent aims were not incompatible. But dissemination would be made easier by underscoring how the program met parent (adopter) needs. This demands a functional analysis of sorts to discover what outcomes are an adopter’s reinforcers (Pinkelman et al., 2022), and sometimes those outcomes are not behaviors at all. Lovaas (1987) understood this when sharing the results of his groundbreaking early intervention program for children with autism. He emphasized that the program raised IQ scores and resulted in age-appropriate school placements, things that are valued by educators and parents. The program itself focused on changing a constellation of discrete behaviors that, once built up, were responsible for these general outcomes. But the published report minimized attention to such details, focusing instead on the outcomes that would show, transparently and without any tortured explanation, that the intervention was of practical importance to adopters.

A good rule of thumb is that, to the maximum extent possible, outcome evidence should accentuate the stakes involved in solving a particular problem—highlighting how, from the potential adopter’s perspective, the intervention makes the world a better place. Different adopters care about different outcomes. For instance, parents may prioritize completing daily routines whereas policy makers may care about cost-effectiveness (as per the demonstration that each case of treated autism can yield huge economic benefits; Smith & Lovaas, 1998), and population-level prevention (as per the finding that early exposure to the Good Behavior Game of Barrish et al. (1969) reduces risk of adult substance abuse and psychiatric problems; Kellam et al., 2008). And whatever big-picture outcome matters, you have to plan to measure it. We don’t suggest changing how behavior analysts evaluate whether interventions work at the behavioral level of the seven-dimension framework (Baer et al., 1968), but in the intervention development phase an extra layer of measurement may be required to drive later dissemination.

Designing Interventions for Scaling Up

Also critical to the success of dissemination efforts is creating interventions that are structurally compatible with adoption settings (Detrich, 1999; Detrich et al., 2016). Many interventions described in behavior analysis journals require resources and staff expertise that aren’t available in the field. To make the leap from lab or teaching clinic to the everyday world, an intervention must work within the constraints that exist where it will be used. This is a major consideration—perhaps the primary one—in scaling up, the process of shifting from small-scale demonstration project to large-scale intervention rollout.5

A familiar example of scaling up in behavior analysis is the installation of Positive Behavior Interventions and Supports (PBIS) in tens of thousands of schools (Kincaid & Horner, 2017). It’s an apt example because PBIS has sometimes gotten the side-eye from behavior analysts for diverging from standard practices in ABA6; however, much of what makes it different is precisely what’s necessary for transportability to tens of thousands of settings (e.g., Critchfield, 2015). Recognizing that a scaled-up intervention like PBIS will not be an exact replica of your typical small demonstration project dramatically changes the goals of initial intervention development. Demonstration projects tend to test what works in a best-case scenario, but before attempting dissemination it’s critical to know what works given the myriad exigencies of adoption environments (Slocum et al., 2014). An intervention must be purpose-built for scaling up, that is, designed from the outset to work under conditions that are less than ideal but unavoidable. This is a complex matter (Fixsen & Blase, 2019; Fixsen et al., 2024), so for present purposes suffice it to say that without special precautions the results of scaling up will be haphazard, as illustrated in a case study described by Araujo et al. (2021). In the 1980s, a small home-visit based program was shown to improve developmental outcomes for a few dozen impoverished Jamaican children. The program was then replicated in Columbia (N = ~700 children) and Peru (N = ~70,000), though with depressingly watered-down results. Effect sizes dropped by 70% or more, mirroring a familiar pattern of effect sizes decreasing as interventions scale up (Gupta et al., 2021). The cause, however, might not really be intervention size. Rather, as interventions become implemented more widely, implementation integrity tends to degrade (Detrich & Keyworth, 2010) as adopters decide how to tailor an intervention to their resources and other local constraints. For instance, compared to the original Jamaica project, the replications involved different types of children; home visitors with less education, training, and supervision; and a less reliable schedule of home visits. Implementation drift thus could explain mediocre results, and is always a risk unless interventions are designed to fit easily into implementation settings. ABA demonstration projects, which so often represent best-case scenarios, need to look more like what is likely to be done out in the wide world.

Persuading

Active dissemination is what it sounds like: identifying potential adopters, discovering their priorities and goals, and engaging them directly rather than hoping they will discover on their own what ABA has to offer. What this means specifically will vary depending on the context. Behavior analysts might personally climb through the ranks in some system until they gain some control over it (Baer et al., 1987), as occurred in a number of state developmental disabilities agencies. They might “market” directly to individual adopters as per Azrin and Foxx’s (2019) Toilet Training in Less Than a Day. They might find ways to harness the power of public media (e.g., Critchfield, 2023). Or they might show policy makers and public officials how an intervention would benefit them, as occurred with PBIS (e.g., Bradshaw & Pas, 2011; Lohrmann et al., 2008).

Every adoption environment presents unique demands, but in all cases dissemination requires clear, persuasive communication. Behavior analysts often unintentionally obfuscate while attempting to inform, in part because the technical language on which we rely can drive away potential adopters (e.g., Critchfield et al., 2017) and impair their ability to properly implement interventions (Jarmolowicz et al., 2008; Marshall et al., 2025). In the verbal behavior equivalent of our discipline’s oldest maxim—The organism is always right—it’s essential to use words that work for the adopter.

And although only empirical evidence can define a superior mousetrap, persuasive efforts should not necessarily present that evidence in unadulterated form. The Peruvian government discovered as much when they launched an initiative to promote the boiling of drinking water in rural villages. Government agents were dispersed to the countryside to talk about what research teaches us, that boiling destroys harmful pathogens. Such a lesson that might sell in suburban America or any other place with an educated populace, but in rural Peru formal education is limited. Villagers, having no understanding of microorganisms, were not in a position to appreciate how boiling defeats them. Sometimes hard evidence confuses rather than enlightens.7

Adoption may be an expression of behavioral choice, but in choice there are biasing factors that can yield more or less preference than expected based on reinforcement rate alone (Baum, 1974). In a social world, one prominent biasing factor is the influence of other people. In Peru, once it became clear that water boiling would be a difficult sell, officials worked closely with community leaders, because their buy-in would make boiling seem socially normative (Rogers, 2003). The State of Texas did something similar with a 1980s public service campaign to reduce roadside littering (Nodjimbadem, 2017). A television ad called litter unattractive and expensive to clean up, but arguably its critical component was iconic Texas musician Steve Ray Vaughn invoking Lonestar pride by intoning, “Don’t Mess With Texas.” Texans may not always feel warm and fuzzy about the government but Vaughn’s pronouncement carried roughly the same weight as Moses delivering God’s commandments. Within 4 years litter was reduced by over 70%.

ABA, as applied to autism, was gifted its own Stevie Ray Vaughn when parent Catherine Maurice (1994) wrote Let Me Hear Your Voice to chronicle her family’s quest to find effective services. Parents of children with autism, it turned out, trusted a fellow parent the way Texans trusted Stevie Ray Vaughn, and arguably Let Me Hear Your Voice had more to do with the public’s current enthusiasm for ABA than anything special about ABA itself (Detrich, 2018). It was our dumb luck that a satisfied customer happened to be a professional writer who was adept at telling an emotional story.

Regarding emotion: it’s one of those private events that behavior analysts tend to shy away from when designing interventions, but it can be conceptualized in terms of fluctuating motivating operations (Critchfield, 2018; Skinner, 1953), and is a potential biasing factor in choice.8 Stories that tug on the heart strings move people. Politicians know this (Polletta, 2008). Advertisers know this (Kang et al., 2020). Behavior analysts, not so much. We tend to tell the stories of our discipline‘s successes with graphs and technical terms and journal articles, but none of that, in the eyes of most observers, counts as a compelling plot (e.g., Bailey, 1991; Foxx, 1996).

Regarding plot: as effective as the Lovaas (1987) report was, a better story was told in the companion documentary Behavioral Treatment of Autistic Children.9 That film’s capstone segment juxtaposed before-and-after descriptions of selected children with everyday video footage and interviews with the kids and their parents. We have seen this film move viewers to tears, highlighting the distinction between savvy scientific communicators like Lovaas, who clarify evidence capably, and gifted story tellers, who provocatively trace the twists and turns in a sympathetic character’s experience. A good story synchronizes listener emotions to the rise and fall of protagonist fortunes (Vonnegut, 2005). This holds attention and can catalyze action (Critchfield, 2018; Hineline, 2018; Kang et al., 2020; Livas et al., 2018). In the latter case, Behavioral Treatment and Let Me Hear Your Voice nominally tacted selected individuals’ experiences, but to a parent of a child with autism functioned as a mand that screamed, “Get your kid into ABA!” Most behavior analysts have a long way to go when it comes to recognizing when an emotional story is called for, not to mention knowing how to tell one.

Toward Better Living

ABA has had a pretty great run so far, but outside of autism and a few other domains this performance has largely gone unnoticed. Our many promising interventions justify the optimism of Skinner (1948, 1953) and other early observers that behavioral technology can be good for what ails the human condition. But even though we now possess a sizeable arsenal of relevant technology the human condition is still in sorry shape. The question for the future is whether ABA can change in ways that will really facilitate “better living through behaviorism.”

ABA arrived at its current ambiguous state (wildly effective but lacking impact) because it was founded on a too-narrow conception of applied, one that demands close attention to the socially important behavior of clients but not to the equally important behavior of adopters. That is why it’s time to add active dissemination to the formal definition of ABA. Aside from the occasional happy accident in which a Catherine Maurice steps up to promote on our behalf, the only way for ABA to become integral to everyday life is for us to make that happen. Progress will depend on developing interventions that are purpose-built for disseminating; on researching the adoption process itself to guide dissemination efforts; and most certainly on spending more time talking to potential adopters and less time talking to one another.

Fortunately, there are signs of growing interest among behavior analysts in the dynamics of dissemination (e.g., Geller, 2022; Griffith & Ré, 2022; Kelly et al., 2019; Lerman, 2024), and it’s obvious that behavior analysts can do what’s needed to spread the good word about ABA because in a few instances that has already happened. Think of the proliferation of PBIS in schools, or the 7+ million copies sold of Toilet Training in Less Than a Day, or of the mainstream embrace of the Good Behavior Game. Any instance in which behavioral technology has claimed a place on society’s larger stage—gaining adoption, demonstrating effectiveness at scale, and promoting better living in ways that regular people can understand—is an exemplar to celebrate and learn from. But to speed the learning curve we should also embrace as colleagues people in various disciplines who are already unraveling Eighth Dimension mysteries (e.g., Fixsen et al., 2019; List et al., 2021; Rogers, 2003), because they have much to teach us.

Second Author’s Postscript

Ronnie Detrich (Fig. 1) passed away on September 9, 2023, bringing to a close a career of over 50 years delivering behavioral services and crafting support systems in a host of settings. Ronnie was a committed practitioner of ABA and a champion of actively disseminating it (Twyman et al., 2024). In the latter capacity he may be best known for his efforts at deciphering the implications of the evidence-based practice movement for ABA and educational systems (e.g., Detrich, 2013, 2020; Detrich et al., 2016; Slocum et al., 2014), and as a founder of the Wing Institute for Evidence-Based Education (now Wing Institute at Morningside Academy). This article is based on Ronnie’s Don Baer Memorial Address at the 2019 Association for Behavior Analysis Convention. As illness gradually overtook him, one of Ronnie’s final wishes was that the presentation be written up for publication, and I’ve given that my best shot using notes, slides, and context kindly provided by Cheryl Detrich. But mea culpa: Few people could craft a thoughtful argument like Ronnie, so I apologize for any place in the article where I failed to capture his voice.

Fig. 1.

Ronnie Detrich delivering his Don Baer Memorial Address. Photo courtesy of Cheryl Detrich

Author Contributions

Ronnie Detrich created the broad outline of this article and identified many of the examples presented herein. With Cheryl Detrich’s blessing, Tom Critchfield developed the formal manuscript.

Funding

No extramural funding supported this work.

Data Availability

No datasets were generated or analysed during the current study.

Declarations

Conflicts of Interest

The authors declare no financial or nonfinancial conflicts of interest related to this work.

Ethical Oversight

Ethics board oversight does not apply because this work does not present original research data.