A Thought for the Scary Season #4: More Concerns About Equity and Inclusion in the Science of Behavior

Posted on by
2

PREVIOUS POSTS IN THIS SERIES

Disciplines do not, and cannot, remain static. Behavior analysis was once the province almost exclusively of middle-class white males, and a positive step in our evolution has been a growing awareness that the discipline must attend to matters of equity and inclusion.

Put simply, to fully achieve our collective potential as a positive force in society requires directly confronting questions about who is able to join the behavior analysis club, and who is able to benefit from the science of behavior.

Not that these questions always are comfortable to contemplate. By some assessments:

Where equity and inclusion are concerned, then, our discipline is a work in progress. Confronting the work yet to be done can be daunting, but the whole point of baselines is to define the opportunity (and imperative) for change. It’s better to know to know our shortcomings than to passively perpetuate historical traditions of differential privilege.

A further reminder of this point can be found in a recent critique of inclusion, or rather the lack thereof, in Contextual Behavioral Science, titled, “The illusion of inclusion: Contextual Behavioral Science and the Black community.” The article, published in Frontiers in Psychology, argues that mental health needs may be different for whites and people of color, and asserts that the evidence base of Acceptance and Commitment Therapy (ACT) does not currently support claims of effectiveness with marginalized subpopulations. For example:

Although inclusion of Black participants as per the US Census is perhaps adequate, the participation of marginalized ethnic groups per study is too low to assess racial differences in treatment response, and the studies were not designed to do so. This makes the generalizability of the pooled data questionable.

Here’s a key point being advanced: Although demographically balanced samples are usually considered desirable for randomized controlled trials (RCTs) and other effectiveness studies, results from such studies generally can support a claim of effectiveness on average. Too few individuals representing a subgroup may be available to permit analyses of the effectiveness for that subgroup.

Consider an illustrative example. Let’s imagine that a given therapy has been evaluated in 8 RCTs, each with groups of about N = 100. Now, in each group, assuming demographically representative samples from the U.S. population, we would expect to find about 1 person identifying as American Indian or Alaska native (1.1% of U.S. residents according to the 2020 Census). All told, that’s about 8 such individuals in treatment groups and 8 in control groups — far too few to support any trustworthy claims of subpopulation effectiveness according to the analytical conventions of group-comparison studies (which are necessary to answer certain kinds of effectiveness research questions).

The paper does mention several ACT studies with majority-minority samples, but argues, correctly, that these comprise a tiny slice of the overall evidence base. 

The paper also claims that inequity extends beyond research programs and into the contextual behavioral science professional community itself, with most Association for Contextual Behavioral Science (ACBS) governing board members being white, and only about 1% of Journal of Contextual Behavioral Science editorial board members being Black. Those are objective data. Somewhat harder to verify is the paper’s claim, elaborated at some length, that ACBS and the contextual behavioral science community have been resistant to calls for greater inclusiveness.

What do I think about all of this? Remember that I am a white male of privilege and therefore not necessarily the best evaluator, but here goes anyway:

  • I’m disappointed that the article said little about equitable access to ACT services. ACT, like all contemporary therapies, emerged into a world where service delivery is funded by health insurance and cash payments, which in the U.S. virtually assures that certain subpopulations will be underserved. As I wrote in a previous post, when something works (and ACT does), we’re ethically bound to spread its benefits equitably across people in need. The paper’s authors might disagree with me on the basis that limited minority-specific effectiveness evidence means that we don’t actually know whether, for those people, ACT works. And they have a point.
  • To be fair to the ACT evidence base, it mirrors what we see for most therapies that meet contemporary standards of evidence-based practice, which tend to emphasize the greatest good for the greatest number. The approach requires that subpopulations be part of the evidence base, but in proportion to their prevalence in the population. Thus, in numerical terms, subpopulations are de-emphasized relative to majority groups. This problem is real, but it’s baked into the general evidence-based practice movement, and not unique to ACT. [Also see the Postscript]
  • The article’s reference list suggests that its authors haven’t conducted the kind of research they believe is necessary to properly validate ACT for Black clients. This is a pretty generic foible of literature critiques: It’s easy to fault other people for not doing the studies you would like to see. The most direct way to address a gap in the literature is to fill it.
  • Regarding equity and inclusion in ACBS, not everything in the article surprised me. Some ACBS members have independently described to me their sense of an entrenched white male hegemony in the organization that devalues their views. I’m reporting hearsay, of course, but let’s be honest about the big picture: A behavioral organization dominated by white people of privilege would not exactly count as an outlier.

As with all published papers, I urge you to read this one for yourself, and cast an appropriately skeptical eye at methods and evidence. My own opinion: Not everything asserted is backed by evidence that rises to standards usually required in peer-reviewed publications. But here I’m simply applying a normal tenet of science dissemination: We readers are supposed to think critically rather than take as gospel what we read. So draw your own conclusions about this article’s particulars, while keeping in mind the concept of Bayesian priors. There is lots other evidence of racial/ethnic/cultural/gender/linguistic skews in the science of behavior. No reasonable consideration of this paper is likely to paint a different picture..

Author Note: I’m all for open dialogue. If someone representing the CBS research community would like to respond in a guest post, contact me.

Postscript (added 10/27/24)

The passage below (undated) is reproduced from Steve Hayes’ blog. I agree with Steve’s general point, that the more broadly applicable an intervention proves to be, the more likely it is to generalize to further settings and types of clients. But if we take seriously the tenets of the evidence-based practice movement, we can’t assume generality to a given setting or type of client until empirical evidence of this exists (as the old saying goes, “absence of evidence is like evidence of absence” of an effect). Do note that, as described in the “Illusion of Inclusion” paper, the limited available evidence specifically addressing Black Americans provides no reason to think that ACT is not effective with that subpopulation. So, in a way, perhaps everyone is right here. ACT has shown impressive generality across problems and types of people, and we need more data specific to its efficacy when applied to the experiences of Black Americans. 

Most of the randomized controlled trials on Acceptance and Commitment Therapy come from high-income countries but a surprisingly large percentage (45%) come from lower and middle-income countries (LMICs). There ACT research has taken on a different character, often prioritizing well-being and quality of life over symptom reduction. 

The global expansion of ACT research has highlighted the importance of understanding how psychological interventions function in diverse cultural contexts. The careful and and at times slow effort to built out the psychological flexibility model has paid off because it has made it easy to culturally modify ACT. For example, ACT principles such as being more open, aware, and focused on your values resonate with all major religions. Thus, in Iran ACT interventions can make their points using stories drawn from Islamic teaching, while in China they may be drawn from Daoism, Confucianism, or Buddhism.

Some important problems that are commonly studied in LMICs such as the distress of infertility in couples who wish to have children are rarely studied in high-income countries because funding agencies do not view them as “disorders.” We as a human community need to learn from LMICs and ACT research has led the way by providing ideas and methods that have spread globally. This holds out hope that we can learn to better serve people from various cultural backgrounds in a world that is becoming increasingly diverse.