Squashing the beef?
Revisiting my beef with behavioral science
Nearly a year ago I wrote a post here about my beef with behaviorial science. It’s good, you should read it! With the news that one of the discipline’s highest profile practitioners, Francesca Gino, has been stripped of tenure and fired by Harvard, I wanted to briefly revisit and update the piece.
While I’m the first to admit that I am often (irrationally) annoyed by narrow academic arguments, behaviorial science is unique in its ability to drive me nuts. Its capacity to infuriate rests primarily on the ways it’s facile assumptions about how the world works intersect so nicely with equally shallow lay understandings of the mechanics of social life. The result is a autocatalytic network of nonsense. Where the worst people in the corporate world are armed with “science” that substantiates their dumb ideas, reinforcing their claims for authority while they implement activities that range from the broadly ineffective to the downright destructive. From nudges, to power poses, to social priming, the nonsense propagated by behaviorial science is like corporate kudzu slowly covering and ultimately strangling organizations.
But as a former academic current sell-out working in the private sector, I remain convinced that we can make corporate life better with more social science. And to that end, I want to share some tips for identifying and using better social science frameworks.
Tl;dr: avoid just-so stories.
Just-so stories are traditionally defined as untestable assertions about some set of events. They are because they are and that just how they are. The allure of behaviorial science is that it provides a veneer of statistical reasoning that proves to you, dear reader, that it is in fact not a just-so story. Look! There are calculations, sometimes quite complicated ones, and we can see that these calculations point to “real”effects between treatment groups. There are p-numbers, for Chrissakes! And these numbers provide a shorthand for reassurance. They confirm that we are in the presence of science. And, by extension, are not being feed a just-so story.
But as any undergrad who’s completed their statistics requirements can tell you, while ice cream sales are highly correlated with drownings, dairy isn’t the silent killer. It an intervening event - it’s summer and hot out - that driving both events. In short, it’s not the statistics that give us the insight but the theoretical framework that shapes their use. And behaviorial science’s theoretical frameworks, while reassuringly comforting in their correspondence with lay understandings are primarily a set of just-so stories told through regression analysis. What is to be done?
In the 1988 classic, “They Live”, “Rowdy” Roddy Piper slips on a special set of sunglasses that enable him to see how the Earth is being dominated by hostile aliens who hold us in submission through subliminal messaging and class conflict. While I have nothing as bad-ass, we can escape just-so stories and develop better understandings about what’s happening and how to explain and intervene in it by drawing on and reversing three of the core assumptions of popular behaviorial science.
Avoiding reductionism. A core tenet of behaviorial science is the positioning of individuals as the inviolable foundations of social life. In this world view, things that happen - political changes, shifts in consumption patterns, emergence, growth, decline, or transformation of entire industries is explainable, ultimately, by examination of individual action or related mental events. To which I ask: how? What are the mechanisms connecting individual actions or personal mental states to anything? How does it connect? How does it aggregate? Is it consistent across levels of aggregation? How? It’s a frankly nonsensical explanatory framework that breaks under basic questioning. Something else is clearly happening (and it’s organizational but I will save this for another time). My point here is that the absence of identifiable, stable mechanisms that get us from individual to wider social context calls the whole argument into question, freeing us from a just-so dead-end.
Questioning universal claims. Behaviorial science in the wild draws much of its claim to scientific legitimacy from work done in experimental contexts. Settling aside the ongoing failures to replicate and the blatant frauds, this is hardly a one-to-one match for the wider world. But given experimentation’s position atop the scientific hierarchy, there are strong inducements to treat experimental findings as universal and law-like. The just-so story here is that we have access to universal insights that we only need to apply because it works everywhere, the same every time. But, again, if activities like priming worked in the same way across time and space then why haven’t all actors (people, orgs, states, whatever) converged on this optimal solution? Clearly something else is happening that prevents effective use of these “universally prescriptive” - do this, get that.
Challenge unambiguous causality. Universal claims about the behavior of discrete individuals and the actions that supposedly flow from their head quickly lead to assumptions about how to shape them. Like get rich quick schemes, they offer generic recipes that promise if we follow these steps, we can directly achieve these outcome. Again, this assumption raises the question, “if you’re so smart, then why aren’t you rich?” If causality is readily divined then why do we struggle so mightily to replicate it across contexts? We once more find ourselves circling another just-so story: it is because it is and that’s just how it is
It feels trite, almost silly, writing up a three-step guide to decoding obvious nonsense that reads like advice from your mother when you were a kid. But she was right: if it sounds too good to be true, it probably is.