I love behavioural science. I mean, sunk cost fallacy dictates I must, as a behavioural scientist. However, a good behavioural scientist, or any scientist, is critical. Towards all things that make for good science. Including the science itself.
We’ve had several decades of behavioural science now. We still claim to be a young field, but really, we aren’t anymore. The kid’s shoes no longer fit. Our rebellious phase is also most definitely over. We’re increasingly going mainstream, like a 20-something year old trying to figure out their first steps into a new corporate job.
We try to learn new things and shed the foolishness of the past. So it’s time that we start to figure out what we need to shed, and how we need to pivot, if we want to make ‘our twenties’ a killer decade.
The rationality debate This is essentially what got the field started. The entire debate on whether we were, or were not, rational agents. Hint: we aren’t rational per se, but there’s a rational to our irrationality. The issue with this debate is that it shaped two different streams of thinking: the dual system thinking associated with Kahneman and the bounded rationality associated with Gigerenzer. Now there’s nothing wrong with that. The issue is that it took us decades of ‘in-fighting’ to get there and now one stream of thought is a lot more prominent than the other (dual system holds dominance currently). Is there a point to this? Not really. Was there a point to it? Given that behavioural economics got its traction for almost being anti-economics, meaning anti-rational-choice-agent, sure. Once upon a time. But that time is over. This debate has become academic, because it always was academic. In the literal sense. No one outside of the ivory tower cares about (ir)rationality. They care about behaviour change.
The nudge debate
What’s the first thing that comes to mind when discussing behavioural science? For a lot of people it’s nudging. Now 2022 was not the year for poor old nudge (the concept, not the book, the revised version did very well!). Several papers came out indicating that nudges have a small to no effect whatsoever, and that’s on the short-term (the results on the longer-term have always been questionable). This sparked a debate around the usefulness of nudges to begin with. Which quickly lead to an existential questioning of the entire field of behavioural science. Bloody hell.
Have we learned our lesson in terms of debating useless artefacts? I think so. I think the debate surrounding the ‘death of a nudge’ was had much more effectively and efficiently. First, it produced several good journal articles and popular medium thought pieces. Second, it was kept within its realm of magnitude. With this I mean that the debate on the ‘usefulness of behavioural science as a whole’ got shut down very quickly. Even more thought pieces came out about the much broader methods of behavioural science as a whole, and the many things that fall under its remit. The counter noise being made was good. Almost like a motivational quick up the arse for the field.
Also, the debate sparked a conversation about better methods of nudging, mainly smart nudging, in which data, AI and ML methods are being applied to make nudges much more applied and personalized. And I, quite frankly, cannot wait for that era of nudging!
The aversion to mixed methods
You’d think that as behavioural economics (a subset of the behavioural sciences) was such a counter movement to neoclassical economics which was heavily reliant on mathematical modelling (and little else) would have opened its arms far and wide to capture all research methodologies to get the most accurate picture of what goes on in the human condition. You’d think so.
But if you had thought so, you would’ve thought wrong.
I personally don’t know and struggle to understand what the issue has been for behavioural science to adopt the qualitative methods. Because we haven’t prioritized this at all. Actually, we have done quite the opposite, trying to exclude them for as long as possible and putting them in a lonely dark corner with the argument that they’re ‘not that scientific’.
Is there a difference between observing someone in real life versus observing hundreds of data points about them through a third party data collector (such as a bank)? Of course. Is there a difference between conducting large scale quantitative surveys or sitting down with a single person at their kitchen table and asking them about their financial situation. Naturally. Does one have more methodological problems than the other? No.
Each of these methods can come with a plethora of advantages, and an equally large plethora of limitations and concerns. Sure, ethnographers and anthropologists can be biased in their observations, notes and recollections. But surveys and surveillance systems are biased in how they’re built, because humans built them. Sure, it is difficult to scale individual qualitative interviews, but running large scale surveys often misses a depth and due its generality often misses important points, thereby rendering the data incomplete. Somehow, we have become very comfortable with the limitations of one, but not the other. Thereby limiting the level of depth of understanding that we need to truly drive behavioural change.
Almost ironically, everything that continues to fit our obsession with modelling, predicting and affecting behaviour at scale, such as data science, AI and ML get given all the time of day. But these methods have issues too, that we are somehow blind too.
Now picture this: the wealth of information that you get if you apply all these methods at once. I personally think the interaction between qualitative data and text analytics is going to be amazing. Who’s with me?!
The WEIRDness
Last but never least; the focus area of behavioural science has been the Western world for far too long. Similar to the prior issue discussed, this is largely driven by the incentives in the academic system. If you’re not running econometric models you need to explain yourself. If you’re not testing on white Western samples, you need to explain yourself.
It is a very clear sign of (neo)colonialism that white and Western continues to be heralded as the default. We should not be. The default should be human. This out to capture all aspects of the human condition, whether they are privileged, or not.
Amazing work has already been done in understanding samples that don’t come from WEIRD (Western, Educated, Industrialised, Rich and Democratic) countries. I am a massive fan of Cristina Bicchieri who’s work on driving behaviour change through understanding social norms in a variety of cultures is nothing short of revolutionary. I think all behavioural scientists need to better understand the work being undertaken by Nudge units that aren’t based in the West (a great example in this area is Fadi Makki, who founded Nudge Lebanon and was inaugural head of the first behavioral insights and nudge unit in the Middle East – B4Development (formerly Qatar Behavioural Insights Unit)).
As it happens, these tend to be two very successful and well published behavioural scientists. But there are so many more. There is so much non-WEIRD work to still be done and rather than having to explain why your sample isn’t white, I think it’s time that we start arguing why the world needs another academic paper tested on white people…
As this article is getting longer and longer I thought I’d leave it there. I’m sure there’s many more things in behavioural science that make little to no sense. What do you think they are?
