Hello lovely people and welcome to this blog in 2020! For the blog 2019 was a pretty good year! We have seen steady growth, especially with the introduction of the Interview series, where I ask prominent members in the field of (or adjacent to) Behavioural Science seven questions. One of those questions is: “How do you think behavioural science will develop (in the next 10 years)?” And we have seen very different answers to that question.
As we closed down 2019, I had uploaded 20 interviews (don’t worry, there’s more to come!). But those interviews revealed a lot: there’s many different ways behavioural science could go. What did they all think? And where do you think we might end up?
Increased Popularity
The interview with Christian Hunt really did give me hope. According to him, Behavioural Science will become a skill that is far more widely deployed across organisations than it is now. He argues it to be a necessity to be effective in the knowledge economy. He does put an emphasis on data: “we have unprecedented knowledge about what drives human decision-making and the data to support that understanding. Yet we’re still some way off using that knowledge to solve the obvious problems we are facing. I’m really excited about what is yet to come. I think we’re really starting to discover exciting things about how we think. But we’ve got a long way to go.”
Jez Groom has mentioned a similar position within the industry: “We're driving an agenda to make Behavioural Science more mainstream in business and building up capability in some of the biggest organisations in the world. In time, a lot of the smaller problems will be solved by people in these organisations and not specialist behavioural scientists.” Well, as a specialist myself, I’m not too sure I’m keen on being replaced, but I’ll take the increased popularity of behavioural science in business!
Ralph Hertwig remains slightly concerned. He mainly hopes that “the current interest in behavioural science evidence for public policy making will prove not to be a fad, but will be sustained far into the future.” He believes that behavioural science evidence will be needed to address many of the great problems of our time—including climate change, polarization, obesity, and misinformation, to name just a few. Moreover, he wants behavioural science to get the chance to dive much deeper: into bounded rationality, because it is so misunderstood, and the massive role the environment (rather than just cognition alone) plays in this. Daniel P. Egan is also more concerned. He thinks we’ll reach a saturation of ‘cool stories’ and organizations will want to see impact and value delivered, so cool results that won't reproduce will have to fade out from the limelight. On a more positive note, he does think that we’ll see more and more ‘applied’ examples, though they’ll be smaller and more focused than some of the extant examples.
The Machines are Taking Over!
A lot of my interviewees have mentioned, and put a clear emphasis on the availability of data and what we can draw from this as a resource. According to Neela Saldanha, there are very interesting interactions between data and behavioral science and she thinks that in the future nudges will be much more segmented than they are now. Data will tell us where to play and behavioral science, how to play. Well said!
Ganna Progrebna is also looking towards data and machine learning for the next development, her being a behavioural data scientist, she has an insight that most of us do not: “In my opinion, Behavioural Data Science and Behavioural AI are the future. There are many important questions to answer. How do we increase human wellbeing at scale? How do we use technology to decrease inequality? How do we deliver personalized goods and services to customers, and yet use their personal data in responsible ways? How do we equip people with necessary skills to recognize cyberattacks? I also hope that behavioural science of the future will concentrate on making sure that technology and people can co-exist in harmony.”
Dan Goldstein is also optimistic when it comes to data: He thinks we will have greatly reduced the publication of false positives and will have a much better sense of the size of effects. Moreover, he argues that the predictive quality of psychological models will also be improved because we have more data, more data sharing, and better model selection practices. Fair play to you Dan!
Chris Starmer (STARMAN!) is the only one that acknowledges data and its possibilities, but dares mentioning experimentation as well: “my impression is that experimentation has now become well established as part of the economics profession’s tool kit and I think that’s here to stay. What’s less clear is how ‘behavioural’ approaches will develop (or stagnate or decay) within theoretical economics – time will tell.” He is hoping that the field of economics will open up to broader ideas, but that will be a slow process.
But Data continues to be a big driver. Andrew Oswald think this will have the biggest impact on the development of behavioural science as a field. But he also mentions the merger of behavioural science into the natural science, something Nattavudh Powdthavee thinks is likely to happen as well: “I believe that researchers in our field will start to collaborate more with researchers from the natural sciences in an attempt to solve many of the world’s problems that today might seem unsolvable — like climate change, for example.”
Graham Loomes also looks towards data: “One fairly safe prediction, I think, is that ‘big data’ will feature increasingly prominently (although I fear that much of the really interesting and useful stuff will be difficult to access and then publish, due to its commercial value to the companies that own it). I should also like to see some part of our massive and ever-increasing computing power being harnessed to build richer, more complex (and thereby hopefully more realistic) models of the processes by which individuals and groups arrive at decisions and to explore (via simulations) the imprecise and probabilistic nature of people’s preferences, judgments, choices and values.”
Gordon Brown (not the politican!) agrees with Graham, he too hopes for better modelling: “I think we will see more and better models of the individual-level psychological effects of inequality. I hope we will see models that go beyond the “critiquing and tweaking” approach; there is still a tendency to identify problems with the “standard model” and then modify that model in minimal ways, rather than looking for a positive alternative approach. I predict that economic models of identity will come much more to the foreground.”
Fading Away ?
George Loewenstein puts forward that some of his colleagues also think that behavioural science (or economics) is likely to fade back into economics, and that this would be a success story for the field. However, he himself doesn’t buy into this. According to him we have barely begun to scratch the surface, and will spend many more decades (or centuries?) diving in deep into the human mind. I sure hope so George!
George’s colleagues aren’t the only ones mentioning the merging of behavioural back into neoclassical economics, Koen Smets has also mentioned this point, quoting Richard Thaler: “I think behavioural economics as a discipline is at a point where it could go into different directions. Richard Thaler has said it will eventually join (or rejoin, depending on your viewpoint) with mainstream economics, making it just as behavioural as it needs to be. I can see that logic, but I am not sure it really will happen. I think it will continue to exist in its own right, connecting economics with other social science disciplines like psychology, sociology and anthropology. I think we will also see more detail and nuance emerge in observed effects. Rather than simple, binary "doesn't replicate" or "does replicate", we will see what works (or doesn't work) in which contexts, with which demographics, in which cultures and so on. We may evolve from today's broad brush insights to a much more granular view.”
Magda Osman has ideas similar to those of Koen, she suspects that, whatever combination of disciplines will form, they will form under a banner that won’t be referred to as Behavioural science. Whatever shape it will take, it will reflect the new demands that society and technology place on the research community to respond to. Magda takes a very practical viewpoint to behavioural science, one which I think fits those of both Daniel P. Egan and Jez very well.
Daniel Read – ever the contrarian – has even more shocking ideas for the field. Indeed, in his case an entire overhaul might be due soon! He, in contrast to Ralph Hertwig, wants to get out of studying rationality and irrationality all together: “Irrationality was a useful concept to get the field moving, but it is impossible to say that a given behaviour is irrational except by invoking idiosyncratic definitions of irrationality. We learn nothing from the term, and it distracts us by making us think that a kind of “irrationality” is a behaviour or bias, when it is in fact only an interaction between some behaviours and some arbitrary normative standards.”
Moreover, he thinks many categories of study will vanish within behavioural science will vanish as we get a better understanding of what some of these concepts mean. He refers to one of his own fields, intertemporal choice, as well: “I predict that very soon we will recognise that time is just one of countless attributes that influence choice and that time is no different than all those other attributes. It is not even clear that “time” is one concept.”
Lastly, Daniel mentions the revolution that has come from p-hacking and the replication crisis: “The revolution is, I think, more devastating than we have yet realised. We have consistently oversold our research, drawing conclusions based on small samples and ungeneralisable designs, we have kept almost countless “uninteresting” failures to replicate in the file drawer, and we have engaged in a great deal of p-hacking without even realising that is what we were doing. My prediction is that when the dust settles, we will have far fewer phenomena in behavioural economics, and these will tend to be phenomena that match our intuitions. The “sexy” counterintuitive findings that filled the journals for a while (I think since the 1960s) will start to wither away -- indeed that is already happening.”
And as hardcore as Daniel makes all of this sound, I can’t help but think this is the healthy way for the field to go. I too, fail to see much interest, or even practical applicability in the rationality debate. Moroever, small samples and bad statistics, with the technology and data we have available should be a thing of the past. They are just bad science now.
Ironically, Daniel Read’s colleague, Nick Chater, thinks the debate on rationality might continue just fine: “I wonder, too, whether we'll start to think about rationality in more social terms---the idealization of the individual as a lone theorist, attempting to predict and control the world, often seems unhelpful, and perhaps we can do better.” So we might not be leaving rationality behind after all! More up my street, Nick hopes that richer models of how people interact (the hidden promises, contracts, threats, and so on, that underpin social life) will increasingly be a focus.
Let’s end on a beautiful note: take it away Bob Sugden: “Economics is very fashion-driven, with a short life-cycle for new approaches. The history of economics suggests to me that behavioural economics is now at about the peak of the cycle. As such, behavioural economics being replaced by something else is the fashionable thing for economists to do.” Thanks Bob…
That was definitely not the note I wanted to end on.
Let’s look at another titan then: Dilip Soman! What does he have to say about behavioural science, are we going to fade away?! “I do expect three things to happen. First, in the theoretical realm, I expect we will see a lot more work on the marketplace effects of human behaviours, not just empirical demonstrations of "humanness." Second, on the applied realm, I suspect we will see more of a convergence between our field and other, related applied fields (design, human-computer interaction, machine learning). Finally, I hope we get to a place where we won't need to repeatedly keep using the "behavioural" term in our field; that people realize the centrality of human behaviour in developing both economic theory and practice.”
We will fade away into becoming the default. Now that is an ending folks!
I know, I KNOW. This was a lengthy article! I thought it would be good to have all our findings so far summarised, as sometimes it is difficult to see where 20 separate articles might take you. The take way message? Predictions are hard. Most interviewees carefully mentioned as much, and told me to not put too much stock with their predictions, but I just think it makes for an interesting debate.
Having read all of this, where do you think behavioural science is going? Are we going to focus more on data, and will this method lead us to the promised land? Will rationality still be a topic of interest? Is behavioural science simply a fad or do we think it is here to stay? What will the business sector do with it all? So many questions! Please do let me know what you think!
Quick administrative note: now that we are in the New Year and the break is effectively over, the normal publishing schedule will resume: articles will be published on Mondays, Wednesdays and Fridays, at 8 AM GMT.
Thanks for all the support throughout 2019, it means a lot (much more than you know)!
Comments