Nudging is a buzzword that floats around places where consultants or policy-makers can be found. In their mouths it refers to the act of encouraging some meaningful change in behaviour by making a small change to... you know... something or other. And then everyone kind of trails off.
This laissez-faire attitude to understanding nudges hasn't done much to curb the enthusiasm around them. In fact, maybe it's the reason the word 'nudging' is still so popular, because frankly, nudges haven't really turned out to be that good.
In a first year psychology course, you'll be taught about them. They'll use the same examples now as they did about a decade ago when I went through the same class. They'll tell you:
- How making people opt out of organ donation rather than opt in when filling out their driver licence paperwork increases the pool of organ donors quite dramatically.
- How putting a fly at the bottom of a urinal makes men more likely to actually pee in the urinal, with amusingly noticeable effects on cleaning bills.
- How, one time, Facebook showed people which of their friends had clicked an 'I voted' button, which made people fractionally more likely to vote.
And that's kind of it, really.
There are other examples of nudging that seem to work, but not many. Few that are tangible enough, or dramatic enough, to include in a first year psychology course certainly.
Earlier this year, the obvious conclusion that follows from my little anecdotal observation was borne out. A review of 'nudging' demonstrated that, fundamentally, there is no real evidence for it. The scientific enthusiasm on the subject has been the product of the fact that academia almost exclusively publishes articles that show an effect, and rarely those that show no effect.
So we see only the occasional outlier cases where nudges appear to work simply because we don't see the hundreds or thousands of times nudges didn't work.
The actual value of nudge theory
I will leave you to read around about the criticisms of nudging. There are plenty of hot-takes on the subject. But I think these criticisms often miss something that really is quite interesting.
You'll notice that, if you've never heard of nudging before, you're probably not much closer to understanding what it is now than before. That's because we haven't yet---and few people seriously do---talked about what lies beneath the concept of nudging: choice architecture.
Choice architecture refers very straightforwardly to the way that choices are presented to people, and how that affects decisions about those choices.
In getting people to 'opt out' rather than 'opt in' to organ donation, we're creating a choice architecture that harnesses our implicit laziness---the so-called 'status quo' bias that means we are much more likely to leave things as they are than change them, particularly when the decisions feel momentous or difficult.
In showing people which of our friends voted, we're creating a choice architecture that harnesses our socialisation---the so-called 'social proof' bias that captures that fact that we tend to do what other people are doing so we don't stand out from the crowd or miss out on something other people appear to think is important.
The word 'nudge' is just a way of referring to a change we have made to the underlying choice architecture that changes how people make their choices.
Perhaps you've observed something a little funny from my examples above. A nudge that rearranges the 'opt in' and 'opt out' boxes on a form is pretty different from a nudge that figures out who your friends are, which of them have voted, when the best time to show this information to you is, and how to change the UI of a program to account for all of this.
Indeed, the 'choice architecture' that underlies these two things are almost incomparably different. One is about straightforward patterns of human behaviour on being presented with a stimulus. One is a complicated network of social connections and the value that they hold for us, along with the various beliefs and attitudes of everyone involved.
But somehow, in reducing them to the words 'choice architecture' and 'nudge', we produce an illusion that they are equatable phenomena. You wouldn't be nearly so confident to teach them to a class as a pair if you referred to them as 'the environment' and 'changes in the environment'. It would be too broad to be interesting or tell you anything about how humans change behaviour. This is sort of the core problem of the discipline.
The distraction of nudge theory
The research into nudging comes more-or-less directly out of the research into human biases. This was the big, sexy contribution of behavioural economics. They deviated from the 'rational actor' model of humans that traditional economists had been working with, recognising instead that we are often irrational. We make decisions according to various biases.
If you go back to my examples, you'll notice that the nudges come with their associated bias. Organ donation with the 'status quo' bias. Social pressure to vote with the 'social proof' bias.
It's really this feature of the enterprise that trips everyone up. In the enthusiasm to document all the biases evident in human behaviour, all we've ended up is a hundred loosely collected biases. As that article writes:
The dominant model of human decision-making across many disciplines ... is the rational-actor model. People make decisions based on their preferences and the constraints that they face. Whether implicitly or explicitly, they typically have the computational power to calculate the best decision and the willpower to carry it out. It’s a fiction but a useful one.
As has become broadly known through the growth of behavioral economics, there are many deviations from this model ... This list of deviations has grown to the extent that if you visit the Wikipedia page ‘List of Cognitive Biases’ you will now see in excess of 200 biases and ‘effects’. These range from the classics described in the seminal papers of Amos Tversky and Daniel Kahneman through to the obscure.
We are still at the collection-of-deviations stage. There are not 200 human biases. There are 200 deviations from the wrong model.
This is a patently unhelpful way to go about thinking about human behaviour, and the changing of it. Human's aren't 'rational' with an enormous list of 'deviations'. It's not exactly clear what the right model is, but as the Wikipedia page begrudgingly observes, "cognitive biases are presumably adaptive", and sometimes1 we just notice when they aren't. And so, bias research has suffered mightily during the replication crisis because listing ever-more-specific deviations from some kind of hypothetical optimal behaviour tells us nothing fundamental about human nature.2 This is no more evident than in the way 'nudging' and 'choice architecture' are usually deployed.
For example, I haven't talked much about my fly in the urinal example. I ask you, what bias does that 'choice architecture' embody? The bias to have fun? Are we nudging fun? Is that a nudge? It's hard to see, really, how this is a helpful way to characterise the situation.
But we don't need to abandon choice architecture. In fact, if we go back to the original authors of the thing, they suggest a much more sensible characterisation.
Everything is choice architecture
We could actually just let Richard Thaler explain this himself:
Who is a choice architect? Everyone in this room is a choice architect. Anyone who designs the environment in which you choose is a choice architect. If you go to a restaurant, there is a menu. Somebody thought about how to structure that menu. In many restaurants you have appetizers, then main courses. In some restaurants the main courses are divided into meat, fish and pasta. In others they are all mixed up. Sometimes they are arranged in order of price. Sometimes there is no apparent order. Everything we know about psychology tells us that all of those things matter. Everything matters.3
In the sciences of mind, the subject of free will is an uncomfortable one, particularly in my field. It really does seem like many, many more of our actions that we are comfortable admitting are the product of the environment. We aren't just independent creatures, but systems that depend crucially on the systems around us.
The right model, when it comes to nudging, is that everything is a choice architecture.4 If we start from here, we realise that the kinds of nudges that people are attracted to are doomed to failure. As I wrote at the start of this article, if we think that 'nudging' is
the act of encouraging some meaningful change in behaviour by making a small change to... you know... something or other
Then we'll obviously fail to encourage something meaningful because our 'small change' has to contend with everything else. It's a rather daunting task.
In particular, we can now explain the most depressing finding from the many government institutes set up to 'nudge' their populations: even successful nudges usually do not encourage lasting behaviour change. A momentary nudge does very little to engage with the choice architectures we face in every other aspect of our lives.
But it also teaches us something else about successful nudges.
Blank slates and community
There are some nudges that stand out. They are, in fact, in the same family as our persistent first-year psychology examples from earlier.
The first is the idea of 'defaults', or our 'status quo' bias from earlier. Making a change to the default option seems to hugely influence the chances that people will adopt the new default. It's no surprise really. We are creatures borne of patterns, down to our very core. It makes sense that we would adhere to a pattern presented to us.
But remember that everything is a choice architecture, not the default. Here, it's the pattern that's the crucial feature. But the pattern we're following is under no obligation to be the one you're appealing to, because everything else is just waiting on the wings. For example, if people don't choose the default the first time, they're much less likely to choose the default in subsequent interactions.
More pointedly, the fact that defaults are the obvious standout makes one rather skeptical about the hopes of policymakers and consultants. Defaults are easiest to build in when creating something, not when changing them. This is not usually what policy-makers and consultants are charged to do.
Not all hope is lost for those poor souls however. There is another standout. The most enduring behaviour change comes from nudges that harness our social identity and the norms that inform it. Attending to the values of communities and the people within it, and using those values to inform individuals about where their actions do and do not match. Or simply tapping in to our deep need for belonging. These are techniques that at least work within the community they are relevant to.5
One is still left a little skeptical about their utility to policy-wonks and strategy gurus. Engaging with communities and individuals at the level of their values doesn't really seem like their strong-suit.
For the rest of us, I'm a little more hopeful. If everything is a choice architecture---if everything matters---we are encouraged to pay far closer attention to the world around us. If defaults are the most influential features of our environment, then we are encouraged to build new things into our lives with care. And if the easiest levers to change this environment are our communities, then we know where to turn first.
200 times... ↩
If you pay attention to this kind of subject, you'll notice that Tversky and Kahneman are really the progenitors of the whole thing. Their 1966 paper kicked it off, and their 2011 book brought it to the people. It's pretty noteable, then, that the domain has proven so shaky that Kahneman himself has had a go at it. ↩
It's probably worth pointing out that this model is actually the premise for the biggest critique of choice architecture. Thaler and his co-author Sunstein, endorse 'libertarian paternalism' for this reason. If everything matters when it comes to the way people make choices, they think we have an obligation to help people make the 'right' choices---the ones we feel are in their best interest. To quote Thaler again "If somebody asks you, what are the directions to get to the museum and you point them in the right direction, you are a paternalist according to us. We're not saying you should go to that museum, but if you want to go to the museum, it's over there, it's not over here.". But of course, as I wrote elsewhere "no one is asking and who is to say his directions are the right ones". ↩
I am failing to find a place to shoehorn this in here, there's a nice example that brought home for me just how ubiquitous 'choice architecture' is, when you look for it. A waiter friend of mine pointed out that she was taught to ask her customers if the food was good when they were chewing to reduce the chance of a complaint. ↩
But, of course, no broader. And as the article I link points out, the closer the 'nudge' is to the individual the more effective. The corollary being, the less close the nudge, the less effective. With a dynamic like that, what community size is really likely to be effective? ↩