
Dr Michael Hallsworth
Chief Behavioural Scientist
This is the second piece in our series explaining the four big shifts in behavioural science that BIT is aiming to achieve.
In 2023, BIT published a manifesto setting out ten proposals for the future of behavioural science. Now, BIT and Nesta are setting out how we can advance these proposals through our own work. The second ‘big shift’ we want to share is:
“Be more attentive to the social drivers of behaviour. Seek to understand how cultures, identities and relationships underpin behaviour.”
Applied behavioural science has often drawn heavily from social psychology: how people perceive and interact with each other. For example, the need to make a behavior ‘social’ is one of the four main principles in our EAST framework. But I want to show you why we need to go further and draw in new insights about how behaviour is strongly shaped by social factors.
Consider a famous explanation from Robert Cialdini of why people litter. It goes like this. When a person sees someone else littering in a messy environment, that gives information about the ‘descriptive social norm’ – many other people are doing it, so it’s OK. When they see litter dropped in a clean place, it has the opposite effect, because it triggers an ‘injunctive social norm’ instead.
This view treats social norms primarily as data that individuals process – offering what Cialdini explicitly calls an ‘information-processing advantage’ or ‘decisional shortcut’. It’s a wonderful insight, but it’s also pretty ‘narrow’: it leaves important social aspects of littering unexplored.
For example, who is the litterer – a local resident, visitor, or authority figure – and what is their relationship to the observer(s), if there are any? How do power dynamics and social status influence the act, the interpretation and the response? In some contexts, littering might be seen as a protest, while in others it might signal community disconnection. How do cultural practices shape views of littering? For example, in one study, religious convictions emerged as a strong factor.
These kinds of questions have led critics to claim that applied behavioural science focuses too much on individuals as the unit of analysis and cognition as a driver of their behaviour. Often these critics are attacking straw men, but overall there is a need to broaden and strengthen the social dimension of applied behavioural science. Otherwise, it will increasingly fail to grasp three main issues relating to knowledge, values and processes.
The first risk is that behavioural scientists don’t register the knowledge or expectations that people have about others in their societies.
Take the famous marshmallow test. Waiting a few minutes to get two marshmallows instead of one only makes sense if you are sure that the person making the promise will follow through. If they’re from a group you mistrust (perhaps because of past experience), you may ‘fail’ the test for good reasons. Levels of trust and expectations of reciprocity are an essential factor to consider.
To keep with the littering example, you can imagine a society where there is a strong inbuilt expectation that people pick up the litter of others. Realising that people leaving litter will have cleaned up after others in the past would completely change the analysis. If you don’t grasp the relevant social expectations, you may design institutions that look good to you but which fail to work.
The second issue is not understanding the values that societies or groups attach to behaviours. Alcohol consumption offers a simple example. A narrow view might frame drinking as a health issue, wherein people downplay future health costs for present pleasure. But that would neglect how people see alcohol bound up in their social identity and social support networks; it would fail to see the social value that people assign and derive from drinking.
Instead, some argue for a greater focus on how people’s social identity shapes their values and, in turn, their actions. Recognising the role of ‘identity-affirming behaviour’ can provide a crucial insight, and a few simple steps can get us closer to this perspective. For example, does the observer socially identify with the person littering, or do they reject them as different? A range of studies show that this factor can lead to completely opposing reactions.
Yet the implications go beyond simple fixes. They can require a full reassessment of ‘biases’ and ‘errors’. If you are making an ‘error’ that is approved of by a social group with which you identify strongly, is that an error at all? It may actually be ‘reputationally rational’.
The final issue is how this knowledge and these values are created and maintained through groups and cultures. The risk is that a narrow view neglects how social factors shape the ways we perceive, prefer and choose.
For example, how do people see some items as ‘litter’ in the first place? If a resident leaves a useful item like a wardrobe outside their property, people might see it as an act of charity rather than littering. Leaving a damaged or less valuable item may edge closer to the category of ‘litter’ – as will the wardrobe, if it stays there for a certain length of time.
Society strongly shapes the way we make these judgements about categories and acceptability. Cultures create the mental models that we use to understand the world:
“In a fundamental sense, what we perceive is itself socially determined… Not only does ‘society’ create individuals, it even creates the categories that one uses to think about oneself and others.”
If cultures play this formative role in cognition, and cultures vary, then obviously the idea of ‘universal’ heuristics and biases is in trouble. Over the last decade, that insight has gained enough traction to start reshaping the field. Now we have studies that examine if and how ideas like prospect theory or temporal discounting vary between cultures. And, I think, they are showing us where to go.
You might be thinking, with frustration, isn’t this obvious? People have always identified with social groups. It can only be news if you started with a very thin view of human nature in the first place. Or maybe, with a wry smile, so behavioural scientists have discovered sociology/anthropology/cultural psychology.
Maybe you’d have a point. But my hope is that behavioural science can approach these social issues in a way that creates something new and useful. First, rather than discarding the idea of universal cognitive processes, there’s an opportunity to map how they are shaped by practices in different groups and cultures.
Take the endowment effect. The first, famous study showed that when students from the US were given a mug, they would only sell it for about twice the amount of money they would pay to buy it in the first place.
Now, behavioural scientists should also be thinking about how cultural concepts of property and possession influence how the effect plays out. In some societies, ‘ownership’ may be a less powerful motivating factor. And, in turn, the endowment effect may shape the cultural practices of countries like the US and elsewhere.
Understanding how this two-way exchange happens should be a priority. An obvious starting point is methods: doubling down on existing ways of exploring the context; using ethnography to uncover implicit cultural norms; and borrowing concepts for explaining culture from other fields.
A change of mindset is also needed. There’s an implicit assumption that culture is somehow secondary, just a way that the main factor – say, the endowment effect – is realised. That mental model of dominance should be replaced by one of equal exchange. As we say in our manifesto, people’s experiences of culture and group identity profoundly influence the way that social norms function, rather than being just interesting variations on a central concept that remains unaffected.
Second, we can take these enriched concepts and analyse how they translate into behaviour – or not. One influential view of culture is that it influences action ‘by shaping a repertoire or “tool-kit” of habits, skills, and styles’.
As we saw with Cialdini and littering, the traditional behavioural science approach is to study how parts of the ‘tool-kit’ of heuristics are triggered by the immediate context. That approach could be adapted to discover how societies make some parts of the cultural ‘toolkit’ more prominent on a larger scale (e.g. social norms of consumption, like buying cards on Valentine’s Day), as well as the contextual cues that mean these ideas result in behaviour (e.g. why someone buys this particular card now).
Developing this kind of insight could be a unique selling point for behavioural science. It could create wholly new explanations of how individuals and societal structures influence each other. That’s a much better idea than dividing the two levels up into a separate ‘i-frame and s-frame’.
Making this happen will require a few things. Behavioural scientists will need to be prompted to widen their lens away from the immediate context they are looking at – and consider these underlying cultural factors as well. Participatory approaches like co-design could uncover what elements of the cultural toolkit seem important to people at different points. And experiments can test how various stories or frames can activate various parts of that toolkit – as we started to do with our work on cultural narratives, mentioned below.
We’re already working to make behavioural science more socially aware:
The shift to a more social form of behavioural science brings tangible benefits. It doesn’t just open up new routes to change, but also increases the chances that a change endures.
One person’s change is more likely to stick if it is reflected and supported by their social groups, institutions and cultures. It also acts as a crucial prompt for behavioural scientists to reflect on how their own social norms and connections influence their judgements.
While these changes may seem like breaking new ground, they also mark a return to some of the origins of behavioural science. In 2004, a report that informed the creation of BIT stressed the importance of taking an ‘ecological approach’ that considers different levels of change.
Further back, pioneers like Herbert Simon envisioned an approach to understanding behaviour that fully integrated the role of social influences. I think we are much better placed now to make that vision a reality.
Chief Behavioural Scientist
Get in touch to see how our deep understanding of human behaviour and evidence-led problem solving can help you.
Get in touch