What influences behaviour?
Authored by Peter Slattery and Jim Curtis (our environmental portfolio lead)
Throughout 2021 BehaviourWorks Australia (BWA) is publishing a book to help policymakers and program managers use tools within our ‘Method’ to design and deliver more effective behaviour change programs. We are releasing one free chapter each month.
In Chapter 6 (published 1 June 2021), we discussed influences on behaviour. In this post we are going to revisit some of the content from the chapter and explore other useful approaches to understanding the influences on behaviour.
What is an influence on ‘behaviour’?
From our perspective an influence on behaviour is anything that affects whether a behaviour occurs. That’s a pretty broad definition but it’s also about as precise as you can get!
As Robert Sapolsky points out in his fantastic book, Behave, behavioural influences range from changes in our brains immediately before we act all the way back to evolutionary factors from centuries before we were born. Similarly, in health the biopsychosocial model suggests that behaviour has biological, psychological and social influences. That’s a lot of influences!
Both here and in our work we will mainly focus on the influences on behaviour that are relatively easy to change and that occur soon before the behaviour. We usually can’t do anything about our audience’s brains or evolutionary traits (and the FBI also forbids us to discuss our time travel work) so there’s not as much point in discussing those here.
Check out Behave if you’d like to learn more about them.
Why categorise influences on behaviour?
We categorise influences on behaviours because categorisation improves our communication and coordination. For instance, it is faster and easier to establish that a set of influences have a shared characteristic than to discuss them all individually.
Once there is such a shared understanding, these categories can then be used to improve communication: you can say things ‘smoking is mainly driven by social factors’ rather than listing a large set of related social influences.
Categories of influences can be used as diagnostic tools (i.e., to see what’s wrong). For example, we could use the categories from the biopsychosocial model (biological, psychological and social) to identify types of barriers that impede desired behavioural outcomes (e.g., not smoking) for a relevant group.
Categories of influence factors help us to explain results more clearly. For instance, if you might categorize interventions by ‘who did, what, where and when to whom’ you can clearly better how explain the precise interventions which have addressed smoking.
Categorisations also enable us to better segment and compare influence factors and combinations of factors. For instance, a future researcher might use your categories to compare the effects of government driven ‘social’, ‘packaging’ and ‘price’ focused interventions on the behaviour of ‘smoking’.
How do BWA categorise influences on behaviour?
In Chapter six, Jim discusses several approaches for categorising behaviour while explaining his journey towards behaviour change enlightenment. Let’s recap some of these.
Robert Cialdini’s principles of persuasion are amongst the most widely known and used categories of influences. Though very useful to know and widely used, these are all related to social influence and thus overlook a range of other influences.
Andrew Darnton’s synthesis of behavioural influences
Jim then discusses the role of other influences, including habits (noting that research suggests that it takes 66 days to form one), emotions, unconscious biases and context.
This all leads into a discussion of a synthesis of behavioural influences by Andrew Darnton and the UK’s Government Social Research Service (Darnton, 2008) which reviewed over 60 models of behaviour (there are now at least 80!) to suggest the following eight categories of influence factors.
As Jim mentions, this model has been particularly influential at BWA and informs a lot of our research and training, much of which Jim has developed and delivered.
What other approaches are useful?
The models discussed in Chapter 6 are some of several approaches which we BWA rely on and just a few of the many the approaches used across the behaviour science community. The following are a just a few of these. I won’t even get close to covering all the options (there are at least 80, as I mentioned), so please just see these as a taste of what’s out there.
If you want to see more then check out Spark Wave who have a very nice review of many models here (I discussed their model in an earlier post). The Behaviour Institute also discuss several different approaches.
COM-B is one of the most rigorous and well researched categorisations of behavioural influences. It argues that there are three main categories of behavioural influences: capability, motivation and opportunity. The supporting theoretical domains framework (TDF) also outlines many secondary categories that sit within the larger factors (e.g., that knowledge drives ability).
The ISM was created by Andrew Darnton and his colleagues at the University of Manchester, and is used by Scottish Government. It categorises influence factors based on whether they are inside of the individual, in their social environment, or about material factors that influence them.
Within the persuasion literature, behavioral influences are often categorized by the ‘source, content, channel, receiver, outcome and effect’ or ‘who did, what, where and when, to whom’. Processing and context factors are also considered.
Which categorisation of behavioural influences is the best?
There is no single best behavioural influences framework. What’s best depends, for instance, on the context, relevant factors at play (e.g., cost, time, optics), time available, and, of course, personal preference and familiarity.
For instance, some models are simpler than others. I find that core categorisation of influences in the FBM (ability, motivation and prompts) are particularly easy to understand and explain. I therefore find that non-experts understand the model more quickly than alternatives such as COM-B .
This simplicity means that FBM has worked well as a diagnostic tool in the SCRUB project where we needed people to explain their own barriers to action, and in quicker engagements, such as our design sprints with practitioners.
However, the FBM approach lacks the depth of academic support and theory of something like COM-B. I generally find COM-B better to use if I have more time, want to build on prior work, publish the findings, or work alongside academic researchers and experts.
How might you use categorisation of behavioural influences?
In our work on Covid-19 related behaviour we used a mixture of Fogg, COM-B and other models in our survey to diagnose which behavioural influences were driving relevant health behaviours. This information was then used to inform priorities for future interventions. See below for one of these outputs (taken from this post)
For an example of a similar diagnosis approach that uses the COM-B categories, see this COM-B Hand Hygiene Behaviour Questionnaire).
Using categorisations of behavioural influences to order outputs
In this anonymised example, BWA used categories from the COM-B theoretical domains framework to summarise evidence for encouraging a particular type of behaviour.
In this post we have explained what a behavioural influence is, why we categorise them, some approaches to do it and some examples of how to use those approaches. Hopefully it was useful!
In Chapter 7 of our book we will discuss audience research methods. You can read about that here. You can find our previous chapters on our website.