Behaviour change 101 series: Five steps to select the right behaviour/s to target

By Research Fellow, Peter Slattery

This is part of a series ‘Behaviour change 101’ where we will draw on our educational materials to approaches that can be used to plan and execute behaviour change research and projects.

Take the example of a company wanting to encourage its employees to exercise more to improve their health. They come to us asking how they should go about doing it. They realise it’s not a particularly prescriptive goal; that a broad range of behavioural changes lead to employees ‘exercising more’. They also realise that some behaviours will be more efficient to target than others.

To help the company succeed, we need to assist them to go from, i) having a broad and abstract understanding of the outcomes they desire to, ii) having a clear set of specific behaviours they should target (and later, interventions to use).

This article provides five steps for identifying ideal target behaviour(s) based on the Exploration phase of the BehaviourWorks Method.

Readers wanting to apply the following tips might not need to consider all of the steps; sometimes you might want to add a few more and sometimes you might want to do them in a different order.

Whatever way you go about it, it will nearly always be important to consider at least a few of them.

When faced with a complex issue, there can be considerable value in doing a system map. This involves systems thinking, a process that aims to consider the entirety of an issue, such as the different parts and players involved and the interrelationships between them.

A variety of different tools are used for systems thinking; most of which attempt to map the relevant actors, influences and relationships.

BWA often combines actor mapping with influence diagrams to identify and understand target audiences.

For example, we have done this with clients for gambling and waste management behaviours.

In the context of exercise behaviour, a system map might highlight how, for example, exercise behaviour is influenced by home, work and social environments, competing demands on time, flexible work arrangements and the demands of workloads, school schedules and children.

Different audiences engage in the same types of behaviour to differing extents, at different times, and face different drivers and barriers when doing so. It is therefore important to develop a detailed definition of the target audience beyond generic terms such as ‘employees’, ‘community’, ‘householders’, ‘employees’, ‘students’, etc.

Clearly identifying the target audience makes it more likely that you will pick the behaviour that is optimal for the people you target, rather than one that is optimal for a more generic group. For example, the optimal behaviour to target for encouraging older male employees to exercise might be very different from what is optimal for younger female employees.

Additionally, picking the correct audience ensures that resources will be used more efficiently and increases the probability of having a positive impact.

Rather than waste your resources targeting something that your audience is unlikely to do, you can instead channel that effort into promoting behaviour that is more likely to be adopted.

It is unlikely you’ll be the first person wanting to change a particular behaviour, especially if that behaviour is a common one like exercise. It is therefore useful to search for answers in relevant academic and practitioner (grey) literature.

The literature might help you identify useful theories, or explanations of the problem, as well as some influences that might be relevant to your chosen audience.

It might also help you to better understand the impact of certain exercise behaviours; for instance, whether one type of exercise is likely to stick or have a positive impact on health.

Depending on the time and resources available, there are several approaches to consider: exhaustive, and comprehensive systematic reviews, rapid reviews (which focus on already synthesised research evidence or high-quality primary studies) and/or unstructured explorations using tools like Google Scholar.

Practice reviews, which involve interviews or other types of engagement with experts working in the field, can also be useful, particularly in cases where there is a lack of relevant literature. A practice review can also be combined with a rapid review of the literature to gain multiple perspectives on the problem.

A better idea of the specific audience and relevant evidence will usually lead to the identification of a range of potentially-relevant behaviours.

It is important that these are clearly specified. At BehaviourWorks, we use an approach (based on work by Fishbein & Ajzen) where we define behaviour by considering who should perform an action, directed at a target, in a given context, at a certain point of time.

In the case of our example problem, we might define the following as our target behaviour:

• Employees (who) exercise (action) at the gym (target) in the office building (context) before the start of each working day (time).

Depending on the segment of our audience and our need/ability to tailor interventions, we could further refine this target behaviour. For example, we might want to tailor to different target audiences or focus on different exercise actions or locations, such as:

• Employees over 50 (who) participate (action) in a gym class (target) at their local gym (context) three times a week after work (time).
• Employees under 50 (who) participate (action) in team sports (target) at their local facilities (context) three times a week after work (time).
• Employees (who) workout for at least 15 minutes (action) at the gym (target) at work (context) three times a week after work (time).
• Employees (who) workout for at least 45 minutes (action) at the gym (target) in their local area (context) three times a week after work (time).

Often there will be value in using a prioritisation process to identify the most efficient and impactful behaviour(s) to focus on for the intervention.

At BehaviourWorks, we often prioritise behaviours using the Impact-Likelihood Matrix (figure below).

In this approach, behaviours are prioritised by mapping them based on:

1. The impact they have on the problem they are intended to address.
2. The likelihood of the target audience adopting the behaviour.

In one application of the approach, the specific behaviours of interest are evaluated by rating them out of 5 for likely impact and adoption.

When mapped to the matrix the result for each will fall into one of four quadrants, as shown below. See here for a published example.

A range of other factors might also be considered during prioritisation. For example, we might consider the scope for change, the political will or the number of resources required for each unit of change.

We might also need to consider the feasibility of measuring, i) changes in behaviour and, ii) the effect of these behaviour changes on important outcomes (e.g., health).

In short, when trying to choose a behaviour, we recommend five steps:

1. Understand the big picture
2. Identify the target audience(s)
3. Understand the evidence
4. Identify relevant behaviour(s)
5. Choose behaviour(s) to focus on

Of course, there’s a lot more to be considered than we have covered here, and it is usually possible – and desirable – to vary the process and to go into greater depth than discussed.

In our training, we cover each of these steps in more detail and discuss a wide range of additional considerations.

In a future post, we will cover steps for selecting the right intervention.

In the meantime, please let us know if you do anything differently and/or if we can help you address your particular problem. Email: Peter.slattery@monash.edu