Long gone is the notion that Healthcare Professionals (HCPs) are simply robotic assessors of benefit and risk. Likewise, behavioural science isn’t the buzzword it was 10 years ago. Today, behavioural science is prominently discussed and factored into healthcare consultancy, yet many automatically associate it with qualitative research; with the richer, more agile methodology where we can get accurate answers and probe around the biases and cognitions that truly drive attitude and behaviour. This blog post seeks to offer an additional solution. Here are our guiding principles for building behavioural science models and thinking into quantitative research.
1. Gain buy-in by providing a flexible structure
Planning and preparation are vital to the successful deployment of behavioural science in any research methodology, but especially in quantitative. With even less wiggle room to manoeuvre in the field there are two fundamental purposes for the set-up: expectation setting and utilising a critical resource to guide the implementation.
So how can we establish credibility with clients and give the research structure and something to build around?
One answer is through the integration of behavioural science models. There are many models available, each with a specific relevance and suitability. By selecting a model suitable to your business need, you should have a good guideline to think about behaviour change from the outset. Whether it’s COM-B, ABC, Health Action Process Approach, or another model, putting in place a guideline of change will give you the building blocks you need to justify the inclusion of exploratory questions and guide which topics are likely to be salient and insightful. Once the model is in place, you can ask questions that directly speak to it. Take COM-B for example; by programming questions that uncover the extent of respondents’ capabilities and opportunities, we can then understand the relationship that both aspects have on motivation, and then in turn, behaviour. This process allows us to identify the influence of each behavioural factor in a quantitative manner.
One final point: don’t be afraid to change the model you are working with.
If your model isn’t quite matching the content you’ve arrived at with your client, then detour before survey design. These models aren’t one size fits all, you should choose the one that best fits your business issue and your understanding of the scope of the intended behavioural shift, designing your questionnaire around that. At this point, we think it’s important to note that this methodology isn’t suitable for every project; utilisation of behavioural science in quantitative needs both client buy-in and substantial knowledge to guide design and reap successful benefits.
2. Behaviour change experts are agile, make things simple and are deeply inquisitive
Once your model is in place, it is time to think about cognitive biases. Essentially, cognitive biases are systematic errors in decision making which arise from shortcuts that cut down our cognitive processing to make our lives easier. As these biases are automatic rather than rational, they can wreak havoc on rational models of expected behaviour. As there are well over 180 documented biases, there can be very extensive interference with our lovely models of change. Our advice is simple when it comes to these biases: keep it simple but agile. 180 biases are not a realistic number to discuss with your client or consider when analysing data. Frameworks of behaviour change attempt to summarise human decision making into memorable chunks. Consider the mnemonic MINDSPACE (Messenger, Incentive, Norms, Defaults, Salience, Priming, Affect, Commitments, Ego) which helps focus on areas of potential influence. Incorporating a framework like this into your thinking allows you to identify a smaller number of biases that are likely to be relevant in this situation.
Your client has a wealth of information on the behaviour of their customers – utilise this. You, as behaviour change experts, can present several biases that may be at play. Your clients can then help you whittle these down to a sensible list which can then be analysed within your data set. Be guided by the frameworks in place, but don’t be a prisoner to them.
As cognitive biases are unconscious, they can be tricky to reveal directly. However, we can ask questions to reveal the influence of different biases. Take the first two letters of MINDSPACE as an example. For ‘Messenger’ we can explore the influence on credibility or believability if information comes from one source or another. For ‘Incentive’ we can explore the extent of additional benefit required to drive change. Essentially, we can uncover these potential biases through a greater understanding of our respondents’ worldviews.
3. Detect, diagnose and plan for the impact of cognitive biases
Now we have a list of potentially relevant biases to watch for, the next step is to figure out where these biases may be having a substantial impact. Here, we propose an alarm system to detect unwelcome cognitive biases. By using a MaxDiff approach we can tripwire our surveys, showing asset features in both a positive and negative light (the same data but flipped). If decision making were entirely rational, we could assume that there would be no difference in the perceived importance of these two data points.
Please see the example below:
- 18% of respondents experienced headaches in a month
- 82% of respondents were headache free in a month
- 5% of patients relapsed and were discontinued on their medication
- 95% of patients completed the course of medication with no efficacy concerns
We suggest half your survey respondents would see the positive framing, and half would see the negative, allowing you to identify the delta between these points. From this we can deduce the extent to which the identified barrier (e.g. 18% have experienced headaches) is open to framing. The larger the extent of the cognitive bias you reveal, the more opportunity there is for framing. Referring to your list you can then begin to hypothesise with the help of your client which bias(es) may be at play and begin to devise solutions to combat them. At the very least, you should have a better idea of what you’re up against.
To summarise, we believe the application of behavioural science within quantitative research provides tangible numbers and statistics around behaviour change and biases among a broad population. To achieve this, we propose the following steps:
- Do most of your thinking upfront. Apply a behaviour change model early and refine the process as you go, bringing your client along with you.
- Build insights that are grounded in and leverage the actionability of behaviour science by asking questions that speak directly to your model and analysing data with a clear behaviour change focus.
- Consider cognitive biases that fit for this situation specifically. Don’t attempt to fit the same cognitive biases into the same solution that worked last time.
- Finally, help your client achieve maximum actionability by outlining which elements of a product or a communication is creating the most interference.
To discuss in more detail, get in contact with Mike Pepp or Phil Barnes.