Solutions

Behavioural Science

Behavioural science and economics offer an evidence-based approach to applying ‘choice architecture’ to processes and policies. Understanding the behaviours that guide us can help us to apply some of the most rewarding and cost effective changes in human behaviour. A great introduction to behavioural science is MINDSPACE. The MINDSPACE Report was commissioned by the UK Cabinet Office and released in 2010. MINDSPACE offers a non-exclusive reference to some of the most effective behavioural ‘levers’ which can help maximise desired outcomes  in a considered, referenced and objective fashion. The Behaviouralist is in a unique position to apply MINDSPACE with Professor Paul Dolan who spearheaded the framework as our advisor, and Doctor Robert Metcalfe, the co-author for the MINDSPACE academic paper as our co-founder.

See Example

Randomised Controlled Trials

Traditionally, organisations implement policies, interventions, or significant changes in their operations without first testing whether and how it actually brings about the desired behaviour change or indeed a cost effective outcome. A Randomised Controlled Trial (RCT) is the gold standard approach to understanding whether something works or not. In a nutshell, it involves assigning customers or individuals to a treatment group (the policy change) and other customers or individuals to a control group (the counterfactual).  In addition to commercial settings, RCTs are  increasingly used  in areas such as tax collection, energy efficiency, transportation, charitable giving, physical health, and health services to indicate the absolute or causal effect of a change or intervention. Our team at The Behaviouralist has pioneering scientists that have a track record in implementing RCTs in for policymakers, companies, and charities, and who ensure the highest standards of research and cost-effectiveness.

See Example

Data Science

As behavioural scientists, we are most interested in what people are doing and not what they are saying. People’s intentions are a poor guide to future behaviour change.  As a result, we search for and utilise quality data on objectively measured individual behaviours. Our ability to generate high quality results depends on our ability to identify and work with a variety of sources of objective data. Within The Behaviouralist, we have highly experienced econometricians and experimentalists who can carry out data elicitation, collection, and estimation methods to world class standards. We believe quality data is crucial to high analytical power, which in turn allows us to demonstrate exactly how successful our interventions have been. 

See Example

How We Do It...

01

Showing an effect

Behavioural science practitioners place a strong emphasis on evidence-based research and identifying causality. With the best intentions, organisations often implement policies, product, interventions, or significant changes in their operations without first testing whether it actually brings about the desired behaviour change or indeed a cost effective outcome. Although the process of experimentation may seem costly, often not experimenting can be even more costly and every minute not understanding the impact of communications or processes could well be wasted money. At The Behaviouralist, we understand the exact requirements that enable organisations to deliver causal experiments so that they can be confident in their application of changes. See our glossary for some of the issues which can negatively affect one of the best-intentions.

02

Field Experiments

Applying a randomised control trial in the ‘field’ is known in many circles as a ‘field experiment’. Field experiments are important as they give the highest degree of validity of the evaluation for our interventions, which in turn allows us to show whether an intervention was successful and if so by how much . Field experiments can be further classified into sub-categories, depending on the study design.  Our team at The Behaviouralist includes world leaders and pioneers in the use and development of field experiments.

03

Correlation vs Causation

Until recently many would draw conclusions through correlations – identifying a link between two independent variables. However, such correlations do not indicate causation. To show that something has led to a change, we must be extremely confident of a causal effect. Where possible this means keeping all else equal and only making a change to one thing (the intervention). This requires us to satisfy a number of experimental conditions which include: randomisation, the correct sample size, removing selection bias, etc.

  • Showing an effect
  • Field Experiments
  • Correlation vs Causation
01

Showing an effect

Behavioural science practitioners place a strong emphasis on evidence-based research and identifying causality. With the best intentions, organisations often implement policies, product, interventions, or significant changes in their operations without first testing whether it actually brings about the desired behaviour change or indeed a cost effective outcome. Although the process of experimentation may seem costly, often not experimenting can be even more costly and every minute not understanding the impact of communications or processes could well be wasted money. At The Behaviouralist, we understand the exact requirements that enable organisations to deliver causal experiments so that they can be confident in their application of changes. See our glossary for some of the issues which can negatively affect one of the best-intentions.

Case Studies
02

Field Experiments

Applying a randomised control trial in the ‘field’ is known in many circles as a ‘field experiment’. Field experiments are important as they give the highest degree of validity of the evaluation for our interventions, which in turn allows us to show whether an intervention was successful and if so by how much . Field experiments can be further classified into sub-categories, depending on the study design.  Our team at The Behaviouralist includes world leaders and pioneers in the use and development of field experiments.

Case Studies
03

Correlation vs Causation

Until recently many would draw conclusions through correlations – identifying a link between two independent variables. However, such correlations do not indicate causation. To show that something has led to a change, we must be extremely confident of a causal effect. Where possible this means keeping all else equal and only making a change to one thing (the intervention). This requires us to satisfy a number of experimental conditions which include: randomisation, the correct sample size, removing selection bias, etc.

Case Studies
Revealed Preference
Revealed Preference
Dual Processing: System 1 & 2
Nudge
Hawthorne effects
Control Group
Sample Size
Selection Bias
Randomisation
Heuristics and Biases
Stated Preference
Treatment Group
MINDSPACE
Peak-End Effect
Loss Framing
Prosocial Incentives

Glossary

Revealed Preference

What we actually do i.e. our actions – are our ‘revealed preferences’. These are different from our stated preferences (what we say). Measuring and understanding revealed preferences is key to understanding behaviour change.

 

Dual Processing: System 1 & 2

In recent years, behavioural scientists have converged on the understanding that thought processing occurs through two distinct ‘systems’ – referred to as the dual-processing model. The two systems – System 1 and System 2 – exhibit different capabilities. System 1 refers to automatic, fast and unconscious behaviour, and System 2 reflective, slow, and conscious behaviour. The concept of dual-processing is central to developing a better understanding of behaviour, in particular the functions of System 1. That is, whilst System 1 makes life easier for us by processing information automatically, it sometimes makes ‘mistakes.’  

Nudge

It is now widely agreed amongst behavioural science practitioners that human behaviour is influenced more by context than cognition.  Over the past few decades, a growing body of literature has illustrated that given this effect, changing our surrounding environments has the power to influence human behaviour in mostly automatic or unconscious ways. In other words, we experience System 1 responses to changes in context or environment. Drawing on this field of research, Richard Thaler and Cass Sunstein brought the application of behavioural insights into government and policy discourse through their book ‘Nudge’.  The book introduces the concept of nudge interventions.  Nudges aim to change behaviour through largely automatic (System 1) responses to changes in context or environment.  Thaler and Sunstein state that nudges should not reduce or restrict choice, but rather guide people towards a certain behaviour or decision.

Hawthorne effects

People behave differently when they know they are part of a study or test. This is known as a Hawthorne effect named after a series of experiments conducted in the Hawthorne suburb of Chicago at the Western Electric Factory in order to study the effects of changing physical conditions on productivity. Although the changes did show an increase in productivity, the conclusion was that the association of people rather than physical changes led to the change.

Control Group

What would have happened. Any improvement requires a baseline, also known as a counterfactual or ‘business as usual’; something to measure outcomes against with all else being equal. This is called the control group in any given experiment.

Sample Size

It is important to undertake a study on a large enough sample in order to make causal inferences. Calculating the correct/minimum size for a sample requires that a number of statistical conditions be met so that we can be confident our intervention has caused a change. Providing this advice is part of our service.

Selection Bias

Our rigorous scientific expertise makes us mindful about selectively biased information. In examining the population groups, behaviours of which we are aiming to understand and potentially change, we are careful about the underlying characteristics of our sample. It is common among extant practices to overlook this selectivity issue and derive far-fetched conclusions about behaviours of the groups of individuals, which may not be representative.  

Randomisation

How can we be certain that our intervention will cause a change? Randomisation allows us to control for many of the variables that occur in the world, balancing the intervention versus all-else-being equal. Randomly allocating an intervention means we can be so much more certain that changes to a treatment group occurred because of our intervention and not through differences in allocation of our sample.

 

Heuristics and Biases

Mental shortcuts allow us to focus in on specific areas of choice, whilst ignoring others. These cognitive biases or heuristics lead to irrational decision-making through the governing of automatic choice – aligned with our need to make decisions quickly and with limited information. Understanding these heuristics is the foundation of behavioural economics.

Stated Preference

A stated preference is what we say we (would) do in a given situation. There is a difference between what we think we do and what we actually do due to distorted perceptions of ourselves and Hawthorne effects.

 

Treatment Group

This is the group of people that receive an intervention versus a control. Through the process of randomisation, the sample group is divided into the control group (that does not receive the treatment) and the treatment group. In this way, the characteristics (age, sex, location, socio-economic background etc.) of the two groups must be comparable. Ideally, the sole difference between control and treatment groups is the outcome variable being studied.

MINDSPACE

MINDSPACE is a mnemonic for a checklist of nine of the most robust contextual influences on behaviour. MINDSPACE stands for Messenger, Incentives, Norms, Defaults, Salience, Priming, Affect, Commitments, and Ego. The influences of each of these are outlined in the ‘MINDSPACE’ report as follows:

  1. Messenger: We are heavily influenced by who communicates information.
  2. Incentives: Our responses to incentives are shaped by predictable mental shortcuts such as strongly avoiding losses.
  3. Norms: We are strongly influenced by what others do.
  4. Defaults: We ‘go with the flow’ of pre-set options.
  5. Salience: Our attention is drawn to what is novel and seems relevant to us.
  6. Priming: Our acts are often influenced by sub-conscious cues.
  7. Affect: Our emotional associations can powerfully shape our actions.
  8. Commitments: We seek to be consistent with our public promises and reciprocate acts.
  9. Ego: We act in ways that make us feel better about ourselves.

The framework was published in 2010 by the Cabinet Office and co-authored by The Behaviouralist’s Senior Advisor Prof. Paul Dolan of the London School of Economics and Political Science.

Peak-End Effect

When we remember an experience, we tend to leave out most parts of it except for the peak i.e., the most extreme positive or negative event, and the end. The duration of the experience does not hold as much importance (duration neglect) as the average of the peak and end events. This is called peak-end effect.

 

Loss Framing

People are more likely to make significant changes to their behaviour due to the fear of losing what (they think) they already have than if they stand to gain something of the same value.

Prosocial Incentives

When the incentive for behaving a certain way rewards the team, community, or society, the incentive is prosocial. Examples of prosocial behaviours are volunteering and charitable giving.