Why nuance is needed when evaluating wellbeing measures

The Centre launched five years ago and has been hugely successful in making the case for why wellbeing should be a key objective of policy. While all agree improving wellbeing matters – what that means, how we achieve it, and how we measure it, isn’t always straightforward. 

Here we revisit six key complexities that policy makers and evaluators embedding wellbeing in their work need to be aware of.

1. Reduce low wellbeing or increase high wellbeing?

Research often focuses on averages, which can give the impression that one intervention will have a similar impact on all individuals’ wellbeing. However, the things that will improve wellbeing aren’t the same for everyone. This study, for example, found that the interventions likely to lift someone with ‘low’ wellbeing up to ‘moderate’, were different from those likely to lift someone from ‘moderate’ to ‘high’ wellbeing. 

So if you’re deciding on what interventions to support, consider whether your priority is to reduce the prevalence of low wellbeing or increase the prevalence of high. The interventions that follow are likely to be different.

2. Treat mental illness to improve wellbeing?

Mental health is at the heart of wellbeing [link to mental and physical health topic page]. Some argue there’s a ‘dual continuum’, where mental illness and wellbeing are two ends of one spectrum); others don’t. Either way, it’s clear they closely align. The issue, then, is how best to improve mental health? 

One perspective is to ‘treat’ individuals with medication and therapy. For many people, that is very much needed. But others argue that treatment alone won’t work unless people are safe, warm, occupied, connected and secure. This review, for example, found that upstream prevention work is likely to be more effective at reducing anxiety and depression in children then interventions in school. 

3. Evaluate impact on wellbeing quickly, and alongside other measures

People adapt to change. After a pay rise someone’s quality of life may permanently improve, but their wellbeing won’t be elevated as a result forever. Downplaying ‘adaptation’ risks is setting up good interventions to fail. 

If wellbeing is the only evaluation outcome captured, interventions that improve ‘objective’ outcomes might not be recognised as successful. So make sure you evaluate wellbeing soon after an intervention [link to measuring wellbeing topic page] – before participants adapt to new circumstances – and evaluate wellbeing alongside relevant ‘objective’ measures.

4. Expectations underpin our subjective evaluations 

How satisfied we are depends on what we expect: different people have different expectations and expectations can change over time. For example, British Social Attitudes data showed ‘objective’ measures of job quality deteriorated between 2006 and 2010, with more unsocial working hours, increased work intensity, and less job security, variety of tasks, and reduced pay. 

And yet, over this period job satisfaction went up. Recession meant employees were relieved to have a job; what they expected went down and their satisfaction with what they had went up. So when looking at temporal trends, interpret wellbeing alongside other measures and consider the wider context.

5. Satisfaction isn’t always the most important outcome

There are many determinants of good health – but one of the largest US studies on the topic found that ’satisfaction with health care’ was not one of them. Higher satisfaction with healthcare, after adjustment for other factors, actually predicted dying earlier. This may have been because more ‘satisfied’ patients had more unnecessary interventions and medication. The things people want from healthcare aren’t always the things that are best for health. 

6. Consider satisfaction alongside other wellbeing measures

Not everyone has entirely high or low wellbeing. For example, one person in six has been found to have a distinctive form of mixed wellbeing. People in this group were satisfied and felt that what they did was worthwhile, but they also were anxious and low. They were more likely to be female, and live in deprived areas where they didn’t feel safe. Many carers, parents, teachers, nurses, and public sector workers may be in this group. Use different wellbeing indicators to identify need in different groups. 

The concept of wellbeing has become central to how we monitor society’s progress, evaluate interventions, and agree priorities. While this perspective is established and clear, it is essential that we remain nuanced and thoughtful in how wellbeing data are collected, interpreted and applied.