Understanding Well-being Data

chapter 9 Understanding

Following the data: how we have come to understand well-being data in this book

We have covered a number of different understandings of well-being and data in this book, as well as considered their impact on, and relevance to, culture and society. We have identified how ideas of well-being differ and transcend time, place, culture and religion. We have encountered how people feel about well-being in their everyday life, and projects to try and understand this phenomenon, as well as the understandings of those responsible for people’s well-being, such as those in government. We have also considered how people interact, even live with data in their everyday lives, but are not always sure they understand them.

We have followed the data into ‘disciplines’, as groups of academics and professionals who look at the world in a particular way, and tend to agree on certain methods to understand it. We have considered how experts understand well-being across research disciplines (including economics, social and cultural policy, social statistics and philosophy), and how they work together in sub-disciplines, and in practice. For example, many economists look for trends in what people value over time and what that means for well-being. This book has presented documents as data to analyse what well-being economists (and other disciplines) value and how that changes over time. We found indications that happiness psychology as a new discipline suited the ends of those eager for ‘a new science of happiness’, but that when it came to deciding on data processes, some psychologists felt

their expertise was overlooked. We found that economics has traditionally held much sway with policy-making institutions, but not necessarily made their ideas and principles accessible to all. Of course, these issues are not specific to economics, but most disciplines using data to understand well-being can lose sight of shared understanding, or being understanding.

In Chap. 5, some of the pros and cons for writers on different Big Data approaches were synthesised. Notably, Tables 5.1, 5.2 and 5.3 indicate that these concerns tended to reflect on the utility of data for the data scientists, or whoever else might want to use them. They did not account for whose data they were and how ethical these approaches might be. Given that Big Data are often collected in ways that are not obvious to people, what could be done better to ensure shared understanding?

There are moves towards greater fairness, accountability and transparency in data uses. Yet, following a data controversy and watching how these principles work in practice demonstrate that much effort remains to establish what a shared understanding of these values look like in practice. We briefly considered the case of the algorithm that decided on students’ A level grades, in lieu of an exam under COVID-19 restrictions in the UK. The outcome was contentious, but the regulator (Ofsted) insisted this was the fairest way to approach the problem. Yet as the headline premises behind the decision-making method emerged in the press, the process became a national scandal; notably, because of the impact on young people’s futures and current well-being. There were then calls for transparency and accountability, but when the algorithm’s methodology was published, the 319-page document was not legible to many and was only even manageable to a very select few.

Transparency could involve showing everyone everything, but how does this compromise understanding? What does that mean when the data and the actions surrounding their use are complex, highly detailed and outside of everyday understandings? Chapter 8 reviewed one research project using valuation with well-being data, step by step. It followed the data backwards, to understand the contexts from which they and the study originated. It also followed the findings forwards to understand how the research was interpreted in other contexts. The report explained that these methods were accepted by experts in government. However, the chapter found that when the methods were reproduced, using the same data, the findings differed, so what does that mean for shared understanding of experts. The chapter also showed that when the findings were reproduced in the media, they were misinterpreted to say that museums boost

happiness, which was not how the research was presented in the report. What does that mean for shared understanding with non-experts in data?

Shared understandings are difficult, when within the same field. In Chap. 7, we encountered two research projects which were ostensibly looking at ‘subjective well-being’ in a similar population: people with an artistic practice and people with a creative occupation. We found that while the term ‘artistic practice’ indicates a level of professionalisation, this was not what the research was necessarily looking at. Similarly, that creative occupants didn’t need to be creative at all—as we might understand the word—according to the UK government’s Department for Culture, Media and Sport. We also found very different data were used to understand the concept of subjective well-being in these studies. What does that mean for how we join-up and share understanding of the well-being of different groups?

We have discovered that the meaning of well-being changes as the nature of data changes, and desire for data evolves and demands for data analytics increase. We have looked at well-being as it is understood as various measurements, and the benefits of understanding well-being at scale and over time, and have witnessed how knowledge and information can be gained, but also how some meanings can be lost by these exercises. Context that ties the data to the people it is about is removed, to enable patterns to become visible at scale, and yet context is rarely accounted for in narratives of the benefits of these data and their uses.

We have seen how well-being data are data about us—they are our data. They require our interactions, often our time, and are used to make decisions that are ostensibly on our behalf, but we may disagree with. We have seen how they change the workplace, how people were managed in COVID-19 and even the TV programmes we end up watching or the music we listen to. We have seen the growth of apps to track our well-being and tell us how to live better or walk more steps, and the market value of these apps increased considerably in this last decade. We have also witnessed how lucrative well-being data can be when their analysis has value to a policy sector, government or, in the case of a pandemic, the whole world.

We have also found indications that despite the fact we are ‘living with data’ (Living With Data), we don’t all necessarily grasp what is happening with our data and what they do for and against us in our day-to-day lives. Unpacking various types and forms of well-being data (data about well-being) and touching on the possible impacts that data and their uses have

on our own well-being, and society more generally, is crucial to grasping some of the contexts of data that get obscured. So, understanding well-being data can help us understand data better. But more than that, contextualising well-being data—discovering the whos, whats, wheres, whens, hows and whys, as well as the so whats and the what nexts—offers insight into politics and policy. It also helps us understand how research and knowledge may claim to know things, but that these claims may have limits.

There are limits to the promise of data: what they can achieve for society is not always good. Technical progress in data and their handling are not always a development for good. The fetishisation of data and proof of value (as with the case studies of social and cultural policy) prove that attachments to data in society are flawed, opening up a market for data practices that shifts the relationship between researcher and data. Our attachment to ideas of novelty and innovation, as with the case of ‘the new sciences’ and Big Data also blindside us to their limits. These are a few of the growing concerns in critical data studies, but need to be a bigger concern in all studies of well-being, across social policy, social statistics, sociology, economics, psychology and so on. There is an opportunity to take what we are learning in critical data studies and well-being studies to help the social sciences consider how it might adopt a more understanding position.

We need to return to how we understand how data are understood and how they can make us a more understanding society. Context matters: where data come from, who they are for and about, where they go and for what purpose. But context matters for more than researchers and more effort should be placed on how it can improve shared understanding, and being a more understanding society. Without acknowledging the limits in capacity, or indeed possibilities for understanding, the What Next? or How can we do it better? questions will not be answered properly for well-being.