chapter 3 Looking at Well-being Data in Context
Everyday well-being data: asking people questions about their lives
Well-being data are not only for policy-makers or international economic development agencies. They can be collected in various ways available to us in everyday situations. Many of us have seen an increase in emails popping into our inboxes or Facebook timelines asking us to complete some kind of questionnaire about our well-being. COVID-19 has seen collection of these kinds of data increase.
These are well-being data collected through a questionnaire not so different from a national-level survey, but on a smaller scale. Although most require good planning and ethical consideration of how the questions you ask people may have some negative impact on them. In short, could your research negatively impact on people’s well-being?
The following section offers a brief overview of methods that can collect ‘smaller data’ for different ends. Vignettes from my own research, a hypothetical questionnaire scenario, and the ethics of ethnography are presented to help you consider the different contexts and considerations of well-being data. Table 3.1 shows the advantages and challenges of different kinds of data for understanding well-being. While this section might help you design your own research, there are countless exhaustive textbooks out there that devote more space to that. Here the goal is to help you to imagine data contexts, and so better evaluate how other people have used data.
Table 3.1 Data sources and their uses
|What kind of
|Existing administrative and monitoring data
|Data gathered as part of operations, routine surveys.
|Equality monitoring data (organisation level)
Births (national level)
|Firms increasingly asking well-being questions as part of monitoring.
|Closed multiple- choice Qs,
for example, nationality, gender identity.
|To monitor or understand whole populations (i.e. of countries or organisations). Can help understand how specific demographics experience ill-being, for example.
|Data access issues (e.g. legal issues, internal procedures, identifying the target group(s), collecting comparator data)
|Existing large-scale survey data
|Long-term, large-scale survey data, administered by international bodies (i.e. OECD), central governments, the ONS (in the UK).
|Labour Force Survey (LFS, ONS)
|‘ONS4’ Personal well-being Qs are now in LFS.
|Closed multiple- choice Qs.
Do you do shift work in your (main) job?
|Nationally representative sample, claims can be made of the whole nation. One example might be understanding if shift work is linked to anxiety at population level.
|Quantitative modelling most likely required. Access can be difficult for some kinds of data and individual researchers.
|Existing qualitative data
|Previous research projects may archive their qualitative data for re-use. This could be from a survey, interviews, diary submissions or free text from surveys.
|1938, Mass Observation project: ‘what is happiness?’
|This example is a project about happiness, but one well-being- related question could be added to a questionnaire on something else.
|Open Question, ‘what is happiness?’ Participants answer with short or long descriptions.
|Can be used to understand one or many people’s depths of experience of well-being at a particular moment in time. For example, asking people to keep mental health diaries in COVID-19.
Not all open access/available to all.
Long descriptions are often time-consuming to analyse and may require specialist knowledge.
|New data collected with a specific purpose (quantitative)
|Small-scale or one-off surveys
|‘BBC Loneliness Experiment’
|This example is a one-off loneliness survey, but one or more well-being- related questions could be added to a questionnaire on something else.
|How would you define loneliness? Options included: having no one to talk to; feeling disconnected from the world; feeling left out; sadness; feeling misunderstood.
|This survey was used to interrogate how whole populations experience and define loneliness. It addressed a gap in knowledge, which is what a large-scale one-off survey would be for.
|Surveys are expensive and harder to design properly than people imagine. Small-scale surveys in organisations are notoriously poorly designed, compromising the data collected.
collected with a specific
|Measuring National Well-being debate
|What matters to you?
What things make life worthwhile?
|The MNW Debate also used quantitative data from an online questionnaire. Qualitative data were collected via group discussions and free text in the online questionnaire.
|This was incredibly resource-heavy to collect the data, costing millions of pounds, and took months to analyse. Analysing focus groups correctly is more complicated than sometimes imagined.
|Social media data
|Analysing geo-located tweets using sentiment analysis could help begin to understand how people feel in: (1) parks, or
(2) public transport.
|People aren’t asked questions. In fact they often don’t know their tweet could be used in research.
|Social media data are mainly used as qualitative data where you ask research questions of the large dataset.
|Ethics of data use. Not everyone can use scraping software or sentiment analysis software. As with all qualitative data, you can analyse by hand, which is resource-heavy.
- For more discussion on Mass Observation and two examples of their qualitative data on the meaning of happiness, please refer to Chap. 5.
- For more information on these approaches, please see Chap. 5.