Reporting subjective well-being data
Using subjective well-being data to complement other measures of well-being requires producers of statistical information to regularly collect and release high-quality nationwide data from large and representative samples. Key audiences include policy-makers, public service providers, private businesses and voluntary sector organisations, researchers and the wider public - all of whom may have an interest in whether, where and when conditions in society are improving. For monitoring exercises in particular, it is important that the figures released mean something to the general public, as well as to more specialist audiences (New Economics Foundation, 2009).
Many of these audiences will not read statistical releases directly, but rather will rely on how these are reported in a variety of media. It is therefore important to consider how to package the data in a concise yet precise manner to ensure that the necessary information can be easily located and conveyed with accuracy by other parties.
The language used to describe measures is also important. The term “happiness” is often used as convenient shorthand for subjective well-being, in both popular media and parts of the academic literature - not least because happiness may be more attention-grabbing and intuitively appealing. The key risk surrounding the term “happiness” is conceptual confusion: whilst the experience of positive emotion (or positive affect) is an important part of subjective well-being, it represents only part of the over-arching concept, and the term “happiness” underplays the evaluative and eudaimonic aspects of subjective well-being as well as the experience of negative affect (pain, sadness, anxiety, etc.), all of which may be of interest to policy-makers.5 We therefore recommend against describing results only in terms of “happiness”, particularly for data releases from national statistics agencies.
Several authors have also shown a tendency to drop the term “subjective” from their reporting, simply describing results in terms of “well-being”. This is also a potential source of confusion. Whilst subjective measures of well-being offer an important insight into respondents’ views about their own well-being, the OECD regards subjective measures as only one of several measures required to develop a balanced view of well-being overall (OECD 2011a; Stiglitz, Sen and Fitoussi, 2009). This concurs with the outcome of the UK ONS’s recent public consultation on what matters for measuring national well-being (ONS, 2011a). For both the ONS and OECD, measuring well-being requires a mix of subjective and objective indicators, and measures across a variety of other dimensions (e.g. education, health, income and wealth, social connections and the environment, to name just a few) are viewed as an essential part of the overall well-being picture.
These considerations mean it will be important, especially when reporting the results of national surveys, to provide a full description of the indicators used - including the underlying constructs of interest, and what they might reflect in addition to “happiness”. This could be accompanied by a brief explanation of the rationale for measuring subjective aspects of well-being and their role in complementing (rather than replacing) other well-being indicators. Chapter 1 also discusses these issues.
For the purposes of high-level communication about subjective well-being results, particularly with non-specialist audiences, it is desirable to identify a small set of key measures and figures. These guidelines recommend that this set should include one primary measure of life evaluation and its dispersion, as well as a limited number of affect measures if possible (see Chapter 3). Eudaimonia and domain-specific life evaluations may also be of interest, although, as multi-dimensional constructs, they can be more challenging to convey in single headline figures. There are several different ways in which current levels of subjective well-being data can be presented for the purposes of monitoring progress - and the choice of method should ultimately be driven by user need and demand. Recent examples are available from France’s National Institute of Statistics and Economic Studies (INSEE - Godefroy, 2011) and the UK’s Office for National Statistics (ONS, 2012). Chapter 3 provides recommendations for the basic output associated with the different question modules proposed as part of these guidelines.
Because of the range of possible approaches to presenting and reporting on subjective well-being data, it is useful to consider the issue within some sort of organising framework. At the most general level, the question of how to report subjective well-being data for the purposes of monitoring progress has four elements:
- • How to report central tendency and level.
- • How to report distribution.
- • Whether and how to aggregate responses.
- • How to report change over time and differences between groups.