EUROSTAT
Directorate D: Economic and Regional Statistics
Unit D-1: Key indicators for European policies /
Doc. Eurostat/F/08/SDG/02/02.1EN
Point 2 :Feasibility study on the measure of wellbeing
Presentation of Eurostat work
Strategic Development Group
Brussels, 3 July 2008
L-41 7/23 (rue de la Loi, nr 41 – 7th floor, room 23).
Feasibility study on the measure of wellbeing
Presentation of Eurostat work
Eurostat commissioned in December 2007 a two-year contract regarding a feasibility study on the measure of wellbeing to a consortium[1] led by IDEA Consult (Belgium). In order to run efficiently the project Eurostat is advised by a Steering Committee composed of experts from various Commission services[2].
Given the abundance of existing approaches to measure wellbeing, the contractor was not requested to develop a new approach but to examine existing ones in order to select three of them which would be afterwards fully implemented for the EU and its MemberStates.
1.Background of the study
Sustainable development is a fundamental and overarching objective of the European Union, which is made operational through the EU Sustainable Development Strategy (SDS). Well-being is a key concept in the SDS, as the renewed strategy (adopted by the European Council in June 2006) states that sustainable development ‘aims at the continuous improvement of the quality of life and well-being on Earth for present and future generations’. This global objective is multidimensional, as there are numerous determinants of human well-being. It is clear from the extensive literature on the concept and different aspects of well-being, but also from the SDstrategy itself, that the multidimensional content of the notion of well-being follows the interdisciplinary approach of sustainable development.
However, both concepts can (and must) be treated methodologically separately from each other. After all, well-being is centred on the (individual) human capabilities to be or achieve something, whereas sustainable development covers all aspects of (intertwined) economic, social and environmental development. Well-being thus forms, in a manner of speaking, a part of the global picture of sustainable development; and so do the policies that aim to improve well-being. Policy plans and actions, aiming to improve well-being, form part of a global strategy for sustainable development, but need to concentrate on those aspects directly or indirectly affecting human well-being.
2.Objectives of the study
Eurostat is confronted to a highly growing demand for measuring well-being, happiness, or other related topics such as progress of societies. Without taking any commitment concerning the possible future use of this study, the project attempts to respond to this request. As an example of possible use, the 2006 European Sustainable Development Strategy gave to Eurostat the mandate to report every two years on its implementation and the wellbeing measures developed in this project could find – if assessed as feasible – a possible implementation in this report.
Within this global objective, the project aims to analyse the feasibility of some promising indicators or indicator sets of well-being. This involves 3 main ‘operational’ objectives:
-Reviewing existing well-being approaches and selecting the most ‘promising’ indicator(set) (s) / methodologies / models.
-Gathering, systematically bundling and analysing the concrete data so as to give empirical support and policy utility to indicators that promise high policy relevance.
-Summarizing lessons and recommendations from the practical implementation of an indicator(set) at European level.
2.1.Focus of the study
Given the relationship to sustainable development, the focus of this study is being given in terms of present and experienced future human well-being of individuals. This means that:
-Factors such as economic performance (e.g. GDP growth), social systems (e.g. social security system), or environmental quality (e.g. air pollution) will be considered to the extent that they function as determinants of well-being, but will not be regarded as ends in themselves.
-Similarly, emergent properties of socio-economic systems (e.g. income inequality, social cohesion) will only be considered to the extent that they influence the well-being of individuals.
-The indicator or set of indicators to be developed will only try to capture the well-being of humans today or what could influence the well-being of next generations, without trying to predict the well-being of future generations. Consequently, the indicator or set of indicators will only take into account the (possible negative) impact of current activities on future well-being where these actions affect our present-day experienced well-being.
In other words: environmental or economic issues will be considered only, and only as a means, if they affect human well-being today. Exploring these determinants, the value of human’s subjective experiences/feelings will be acknowledged in the search for the best (set of) indicator(s).
2.2.Timetable
The project started in January 2008, and is planned to finish at the end of 2009. To reach the above-mentioned goals of developing, feeding and optimizing a policy relevant and clear indicator on well-being, the project is broken down in actions, phases, milestones and continuous internal and external reflections. These are structured around 4 ‘tasks’:
Task 1- Review of current approaches in measuring well-being;
Task 2 - Data compilation for the selected indicator(set) (s);
Task 3 - Data analysis for the selected indicator(set) (s);
Task 4 - Critical analysis of the work and further recommendations.
With the following indicative timetable:
Jun-08 / Dec-08 / Jun-09 / Dec-09Task 1
Task 2
Task 3
Task 4
3.First results
3.1.Literature Review
The basis of the first phase was the screening and systematically bundling of the work already done in the field of well-being; both at conceptual level and with respect to the measurement of it. This subtask has resulted in:
-an extensive overview of possible interesting indicators, indicating their main components and the methodological/conceptual behind them (see list in annex);
-an accompanying conceptual note.
This note explains the scope and focus of the concept of well-being used, clarifies the terminology used and establishes a transparent overview of existing well-being/quality of life indicators with their main characteristics.
3.2.Development of evaluation criteria
Subsequently, a framework to systematically assess the indicator list was developed. This exercise has resulted in a note with possible relevant evaluation criteria, including an operational instrument for the actual screening of the indicators, indicator components and their variables.
3.3.Selection of indicators for further testing
A screening template was used to systematically analyse the existing WB-indicators on their merits and shortcomings on the basis of the list of evaluation criteria for the indicators and the underlying variables/data. This set of evaluation criteria is symmetric to the ESS definition of quality and includes issues such as relevance and coverage, comparability and country specificity, data availability and accuracy.
3.4.Workshop with well-being experts
The consortium and Eurostat organised on 9 June 2008 a workshop with academic experts and representatives from policy DGs and Eurostat. The aim of the workshop was to have a structured discussion and reflection about the set of potentially promising indicators and their relative merits in measuring well-being.The following 9approaches were discussed at the workshop:
Subjective approaches:
–Life satisfaction: based on a single question on life satisfaction. This indicator/approach is considered as too simplistic and misses some objective considerations.
–Positive and negative affect schedule (University of Iowa): the components are positive/negative affects. This approach has a limited scope and policy relevance.It does not say anything about the drivers for well-being.
–Psychological needs scale (University of Rochester): the components are competence, relatedness, and autonomy. Based on individual data, the approach covers a number of relevant subjective issues, but is not complete (e.g. health issues missing). The approach is based on a psychological conceptwhich has not been empirically tested. The authors claimthat it is more culturally neutral than other measures, but from the questions there do seem to be cultural biases. The approach is not very policy relevant.
Objective approaches:
–Human Development Index (UNDP): the components are life expectancy, knowledge, and standard of living. Although it comprises only objective measures, the indicator is not comprehensive enough and not adequate for developed countries. The weighting is artificial as for most composite indicators.
–Index of Human Well-Being (R. Prescott-Allen):it combines 5 components which are health and population, wealth, knowledge and culture, community, and equity. There is little information on the aggregation of variables. The list of components and indicators can be considered as arbitrary. The treatment of equity is problematic.
–Quality of Life Index(T. Rahman et al, UN University): it combines 8 domains which are relationship with family and friends, emotional well-being, health, work and productive activity, material well-being, feeling part of one's local community, personal safety, and quality of the environment. The list of domains seems well chosen and rather comprehensive. The individual indicators, however, appear to have been chosen for their wide availability rather than their suitability (e.g. suicide rate as a measure of emotional well-being).
Combined (objective and subjective) approaches:
–Happy Life–Expectancy (R. Veenhoven): the indicator combines life expectancy with average declared happiness converted to a 0-1 scale. The index is over-simplistic. Similar scores would be obtained with a short happy life or a long unhappy one. However the two variables are supposedly positively correlated.
–The 'Economist'Quality of life Index (The Economist Intelligence Unit): the indicator comprises 9 components: material well-being, health, political stability and security, family life, community life, climate and geography, job security, political freedom, and gender equality. interesting approach, except very doubtful (statistical instead of conceptual) method of choosing variables,
–Set of well-being measures (DEFRA-UK approach):the set of indicators covers a wide range of domains: well-being, society, employment and poverty, education, health, mobility and access, social justice and environmental equality. The list of indicators is based on research and split by age, social class etc. There has been no attempt to aggregate all elements into a single indicator.
Following the workshop discussions, the consortium and Eurostat will review the advantages and disadvantagesof each approach, and consider possible adaptations (contents, methodology, etc…) which could improve the approach.The selection of the 3 indicators to be further analysed and developed will be taken in close cooperation between the consortium and Eurostat, after consultation of the Steering Committee.
4.Conclusion
It is obviously too early to draw any conclusion from this promising project, which will deliver a feasibility study that Eurostat is free to endorse or not. The project attempts to respond to a highly growing demand for measuring well-being, happiness, or other related topics such as progress of societies. Two basic questions for Eurostat and the ESS can be anticipated from the outcome of this project:
–Should official statistics be further involved in subjective matters, which are inevitable when talking about well-being?
–Should official statistics deal with composite indicators which are a way to synthesize the information and could attract more easily the attention of media, policy-makers and the general public, although being based on an uncertain technical basis?
1
Annex:List of approaches reviewed by the project
Nr / Index / Description / Source / Advantages / Disadvantages1 / Life Satisfaction (single question) / Self-reported satisfaction with 'life as a whole', typically on a scale of 0-10. Research demonstrates high validity in terms of correlations with other measures of life satisfaction, such as other people's reports and day reconstruction methodologies. Also strong relationships with health, income, employment status, relationship status, intentional activities demonstrate that this is a sensitive variable in comparing between individuals, between countries, and for tracking change over time for individuals. / several (e.g. World Values Survey) / - Most widespread SWB question,
- data available from almost all developed countries and many developing countries,
- time series available
- high validity
- good outcome indicator / - possibly no/little model of change
2 / Psychological well-being scale / 6 aspects: self-acceptance; positive relations with others; autonomy; environmental mastery; purpose in life; personal growth. Each assessed using a 20-item self-report instrument, though a 3-item version has also been developed (Midlife in the US study). Theoretically billed as an amalgam of approaches from humanistic, developmental and clinical psychology. Some doubts have been raised over the factor structure - limited support for 6 discrete factors - and internal reliability / Ryff CD (1989): Happiness is everything, or is it? Explorations on the meaning of psychological well-being. Journal of Personality and Social Psychology 57:1069-1081. faculty/bio/ryff.html / - Potentially operationalisable from the ESS module. / - Poor factor structure (amalgam of approaches)
- limited support for choice of variables
- unsure internal reliability
- No/little model of change
3 / Orientations to Happiness Scale / 3 aspects: pleasure, meaning, engagement. Assessed using 18 item-scale with 6 items for each dimension. Theoretically background is Seligman's tripartite concept of "the good life". Engagement component partially based on Csikszentmihalyi's "flow". Seems to be internally consistent although hasn't really been used for anything outside of Seligman's Authentic Happiness online survey. / / - Potentially interesting content/model
- good outcome / - No data for Europe
- poor factor structure
4 / Psychological Needs scale / 3 aspects: autonomy, competence, relatedness. Assessed using 21 item-scale with 7 items for each dimension. Based on Self-Determination theory, which is a modern development of humanistic "needs" theories (e.g. Maslow). Posits that the needs for Aut, Comp and Rel are innate, universal and fundamental prerequisites to flourishing. / Deci & Ryan - / - Good outcome indicator
- Good model of change
- Scales seem to be robust / unclear link with WB
5 / Positive and Negative Affect Schedule (PANAS) / Assessed the relative balance of positive and negative feelings over a given period of time. 20 item scale with 10 positive and 10 negative. Can be used over different time frames but typically between the last week and the last month. Seems to be robust. However, only captures hedonic dimension, which few would claim is synonymous with overall well-being. / / - Validated
- Seems to be robust
- good outcome
- can be combined with others / Very narrow scope (only hedonic)
Nr / Index / Description / Source / Advantages / Disadvantages
6 / Affectometer 2 / 10 aspects ("mnemonic qualities of happiness"): Confluence, social interest, optimism, freedom, self esteem, energy, self efficacy, cheerfulness, social support, thought clarity. Each assessed through 4 items. Designed explicitly as a longitudinal measure / / Has been shown to have test-retest reliability / - Not strongly grounded in any particular theory
- not really used
7 / Satisfaction with Life Scale / 5 item scale intended to capture overall cognitive evaluations of life (rather than hedonic / affective components). Essentially, a more psychometrically sound version of single-item Life Satisfaction. Calls for respondents to make a counterfactual comparison with their "ideal" life - some versions compare past, present and future satisfaction. / Diener - / - Validated: has been used as a benchmark for others (e.g. Ryff) to validate their own scales
- more robust than number 1 / - Outcome measure
- a few questions not valid
8 / U-index / Derived from Day Reconstruction data which asks people to think back over the last 24 hours, think about what they were doing and how they felt in terms of positive and negative affect. U-index is the % of time for which the strongest felt feeling was negative (~ highlights negative feelings). Does correlate with life satisfaction, but not that highly. / Kahneman D & Krueger A (2006). Developments in the Measurement of Subjective Well-Being.
/ - new methodology
- unhappiness point of view: good idea / New methodology
9 / Personal Well-being Index / 8 aspects: standard of living, health, achievement, relationships, safety, community, future security, spirituality. Each assessed in terms of satisfaction. Optional "overall" satisfaction item is also included - typically, the domains explain 30-60% of variance In overall measure. / / Variables can be approximated with other data / Data not available
10 / CASP-19 / 19 questions, on Competence, Autonomy, Self-Realisation and Pleasure. However, factor analysis reveals a single latent factor, so structure doesn't come through. Also, negative skew reveals a ceiling effect. Designed for elderly people, used in the British Household Panel Survey. / / - to much overlap between domains
- poor factor structure
11 / European Quality of Life Survey / Survey includes life satisfaction, domain satisfactions, happiness, time, loving, being and having deficits and alienation. Explores interrelations and predictors in great length. No top line index created, but implicit in methodology is that life satisfaction is considered as gold standard. / Bohnke P (2005)
12 / Index of Individual Living Conditions / Based on European Community Household Panel survey (1995-2001). Based on 'objective condition' questions on living standards, housing, education, health, social relations and work. About 25 questions in total. Not entirely economic, but economic factors dominate. / GESIS, social indicators research (Heinz Herbert Noll) -
Nr / Index / Description / Source / Advantages / Disadvantages
13 / Gallup Well-Being Index / Composite of seven questions, three of which are evaluative, and four are yes/no questions that can be seen as hedonic. / Gallup / - has been calculated for more countries than any other subjective index
- high quality data/survey (in terms of frequency and variables <=> ESS) / - Not very transparent methodology,
- nor any theory behind it
- private - very expensive
14 / European Social Survey / 55-question survey carried out across most EU nations, plus a few others. Has not yet been developed into a single indicator of well-being. / European Social Survey