Category: Measuring

Further developing the framework

Our working model highlights a number of questions about how National Accounts of Well-being should best be composed, structured and measured. We hope that this will generate a new wave of research to produce a robust and well-validated framework that governments will be able to use to reliably measure population well-being. Issues for further exploration include:

  • Refining the components of well-being included in the framework. The components in our National Accounts of Well-being framework were designed based on considerations of the best current theory about the elements that constitute human well-being together with the possibilities offered by the data available in the European Social Survey. New attempts to construct similar frameworks might consider including additional components or grouping up measures in different ways to form new components.
  • Improving survey measures. The well-being module of the European Social Survey on which our working model for National Accounts of Well-being is based was designed by a cross-national group of well-being experts. But there are likely to be a number of possible improvements to the measures that were used, including: changes to question wording to improve clarity, comprehensibility and cross-country comparability, developing additional measures so that no construct is measured by a single question and adding measures to explore further components of well-being or to examine well-being within specific life domains.
  • Exploring geographically nested and detailed sub-group measurements. The sampling methodology of the European Social Survey is based on randomly selecting representative samples within a country as a whole so that the resulting data provide a reliable indication of average well-being at national level, as well as of well-being among key population subgroups. However, with more complex sample designs, well-being could be measured at local and regional levels, providing well-being accounts at local authority and regional level, which could then be grouped up to the national level. Alternatively, over-sampling of particular sub-groups in the population, for example particular minority ethnic groups, would allow more detailed exploration of the well-being of subgroups for which otherwise there would not be sufficient numbers in a random population sample.

Indicator scores

National Accounts of Well-being indicators are designed so that they are measured on 0-10 scales, calibrated so that 5 always represents the average score across Europe. This average is calculated from the data for all the respondents in the 22 countries included in these National Accounts. Scores above 5 therefore show that well-being is higher than this average level and scores below 5 that is below the average.

The rest of this page describes in detail how indicator scores were calculated.

Calculating indicator scores

The subcomponent and component indicators within the framework are a set of composites which each combine responses to several questions (apart from two single item subcomponents). Subcomponents and components are then aggregated to create overall personal and social well-being scores. The groupings in which measures are combined derive from an iterative process based on analysis of well-being theory, the substantive content of the question wordings, and the statistical structure of the data.

After deciding on the groupings, indicators were calculated in a three-stage process.

First, scores for the original survey questions were standardised to allow meaningful comparison using standardised z-scores, measured in standard deviation units. Second, questions were aggregated to produce subcomponent and component scores by taking the unweighted mean of the z-scores for the lower level indicators or questions.

Third, the results were transformed on to 0–10 scales, calibrated so that 5 always represents the average score for respondents across Europe (the 22 included countries). In order to achieve this, the transformation was carried out using a curvilinear relationship. This allowed the scaling factor by which the z-scores are multiplied alter across the range of possible scores, and so that responses at the upper end of the distribution are spread out. The curvilinear relationship is described by the formula:
where zi is the z-score that we want to transform and ti is the transformed score.

mi and ci are determined from the theoretical minimum and maximum z-scores for the indicator in question, according to the following formulae. Note that the formulae shown below are correct and revise an error in those printed in Appendix 2 of the National Accounts of Well-being report.

where min is a negative value and max a positive value.

The resulting indicator scores allow for comparison on a single indicator between countries and demographic groups and would allow comparison over time should more data be collected. They also allow meaningful comparison between performance on different components and subcomponents. What they do not do is allow the comparison of performance with absolute targets or thresholds. Whilst 10 is the theoretical maximum score for any of the national accounts indicators, there is no definition of how high you have to score to have high well-being, other than the fact that scoring above 5 means scoring above the European average.

While their design means that the scores presented here are intrinsically relative to the average well-being levels measured in 2006/2007 in the group of 22 European countries included, this sort of relativity is in effect a feature of any national accounting system. In the same way, knowing the level of a country’s GDP in a particular quarter has little meaning, without knowing how its GDP has changed over the preceding period and how it compares to the GDP of other countries. The indicators therefore gain meaning when they are used to compare the experiences of groups located in different social and geographical space, and over different times.

You can download the country scores for each of the National Accounts of Well-being indicators in standardised units.

European Social Survey data

Since 2001, the European Social Survey (ESS) has begun mapping long-term attitudinal and behavioural change in Europe. Covering attitudes to religion, politics, discrimination and pressing policy concerns, the data reveal intriguing contrasts and similarities between amongst over 30 European countries. It is the first social science project to win Europe’s prestigious Descartes Prize “for excellence in collaborative scientific research”, and is also one of the first to become a European Commission ‘Infrastructure’, in recognition of its high technical and academic standards and their impact on advancing the field of comparative social measurement.

In 2005, Professor Felicia Huppert of the University of Cambridge, asked nef, together with four other research centres, to join an ultimately successful application to the ESS to develop a 50-item questionnaire module to assess ‘personal and social well-being’ across Europe. A key aim in the design of the module was to measure both the ‘feelings’ and ‘functionings’ aspects of well-being, as well as psychological resources, such as resilience and self-esteem. A further aim was to go beyond individualistic aspects of well-being, by also incorporating measures of interpersonal, social well-being.

The survey fieldwork was carried out using face-to-face interviewing across Europe from autumn 2006 onwards with data released in autumn 2007 (Round 3, edition 3.1). In each country a sample of over 1500 adults was drawn using random probability methods. The resulting dataset therefore contains detailed measures of the individual experiences of just under 45,000 people. These data have been used to construct the national accounts framework reported here, for 22 European countries participating in the survey, covering both EU and non-EU members. The dataset can be downloaded (Round 3, edition 3.1)., and the individual survey questions explored online, at the ESS data website.

We used the weighted dataset for 22 European countries, excluding Romania and Lativa, for whom full survey weights were not available, and Russia. In addition respondents who had missing data on any of the questions included in the accounts (except for one question not asked in Hungary – see Appendix 2 of the report for details) were also excluded. Russia was excluded because its large population means that the data have to have large weights applied (a quarter of the total weighted count for the combined dataset). This high weight could have lead to distortion in analysis. Furthermore as conditions in Russia are not typical in ‘Europe’, and indeed much of Russia is not geographically in Europe, we decided not to include it in the national accounts.

Read more about how we used ESS data to calculate indicator scores .

The indicators: an overview

The National Accounts of Well-being indicators on this website represent a working model of measures which governments can use to monitor the well-being of their citizens.

The indicators have been constructed by nef, using data from the 2006-2007 European Social Survey . This is the most comprehensive and detailed international survey of well-being ever undertaken.

The resulting survey dataset was used to develop an indicator set around two headline measures of personal well-being and social well-being : the two crucial aspects of how people experience their lives. As the diagram shows, each headline indicator is broken down into component indicators (and in places also subcomponent indicators) which reflect the different aspects which together comprise experienced well-being. The seven main components of personal and social well-being are also the elements used to create national Well-being Profiles (with emotional well-being split into its subcomponents).

In addition, the personal well-being and social well-being indicators can be brought together in different ways to form versions of a combined well-being indicator . As an example of a well-being indicator within a specific life domain, we have also created a satellite indicator of well-being at work . In the report we also analyse indicators of the dispersion of personal well-being and of social well-being, which reveal how equally levels of well-being are distributed within group.

The indicators were created by standardising and transforming the survey data so that all results are presented on 0–10 scales, with a score of 5 always representing the average score across the 22 European countries included.

Read more about how the indicators were constructed.

Developing a framework

The challenge in devising a framework for National Accounts of Well-being is to match the multiplicity and dynamism of what constitutes and contributes to people’s well-being with what gets measured. Our recommended framework for National Accounts of Well-being is therefore based on capturing:

  • More than life satisfaction. Understanding subjective well-being as a multifaceted, dynamic combination of different factors has important implications for the way in which it is measured. This requires indicators which look beyond single item questions and capture more than simply life satisfaction, the survey question on which much well-being research to date has been based. There are a number of reasons for this, including the multi-dimensional nature of well-being and the errors which are known to arise from using a single question to measure a psychological state. There is also evidence that the specific question on life satisfaction is problematic because it leads people to focus on some aspects of their lives but leave out others, is subject to bias because of its general nature, and is not very sensitive to policy changes. This makes it a very blunt tool on which to base government decisions, and suggests advantages to using a broader set of well-being indicators, of which the constituent parts have clear links to defined policy areas.
  • Personal and social dimensions. A focus on the quality of people’s experiences of their lives might suggest that the predominant concern of National Accounts of Well-being is with people as individuals, aggregating different people’s reports of how they are feeling within themselves and experiencing life from their personal standpoint. But research shows that a crucial factor in affecting the quality of people’s experience of life is the strength of their relationships with others. There is plentiful evidence that feeling close to, and valued by, other people is a fundamental human need and one that contributes to functioning well in the world. Our approach, therefore, advocates a national accounting system which measures the social dimension of well-being (in terms of individuals’ subjective reports about how they feel they relate to others) as well as the personal dimension.
  • Feelings, functioning and psychological resources. The traditional focus on happiness and life satisfaction measures in well-being research has often led to an identification of well-being with experiencing good feelings and making positive judgements about how life is going. Our framework for National Accounts of Well-being moves beyond that to also measure how well people are doing, in terms of their functioning and the realisation of their potential. Psychological resources, such as resilience, should also be included in any national accounts framework and reflect growing recognition of ‘mental capital’ as a key component of well-being.