Recently, the debate on how to identify and select new measures of wellbeing reached a wide audience especially thanks to the big media’s “ballyhoo”. That debate, very often accompanied by Robert Kennedy’s words (March 18, 1968, speech at Kansas University) has been urged also thanks to many prestigious initiatives, like the commission appointed by the French President in 2008 and now known through the chairs’ names (Stiglitz, Sen e Fitoussi). In many cases, the debate has been trivialized to the simple concern «what indicator can replace GDP?».
Since several decades, many researchers all over the world have been continuously working on defining concepts and measures able to measure and monitor the wellbeing of nations.1
Actually what emerges clearly is that the change of paradigm introduces several methodological implications in identifying indicators. Actually, monitoring wellbeing through indicators introduces some issues representing at the same time a challenge (given by the complexity), a need (represented by the relativization) and a risk (given by the over-reductionism).
This work, fruit of previous reflections, aims at clarifying and disentangling these issues:
(i) clarifying different issues concerning wellbeing of societies by providing conceptual instrument allowing anyone to orient oneself among all the emerging proposals and to distinguish between serious and propagandistic ones; (ii) unravelling some important methodological aspects and issues that should be considered in measuring it and constructing indicators.
The key allowing for the proper identification of new measures lies in the players’ (statisticians, researchers, analysts, policy makers, and so on) capacity and awareness in considering complexity, avoiding over-reductionism and investigating relativization.
A need: making relative
Selecting indicators implies a reflection about the objectives of their adoption (monitoring, comparing and benchmarking among territories, supporting and evaluating policy decisions, etc.). In that reflection two indicators’ characteristics should be considered, consistency with reference to concepts and adequacy with reference to the observed reality. Adequacy includes the idea to observe in a relative way.
Even though the definition of well-being may find a wide agreement, its operationalization (in terms of indicators) should take into account that different areas could adopt different indicators in order to measure the same concept. Making relative expressed in this terms represents a delicate issue, especially when different realities need to be compared. With reference to this, it should be clarified that comparing different realities (cultures, values, etc.) could require differentiated choices (Stiglitz et al., 2009). In fact, variables’ choice depends on shared values, which are functions of time and place. Consequently, transferring a wellbeing definition [table A]developed in a certain context could be misleading. With this respect, a good example is the variable “leisure time” whose definition can be different also from one individual to another. So, the process of comparing different areas would refer to the concept as a whole and not to single indicators (comparing synthesis tic indicators).
Making relative concerns also the interpretation of outcomes. This can be illustrated through a simple example: how to interpret a region’s high value produced by the ratio number of hospital bed / dimension of population? At a first glance, a high level could reveal a region paying attention to needs and requirements of population’s health. A later look could be alarming: does the high rate answers a concrete need of that territory? If so, the outcome showed by the indicator would point out an inadequate answer to a concrete problem existing, for example, at environmental level causing particular pathologies. In other words we would need other indicators redirecting the policy action towards other domains (e.g., environment). Maybe, proposing city mobility compatible with a healthy environment produces a better quality of the air and life style allowing a healthy life and hopefully a lower need of hospital beds.
Making relative through analyses
Monitoring wellbeing through indicators requires controlling values and levels through different perspectives allowing variations and differences to be detected. This can be analytically done through:
- Distributions analysis: appropriate representations of distributions allow flattening provided by average figures to be avoided.
- Correlation analysis, in order to explore also hypotheses concerning possible causal relations; this is crucial also in policy perspective: before intervention, one must know what causes what, which requires relatively mainstream scientific research, which may not be available yet.
- Break downs analysis, according through population composition (e.g., analyses by language, sex, age, education, ethnic background, income, etc.).
- Trend analysis, which can reveal particular tendencies and evolutions (linear, non-linear, chaotic, and so on).
A risk: reductionism
As we have seen, indicators in themselves are reduced descriptions of the reality. In other words, in their essence, they represent the idea of reductionism. Reductionism cannot be avoided, since it is actually impossible pull an image and a story from a pure observation of the reality.
The systematic identification of elementary indicators, in terms of concepts, domains and cases, allows a downright system of indicators to be constructed (more complex than a simple set of indicators, which are not always related to a conceptual framework).
In other words, the system of indicators represents a way to reduce an excessive reductionism at observational level. [Table B]
On the other hand, the consistent application of the hierarchical design produces a complex structure, difficult to be managed. In order to obtain a meaningful and interpretable picture, data should be managed with particular care to reduce the dimensions of the complexity.
From the methodological point of view, the reduction of the complexity can find an instrument in the synthesis, which may concern different aspects of the system (Maggino, 2009):
- synthesis of units (cases), when values observed for micro units are aggregated; the synthesis is compared at macro level (social groups, age groups, geographic areas). Generally, this kind of synthesis is accomplished by applying very simple even statistical instruments (e.g., average), which turn out to be very insufficient and unsatisfying since they do not allow the phenomenon’s distribution to be correctly represented. A possible (not necessarily the best) solution is to report a value related to the shape of the distribution (interquartile range).
- synthesis of indicators, when values observed for several indicators are aggregated for each case; the synthesis can be carried out at both micro (individual level) and macro level (regional, national, group level).
Reductionism applied to indicators can find essentially two solutions: (i) reducing the number of indicators, (ii) synthesising indicators.
The former approach needs a solid conceptual support. From the statistical point of view, the only evidence supporting the exclusion of one between two indicators is interrelation. A high level of relation between two indicators allows us to consider just one of them, assuming that indicators showing high relation are actually measuring the same concept’s component. However, this assumption is not necessary always true. The degree of freedom for such decisions is in the reality: the relationship between two indicators (e.g., number of firemen and amount of damages in a fire) can be high but mediated by a third one (e.g., dimension of the fire). If the nature of the third indicator changes, the relationship between the other two indicators changes or disappears, even though they will continue to describe, autonomously, the reality. If, by observing the previous high correlation, we excluded one of the two indicators, doing without one of them could deny ourselves precious pieces of the whole picture (as represented by the indicators). This means that having a solid conceptual model allowing indicators concepts’ relationships to be identified and interpreted.
On the other hand, it is dangerous to concentrate just on a few elements and statistically infer the sufficiency of the reduced observation from them
The latter approach consists in combining the indicators in a meaningful way.
a. the conceptual definition: carrying out a synthesis of indicators should take into account the conceptual process which led to their selection. As we have seen, the complexity of the phenomena to be described by the indicators requires a definition which follows the classical Lazarsfeld’s hierarchical design (1958). The hierarchical design is drawn starting from the concepts, the areas defining the concept, the latent variables covering each area. It can be drawn also through sub-designs (e.g. each area could require sub-areas) and its logic can be applied both at micro and macro level. With reference to synthesising indicators, the main guiding conceptual definition to be taken onto account is the relationship between each latent variable and its indicators. This relationship (defined model of measurement) can be reflective (where the indicators are seen as functions of the latent variable, whereby changes in the latent variable are reflected in changes in the observable indicators) or formative (where a concept is assumed to be defined by, or to be a function of, a group of indicators, identified in order to define it).
b. the perspectives of synthesis: the consistent application of the illustrated hierarchical design leads to a complex structure which can be represented through a conceptual matrix. In this conceptual matrix, each row represents a dimension/sub-dimension, while each column represents a domain. Indicators are made concrete in each cell. Table “C” represents the Conceptual Matrix for Indicators (CMI) referring to the previous description of possible aspects.
The synthesis can be achieved through the different perspectives:
for each uni-dimensional variable (for a single domain or across domains);
for each multidimensional variable (for a single domain or across domains);
across variables (for a single domain).
c. the approaches to synthesis
From the technical point of view, synthesis can be faced through:
• aggregative-compensative approaches:
– based on correlations (reflective approach): in the presence of high correlations (difficult to observe), the indicators can be aggregated and the aggregated score is easily interpretable;
– based on weights (formative approach): in this case, not-correlated indicators are preferable however are difficult to interpret with reference to the concept, which turns out to have a normative meaning.
Aggregation of indicators should take into account the nature of indicators (Maggino, 2009) which can be:
– reflective, when indicators are seen as functions of the conceptual (latent) variable
– formative, when indicators are viewed as causing – rather than being effect of – the conceptual variable
Reflective indicators are linearly related and interchangeable (the removal of an indicator does not change the essential nature of the underlying construct).
• non-aggregative synthetic approaches, based upon discrete mathematics.
As said, in order to measure and monitor wellbeing, a complex approach is needed, which requires many indicators, designed and organized in consistent conceptual structure. The obtained system provides all knowledge instruments allowing decisions to be taken more consciously.
Dealing with wellbeing measurement by taking into account its multidimensionality not only involves philosophical and political issues but concerns each individual’s and community’s real life. Consequently, the concepts are taken into account at different levels (individual, family, local, national, etc.), which interact between each other. In this perspective, in order to reach positive changes and trends, the decision-making process should be supported by a monitoring system, seen as a continuous observation of the well-being aimed at observing changes, evaluating effects of policies, and planning future activities.
At the same time, the monitoring system should be grounded on a solid democratic system and transparent media system. In this, a strategic role is important roles is played by the education and research system (school, university, etc.) and the official statistics, both meeting social consensus.
In this frame, we could image the policy maker like a pilot sitting at the flight desk (Maggino, 2009). Statistics has the task of defining, constructing and developing the instruments located in the cockpit. However, that activity needs to take into account that it needs:
• a clear definition of destination (referring to goals)
• a fair procedure allowing the community to take a shared decision concerning destination (democracy)
• a deepen knowledge of pre-conditions (resources)
• a constant observation of flight conditions (monitoring system)
• a continuous transmission and sharing of information on flight conditions (information system)
• a cultural environment available to the improvement of the whole system’s conditions (scientific research)
• a system allowing the community to face and manage emergencies (welfare and social security)
If even just one of these items is missed, the achievement of a good society is seriously damaged.
1. Particular attention should be devoted to the term “happiness”, which assumes different meanings according to different authors. Many scholars refer “happiness” to the affective component of subjective well-being (Nuvolati 2002; Diener et al. 2008). Others consider happiness as a synonymous of life satisfaction Veenhoven (1994).
Besides the different conceptual views, the statistical evidences can tell different stories. The highest rank correlation value between “level of satisfaction with life as a whole” and “level of happiness” by country in round 4 of the European Social Survey data is 0.6 (registered for the United Kingdom sample), revealing not only that the two components are not coinciding but also that a linguistic problem underlies the definition of happiness. Just an example concerning that. “Happiness” (one of the dimensions of the subjective wellbeing) is translated in Italian “felicità”. Actually, if we look at the linguistic roots of them we could realize that they are dealing with two different worlds, since the former comes from “to happen” and latter from the Latin “felix” which can be translated as “chance”. In the rest of this work this definition of subjective wellbeing will be adopted, in which happiness represents just one of the aspects (the positive affect).
Allison, D. B., Alfonso, V. C., & Dunn, G. M. (1991, January) „The extended satisfaction with life scale”, The Behavioral Therapist, 15-16
Anand, S., Sen A. (1997) “Concepts of Human Development and Poverty: A Multidimensional Perspective” Human Development Papers 1997, UNDP.
Andrews, F. M. & Withey, S. B. (1976) Social Indicators of Well-Being: Americans’ Perceptions of Life Quality, New York, Plenum Press.
Argyle, M., (1987). The Psychology of Happiness, Methuen, London (trad. it. Psicologia della felicità, Cortina Editore, Milano).
Berger-Schmitt R. and H.-H. Noll (2000) Conceptual Framework and Structure of a European System of Social Indicators, EuReporting Working Paper No. 9, Centre for Survey Research and Methodology (ZUMA) – Social Indicators Department, Mannheim.
Bradburn N.M. (1969) The Structure of Psychological Well-being. Aldine, Chicago.
Diener E., E. Suh (1997) “Measuring quality of life: economic, social, and subjective indicators”, Social Indicators Research, vol. 40, 189–216.
Diener E. and Emmons R.A. (1984). “The independence of positive and negative affect”. Journal of Personality and Social Psychology, 47 (5).
Diener, E. and Seligman, M. E. P. (2004). “Beyond money: Toward an economy of well-being”. Psychological Science in the Public Interest, 5.
Diener, E., Lucas R.E., Schimmack U., Helliwell J. (2008) Well-Being for Public Policy, Oxford University Press, Oxford.
Eurofound – European Foundation for the Improvement of Living and Working Conditions (2005) First European Quality of Life Survey: Life satisfaction, happiness and sense of belonging.
Eurostat (2000a) Definition of quality in Statistics, Eurostat Working Group on Assessment of Quality in Statistics, Eurostat/A4/Quality/00/General/Definition, Luxembourg, April 4-5.
Eurostat (2000b) Standard Quality Report, Eurostat Working Group on Assessment of Quality in Statistics, Eurostat/A4/Quality/00/General/Standard Report, Luxembourg, April 4-5.
Felce D., Perry J. (1995). “Quality of Life: Its Definition and Measurement”. Research in Developmental Disabilities, 16 (1), pp. 51-74
Giovannini E., A. Morrone, T. Rondinella, L.L. Sabbadini (2012) “L’iniziativa CNEL-ISTAT per la misurazione del Benessere Equo e Sostenibile in Italia” in Autonomie locali e servizi sociali, n. 1, Il Mulino, Bologna.
Horn R.V., (1993) Statistical Indicators, Cambridge University Press, Cambridge.
Johansson S. (2002). “Conceptualizing and Measuring Quality of Life for National Policy”. Social Indicators Research, 58, pp. 13-32.
Kahneman, D. and Krueger, A. B. (2006) “Developments in the measurement of subjective well-being” Journal of Economic Perspectives, 20(1).
Maggino F. (2009) The state of the art in indicators construction in the perspective of a comprehensive approach in measuring well-being of societies, Firenze University Press, Archivio E-Prints, Firenze.
Maggino F. e E. Ruviglioni (2011) “Preaching to the Choir: Are the Commission’s Recommendations Already Applied?” in Social Indicators Research, Vol. 102, Issue 1, pp. 131-156.
Noll H.-H. (2004) Social indicators and indicators systems: tools for social monitoring and reporting, paper presented at OECD, World Forum “Statistics, knowledge and policy”, Palermo, 10-13 November 2004.
Nuvolati G. (1997). Uno specifico settore di applicazione degli indicatori sociali: La qualità della vita. In: Zajczyk F. Il mondo degli indicatori sociali, una guida alla ricerca sulla qualità della vita. La Nuova Italia Scientifica, Roma, pp. 69-94.
Nuvolati G. (2002). Qualità della vita e indicatori sociali. Seminar held at the PhD degree programme “Scienza tecnologia e società”, aprile, Dipartimento di Sociologia e di Scienza Politica, Università della Calabria. Available on http://www.sociologia.unical.it/convdottorati/nuvolati.pdf.
Patel S., M. Hiraga, and L. Wang (World Bank) D. Drew and D. Lynd (Unesco) (2003) A Framework for Assessing the Quality of Education Statistics, Development Data Group and Human Development Network, World Bank, Washington, D.C.
Sharpe A., J. Salzman (2004) Methodological Choices Encountered in the Construction of Composite Indices of Economic and Social Well-Being, Center for the Study of Living Standards, Ottawa.
Sirgy M.J. (2011) “Theoretical Perspectives Guiding QoL Indicators Project”, Social Indicators Research, vol. 103, 1–22.
Sirgy M.J., A.C. Michalos, A.L. Ferriss, R.A. Easterlin, D. Patrick and W. Pavot (2006) “The Quality-of-Life (QOL) Research Movement: Past, Present, and Future”, Social Indicators Research, vol. 76, n.3, 343-466.
Śleszyński, J. (2012) Prospects for synthetic sustainable development indicators, paper presented at the conference “Quality of Life and Sustainable Development”, September 20-21, Wroclaw (Poland)
Stiglitz J. E., A. Sen & J.-P. Fitoussi eds (2009) Report by the Commission on the Measurement of Economic Performance and Social Progress, Paris. http://www.stiglitz-sen-fitoussi.fr/en/index.htm
Veenhoven, R. (1994) “Is happiness a trait? Tests of the theory that a better society does not make people any happier” Social Indicators Research, 32(2).
Zapf W. (1975) “Le système d’indicateurs sociaux: approches et problèmes”, Revue Internationale des Sciences Sociales, Vol. XXVII, n. 3.
Zapf W. (1984) “Individuelle Wohlfahrt: Lebensbedingungen und Wahrgenommene Lebensqualität”, in W. Glatzer e Zapf W. (eds.) Lebensqualität in der Bundesrepublik, Frankfurt a. M. – New York, Campus, pp. 13-26.
Recentemente, il dibattito sulla definizione delle modalità di identificazione e di selezione delle nuove misure del benessere ha raggiunto una amplissima platea specialmente grazie al “vociare” prodotto dai grandi mezzi di informazione.
Questo dibattito – che molto spesso viene accompagnato dalle parole pronunciate da Robert Kennedy (tratte dal discorso tenuto il 18 marzo 1968 presso l’Università del Kansas) – è stato stimolato anche grazie a numerose iniziative di prestigio, come la Commissione incaricata dal presidente francese nel 2008 ed ora meglio conosciuta con i nomi dei coordinatori (Stiglitz, Sen e Fitoussi).
Filomena MAGGINOProfessore associato di Statistica Sociale Università degli Studi di Firenze