Wellbeing Measurement

May 26, 2008

Wealth. Economists have taken the most indirect approach. They infer wellbeing from wealth. The theory behind this approach is that people use money to realize preferences. If you want a house, and you have the money to buy a house, you are probably going to buy a house. As buying a house realizes your preference, your wellbeing increases. In contrast, somebody else may also want a house, but does not have the money to afford one. As a result, this person cannot realize a preference and wellbeing is lower. There are several problems with this approach. One problem is that money can buy a lot of things, but not everything. Moreover, the preferences that cannot be realized with money may be more important for wellbeing, especially in rich countries, than the preferences that money can buy. Another reason why money is an imperfect measure of wellbeing is that most people have to work for their money. This creates two problems. First, some people actually enjoy their work. Thus, their work itself realizes some of their preferences. Others work only to get a paycheck. For these people, some preferences remain unrealized because they have to spend a lot of their time doing things that they would rather not do. To know which type you are, ask yourself whether you would continue working after you won the lottery. Believe it or not, some people say that they would continue working, and some actually do, if they won only $250,000. Wealth is a better indicator of wellbeing at the national level. Economists' measure of nations wealth like Gross Domestic Product, or Purchasing Power Parity are good predictors of nations' wellbeing (World Map of Happiness). However, to make this empirical claim, we need a more direct measure of wellbeing. The main reason economists have relied on wealth to measure wellbeing is an old assumption that there it is impossible to measure wellbeing by other means (Robins, 1938). However, over the past 50 years, social scientists have developed and validated several other measures of wellbeing.

Life-Satisfaction Judgments. The most widely used measures of wellbeing are life satisfaction and global happiness judgments. The vast majority of empirical findings about wellbeing are based on these judgments. The most commonly used life-satisfaction measures are Cantril's ladder, a 10 or 11 point scale ranging from totally dissatisfied to totally satisfied, and the five-item Satisfaction with Life Scale. It is relatively unimportant, which of these measures is used in a particular study, because these measures are highly correlated with each other. In other words, if you score high on one measure, you are likely to score high on the other measures as well.

[more information will be added]

Average Domain Satisfaction. Life Satisfaction Judgments assume that people are able to integrate information about various preferences in a consistent manner (i.e., different preferences are weighted according to their importance) in a reasonable amount of time. It is assumed that people will focus on all important preferences and not focus only on a few preferences that are in the focus of attention during an interview or when they complete a survey. To avoid some of these problems, it is possible to break-up the global question about all preferences into several questions about specific domains of preferences, such as preferences related to work, family, recreation, health, etc. The average satisfaction across these domains is then used as a measure of wellbeing. Average domain satisfaction is highly correlated with global judgments of life satisfaction. The method has two disadvantages. First, some preferences may not be captured by the list of domains that are presented to the respondent. Second, the simple average fails to take the subjective importance of different domains into account. Nevertheless, average domain satisfaction is highly correlated with global life satisfaction (r > .6). Thus, average domain satisfaction is a valid, although imperfect indicator of wellbeing.

Positive Affect Negative Affect Balance. People's feelings are linked to wellbeing in two ways. First, feelings respond to preference-realizations (Arnold, Lazarus, ). Second, people tend to have preferences about feelings; typically they like to feel good and they do not like to feel bad. Although feelings are mental states that cannot be directly observed, it is possible to measure people's feelings using a variety of measures. The most commonly used measure of feelings are self-reports. To obtain measures of wellbeing over extended periods of time, researchers typically ask people to estimate the amount of positive and negative feelings over that time period. Sometimes, reports of feelings over shorter time periods are obtained repeatedly and then averaged to measure wellbeing over longer time periods.

The first measure was Bradburn's (1969) Affective Balance Scale. Today the most widely used measure is the Positive Affect Negative Affect Schedule (Watson, Tellegen, & Clark, 1988). However, there is no generally accepted measure of affective wellbeing and many studies use their own measures. Different measures are highly correlated with each other (Watson, 1988). I have developed a six-item Hedonic Balance Scale with good psychometric properties (Schimmack, Diener, & Oishi, 2002). However, the measures are not identical and sometimes produce conflicting results. In these circumstances, it is important to determine why these inconsistencies occur and to compare findings to other wellbeing measures to make inferences about wellbeing.

Multiple Indicators. All measures of wellbeing are only indirect indicators of wellbeing. None of the measures directly reflects wellbeing. Because each method has different strength and weaknesses, it is likely that a combination of all measures provides a more accurate measure of wellbeing than a single method. Thus, it is important to establish that potential causes of wellbeing are related in the same way to the different measures. For example, if unemployment is related to lower income, lower positive affect, higher negative affect, lower life satisfaction, and lower domain satisfaction, it is fairly save to infer that unemployment is negatively related to wellbeing.

Quality of Life Index (Economist Intelligence Unit). This measure is designed to assess nations' wellbeing. It does not directly rely on subjective measures to assess wellbeing. Instead, life-satisfaction ratings are regressed onto several objective measures, and the regression weights are used to weight objective indicators. The aim of this approach is to remove measurement error from subjective ratings. However, the method is not foolproof. Two things can go wrong. First, the index does not reflect aspects of wellbeing that are not covered in the objective indicators used to predict life-satisfaction. Second, the method will not remove rating biases that happen to be correlated with objective predictors. The correlation between the Quality of Life Index and nations' wealth as measured by purchasing power parity is r = .85 (N = 111). However, it is also highly correlated with purely subjective ratings of wellbeing in the World Value Survey and in a recent Gallup poll (r = .80, r = .79). Thus, the measure is a valid measure of nations' wellbeing, but it is not clear whether it is a better measure than purely subjective measures.

Gallup-Healthways Well-Being Index. This index is based on the World Health Organization's definition of health as “not only the absence of infirmity and disease, but also a state of physical, mental, and social well-being.” Many questions are about physical health. The main wellbeing measures are a question about global life-satisfaction (Cantril's ladder), questions about affective experiences ("Did you experience the following feelings during A LOT OF THE DAY yesterday? How about? Enjoyment, Physical Pain, Worry, Sadness, Stress, Anger, Happiness"), and some evaluations of a number of life-aspects ("the city or area where you live?" "job or work you do" "standard of living" "health?" The most novel aspect of this index is that it is being assessed in national representative samples on a daily basis since January 2008. It will be possible to track fluctuations and long-term changes in American's wellbeing as a function of life events.

Human Development Index. The Human Development Index (HDI) was developed by the United Nations to track nations' development. It combines information about three objective indicators of development, namely health (e.g., life expectancy), education (e.g., literacy), and wealth (GDP per capita). Although the measure is a measure of development, it is commonly seen as a measure of wellbeing (Human Development Reports). However, the measure has a number of limitations. First, education and longevity are objective indicators that do not take individuals' own preferences into account. Second, the measure is a poor measure of wellbeing in affluent and educated societies. Third, longevity per se is an inadequate indicator of wellbeing because it does not take the quality of a life into account. For some people, a longer life with more misery does not increase wellbeing. Despite these shortcomings, the measure is a reasonable indicator of nations' wellbeing, as reflected in high correlations with other wellbeing indicators, including average life-satisfaction (wellbeing and HDI). A plausible explanation of this finding is that heath, education, and wealth are important characteristics of nations that provide more opportunities for individuals to realize their preferences.

Human Wellbeing Index (HWI). The Human Wellbeing Index was developed by Robert Prescott-Allen in his book "The Well Being of Nations." The HWI is an attempt to overcome some of the limitations of GDP and the Human Development Index (HDI) as measures of nations' wellbeing. Its main purpose is to be a component in a wellbeing indicator that addresses issues of sustainability and the "wellbeing" of the ecosystem. HWI is a composite of five domains:
Health and population. How long people may expect to live in good health [1 indicator]. The stability of family size [1 indicator].
Wealth. How well needs are met for income, food, safe water, and sanitation [6 indicators]. The size and condition of the national economy, including inflation, unemployment, and the debt burden [8 indicators].
Knowledge and culture. Education (primary, secondary, and tertiary school enrollment rates) and communication (accessibility and reliability of the telephone system and use of the Internet) [6 indicators]. Lack of a suitable indicator prevented coverage of culture.
Community. Freedom and governance (political rights, civil liberties, press freedom, and corruption) [4 indicators]. Peacefulness (military expenditure and deaths from armed conflicts and terrorism) [2 indicators]. Violent crime rates [4 indicators].
Equity. Household equity: the difference in income share between the richest and poorest fifths of the population [1 indicator]. Gender equity: disparities between males and females in income, education, and parliamentary decision-making [3 indicators]. To prevent a high score for equity from offsetting poor human conditions, equity is included in the HWI only when it does not raise the index.
Source. Prescott-Allen R (2001) The Wellbeing of Nations — A Country-by-Country Index of Quality of Life and the Environment. Island Press, Washington, Covelo, London.

The Human wellbeing index can be combined with the ecosystem well-being (EWI) index to form a composite wellbeing index (WI) (see map or ranking).
[I think it is problematic to use the concept of wellbeing in the context of an ecosystem. I think sustainability is an important goal for the wellbeing of future human (and non-human) beings. However, it is not clear in which way the ecosystem has a stake in its existence, or whether the ecosystem benefits from a decrease in the population of polar bears and an increase in the population of other species. I like polar bears, and I do care, but in which way is the eco-system better off with or without polar bears?]
[Another problem is that sustainability is a global problem, whereas wellbeing applies to individuals' lives. Countries with high (human) wellbeing benefit from importing goods that produce pollution in other countries. Furthermore, different countries have different needs in terms of energy consumption for heating. Another problem is that the weights of the two components are arbitrary, but determine the responsiveness of the indicator to policies. Thus, I do not think that it is helpful to combine human wellbeing and ecological/environmental indicators in a common indicator of wellbeing. At a minimum, it should be clear that the indicator could go up, even if most people become less satisfied with their lives. The reason is that objective normative criteria can trump subjective evaluations of one's own life.]