Research on subjective or self-reported well-being (SWB) has been ongoing for several decades, with the past few years seeing an increased interest by some countries in using SWB measures to evaluate government policies and provide a broader assessment of the health of a society than is provided by such standard economic measures as Gross Domestic Product (see, for example, Stiglitz, Sen, and Fitoussi, 2009). The National Institute on Aging and the United Kingdom Economic and Social Research Council asked a panel of the National Research Council’s Committee on National Statistics to review the current state of research knowledge and evaluate methods for measuring self-reported well-being and to offer guidance about adopting SWB measures in official population surveys (see Box 1-1 for the full charge to the panel). NIA also asked the panel to prepare an interim report on the usefulness of the Subjective Well-Being module of the American Time Use Survey (ATUS), with a view as to the utility of continuing the module in 2013.
The SWB module is the only national data source in the United States that links selfreported well-being information to individuals’ activities and time-use patterns. It provides researchers with unique insights that are only revealed by melding ratings of affect with time use information. The SWB module, overseen by the Bureau of Labor Statistics (BLS) and sponsored by the National Institute on Aging (NIA), was developed with guidance from several noted academics—Angus Deaton, Daniel Kahneman, Alan Krueger, David Schkade, and Arthur Stone among them—working in the field.
Though the SWB module has only been in existence since 2010, it is not too early to begin assessing its potential value to researchers and policy makers. The purpose of this report is to inform planning discussions about the module’s future—it discusses the costs and benefits of a third wave of data collection, whether the survey module should be modified, and whether experiments should be done to improve the module should it become permanent.
This brief report is intended to fulfill only one narrow aspect of the panel’s broader task as described in Box 1-1. It provides (1) an overview of the ATUS and the SWB module; (2) a brief discussion of research applications to date; and (3) preliminary assessment of the value of SWB module data. The panel’s final report will address issues of whether research has advanced to the point that SWB measures—and which kinds of measures—should be regularly included in major surveys of official statistical agencies to help inform government economic and social policies.
1This section draws heavily from a presentation to the panel by Rachel Kranz-Kent of BLS, and from the Federal Register, Volume 76, Number 134 (July 13, 2011): http://webapps.dol.gov/federalregister/HtmlDisplay.aspx?DocId=25169&AgencyId=6&DocumentType=3 (accessed on August 24, 2012).
An ad hoc panel will review the current state of research and evaluate methods for the measurement of subjective well-being (SWB) in population surveys. On the basis of this evaluation, the panel will offer guidance about adopting SWB measures in official government surveys to inform social and economic policies. The study will be carried out in two phases. The first phase, which is the subject of this statement of task, is to consider whether research has advanced to a point that warrants the federal government collecting data that allow aspects of the population’s SWB to be tracked and associated with changing conditions. The study will focus on experienced well-being (e.g., reports of momentary positive and rewarding, or negative and distressing, states) and time-based approaches (some of the most promising of which are oriented toward monitoring misery and pain as opposed to “happiness”), though their connection with life-evaluative measures will also be considered. Although primarily focused on SWB measures for inclusion in U.S. government surveys, the panel will also consider inclusion of SWB measures in surveys in the United Kingdom and European Union, in order to facilitate cross-national comparisons in addition to comparisons over time and for population groups within the United States. The panel will prepare a short interim report on the usefulness of the American Time Use Survey SWB module, and a final report identifying potential indicators and offering recommendations for their measurement. A later, separate second phase will seek to develop a framework modeled on the National Income and Product Accounts to integrate time-based inputs and outputs, and SWB measures, into selected satellite, or experimental, subaccounts.
The ATUS is the first federally administered, continuous survey on time use in the United States (and in the world). It is designed to obtain estimates of the time spent by respondents in childcare, at work, traveling, sleeping, volunteering, engaged in leisure pursuits, and a wide range of other activities. Time-use data augment income and wage data for individuals and families that analysts can use to create a more complete picture of quality of life in a society. Along with income and product data, information about time-use patterns is essential for research that evaluates the contribution of nonmarket work to national economies. The data also enable comparisons between nations that have different mixes of market and nonmarket production modes. To illustrate, the households of two countries may enjoy similar home services and amenities—quality of meals, level of home cleaning and maintenance, elder and child care, etc.—but one may perform more of these tasks themselves (home production) while the other may more typically hire the tasks out in the market. The latter economy will register higher per capita gross domestic product even though the standard of living may be comparable in the two countries. Relatedly, countries may vary in the amount of time that individuals must work to achieve a given material standard of living, resulting in different amounts of leisure. This difference would also not show up directly in market (only) measures of economic activity, yet it is likely that it affects well-being.
The ATUS provides nationally representative estimates of how people spend their time. It has been conducted continuously since 2003. The survey sample is a repeated cross-section of individuals who are drawn from U.S. households completing their eighth and final month of interviews for the Current Population Survey (CPS). One individual from each household is
selected to take part in one computer-assisted telephone interview. Respondents are interviewed for the ATUS between two and five months after they rotate out of the CPS.
Interviewers ask respondents to report all of their activities for one specified 24-hour day, the day prior to the interview. Respondents also report who was with them during activities, where they were, how long each activity lasted, and if they were paid. For the ATUS (following the core time diary questions but prior to the SWB module) some of the CPS information—for example about who is living in the household and labor force status—is confirmed and updated.2 Measurement of socioeconomic well-being based on the ATUS is enhanced by its connection to the CPS which is rich in socio-demographic variables—namely, characteristics of the individual and the household including labor force status, income, state of residence, educational attainment, race and ethnicity, nativity, detailed marital status (divorced, never married, etc.), and disability status.3
The SWB module adds to the substantive content of the ATUS by revealing not only what people are doing with their time, but also how they experience their time—specifically how happy, tired, sad, stressed, and in pain they felt while engaged in specific activities on the day prior to the interview.4 This information has numerous practical applications for sociologists, economists, educators, government policy makers, businesspersons, health researchers, and others. The module follows directly after the core ATUS; it was administered on an ongoing basis during 2010 and is being done again during 2012. The module surveys individuals aged 15 and over from a nationally representative sample of approximately 2,190 households each month.
Respondents are asked questions about three activities selected with equal probability from those reported in the ATUS time diary (the well-being module questions are asked immediately after the core ATUS) (see Box 1-2). A few activities—sleeping, grooming, and private activities—are never included in the SWB module. The time diary refers to the core part of the ATUS, in which respondents report the activities they did from 4 a.m. on txshe day before the interview to 4 a.m. on the day of the interview. The precodes listed in Box 1-2 are for activities that are straightforward to code, but they are in no way representative of the full activity lexicon used by ATUS coders. The vast majority of ATUS activities are typed into the collection instrument (verbatim) and then coded in a separate processing step.5 The module also collects data on whether respondents were interacting with anyone while doing the selected activities and how meaningful the activities were to them.
2Technical details of the sample design and the survey methodology can be found in the American Time Use Survey User’s Guide: Understanding ATUS 2003-2011 Available at: http://www.bls.gov/tus/atususersguide.pdf(accessed on September 3, 2012).
3Information about who is living in the household and about labor force status is updated in the ATUS, which is important since the CPS data are a little dated by the time the ATUS interview takes place.
ATUS Question Identifying an Activity
So let’s begin. Yesterday, Monday, at 4:00 a.m., what were you doing?
• Use the slash key (/) for recording separate/simultaneous activities.
• Do not use precodes for secondary activities.
2. Grooming (self)
3. Watching TV
4. Working at main job
5. Working at other job
6. Preparing meals or snacks
7. Eating and drinking
8. Cleaning kitchen
10. Grocery shopping
11. Attending religious service
12. Paying household bills
13. Caring for animals and pets
14. Don’t know/Can’t remember
15. Refusal/None of your business
Respondents are asked to rate, for each of the three randomly selected activities, six feelings—pain, happy, tired, sad, stressed, and meaningful—on a scale from 0 to 6: 0 means the feeling was not present, and 6 means the feeling was very strong.
ATUS SWB Text Asking Respondents to Rate Strength of
Feeling During Specific Activities
Between 12:00 p.m. and 1:00 p.m. yesterday, you said you were eating and drinking. The next set of questions asks how you felt during that particular time.
Please use a scale from 0 to 6, where 0 means you did not otherwise experience this feeling at all and a 6 means the feeling was very strong. You may choose any number 0, 1, 2, 3, 4, 5, or 6 to reflect how strongly you experienced this feeling during this time.
The following health related questions (paraphrased here) are also asked after the three random activity episodes are chosen:
• Did you take pain medication yesterday?
• When you woke up yesterday, how well rested did you feel?
• Do you have hypertension?
• Would you say your health in general is excellent, very good, good, fair, or poor?
This information creates opportunities to analyze interactions between health states and reported assessments of emotional states. This is important because daily experience is linked to health status and other outcomes via channels such as worry and stress on the one hand, and pleasure and enjoyment on the other.
The ATUS SWB module was initially designed to collect information primarily on experienced (“hedonic”) well-being—that is, about people’s emotions associated with a recent time period and the activities that occurred during that period. The hedonic dimension of wellbeing is directly related to the environment or context in which people live—the quality of their jobs, their immediate state of health, the nature of their commute to work, and the nature of their social networks—and is reflected in positive and negative affective states. These kinds of hedonic measures contrast with self-reported assessments of overall life satisfaction or happiness. Such “evaluative” well-being measures are more likely to reflect people’s attitudes about their lives as a whole.
The first, 2010, module included only hedonic measures. The second wave (conducted in 2012) includes two additional questions, one on overall life satisfaction and one on whether or not recent emotional experience was typical. The life satisfaction responses are collected using the Cantril ladder scale.6 As noted on the BLS supporting statement for the project (p. 2), asking the Cantril ladder question enables researchers “to build a link between time use and day reconstruction methods of measuring well-being on the one hand, and standard life evaluation questions on the other … a direction of research that has not been possible to date.” The life evaluation question enhances the value both of the ATUS supplement and other surveys that use a Cantril ladder question.
Measurement of both experienced well-being (i.e., reports of momentary positive and rewarding or negative and distressing states) and evaluative well-being (i.e., cognitive judgments of overall life satisfaction or dissatisfaction) extends the policy value of the SWB module data. The value added comes from what can be learned from differences between what the two measures show. For example Kahneman and Deaton (2010, p. 1) find that “emotional well being and life evaluation have different correlates in the circumstances of people’s lives” and particularly striking “differences in the relationship of these aspects of well being to income.”
Distinguishing between different dimensions of well-being also allows investigation of psychological changes associated with aging (e.g., reduced mobility) that might affect both these dimensions of well-being. Another area where the two dimensions provide complementary information is job satisfaction. Getting promoted or obtaining a new job that
6The Cantril Self-Anchoring Scale asks respondents to imagine a ladder with steps numbered from 0 at the bottom to 10 at the top, in which the top of the ladder represents the best possible life for them and the bottom of the ladder represents the worst possible life. They are asked which step of the ladder they personally feel they stand on at this time (for a present assessment). For a good description and discussion of the Cantril Scale, see Diener et al. (2009).
entails long hours might raise a worker’s evaluative well-being, but the associated stress might reduce experienced well-being, at least in the short term. Similar comparisons could be made across professions. Respondents’ reported differences between experience and evaluative measures might also help explain why some people attach high meaning to work, career, and related time commitments while others focus more on simple day-to-day contentment and how or if these correlations vary across age, income, and other demographic or cohort factors. For education research, measures of multiple dimensions of subjective well-being may help provide an understanding of why students make (or do not make) the investments in schooling choices that they do (or do not) make.
The second new question for 2012 asks whether the respondents’ emotional experience yesterday (the day before the interview) was typical for that day of the week:
Thinking about yesterday as a whole, how would you say your feelings, both good and bad, compared to a typical Monday? Were they better than a typical Monday, the same as a typical Monday, or worse than a typical Monday (respondents answer “better,” “the same,” or “worse”).
This question may provide insights about day of week effects and day to day variation in reported well-being scores.
Data from the SWB module supports the BLS mission of providing relevant information on economic and social issues. The data provide a richer description of work experience; specifically, these data describe how individuals feel (tired, stressed, in pain) during work episodes compared to non-work episodes, and how often workers interact on the job. Data from the module can also be used to measure whether the amount of physical pain that workers experience varies by occupation and disability status. The fact the SWB module can be linked to demographic characteristics of respondents—labor force status, occupation, earnings, household composition, school enrollment status, and other characteristics captured on the core ATUS and CPS—opens up a wide array of possible studies on the correlates of self-reported well-being.7
Collection of data on subjective well-being also supports the mission of the module's sponsor, the National Institute on Aging (NIA), to improve the health and well-being of older Americans. Examples of questions that can be answered include:
• Do older workers experience more pain than younger workers on and off the job?
• Is the age-pain gradient related to differences in activities or differences in the amount of pain experienced during a given set of activities?
• Do those in poor health spend time in different activities relative to those in good health?
To date, much of the research on nonmarket components of health and well-being has been informed by global assessments of positive or negative affect averaged over time that are divorced from measures of time use or context. Nor has that research typically addressed age differences or age-related changes in these associations. In this vein, data from the SWB
7In addition, because the ATUS is conducted through the year, it is possible to study seasonal effects on well-being—a topic of interest in a number of research areas.
module might inform policies on redesigning cities to support healthy aging, the allocation of funds to programs that affect older populations, and changes to the health care system to support better maintenance of good health. Researchers have already begun to explore which aspects of experienced and evaluative well-being, time use, and context promote or impede healthy aging. Further work can be done to examine the unique correlative and predictive associations of evaluated and experienced well-being with health and with differences related to life stage, retirement status, and individual characteristics.