generally overestimated. Second, consistent with the focusing illusion, the predicted prevalence of a bad mood for people with undesirable circum- stances was grossly exaggerated. The focusing illusion helps explain why the results of well-being research are often counter- intuitive. The false intuitions likely arise from a failure to recognize that people do not continu- ously think about their circumstances, whether positive or negative. Schkade and Kahneman (9) noted that, BNothing in life is quite as important as you think it is while you are thinking about it.[ Individuals who have recently experienced a significant life change (e.g., becoming disabled, winning a lottery, or getting married) surely think of their new circumstances many times each day, but the allocation of attention eventually changes, so that they spend most of their time attending to and drawing pleasure or displeasure from expe- riences such as having breakfast or watching television (10). However, they are likely to be reminded of their status when prompted to answer a global judgment question such as, BHow satisfied are you with your life these days?[ The correlation between household income and reported general life satisfaction on a numeric scale (i.e., global happiness as distinct from experienced happiness over time) in U.S. samples typically ranges from 0.15 to 0.30 (11). The relation between global happiness and income for 2004 with data from the General Social Survey (GSS) is illustrated in Table 2. Those with incomes over $90,000 were nearly twice as likely to report being Bvery happy[ as those with incomes below $20,000, although there is hardly any difference between the highest income group and those in the $50,000 to $89,999 bracket. There are reasons to believe that the cor- relation between income and judgments of life satisfaction overstates the effect of income on subjective well-being. First, increases in income have mostly a transitory effect on individuals_ reported life satisfaction (2, 12). Second, large increases in income for a given country over time are not associated with increases in aver- age subjective well-being. Easterlin (1), for example, found that the fivefold increase in real income in Japan between 1958 and 1987 did not coincide with an increase in the average self-reported happiness level there. Third, al- though average life satisfaction in countries tends to rise with gross domestic product (GDP) per capita at low levels of income, there is little or no further increase in life satisfaction once GDP per capita exceeds $12,000 (3). Fourth, when subjective well-being is mea- sured from moment to moment���either by querying people in real time with the Ecolog- ical Momentary Assessment (EMA) technique (13) or by asking them to recall their feelings for each episode of the previous day with the Day Reconstruction Method (DRM) (14)��� income is more weakly correlated with experi- enced feelings such as momentary happiness averaged over the course of the day (henceforth called duration-weighted or experienced happi- ness) than it is with a global judgment of life satisfaction or overall happiness, or with a global report of yesterday_s mood (Table 3) (15, 16). This pattern is probably not a result of greater noise in the duration-weighted happiness mea- sure than in life satisfaction (17). Other life cir- cumstances, such as marital status, also exhibit a weaker correlation with duration-weighted happiness than with global life satisfaction. An analysis of EMA data also points to a weak and sometimes perverse relation between experienced affect and income. Specifically, we examined EMA data from the Cornell Work- Site Blood Pressure Study of 374 workers at 10 work sites, who were queried about their inten- sity of various feelings on a 0 to 3 scale every 25 min or so during an entire workday (18). The correlation between personal income and the average happiness rating during the day was just 0.01 (P 0 0.84), whereas family income was significantly positively correlated with ratings of angry/hostile (r 0 0.14), anxious/tense (r 0 0.14), and excited (r 0 0.18). Thus, higher income was associated with more intense nega- tive experienced emotions and greater arousal, but not greater experienced happiness. Why does income have such a weak effect on subjective well-being? There are several ex- planations, all of which may contribute to vary- ing degrees. First, Duesenberry (19), Easterlin (2), Frank (20), and others have argued that relative income rather than the level of income affects well-being���earning more or less than others looms larger than how much one earns. Indeed, much evidence indicates that rank within the income distribution influences life satisfac- tion (21���23). As society grows richer, average rank does not change, so the relative income hypothesis could explain the stability of av- erage subjective well-being despite national income growth. The importance placed on rel- ative income may also account for the stronger correlation between income and global life satisfaction than between income and experi- enced affect, as life satisfaction questions prob- ably evoke a reflection on relative status that is not present in moment-to-moment ratings of affect. The relative income hypothesis cannot by itself explain why a permanent increase in an individual_s income has a transitory effect on her well-being, as relative standing would increase. However, the increase in relative standing can be offset by changes in the reference group: After a promotion, the new peers increasingly serve as a reference point, making the improvement rela- tive to one_s previous peers less influential (24). Second, Easterlin (1, 2) argues that individuals adapt to material goods, and Scitovsky (25) argues that material goods yield little joy for Table 2. Distribution of self-reported global happiness by family income, 2004. The GSS question posed was ������Taken all together, how would you say things are these days���would you say that you are very happy, pretty happy, or not too happy?������ Sample size was 1173 individuals. Response Percentage indicating global happiness at family income of Under $20,000 $20,000���$49,999 $50,000���$89,999 $90,000 and over Not too happy 17.2 13.0 7.7 5.3 Pretty happy 60.5 56.8 50.3 51.8 Very happy 22.2 30.2 41.9 42.9 Table 3. Correlations between selected life circumstances and subjective well-being measures. The question posed was ������We would like to know how you feel and what mood you are in when you are at home. When you are at home, what percentage of the time are you in a bad mood____%, a little low or irritable____%, in a mildly pleasant mood____%, in a very good mood____%." The last two response categories were added together to obtain the percentage of time in a good mood. Duration-weighted ������happy������ is the average of each person���s duration-weighted average rating of the feeling happy over episodes of the day, where 0 refers to ������not at all������ and 6 refers to ������very much,������ and each individual���s responses were weighted by the duration of the episode. Sample consists of 740 women from Columbus, Ohio, who completed the DRM in May 2005 (16). Characteristic Life satisfaction Amount of day in good mood (%) Duration- weighted "happy" Household income 0.32*** 0.20*** 0.06 Married 0.21*** 0.15*** 0.03 Years of education 0.16*** 0.13*** 0.03 Employed 0.14*** 0.12** 0.01 Body mass index ���0.13*** ���0.08* ���0.06 *P G 0.05 **P G 0.01 ***P G 0.001. SPECIALSECTION www.sciencemag.org SCIENCE VOL 312 30 JUNE 2006 1909 on December 13, 2006 www.sciencemag.org Downloaded from

most individuals. Thus, increases in income, which are expected to raise well-being by raising consumption opportunities, may in fact have little lasting effect because of hedonic adaptation or because the consumption of material goods has little effect on well-being above a certain level of consumption (26). Moreover, people_s aspirations adapt to their possibilities and the income that people say they need to get along rises with in- come, both in a cross section and over time (27). Finally, we would propose another explana- tion: As income rises, people_s time use does not appear to shift toward activities that are associated with improved affect. Subjective well-being is connected to how people spend their time. In a representative, nationwide sam- ple, people with greater income tend to devote relatively more of their time to work, compul- sory nonwork activities (such as shopping and childcare), and active leisure (such as exercise) and less of their time to passive leisure activities (such as watching TV) (Table 4). The activities that higher-income individuals spend relatively more of their time engaged in are associated with no greater happiness, on average, but with slightly higher tension and stress. The latter finding might help explain why income is more highly correlated with general life satisfaction than with experienced happiness, as tension and stress may accompany goal attainment, which in turn contributes to judgments of life satisfaction more than it does to experienced happiness. The results in Table 4 also highlight the pos- sible role of the focusing illusion. When some- one reflects on how additional income would change subjective well-being, they are probably tempted to think about spending more time in leisurely pursuits such as watching a large- screen plasma TV or playing golf, but in reality they should think of spending a lot more time working and commuting and a lot less time engaged in passive leisure (and perhaps a bit more golf). By itself, this shift in time use is unlikely to lead to much increase in experienced happiness, although it could increase tension and one_s sense of accomplishment and satisfaction. Despite the weak relation between income and global life satisfaction or experienced happiness, many people are highly motivated to increase their income. In some cases, this focusing illusion may lead to a misallocation of time, from ac- cepting lengthy commutes (which are among the worst moments of the day) to sacrificing time spent socializing (which are among the best moments of the day) (28, 29). An emphasis on the role of attention helps to explain both why many people seek high income���because their predictions exaggerate the increase in happiness due to the focusing illusion���and why the long- term effect of income gains become relatively small, because attention eventually shifts to less novel aspects of daily life. References and Notes 1. R. Easterlin, J. Econ. Behav. Organ. 27, 35 (1995). 2. R. Easterlin, ������Building a better theory of well-being,������ Discussion Paper No. 742, IZA, Bonn, Germany, 2003. 3. R. Layard, Happiness: Lessons from a New Science (Penguin Press, London, 2005). 4. D. Gilbert, Stumbling on Happiness (Knopf, New York, 2006). 5. F. Strack, L. Martin, N. Schwarz, Eur. J. Soc. Psychol. 18, 429 (1988). 6. N. Schwarz, F. Strack, H. Mai, Public Opin. Q. 55, 3 (1991). 7. D. Smith, N. Schwarz, T. Roberts, P. Ubel, Qual. Res. 15, 621 (2006). 8. N. Schwarz, F. Strack, in Well-Being: The Foundations of Hedonic Psychology, D. Kahneman, E. Diener, N. Schwarz, Eds. (Russell Sage Foundation, New York, 1999), pp. 61���84. 9. D. Schkade, D. Kahneman, Psychol. Sci. 9, 340 (1998). 10. D. Kahneman, R. H. Thaler, J. Econ. Perspect. 20 (1), 221 (2006). 11. E. Diener, R. Biswas-Diener, Soc. Indic. Res. 57, 119 (2002). 12. B. Frey, A. Stutzer, Happiness and Economics: How the Economy and Institutions Affect Well-Being (Princeton Univ. Press, Princeton, NJ, 2002). 13. A. Stone, S. Shiffman, Ann. Behav. Med. 16, 199 (1994). 14. D. Kahneman, A. Krueger, D. Schkade, N. Schwarz, A. Stone, Science 306, 1776 (2004). 15. In general, we find that the retrospective report of mood on the previous day, which is a global evaluation, shares variance both with the global measures of life satisfaction and with disaggregated measures of emo- tional experience at particular times. 16. D. Kahneman, D. Schkade, C. Fischler, A. Krueger, A. Krilla, ������A study of well-being in two cities,������ Discussion Paper No. 53, Center for Health and Wellbeing, Princeton, NJ, 2006. 17. We conducted a reliability study of the DRM that asked the same questions of 229 women two weeks apart, and found about the same two-week serial correlation in duration-weighted happiness as in life satisfaction for the respondents. 18. P. Schnall, J. Schwartz, P. Landsbergis, K. Warren, T. Pickering, Psychosom. Med. 60, 697 (1998). 19. J. Duesenberry, Income, Saving, and the Theory of Consumer Behavior (Harvard Univ. Press, Cambridge, MA, 1949). 20. R. Frank, Luxury Fever (Princeton Univ. Press, Princeton, NJ, 1999). 21. A. Clark, A. Oswald, J. Public Econ. 61, 359 (1996). 22. A. Ferrer-i-Carbonell, J. Public Econ. 89, 997 (2005). 23. E. Luttmer, Q. J. Econ. 120, 963 (2005). 24. W. Runciman, Relative Deprivation and Social Justice (Univ. of California Press, Berkeley, CA, 1966). 25. T. Scitovsky, The Joyless Economy (Oxford Univ. Press, Oxford, 1976). 26. S. Frederick, G. Loewenstein, in Well-Being: The Founda- tions of Hedonic Psychology, D. Kahneman, E. Diener, N. Schwarz, Eds. (Russell Sage Foundation, New York, 1999), pp. 302���329. 27. B. Van Praag, P. Frijter, in Well-Being: The Foundations of Hedonic Psychology, D. Kahneman, E. Diener, N. Schwarz, Eds. (Russell Sage Foundation, New York, 1999), pp. 413���433. 28. See (31) for evidence on the misallocation of commuting time and (14) on the hedonic experience of commuting and socializing. 29. It goes without saying that happiness is not the only measure of human welfare. Moreover, although income gains may not contribute very much to experienced happiness or life satisfaction, wealthier societies may well enjoy better health care, safer and cleaner environments, cultural benefits and other amenities that improve the quality of life. 30. ������Time-use survey���First results announced by Bureau of Labor Statistics,������ U.S. Department of Labor, USDL 04-1797 (http://www.bls.gov/). 31. A. Stutzer, B. Frey, ������Stress that doesn���t pay: The commuting paradox,������ Discussion Paper No. 127, IZA, Bonn, Germany, 2004. 32. The authors thank M. Connolly, M. Fifer, and A. Krilla for research assistance, and the Hewlett Foundation, the National Institute on Aging, and Princeton University���s Woodrow Wilson School and Center for Economic Policy Studies for financial support. 10.1126/science.1129688 Table 4. How is time spent and do the activities bring happiness? Time allocation is weighted-average percentage of the nonsleep day for each sampled observation from the American Time-Use Survey (30). Weighted average of weekday (5 out of 7) and weekend (2 out of 7) is presented. Sample consists of 3917 men and 4944 women age 18 to 60. Last two rows were computed by authors from a DRM survey of 810 women in Columbus, Ohio, in May 2005 if multiple activities were performed during an episode, the activity refers to the one that was selected as ������most important.������ Family income/Gender Active leisure Eating Passive leisure Compulsory Work and commute Other Men Time allocation (%) G$20,000 6.6 6.6 34.7 20.8 29.1 2.1 $20,000���$99,999 8.1 7.2 26.4 21.8 35.4 1.1 $100,000+ 10.2 8.6 19.9 23.6 36.9 0.8 Women G$20,000 5.3 5.7 33.5 35.6 18.5 1.4 $20,000���$99,999 7.5 6.7 23.8 34.3 26.7 1.0 $100,000+ 9.1 7.0 19.6 35.9 27.3 1.1 Women Feelings (0-6 scale) Happy 4.67 4.45 4.21 4.04 3.94 4.25 Tense/Stressed 0.92 1.17 1.30 1.80 2.00 1.61 L I F E C Y C L E S 30 JUNE 2006 VOL 312 SCIENCE www.sciencemag.org 1910 on December 13, 2006 www.sciencemag.org Downloaded from