Would You Be Happier If You Were Richer? A Focusing Illusion

Daniel Kahneman, Alan B. Krueger, David Schkade, Norbert Schwarz and Arthur A. Stone, Science 312:1908–1910, 30 June 2006 (doi:10.1126/science.1129688). Read in full from the UCLA-hosted copy.

The paper the wiki had been describing second-hand for two weeks. focusing-illusion already carried its conclusion, from an essay about it; this is the primary source, with the numbers.

Note the genre. It is a Perspective in a “Life Cycles” special section, not a research report: part review of the well-being literature, part first presentation of the authors’ own survey data. Four pages, 32 references. That matters for how much weight any single figure here can carry.

The claim

High income and good mood are widely believed to go together, and mostly do not. People earning above average report more satisfaction with their lives, are “barely happier than others in moment-to-moment experience, tend to be more tense, and do not spend more time in particularly enjoyable activities.” The effect of income on life satisfaction is also transient.

The experiment that names the illusion

Working women reported the percentage of the previous day they had spent in a bad mood, and also predicted that percentage for people in various circumstances. Predictions were then set against the reports of respondents actually in those circumstances (n between 59 and 237 per actual cell, 83–87 per predicted cell):

Every predicted gap exceeds the actual one at P < 0.001, and the bad mood of people in the worse circumstance is overestimated across the board. The pattern is not “income does nothing” — the poorest group really does report more bad mood — it is that the difference is roughly a third of what observers expect.

Where the two measures diverge

740 women in Columbus, Ohio completed the Day Reconstruction Method in May 2005. Correlations with household income:

• life satisfaction 0.32
• share of the day in a good mood 0.20
• duration-weighted “happy” 0.06 (not significant)

Marriage, education, employment and body mass index all thin out the same way, so the effect belongs to the measure, not to income. In the Cornell Work-Site Blood Pressure Study (374 workers sampled every ~25 minutes through a workday), personal income against average happiness came out at 0.01 (P = 0.84) while family income correlated with angry/hostile (0.14), anxious/tense (0.14) and excited (0.18). Higher income bought more intense negative feeling and more arousal, not more happiness.

Global self-report keeps a visible income gradient: in the 2004 General Social Survey (n = 1173), 22.2% under $20,000 called themselves “very happy” against 42.9% over $90,000 — but the $50,000–89,999 bracket is at 41.9%, so almost the whole gradient is spent below $90,000.

The mechanism the authors add

Their own contribution, beyond reviewing relative-income and adaptation accounts, is time use. From the American Time-Use Survey (3,917 men, 4,944 women), passive leisure falls from 34.7% to 19.9% of men’s waking day between the bottom and top income bands while work and commuting rise from 29.1% to 36.9%. Rated by 810 women on a 0–6 scale, active leisure scores 4.67 happy / 0.92 tense; work and commute score 3.94 / 2.00. So the day a higher income actually buys is one shifted toward the hours that rate worst.

Imagining a raise, the authors write, people picture the plasma TV and the golf; “in reality they should think of spending a lot more time working and commuting and a lot less time engaged in passive leisure.”

Two other numbers, both borrowed

• Easterlin’s finding that a fivefold increase in Japanese real income between 1958 and 1987 came with no rise in average reported happiness.
• Layard’s cross-country claim that life satisfaction stops climbing with GDP per capita above about $12,000.

Both are cited, not measured here.

What it does and does not settle

It supplies the primary evidence focusing-illusion was resting on someone else’s summary of, and a specific instrument (the prediction-versus-report design) for the spoke’s general question about how a question shapes its own answer.

It does not settle the income question. The satiation claim in this literature was contested and then revised by Kahneman himself — see income-happiness-conflict-resolved, which finds the flattening holds only for the least happy respondents. Read forward, the 2006 correlation of 0.06 between income and duration-weighted happiness is the number that later work disputes.

Related

focusing-illusion · income-happiness-conflict-resolved · daniel-kahneman · thinking-fast-and-slow · heuristics-and-biases · kahneman-focusing-illusion · matthew-killingsworth · synthesis