Most happiness research is based on questions which ask respondents about how happy they are or how satisfied they are with their lives in general. A number of papers over the years have explored the links between these overall ratings and people’s day-to-day emotional states and found that (at least in the survey period) there are only modest correlations between them.
The most recent paper, by Daniel Kahneman and Angus Deaton, using a massive 450,000 person US sample over 2008-09 (bad years economically in the US), again finds that the statistical relationship isn’t strong.
They used three measures of daily emotional well-being: positive affect (reports of happiness, enjoyment and frequent smiling), blue affect (reports of worry and sadness) and stress the day before the survey. The life satisfaction question asked respondents to rate their lives from 0, worst possible life, to 10, best possible life. None of the correlations between emotional well-being and life satisfaction were higher than .31 (0 would be no relationship, 1 would be a complete association).
The relationships between the measures of well-being and income also show different patterns. For daily emotional well-being, the income effect plateaus at about $60,000 household income a year for stress free (proportion of population reporting no stress the previous day) and about $75,000 a year for positive affect and blue affect. But for overall life satisfacton, higher income continues to be associated with higher ratings.
So it seems that the distribution of day-today emotions is more egalitarian than the distribution of life satisfaction.
Drilling down further, there are some more interesting results: being a graduate is beneficial for life satisfaction but is associated with only very small increases in postive affect and reductions in blue affect, and is linked to increases in stress. Being a parent is associated with a negative effect on life satisfaction but a small increase in positive effect (I thought it might be the other way around – having kids fulfils a life goal but does not otherwise make life better). Kids mean more stress.
For the politics and economics of happiness, these findings complicate matters considerably. Most policy recommendations are based on life satisfaction findings. But are we most concerned about day-to-day emotional states or about people’s overall assessment of their lives?