One of the steadier findings in well-being research is that religious people report doing better. The harder question — the one the American Affairs essay on the happiness gap presses — is what kind of “better.” Well-being is not one thing: there is the Evaluative rating you give your life, the Experiential feel of your recent days, and the Eudaimonic sense that what you do is worthwhile. The European Social Survey measures all three, which lets us ask which of them religious practice actually tracks — in Europe, where regular attendance is the exception rather than the rule.
The naive answer is misleading, because the people who fill the pews are, on average, older, poorer and more rural — all of which colour well-being on their own. So every figure here is measured within country and adjusted for age, gender and household income, and each outcome is put on a common standard-deviation scale. The result is a sharp division of labour. Going from never attending to attending every day is associated with +0.36 standard deviations more meaning, +0.30 in feeling life is worthwhile, and +0.28 in life satisfaction. On how the week actually felt — mood, depression, anxiety — it buys almost nothing.
What religious attendance is — and isn’t — associated with
Adjusted within-country association of going from never to daily attendance with each outcome, in standard deviations. Bars are 95% intervals. Colour marks the construct.
The contrast is clean because the experiential estimates are not just small, they are indistinguishable from zero: the intervals for mood, depression and anxiety all cross the no-effect line. Whatever regular worship does for the people who practise it, it is not making their ordinary days feel calmer or less low. What it tracks is the slower, more cognitive registers of well-being — the judgment that life is worthwhile and, to a lesser degree, the overall rating of that life.
You can see why adjustment was necessary by looking at the raw gradient, before any controls. Feeling life is worthwhile climbs steadily with attendance — but the share who avoided feeling depressed actually falls a little among the most observant, because they are older and poorer. Net those things out and the depression line flattens to nothing; the meaning line stays standing.
The raw gradient, before adjustment
Weighted means by attendance band, each measure standardized across the four bands so they share a scale. The confounding the adjusted model removes is visible here.
What this shows — and what it doesn’t
This is association, within country and net of the obvious confounders, not proof of cause — the devout differ from the secular in ways no survey fully captures, and the arrow could run either way (people in search of meaning may seek out religion). The measures are self-reports, and 2012 is a single snapshot. But the pattern is specific enough to be useful, and it lines up with how the well-being constructs are supposed to behave: religion is, among other things, a meaning-making institution, and it is on the meaning register — not the daily-mood register — that its signal is loudest. That it also tracks the evaluative life-rating, while leaving felt mood untouched, is the same lesson the rest of this series keeps returning to: “does religion make you happier” has no single answer until you say which happiness you mean.
Notes & data
- Source. European Social Survey (ESS), round 6 (2012); the round carrying all three construct batteries. Weighted with the analysis weight. 43,679 respondents with complete attendance, age, gender and income data.
- Model. For each outcome, a weighted least-squares regression of the standardized outcome on religious attendance (
attend_freq, 1=never .. 7=every day) with country fixed effects and controls for age, age², gender and household-income decile. The reported figure is the attendance coefficient × 6 (never → every day), in outcome standard deviations, with 95% intervals. - Constructs. Evaluative life satisfaction; Experiential felt-mood index, not-depressed, not-anxious (past-week items); Eudaimonic meaning index and “life feels worthwhile.” Labeled throughout; never merged.
- Why adjust. Frequent attenders are older, poorer and more rural; the raw gradient (second figure) confounds those with religion. Country fixed effects also remove differences between secular and observant nations, so the slope is within-society.
- Limits. Cross-sectional, associational, self-reported; reverse causation (seeking meaning → attending) is not ruled out. Companion pieces map the same constructs across countries.
- Prior art. Putnam & Campbell, American Grace (religious social ties and well-being); Oishi & Diener on meaning; Renn, American Affairs (2023).