Ask a Gallup interviewer in Dhaka or Dakar whether religion is an important part of your daily life and almost everyone says yes. Ask in Stockholm or Prague and almost no one does. That single contrast — how normative faith is — has long been suspected of changing what religion does for a person. The sociologist’s claim, sharpest in the work of Olga Stavrova and colleagues, is that religion repays the believer most where belief is expected, and little or nothing where it is unusual.
To test it cleanly we do not compare a religious person in one country to a secular person in another. Instead we fit a regression inside each country, weighted by Gallup’s within-country weight, that asks: holding age, sex, education, marital status, household size and income rank constant, how much higher is the outcome for someone who says religion matters? Call that one number the country’s religiosity premium, θ. Do it once per country and you get a distribution of premiums — one per nation — whose spread is itself the evidence. Then regress those premiums on how religious each country is. The “slope of the slopes” is the whole argument in a single coefficient.
The premium and the norm meta-regression · one dot per country
x-axis: the national religiosity norm (share for whom religion is important). y-axis: that country’s within-country religiosity premium θ. The flat baseline is “no premium.”
The shape of the spread distribution of θ · the secular negatives are the point
Each tick is a country’s premium, stacked by value and coloured by sign. The vertical line marks the population-weighted “representative person.”
Where the premium lives choropleth · diverging scale centred at zero
The within-country life-rating premium, country by country. Teal where the religious rate life higher; orange where they rate it lower. Grey = suppressed or unmapped.
What this does and doesn’t show
Three cautions keep the reading honest. First, this is association, not causation: religiosity is woven into national context, and the same forces that make a society devout may independently shape its well-being. The premium is what religion marks, not proof of what it causes. Second, cross-cultural response styles differ — people in some places use the high end of every scale — so we lean on within-country contrasts, which difference such styles away, and we treat small gaps cautiously. Third, the effects are modest: even the steepest premiums move a life rating by a fraction of a point, not a wholesale transformation.
The cleanest summary is a split verdict. For the life rating — the evaluative question of how good your life is, all things considered — faith’s payoff is real on average but flat against the national norm; the devout-society advantage the literature predicted does not appear here. For the experiential side — whether yesterday felt heavy, whether someone has your back — the social-norm story holds with force, and it holds even after adjusting for how developed a country is. Religion’s gift, on this reading, is less a verdict on life than a texture of daily life, and that texture is richest where everyone around you shares it.