Quoting Bejamin Beit-Hallahmi’s textbook Psychological Perspectives on Religion and Religiosity, pages 123-125.

The challenge of aggregate measures

The findings on the largely beneficial effects of religiosity on individuals have been challenged by the different picture that emerges from research that looks at individuals in the aggregate. While the literature reports that more religious individuals are healthier, better-behaved, and happier, when we look at large groups of individuals living in a particular territory, the picture is completely reversed.

That is how the “religious engagement paradox” is described:

Curiously, irreligious places (nations, states) and highly religious individuals tend to exhibit high levels of health, wellbeing, and prosociality. Religious engagement correlates negatively with prosociality and well-being across aggregate levels (countries and American states), and positively across individuals (especially, as noted earlier, in more religious countries). (Myers, 2012a, p. 916)

This paradox persists whenever we look at large groups or nations. Religiosity is clearly good for individuals, at least in some ways, but religion as a social force is less beneficial.

As Myers (2012a) noted, irreligious nations exhibit high levels of health, wellbeing, and pro-sociality. Koenig and Larson (2001) noted that the vast majority of Scandinavians are atheists or nonreligious. At the same time, Scandinavian nations tend to experience the highest levels of subjective well-being (SWB) in the world (Diener, Kahneman and Helliwell, 2010; Zuckerman, 2008, 2009). The Columbia University Earth Institute published in 2013 its World Happiness Report. This report identifies the countries with the highest levels of happiness in the 2010–2012 surveys as Denmark, Norway, Switzerland, the Netherlands, Sweden, and Canada. The United Nations Human Development Index, published since 1990, which combines measures of life expectancy, literacy, education, living standards, and GDP per capita, and other global indices of the quality of life, published annually, always list less religious nations (Norway, Iceland, Australia, Luxembourg, Canada) at the top.

Similar lists of “the best country to live in” or “the best country for children,” which use data on public health, longevity, education, health care, and income disparity include the familiar names of less religious nations: Norway, Sweden, Australia, Canada, the Netherlands, Iceland, and Belgium. At the bottom of such lists are African countries, Moslem countries (Bangladesh, Pakistan, Egypt, Morocco, and Yemen) and India. Inside the United States, if the fifty states are ranked in terms of average health, Vermont, Hawaii, New Hampshire, Massachusetts, and Minnesota are the top five. Mississippi, Louisiana, Arkansas, West Virginia, and South Carolina make up the bottom five. It turns out that the bottom five states are among the most religious, while the top five are among the more secular (Gray and Wegner, 2010).

Life expectancy is one clear measure of the quality of life in any nation. Looking at the global ranking, it becomes clear that religiosity is negatively correlated with longevity. If religiosity affects longevity positively, do we expect more religious nations to have longer life expectancies? Do the more religious people of Ethiopia live longer than the less religious people of Japan? The answer is obvious. The United States Central Intelligence Agency publishes an annual life expectancy ranking of 223 world nations and territories. Among the top thirty, are, as expected, Japan, Australia, France, Canada, Sweden, Iceland, the Netherlands, and Norway (the United States was number fifty in 2013). The bottom thirty is made up of African countries and Afghanistan, the last well known for religiosity.

The United States is often looked at as a laboratory for testing crucial hypotheses about the effects of religiosity. If religiosity guarantees well-being, the United States should be a happy place, but the reality is somewhat different. In 1997, Ralph Reed, executive director of the Christian Coalition, a right-wing religious lobby, stated that the United States was,

the most devoutly religious nation in the entire world… a nation undergirded by faith, built by faith. and enlivened by faith. It is not a faith in word alone, but an active, transforming faith. Look around today and what you will see are the fruits of our national faith. (Reed, 1997, p. 27)

Nevertheless, Reed described “[s]ocial pathologies once imagined only in our darkest nightmares are a daily reality” (Reed, 1997, p. 28). Compared to other wealthy nations (per capita GNP over $20,000), or other English-speaking nations, the United States has (in addition to economic insecurity) the highest rates of violence and imprisonment, as well as other social pathologies such as the highest teen pregnancy rate.

The findings by Diener, Tay, and Myers (2011), reported above, seem relevant to the aggregate paradox. They showed that in nations which suffer from poverty, hunger, and low life expectancy, religiosity was correlated with SWB, and religious individuals had a higher SWB in poor nations but not in wealthy and secularized nations.