From Tom Rees of Epiphenom:

Are religious people more or less anxious? The problem is not as simple as it sounds. In general, religion is supposed to make people less anxious. But, partly for this reason, the people who turn to religion are more anxious to start with. What’s more, all religions are not the same, and different aspects of religion might have different effects.
It’s a surprisingly under-researched topic, but a couple of new papers have looked into it, including this one from Northern Ireland.
The political landscape in Northern Ireland is marked by a sharp sectarian divide between Protestants and Catholics. What Chris Lewis and his colleagues found was that female Catholics were the most anxious, and they also went to church the most often. Counterintuitively, they also showed that going to church was most strongly linked to less anxiety in Catholic women!
The study looked at data from the 2001 Health and Social Wellbeing Survey. It included the 12-item General Health Questionnaire, which is the gold-standard measure of anxiety. This survey found that people in Northern Ireland tend to report more anxiety than do people living in the rest of the United Kingdom (here’s a detailed report, if you’re interested).
Lewis and his colleagues split the survey group into four: male and female, Protestant and Catholic. They found that men had lower anxiety scores than women, and Protestants had lower anxiety scores than Catholics. These effects were additive: Male Protestants were the least anxious, and female Catholics were the most anxious. The average differences were small (about 1.5 on a 36-point scale), but statistically significant.
Churchgoing habits matched this pattern exactly. Male Protestants went to church least often (every few months, on average), and female Catholics went the most often (every two weeks, on average).
Then the researchers looked at the correlation within these groups. What they found was that within each group, people who went to church more often were less anxious. Male Protestant churchgoers were less anxious than male Protestant non-churchgoers, and female Catholic churchgoers were less anxious than female Catholic non-churchgoers (although still more anxious than male Protestant non-churchgoers).
Now, the effect was pretty tiny. But what was interesting was that the strength of the effect followed the same pattern as for anxiety and churchgoing across the four groups. In other words, going to church had the biggest effect on reducing anxiety among female Catholics and the smallest effect among male Protestants.
What to make of all this?
The simplest explanation is that being female and/or Catholic in Northern Ireland is a risk factor for anxiety. As a result, many Catholic women turn to the Church, and those who do have their anxiety levels reduced.