So striking is it how perfect the correlation is for a question with nine possible responses that, apropos of nothing, sharing what I inadvertently stumbled onto seemed worthwhile:
The only confound being avoided here is the racial one, though the results similarly stair step up continuously from “never” to “multiple times a week” without exception when respondents of all races are included. Correlation is obviously not necessarily causation. People who are depressed tend not to do much of anything, including church. But if you’re looking for cheerfulness in your social circle, God’s house may be worth checking out.
GSS variables used: ATTEND, RACECEN1(1), HISPANIC(1)