As you may or may not know, when I'm not ranting about my parents, I'm a sociologist. Specifically, I'm a grad student in sociology: I have a MA, and am planning on completing my PhD either next May or August depending on how things go (if I don't get this professorship job I just interviewed for I may end up staying in grad school until 2010).
The specific topic I study is family and gender, and my dissertation is about cohabitation in the United States. I do my research using large scale nationally-representative datasets.
Right now I am working on an index of religiosity- that is, a measure of how important religion is to the daily lives of people in my dataset. I want to account for this in my statistical models, since religiosity is highly correlated with cohabitation.
So here are some brand spanking new numbers I just calculated using the National Survey of Family Growth dataset. This dataset was collected in 1995 and 2002, is nationally representative, and specifically surveyed women age 15-49 (women of childbearing ages). So these numbers are only true for women age 15-49 in the US.
For all women age 15-49:
Religion very important: 57.04
Religion somewhat important: 36.45%
Religion not important: 6.5%
I'm specifically looking at women who have ever been married (so may be divorced or widowed at the time of survey, but were married at least once). For women age 15-49 who have ever been married:
Religion very important:60.90%
Religion somewhat important: 33.97%
Religion not important: 5.13%
For women age 15-49 who have ever been married and didn't cohabit before marriage:
Religion very important:68.03%
Religion somewhat important: 28.31%
Religion not important: 3.66%
For women age 15-49 who have ever been married and who DID cohabit before marriage:
Religion very important:51.29%
Religion somewhat important: 41.60%
Religion not important: 7.12%
It seems that not being religious is a pretty rare state for women of childbearing ages in the United States.