Is Atheism Only for the Upper Class? Socioeconomic Differences Among the Religiously Unaffiliated

[struggle4progress]

42. pdf of whole survey here:

http://publicreligion.org/site/wp-content/uploads/2012/10/AVS-2012-Pre-election-Report-for-Web.pdf

... Today, roughly 1-in-5 (19%) Americans currently report that they are religiously unaffiliated ... Interviews were conducted by telephone among a random sample of 3,003 adults 18 years of age or older in the continental United State ...

So the sample includes about 570 religiously unaffiliated. The HuffPo article asserts

... In 2012, the American Values Survey identified three distinct groups among the unaffiliated: secular Americans (39 percent), self-identified atheists and agnostics (36 percent), and unattached believers (23 percent) ... Nearly half (45 percent) of atheists and agnostics have at least a 4-year college education, compared to 27 percent of secular Americans and just 17 percent of unattached believers. Similarly, 45 percent of atheists and agnostics have household incomes of at least $75,000 a year. Less than 3-in-10 secular Americans (28 percent) and roughly 1-in-5 (21 percent) unattached believers have household incomes in this range ...

This translates approximately as:

The 2012 American Values Survey interviewed 222 secular Americans, 205 self-identified atheists and agnostics, and 131 unattached believers. 92 of the atheists and agnostics had at least a 4-year college education, compared to 60 of the secular Americans and just 22 of the unattached believers. Similarly, 92 of the atheists and agnostics had household incomes of at least $75,000 a year. 62 of the secular Americans and 27 of the unattached believers had household incomes in this range

So the percentage of folk in each category with household income of at least $75K is about the same as the percentage of folk in each category with at least a 4-year college education. That seems credible, and it might suggest fairly little influence on economic success, other than education

I'm not so sure about the cross-class comparisons though: the sample size seems a bit small