Charles Murray’s new book Coming Apart supposedly deals with the increasing divergence within the white community over the period 1950-2010. It compares two groups of white people in the 30-49 age range: upper middle class (defined as those with a least a bachelor’s degree and working as a manager or similar high-status professions) with working class (no more than a high school degree and those who work in blue-collar, service or low-skill jobs).

The book has been criticized but not panned nearly as much as his infamous The Bell Curve (which I too critiqued in some detail here and in my book The Achievement Gap in US Education: Canaries in the Mine (2005)). One major criticism is that the two groups being contrasted may represent just a small and decreasing part of the population. In Jared Bernstein’s detailed review, he points out that while Murray claims that these two groups represent the top 20% and the bottom 30%, that still leaves half the population out. But Bernstein says that in reality, Murray is actually looking at subgroups within these subgroups and it is not clear how big those sub-sub-groups actually are.

I have not read his new book (I decided after The Bell Curve that life was too short to spend much time on Murray’s shoddy scholarship) but was interested in one piece of data that The Plain Dealer of Sunday, March 4 had an infographic on that was supposedly obtained from the book, which in turn was taken from the General Social Survey (GSS). What the infographic showed was that from 1972 to 2010, the number of nonreligious amongst the upper middle class rose from 27% to 40% but that of the working class rose from 35% to 59%. It identified the ‘nonreligious’ as those who do not have a religion or do not attend services more than once a year.

This is good news for three reasons.

First, the number of nonreligious he reports in the US was surprising high. I have not seen any study elsewhere that indicated such high levels of nonreligiosity.

Second, it breaks down the absurd stereotype that nonbelief is some elite affectation, a kind of luxury that only those with advanced education and time to ponder the meaning of life succumb to. This position is sometimes advanced by sophisticated religious people and is usually accompanied by the patronizing and self-serving assumption that we atheists should be cautious about disturbing the religious beliefs of the poor since it is belief in a god that enables them to endure the hardships of their lives.

The third reason to welcome this result is that since the size of the working class is much larger than that of the upper middle class and forms the backbone of religious institutions, this means that their disaffection from religion creates a much more dire situation for those institutions.

In general, one has to be careful about drawing inferences about levels of religiosity based on survey responses because the results tend to be all over the map, depending crucially on how the surveys are done and how the questions are worded. While there is a definite and unmistakable trend to less religiosity over time and to greater skepticism among the young (and I will look more closely in a future post at a recent study that supports that view), the degree and extent of such changes as quoted by different studies is quite variable.

I tend to be more skeptical about surveys that ask questions that require a high-level inference by responders or are open to wide interpretation, such as whether people ‘believe’ in a god or in the afterlife. People can interpret the word ‘believe’ in a variety of ways, all the way from ‘I am sure of it and am certain that I will be playing golf with Jesus in heaven immediately after I die’ all the way down to ‘yeah, kinda-sorta, I guess, I don’t see why not’ equivocation. Since societal pressures tend to make people less willing to publicly declare unbelief, they are likely to take refuge in whatever meaning of the word ‘belief’ they think is likely to be congenial to the questioner and those around them.

Questions that are more concrete and involve simply recall of numbers or require low-level inference (How many times per month do you go to a house of worship? How frequently do you pray? When was the last time you read your holy book? How much do you contribute to support your religious institution?) are likely to provide more accurate measures of religiosity. Asking people to keep a diary or journal where they record such actual practices is even better. Whenever researchers have done so, the results show a lower level of religious fervor than people are willing to admit. As I said before, I am more interested in the level of allegiance that people have to their religious institutions and leaders than to vague beliefs, and these kinds of questions provide a better measure.

So the best results are obtained from those studies that ask people to keep a diary or journal or have some other means of tracking actual behavior. One step up from that are surveys that ask people low-level inference questions that ask questions that require concrete, preferably numerical, responses. The least reliable are vague questions about belief. However even the last kind can be useful as measures of longitudinal trends, if people are asked the same questions over time.