Maybe religion isn’t dying, but it’s definitely not well

In a Wall Street Journal op-ed, Byron R. Johnson and Jeff Levin, referring to their paper in the Interdisciplinary Journal of Research on Religion (IJRR), argue that religiosity is not declining in the US but rather, “Americans are becoming more religious, and religious institutions are thriving.”

This claim is incredible in the “not credible” sense.

The gist of the argument put forward by Johnson and Levin is that some individuals who report no religious affiliation retain some elements of religiosity, like attending religious services and praying. That is true, and something scholars have known for decades. In fact, in their IJRR paper, they cite a study I co-authored in 2009 as support for this claim.

Rather than recognize that religious affiliation (or lack thereof) is not a perfect predictor of other aspects of religiosity and leave it at that, the authors then deploy a preposterous non sequitur to arrive at one of their key points. Here is the direct quote from the IJRR paper:

If nones are not irreligious, then what are the studies on the “rise of the nones” actually capturing?

Let me make it clear what they did here. The authors noted that some individuals without a religious affiliation exhibit some religious characteristics, then turn that into “nones are not irreligious”—an enormous leap that is not grounded in empirical reality. This faulty generalization serves as the basis of their efforts to muddy the waters on religious decline in the US.

What makes Johnson and Levin’s op-ed and their corresponding IJRR article even more astonishing is that they are claiming religiosity is not declining—a claim that is fundamentally about trends over time—without looking at trends over time. All of their data come from relatively recent cross-sectional surveys. Arguments about trends should employ data on trends. Since trend data is readily available to examine the claims made by Johnson and Levin, it is easy to examine it using one of the datasets they referenced, the General Social Survey (GSS), which has data on religiosity in the US going back to the 1970s.

We can also make this very simple. If multiple measures of religiosity—not just a single measure—show declines over time, we can conclude that religiosity is declining over time.

Figure 1 shows religious affiliation from 1972 to 2018 in the US. Two trends are clearly visible in this figure. The percentage of the US population that is Protestant has declined, from 64% in 1972 to 49% in 2018, a 15% decline. During that same time, those who reported no religious affiliation, the “Nones,” increased from 5 to 23%, a gain of 18%.

One of Johnson and Levin’s claims in their IJRR paper is that many of the Nones should really be classified as members of “other” religions and that minority religions in the US are growing. They are correct that “other” and “minority” religions like Islam have seen some growth, but it is not nearly enough to account for Protestant losses over the last 50 years. In 2017, 1.7% of the US population reported an affiliation with some other religion. In 2018, it was 4.9%, an increase of 3.2%. That doesn’t come close to accounting for Protestant declines and the smaller declines seen among Catholics (26% down to 21%) and Jews (3.4% to 1.7%) over the same time frame.

Affiliation, of course, is just one aspect of religiosity. Are Johnson and Levin correct that Americans are becoming more religious? Religious attendance is an important behavioral aspect of religiosity that speaks to this issue. Figure 2 shows religious attendance from 1972 to 2018. To simplify the chart, I included just three options: those who never attend, those who attend once a month or less, and those who attend more than once a month. In 1972, 9% of Americans never attended religious services. In 2018, 31% of Americans never attended religious services, an increase of 22%. Most of that increase came at the expense of Americans who attended more than once a month, which was 50% in 1972 but has declined to 34% as of 2018. Trend data show that Americans are attending religious services less frequently than they used to.

What about belief in a god? The GSS first introduced a question about belief in a god in 1988, so we have a shorter time span to examine (though Gallup recently found that belief in a god has declined substantially since 1945). Figure 3 illustrates the changes that have occurred over the last thirty years. The clearest trend is the decline in people who indicate they “know god exists” (or theists) which was 64% in 1988 but declined to 54% in 2018. Atheists increased from 1.5% in 1988 to 4.7% in 2018; agnostics increased from 3.8% to 6.5%, and those who believe in a higher power increased from 7.9% to 12.9%. Trend data show Americans are decreasingly likely to report that they “know god exists” and instead are choosing other options, like atheism, agnosticism, or non-specific belief in a “higher power.”

Finally, what about views of the bible?

The question asking about perspectives on the bible was added to the GSS in 1984, so we don’t have as long of trend data on this as we do for affiliation or attendance, but Figure 4 is still informative. Believing that the bible is the literal word of a god has declined from 38% in 1984 to 31% in 2018. Viewing the bible as a collection of fables or myths has increased from 14% in 1984 to 21% in 2018. Trend data show that Americans increasingly see the bible not as scripture but as a collection of myths and fables.

When we examine trend data, pretty much all of it leads to one very obvious conclusion: religiosity in the US is on the decline. I will happily concede, just as I did in 2009, that some people who have no religious affiliation (i.e., “nones”) still believe in a god, attend religious services, and consider the bible to be the literal word of god. No scholars who study secularization or the nonreligious are arguing otherwise. The fact that some Nones exhibit some aspects of religiosity does not undermine the trends that are obvious to anyone with eyes willing to see what is really happening.

What about Johnson and Levin’s claim that even some atheists attend religious services and report a religious affiliation? That too is true. In fact, about one-fourth of the people in the US who report not believing in a god (i.e., atheists) also report having a religious affiliation. There are two ways to interpret this fact. The one Johnson and Levin clearly favor is that some atheists are really religious. The other interpretation is that many of the people who claim a religious affiliation are not particularly religious, and some are atheists who may retain a religious affiliation for the culture, tradition, or music, or to attend with a religious spouse, or because they value the community. I favor the second option because all of the figures above and the data support it.

A simple illustration: Roughly 77% of Americans reported a religious affiliation in 2018 (that has since declined), but just 34% of Americans reported attending religious services more than once a month. There is a gap there. Roughly 43% of Americans said they have a religious affiliation but only rarely attend religious services. If that is the case (and it is), that leads me to wonder: Just how religious are Americans? The low level of religious behavior observed among religiously affiliated Americans is never mentioned in Johnson and Levin’s op-ed or their IJRR article. Their focus is exclusively on how supposedly religious the Nones are.

It is quite rare to find outright empirical denialism among scientists, but it does occur. I still remember the first time I observed it in my discipline, the sociology of religion. As it happens, it involved a co-author of the IJRR piece that is the basis of Johnson and Levin’s op-ed, Rodney Stark. Shortly after my 2009 report on the rise of the Nones was released, Rodney Stark published a book that, despite clear evidence from multiple, well-respected surveys, denied that Nones were on the rise.

Rodney Stark appears to no longer be aware of reality, and that seems to be rubbing off on some of his colleagues at Baylor, like Johnson and Levin, who seem to want to deny the reality that is all around them.