the strong program in the sociology of religion

The Immanent Frame has an interesting working paper by David Smilde and Matthew May on the state of the subfield of the sociology of religion. Highlights, quickly, are that there’s a lot of work on the sociology of religion; that most of it focuses on the US and on Christianity, and in particular on Protestants; and that the prevailing paradigm is a “strong program” treating religion as an independent variable predicting other outcomes. Studies that are funded are more likely — by a lot — to find that religion is “good” for people in some way, but the funding source (government vs. foundation, e.g.) is not a factor.

I am not a scholar of religion, though I am interested in the field, so approach this largely from the outside. While I find the independent-variable approach worthwhile, I do worry that religion as a dependent variable deserves lots more consideration too. This is both as a function of time (changes in religious matters over time) and in terms of understanding the origins and development of modalities of religious belief, thought, and practice. It’s too bad that, apparently, the “strong program” seems to have come at the expense of understanding the origins of variation in religiousness. (By the way, the “strong program” term comes from Jeffrey Alexander, who wishes it upon cultural sociology, except that in Alexander’s world culture doesn’t really cause or get caused — it just is, on its own, sui generis. The sociology of religion is to be commended for avoiding that trap.)

I wonder, too, if there’s an institutional explanation. Over the time period covered by the Smilde and May study, there’s been something of an explosion of data available for analysis on religion, particularly in the United States. Much of this data is cross-sectional, which encourages analysts to treat variables as fixed and causal (a.k.a. “independent”). To the extent that religion is a dependent variable, a temporal dimension is a virtual necessity, since one is looking for the etiology of an observed state (whether observed individually or societally).  Religion, like other social measures, is also treated in standardized ways, which again encourages analysts to think of it as relatively static.

I would love to see similar studies in other subfields as well. My guess is that over the same time period sociology in general became more causally oriented, used more secondary data analysis, and considered less the development of social states–thus I would imagine that the sociology of religion probably mirrors much of the rest of the discipline.

Author: andrewperrin

University of North Carolina, Chapel Hill

6 thoughts on “the strong program in the sociology of religion”

  1. Finding positive “effects” of religion in micro models has proved pretty easy, which may be why there was a surge in articles that do it – and they are way more common in non-“top-three” journals. I will point out that, if you reverse the coding on the “religious” variable, and show that *not* being religious has negative effects, then you paint a less “positive” picture of stratification. That could be described as a system in which failure to conform to religious norms is penalized. Is that “good”? (The same can be said about education, marriage, and having a gun during a shootout.)


  2. Conventional use of the phrase “strong program in x” these days seems to be code for “You may think you have some analytical or empirical criticisms of my view but lalalalalalala I can’t heeeeear you”.

    As for the sociology of religion in the United States, it seems obvious that the selection effects on the participant side are enormous and that a lot of the output in the journals is best thought of as Cheering for the Home Team.


    1. It’s funny–I think the “strong” program approach actually signals the secondary importance (from the perspective of “mainstream” sociology) of a subfield. Take networks. For a long time, people treated networks as IVs that predicted things that most sociologists actually cared about (e.g., jobs). But eventually, after lots of research showing that networks matter for a host of such outcomes, they have become “important” enough for mainstream sociologists to take interest in them as outcomes themselves. Right now, I’m not sure how much interest the ASR/AJS audience would have in models of speaking in tongues.


  3. Kieran has it nailed. If you look closely at Smilde’s paper, you’ll see that the “surge” lacks much flow volume. Though, most of the articles in ASR and AJS use longitudinal data (Stolzenberg et al, Axinn and Pearce, Bearman and Bruckner, Keister, Myers, etc…). And, I’d bet the private influence on positive findings is a function of grant size, not grant acknowledgment. Smilde has our (Darnell and Sherkat) ASR paper as a negative result with religious funding only because I acknowledge a very small Lilly award which was not awarded for that particular study. Private religious foundations have funded religious cheerleading, and it has an impact on academic careers and the content of longstanding data collection projects.

    It’ll be interesting to see what happens in the next decade or so. Right now, I think the bulk of the energy and resources are in areas which are not conducive to connections to the broader discipline or to other arenas of social scientific inquiry. But, others disagree.


Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s