Effects of religious funding on sociological research - a critique

When looking into Templeton funding for sociology recently, I came across an article in Inside Higher Ed from last year, “Sociologists Get Religion,” which responded to a Social Science Research Council working paper – “The Emerging Strong Program in the Sociology of Religion” - offering the results of a study describing the increase in sociological articles on religion in recent years and analyzing their content. As some of those interviewed for the article point out, the increase is in significant part driven by the increased availability of funding:

David Yamane, a sociologist at Wake Forest University who is editor of the journal Sociology of Religion, questions the idea that more sociologists are studying religion. He said that he receives very few submissions from sociologists who don't specialize in religion, which suggests to him that the new findings don't reflect more sociologists doing religion, but that those who do study religion "have gotten better and have more resources at their disposal to get their work out into the marketplace of ideas."

He noted a "huge influx of money" -- from groups like the Pew Charitable Trusts, the Lilly Endowment and the John Templeton Foundation -- that have led these sociologists to "have greater resources to pursue their studies than in the past." He also said that data sets created in part with support from such sources have yielded much high-impact research that sociologists could publish in top journals. Yamane views these trends positively.

[Darren] Sherkat, of Southern Illinois, views this funding with much more concern. He said that one reason sociology as a field long dismissed religion was because of an assumption that it was studied by those who wanted to advance a faith, not scholarship. "When the motivation for study is no longer what religion does or how religion works, but a view of religion being normatively appropriate or superior, then secular sociological interest will decline," he said.

The working paper by Smilde and May, worth a read, finds not only an increase in the number of articles on religion and an increase in funded research on the subject, but a rise in “pro-religious” articles, which they view as a potential cause for concern:

[T]he issue of pro-religiousness also needs to be examined. In our view, the idea that religion, in any given instance, contributes to human well-being is an entirely plausible, unobjectionable, and frequently replicated scientific finding. What is more, in some discursive contexts, it provides a vital contribution. While arguing that religion contributes to human well being arguably reinforces common wisdom in societies like the United States, within the academy it frequently amounts to a surprising and courageous position.¹ We also think that having religious motivations in sociological research is epistemologically and scientifically unobjectionable. In the classic Weberian formulation, social scientists inevitably research intellectual problems generated by specific value commitments. And value commitments entail a certain image of the way the world works, which social scientific research can corroborate (without being able to prove the ultimate validity of the values).

However, we are concerned that an increasing tendency towards pro-religiousness in the sub-discipline could saturate it with research that fits this project, setting aside research questions that do not. Research that would seek to show, for example, how religious participation might facilitate racism, insensitivity to inequality, militarism, patriarchy, civil privatism, political polarization, or other social ills could be left undone because of a scarcity of funding; or pushed aside if it doesn’t fit with the interests of scholarly gatekeepers. Likewise, studies of religion that simply see it as a sociologically important phenomena [sic] that is neither universally good nor universally bad, similar to phenomena such as the “state”, “gender,” or “class,” could also have a reduced set of opportunities. Such an imbalance could, in turn, distance the sub-discipline from the concerns and debates of the larger discipline precisely at the time that religion come to the fore in public consciousness as a central social and political issue. (9-10)

This issue is of greatest concern to me, and it’s here that I have the most problems with the working paper. The authors examined all issues of three major sociological journals and one issue per year of two major sociology of religion journals over a thirty-year period (1978-2007), for a sample of 587 articles. They found a rise over that time in the percentage of positive socio-evaluative findings about religion. During the same time, there was a dramatic decline in the percentage of articles with negative socio-evaluative findings (to under 5% in 1998-2002!), with a steep rise in the past several years (there was really nowhere to go but up), although this was matched by a steep rise in the number of positive articles and so remains well below that figure. (This is all shown in Figure 3 on page 13.)

The authors investigate possible relationships between research funding and positive findings about religion. They find a statistically significant relationship between the two, with an unexpected twist: while funding in general is significantly related to positive findings, by far the strongest relationship was with public funding, and the relationship between funding by religious foundations and positive findings was not statistically significant.

Only further research will help us understand the nature of this relationship; but, at minimum, the highly significant relationship between public funding and positive socio-evaluative findings regarding religion complicates common wisdom regarding a presumed antipathy of government bureaucracies and officials to the role of religion in society. (25-6)

This is surprising, but there are some issues to keep in mind. First, while I don’t doubt the sincerity of the authors in developing measures of positivity and negativity or evaluating the articles,² this can be fraught. Cases like the recent article by Ecklund and Long, which I’ll be talking about in an upcoming post, suggest that substantial context beyond the article itself might be needed to evaluate the views being expressed with regard to religion. Moreover, as “religion” isn’t a single, unified entity, it could be the case that articles are implicitly promoting a positive view of some religious forms by looking at negative aspects of others (e.g., pointing to negative aspects of fundamentalist religion with an implicit comparison to more liberal forms, the possibility and some examples of which the authors note themselves).³

Related to this, while it could be argued that the sociology of religion should focus on, well, religion, the dearth of attention to atheism (not to mention more explicitly negative presentations of irreligion or secularism) is itself a form of pro-religious bias worth appreciating; it serves to keep atheism invisible and to divert attention from essential questions of fact and epistemology regarding religious belief, which are relevant to sociology and shouldn’t – in an era in which evidence is of such vital importance – be marginalized.

Second, I’m not sure there is any such common wisdom, including among sociologists, that assumes an “antipathy of government bureaucracies and officials to the role of religion in society.” (Since they include funding not just from government agencies, which is generally awarded on the basis of evaluations not by “bureaucrats” but by others in the applicant’s field anyway, and public universities, this image of “government” funding is a bit odd in any event.) The authors note:

This relationship probably has to do with the support coming from federal institutes specializing in health issues—such as the National Institute for Health, the National Institute for Mental Health, and the National Institution for Child Health and Development—funding research on health that includes religion; however, it cannot be reduced to that. Twenty-six of the fifty-one articles mention financial support from public state universities. In either case, this clear and robust finding complicates frequent suggestions that government institutions and bureaucrats are the main motors of secularization (for a sociological version of this argument see Hechter 2004). Here it would seem that public sector institutions and administrators are quite receptive to scientific research showing a positive role for religion in society. As Winnifred Fallers Sullivan (2007) has argued in her research, it is increasingly the case that “we are all religious” in the eyes of the US Federal Government. (21)

Again, I don’t think that those who view governments as agents of secularization in this context are for the most part arguing that they harbor antipathy toward religion, though I do suspect people generally expect public funding to be more “neutral,” in the sense of not having a strong pro-religious agenda and thus being more open to potentially negative research. In light of this – and more generally in terms of public science funding - I regard this finding and the last sentence of the quoted passage as extremely disturbing, though the authors don’t appear to at all. The notion that “research on health that includes religion,” publicly funded or otherwise, would naturally focus on positive effects of religion is not an explanation but an assumption that should itself be interrogated. (As health is an area where religion comes into direct contact with science, an uncritical privileging here is quite troubling.)

Third, and most important for my purposes here, there are problems with the authors’ proposed explanation for the absence of a statistically significant relationship between positive findings and religious funding:

It suggests that rather than supporting pro-religiousness, it seems to focus its support on classic “religious sociology,” in which religious scholars use sociology to engage religion, warts and all, in order to improve religious institutions and practices. (26)

In addition to what I mentioned above about more subtle and contextualized readings being necessary here,⁴ there are other, serious problems. In terms of the public-private distinction, the researchers don’t appear to distinguish research produced in privately-funded centers at public universities, which, as I’ll discuss in my next post, seem quite common. Therefore, what might appear to be solely publicly funded could have religious money behind it (this mixing itself is a problem). More important, what the authors don’t provide, but would be easy enough to, I assume, is a breakdown by specific foundation.⁵ (Despite the interesting speculation about health-oriented research, they don’t do this for public funding agencies, either.) It could be that substantial differences among foundations in this regard exist, and this would be useful information.

It’s a bit disappointing that the authors don’t present the data broken down in this way, as they recognize Templeton’s strong agenda (18), contrasting it explicitly with that of the Lilly Foundation. This would seem to suggest that their generalizations about foundation funding could well rest on an improper combining of data in which, say, Lilly’s greater let’s-call-it-equanimity hides a more clear pattern with Templeton’s funding. I would bet a significant sum that research funded by Templeton is highly likely to produce positive findings, and that the number of articles based on Templeton-funded research that put forth genuinely negative socio-evaluative results about religion is darn near close to zero.

Which brings me to another key issue. Although the authors state that the findings represented in Table 9 “suggest that there is no perceptible association between funding from religious sources and socio-evaluative findings” generally (21), this table in fact shows only the relationship between different funding sources and positive findings. Unless I missed it, they don’t present any data on the relationship between funding sources and negative findings. This is surprising given that the authors themselves note the possibility for the marginalization or exclusion of critical research in these conditions as a major concern.

In sum, then, my problems with this portion of the research revolve around complexities of evaluating “positive” and “negative” findings concerning religion that are only very partially noted or addressed, the combining of data that may hide significant effects from particular funding sources, and the lack of data on the relationship between funding sources and negative socio-evaluative findings. As I said, the paper is worth a read; it presents interesting data and the relationship the authors find between public funding and pro-religious findings is startling. But with regard to the effects of religious funding in general and that of Templeton in particular I don’t think their conclusions (admittedly preliminary) are supported. In my next post, I’ll present some examples of Templeton-funded projects involving sociologists to provide a better feel for the sort of work they fund.


¹Here the authors insert a bizarre and troubling footnote:

⁶The first author of this article once received an invitation with four other faculty members to a university symposium on evolution with the charge “to show that intelligent design is irrational.” While the two natural scientists fulfilled the request, the two social scientists spoiled the show by arguing that there was nothing inherently irrational about it.

It’s strange to quote directly from a personal communication in this context without citing the source or providing the context of the alleged quotation, and the footnote provides no support whatsoever for the claim that “within the academy [arguing that religion contributes to human well being] frequently amounts to a surprising and courageous position.” Arguments concerning the irrationality of creationism are not about whether religion contributes to human well-being. In any event, the claim is challenged by their own data, which shows a large and steadily increasing number of articles characterized by “positive socio-evaluative findings” about religion, well higher than the number showing negative socio-evaluative findings. But it’s distressing to read sociologists writing so cavalierly about evolution. Understanding what the symposium organizer meant by “irrational” would require context the authors don’t supply, but belief in ID with the knowledge of the strength of the ToE in the contemporary era is in fact profoundly irrational.

² They describe their coding somewhat vaguely in the methodological appendix:

Socio-evaluative Conclusions: We coded an article as having a positive socio-evaluation if its findings clearly showed religion to contribute to human agency or autonomy in general or to concrete outcomes generally considered positive a[t] the micro level (such as physical or mental health, life satisfaction, educational attainment, low deviance rates, participation in civil society, low divorce rates) and at the macro level (such as economic growth or democratic consolidation). We had a separate code for analyses that “debunked the negative” by contradicting a negative stereotype of religious practice. So, for example, an article might show that members of Baptist churches are no less likely to value science than the larger population. For the analysis we grouped these together with positive socio-evaluative findings. We coded an article as having a negative socio-evaluation if its findings clearly showed religious practice to diminish human agency or autonomy in general, or to have concrete outcomes generally considered negative at the micro-level (such as stress, hysteria, maladjustment, passivity, volatility, low educational or career attainment) or at a relatively more macro level (racism, tolerance and bigotry, gender inequality, poverty and authoritarianism). We had a separate code called “debunking the positive” for articles that contradicted widely held positive socio-evaluative findings. So, for example, an article might show that, despite common sociological wisdom, Catholic practice does not reduce levels of suicide. We combined this with negative findings for some of our analysis. Articles that provided both positive and negative socio-evaluative findings were coded as such. In other words, articles could have more than one value of this variable. (27-8)

³ “Furthermore some religious researchers funded by religious organizations often focus on the negative consequences of the practice of specific types of religion—perhaps as part of a confessional search for proper religious practices. These types of findings are not hard to understand given the Lilly Foundation’s interest in the overall health of churches. To such a perspective negative findings are as important as positive findings.14”

⁴And really, it has to be noted that the idea that “religious sociology,” “in which religious scholars use sociology to engage religion…in order to improve religious institutions and practices” is anything other than pro-religious is not at all tenable. This isn't to say that it means the findings are positively skewed, but it certainly contributes to conditions in which sociology is plainly serving a religious agenda.

⁵It appears Templeton and Lilly are the two major private religious funders of interest:

We decided on how to code foundations by consulting their web pages to see whether or not they described their purpose in religious terms. This led us, for example, to code the Pew Foundation as simply a private foundation, while the Lilly Foundation we coded as religious. A different methodology—such as looking at different foundations’ charters—could well have led to a different result. (28)