If the tendency of American religion to stimulate political conservatism is among the most oft-cited religion and politics findings, this book joins a growing chorus of scholarship devoted to revealing the reasons this empirical regularity is misleading. Although it is an unassailable fact that—in the aggregate—frequent worshipers report more conservative political views, Djupe and Calfano’s research highlights a new way of thinking about religious influence. It is not frequency of worship per se that influences public opinion—it is the political communication that occurs in religious settings. Furthermore, political messages in religious settings are not uniformly conservative, nor is influence limited to conservative religious messages. Why has a flawed inference, drawn from analyses of aggregate trends, become so ubiquitous to understanding the influence of American religion on American politics? Overreliance on the survey research method has obscured more careful attention to the underlying causal processes, according to Djupe and Calfano. Accordingly, God Talk is a call to use the experimental method to measure the influence of political communication in religious settings in new ways. There are many things to like about this book, including eight chapters each reporting separate sets of elegantly designed experiments, but I suspect its lasting influence in the field will be an urgent call for theoretical revision and methodological pluralism.
In Chapter 1, Djupe and Calfano distill a voluminous literature about religious influences on public opinion into three major theoretical approaches. Having described each approach, they highlight ongoing controversies and unanswered questions before outlining a path forward using the experimental method to gain new empirical leverage.
The Religious Commitment Approach includes the vast majority of recent religion and politics research. This approach, “now more commonly known as the ‘3Bs’ approach” (p. 21), begins by dividing people according to the place of worship to which they “belong”—generally into a handful of broad religious traditions. Then, “beliefs and behaviors” are used to assess commitment to one’s religious tradition. The key drawback of this approach is lack of attention to the messages religious people receive. The assumption that all “Evangelical Protestants,” for example, receive the same religious communications and only vary in intensity (as measured by religious commitment), is deeply flawed. The experimental method, they demonstrate, can help disentangle message effects from commitment effects.
The Psychological Approach also tends to ignore the messages religious people receive, but avoids placing people into overbroad groupings (religious traditions) by utilizing religious identities and focusing on very specific values and beliefs. Finally, in contrast to the lack of attention to religious communication within the Psychological and 3Bs traditions, the Social Networks Approach argues that, “religious influence may be found primarily in the local house of worship and that political communication is the primary mechanism for exerting this influence” (p. 26). Yet questions linger about whether individuals with distinct political views find like company, or if political communication in social networks influences opinions. Again the experimental method may help disentangle how religion influences public opinion. More generally, Djupe and Calfano call for scholarship that engages multiple theoretical approaches to understanding the way religion influences public opinion and highlight areas where experimental research designs promise to discriminate among them.
The core of the book consists of eight chapters (some written with additional coauthors), each reporting experiments designed to assess the religious effects of political communication in different domains. The experiments are survey-based, but involve a variety of sample types (convenience samples of university students, purposive samples of specific religious groups, and full probability samples of the national electorate). Chapters 2 and 3 concern the use of religion by political elites. Chapters 4 through 7 concern evaluations of the public arguments of religious elites. And Chapters 8 and 9 concern communication from clergy. In each of these chapters there is ample evidence that conservative messages can have conservative effects among religious people. However, there is equally strong evidence that liberal messages can have liberal effects among religious people. For example, in Chapter 2 the authors find that when political candidates use God talk (language that is recognizable as “religious” only to religious people, e.g. biblical phrases that are not explicitly religious) support among Evangelical voters increases—but does so equally well whether or not the candidate in question is described as a Republican (p. 59). And In Chapter 4, when Evangelical elites explain having arrived at pro-environmental positions through careful biblical study or prayerful consideration, support for the need to address global warming increases dramatically among Evangelicals (even though it decreases when Evangelical elites take pro-environmental positions but do not cite a process accepted by the community for arriving at the position, p. 105).
I am undoubtedly making too much of the fact that many of the findings reported in God Talk undermine a unidirectional (e.g. more religion equals more conservative) model of religious influence. While the authors certainly frame things to highlight this contribution, the greater contribution is surely the consistent evidence they produce indicating communication matters in religious contexts. In chapter after chapter, carefully designed experiments find effects observational studies routinely fail to detect. For example, in contrast to the minimal clergy effects found in observational studies, Chapters 8 and 9 report on experiments in which robust clergy effects emerge.
The pessimist worries that life is not like the survey-based experiment. Listening to a sermon is very different than reading a prompt or a vignette and one rarely has occasion to express an opinion about foreign intervention or immigration immediately after hearing about “inclusive” or “exclusive” religious values. More optimistically, one might view the experimental evidence as proof clergy—and communication in religious settings more generally—can matter. Even more optimistically still, God Talk reveals that communication does matter when investigators look in the right places. The experiments are highly nuanced and the designs reflect careful thought about messenger characteristics (e.g. Evangelical identity, clergy status), message content (e.g. was the expressed position of a religious elite arrived at through prayerful consideration, how was the issue framed in the sermon), and receiver characteristics (e.g. religious affiliation, frequency of worship).
Clearly communication in religious contexts is highly nuanced. Accordingly, it is also true that the results are complicated, often involving three-way statistical interactions. The authors do readers a great service by providing results in graphic form, but most of the graphs involve multiple contingent effects and one needs to know which comparisons to make to identify each contingent effect. Readers should lean heavily on the text as they interpret the results. Because the authors have worked hard to explain their results, the complexity of the research is not a defect. On the contrary, I would not be surprised if observational studies uncover similar effects if care is taken to incorporate each of these factors (messenger, message, and receiver characteristics) in the design. In this way, God Talk will benefit the field both by stimulating additional use of the experimental method and by informing design choices when observational methods are used.
This is a great book for religion and politics scholars, it will lead to much discussion in seminars, and will surely stimulate greater use of the Psychological and Social Network approaches. The authors’ conclude, and I agree, “If there is an emerging perspective on religious influence, it clearly will be devoted to careful thinking about religious communication (exposure), exploring the various combinations of message and messenger” (p. 214).
