METHODOLOGY

The methodology for this updated study followed the design of the earlier research by Kettell (Reference Kettell2012). The study began by producing a dataset of the top-ranked political science journals (as of 2021) using five-year impact-factor scores obtained from the Institute for Scientific Information Web of Knowledge (see https://jcr.clarivate.com/jcr/home) (Kettell Reference Kettell2023). The full list of journals used in this dataset, as well as the relative change in their rankings compared to 2009 (i.e., the cutoff point for the original study), is in table 1. The new entrants are designated by an asterisk in the table.

Table 1

Political Science Journals Ranked by Five-Year Impact Factor (2021)

It is notable that there was a substantial turnover between the two periods, with half of the journals from the previous study no longer appearing in the top 20. To ameliorate the potential effects of this turnover for the purposes of comparing the two periods, a second dataset was produced based on the original journal rankings using the top 20 from 2009. In addition to the 10 non-asterisked journals in table 1, the second dataset included Journal of Peace Research, Political Geography, Quarterly Journal of Political Science, Journal of Conflict Resolution, European Union Politics, Journal of Politics, International Studies Quarterly, Studies in Comparative International Development, Review of International Political Economy, and Governance.

To examine the extent to which the outputs of the journals in these two datasets engaged with religious themes, the current study examined all articles that were published from 2011 to 2020. To avoid the sample being distorted by a high volume of short publications, the criteria for inclusion centered on work that was six or more pages in length; shorter book reviews, replies, editorials, and articles were excluded. The focus was on original articles, review papers, commentaries, longer form replies, and symposia. The study then conducted a content analysis based on the title, abstract, and keywords for each output. An article was coded as engaging with religious themes if it included key terms relating to religion (e.g., “church,” “faith,” and “creationism”) or if it used direct religious-identity markers (e.g., “Christian” and “Muslim”). Positively coded items were classified in one of two ways. Outputs for which religious themes formed the core or a principal focus of the research were classified as “primary” items. An example of this was Tilley’s (Reference Tilley2014) article, entitled “We Don’t Do God? Religion and Party Choice in Britain,” in which the central theme of the research was the importance of religious factors in voting preferences. Outputs in which religious themes featured as a more indirect or a minor part of the research were classified as “secondary.” An example of this category was Haggard and Kaufman’s (Reference Haggard and Kaufman2016) article, entitled “Democratization During the Third Wave,” the abstract for which includes a single reference to “ethno-religious cleavages” and indicates that this is part of a wider analysis along with other factors such as underdevelopment, resource endowments, and inequality.

This coding scheme for the analysis was mostly straightforward, although at times a subjective judgment was required to make the classification. To err on the side of caution and to avoid underreporting the extent to which papers had engaged with religious issues, cases in which the thematic link was uncertain, or the key terms were ambiguous (e.g., outputs that referred to “antisemitism” without specifying whether it was intended to denote an ethnic or a religious term) were coded as “religious.”

One drawback to this method is that an examination of the title, abstract, and keywords may not necessarily capture whether a paper referred to religious themes in some way. As such, it is possible that some outputs in the sample had engaged with religious issues but were not positively coded because they did not present this in the opening material. That said, it is reasonable to assume that if religious issues were significantly related to the research topic, then this would have been communicated via the title, abstract, or keywords. Therefore, whereas this possibility cannot be discounted, it is likely to have affected only a few papers and thus would not have influenced the overall results of the study in any meaningful way.

OVERALL RESULTS

The original study encompassed the period 2000–2010 and produced a sample of 7,245 published outputs. The subsequent content analysis found that only 97 outputs (i.e., 1.34% of the total sample) were classified as “primary” items and 87 papers as “secondary.” In total, 184 articles engaged with themes and issues relating to religion during this period—an average of 2.54%.

The results of the updated study are divided between the two datasets. Dataset A (comprising the top 20 political science journals from 2021) produced a sample of 8,253 outputs. Of these, 188 were coded as primary (2.3% of the total) and 113 as secondary (1.4% of the total). In all, 302 papers were classified as engaging with themes and issues related to religion in some way (3.7% of the total). Dataset B (using the top 20 journals from the original study) produced a total of 9,015 outputs. Of these, 232 (2.6% of the total) were classified as primary and 130 (1.4%) as secondary. A total of 362 papers in this dataset were coded as having engaged with religious themes (4.0% overall). These percentages are listed in table 2.

Table 2

Primary and Secondary Outputs

The first key finding of this study is that in terms of the overall numbers and relative proportions, the updated datasets contained twice as many (or almost, in the case of dataset A) primary-coded papers as in the earlier period. The updated analysis also reveals that the total number of political science papers relating to religion increased substantially: by almost two thirds for dataset A (184 to 302) and almost doubling for dataset B (184 to 362). However, the study also shows that the proportional increase between the two periods was far more modest: 1.1 percentage points for dataset A and 1.5 percentage points for dataset B.

The updated study also found a wide variation in the engagement with religious themes among journals, with outputs involving religion concentrated in only a few periodicals. Across both datasets, the journal with the highest overall number of outputs coded as religious (63) as well as the highest proportion of its outputs engaging with religion (11.9%) was Political Psychology. This journal was far ahead of the second-placed journal, American Political Science Review, which published 36 papers that engaged with religion (7.1% of its total output). At the other end of the scale, six journals had an average of less than 1%, and three periodicals each published only one article on religious themes. Almost half of the sample (14 journals from the combined datasets) published fewer than 10 articles on religion during the entire period.

The pattern for primary-coded outputs was similar. The periodical with the highest overall number of primary-coded publications again was Political Psychology, with 42 papers (8% of its total publications). Two journals (i.e., Political Analysis and Journal of Public Administration Research and Theory) did not publish any primary-coded items. Another six journals each published only one primary-coded paper, and 20 journals (two thirds of the combined datasets) published fewer than 10 primary-coded items during the period of analysis. These percentages, ranked according to the overall proportions of papers relating to religion, are listed in table 3 (dataset A) and table 4 (the additional journals included in dataset B).

Table 3

Religious Themes in Political Science Journals (Dataset A)

Table 4

Additional Journals (Dataset B)

ANALYSIS BY KEY THEMES AND SUBFIELDS

The original study further examined the primary-coded items to reveal the topics that were focused on in more detail. This research found a high level of concentration, with most articles clustered around a few themes. Of the 10 most popular themes identified in the sample, most were related to issues of “violence, conflict, and terrorism” (21.7%), “Islam and Muslims” (20.6%), and “US politics and society” (17.5%). In contrast, the topic of “women and gender” featured in only 4.1% of primary-coded papers and “life and death issues” (e.g., abortion and assisted dying) featured in only 3.1%.

Repeating this analysis for the updated sample using the same categories revealed a similar level of concentration for both datasets A and B. The relative weightings remained broadly similar but there were significant changes. The most dramatic change was a substantial increase in the proportion of papers on the theme of Islam and Muslims. Of the 188 primary-coded items in dataset A, 75 (39.9%) were related to this theme. This was double the proportion from the previous study, making it the single largest category by a considerable margin. The percentages for dataset B were even more dramatic: of the 232 primary-coded items, 103 were related to this theme (i.e., 44.4%). The next most popular theme was violence, conflict, and terrorism, which featured in 23.4% and 31% of all of the primary-coded papers in datasets A and B, respectively. A more detailed analysis further revealed strong overlaps between these themes. For dataset A, 33% of all papers coded for Islam and Muslims were related to themes of violence (25 of 75), and 57% of all papers that were coded for themes of violence were related to Islam and Muslims (25 of 44). For dataset B, the respective percentages were 34% (35 of 103) and 48.6% (35 of 72). These percentages show the proportion of papers in each category for each dataset and are listed in table 5.

Table 5

Primary-Coded Articles by Main Themes (Percentage)

The original study combined this thematic analysis with an examination of published outputs according to the political science subfield to which they were related. The papers were classified using the eight disciplinary subdivisions identified by the UK’s Economic and Social Research Council (Wakeling Reference Wakeling2006): political institutions, political theory, comparative politics, political behavior, political economy, public policy, international relations, and political methodology. Although articles could be classified in more than one category, the distribution of primary-coded items again was highly concentrated in only a few disciplinary areas. Political behavior—a broad category including various topics such as voting, public opinion, identity issues, and political violence—accounted for almost two thirds (63.9%) of all primary-coded papers. This was far ahead of the second-placed subfield, political institutions, which accounted for 17.5% of all primary-coded outputs. The lowest-ranked subfields were public policy and political economy.

A similar distributional pattern was found for the updated study. The overwhelmingly dominant subfield remained political behavior, which accounted for more than three quarters (76.6% and 79.7% in datasets A and B, respectively) of all primary-coded papers. This was significantly ahead of the second- and third-placed subfields: comparative politics and political institutions. The lowest-ranked subfields were political methodology and international relations. The breakdown of primary-coded papers by subfield is presented in table 6.

Table 6

Primary-Coded Articles by Subfield (Percentage)

CONCLUDING ANALYSIS

The results of this updated study provide mixed results for scholars who are working at the intersection of political science and religion. In one respect, the overall increase in political science papers engaging with religious themes—particularly the increase in primary-coded outputs—is to be welcomed. Considered as a whole, however, the results suggest that religious issues have yet to be substantially integrated into the mainstream of the discipline—at least insofar as this can be measured by examining the outputs of leading political science journals. One key finding of this study is that the overall proportion of papers engaging with religious themes has remained comparatively static. For dataset A, the increase was only 1.1 percentage points compared to the earlier period; for dataset B, it was 1.5 percentage points. Therefore, the total proportions of papers engaging with religious themes in these datasets were 3.66% and 4.02%, respectively—a small increase from the 2.54% found in the original study.

This finding raises three questions. First, it reasonably could be asked why growth in this area has been so sluggish. Scholars have proposed a range of factors that impede the engagement of political science with religious themes, including the complexity of the subject, a lack of robust datasets, issue-attention cycles, the secular origins and nature of the discipline, and a lack of outlets for political science work in this area (Philpott Reference Philpott2009; Wald and Wilcox Reference Wald and Wilcox2006). These claims are not without merit, but there is little reason to believe that these obstacles should be insurmountable barriers to growth. Moreover, significant inroads have been made during the past 15 years to provide greater opportunities for scholars working at this disciplinary juncture. Thriving sections within the American Political Science Association and the UK’s Political Studies Association (among others) are devoted to promoting research in this area, and a dedicated journal (i.e., Politics and Religion) began publishing in 2008. Clearly, religion is not being “overlooked” by political scientists per se, even if its penetration into the mainstream of the discipline appears to be somewhat muted. Indeed, it is possible that the growth in these areas has contributed to the relatively weak level of growth elsewhere, by directing scholarship away from engagement with more mainstream categories.

This poses the second question of what the appropriate level of mainstream engagement with religion should be. In one respect, this is an impossible question to answer, not least because any percentage necessarily would be subjective and arbitrary. Even if it were possible to determine a technical weighting of disciplinary importance, it would be subject to ongoing revision according to academic trends and real-world political developments. Moreover, focusing on a headline percentage risks obscuring what arguably are more significant issues beneath the surface. One issue is that most publications engaging with religion are from a small number of outlets. Two thirds of all of the journals used for this study published fewer than 10 primary-coded articles during the 2011–2020 period, and only three (i.e., Political Psychology, American Political Science Review, and Political Geography) accounted for a third (32.2%) of all primary-coded papers published in the same period (89 of 276). Here, at least, there is scope for improvement.

A more relevant third question is: “How much is enough?” In this sense, it may be that the growth of scholarly interest in the field of politics and religion already has crossed the threshold needed to develop and sustain a vibrant and active academic community, regardless of any mainstream expansion. Yet, significant challenges also remain. One notable issue is the thematic and disciplinary concentration of scholarly work. Realistically, we would not expect to see an equal proportion of outputs engaging with the fields of political behavior and political methodology, given the range of the two subdivisions. However, scope clearly exists for greater engagement in subfields including public policy, political economy, and international relations, which seem to be significantly underrepresented in terms of their overall proportions of primary-coded work.

This problem is mirrored in the case of thematic concentration, in which the majority of mainstream political science engagement is focused on only two core areas: violence, conflict, and terrorism, and Islam and Muslims. The scale of this concentration increased markedly during the past decade, which may be due to the Western-centric nature of the discipline and its leading periodicals (most of which are centered on Europe and the United States). However, the lack of diversity poses a significant challenge to the development of scholarly work in this area. Focusing predominantly on a narrow range of issues risks skewing the direction of future research and, in this instance, may reproduce political discourses around the idea of religion as a security issue, replete with notions of an inherent link to violent activity—particularly in the case of Islam. The scope for more diversity in the types of themes and issues with which scholars engage is considerable. In addition to the areas of political activity noted previously, religious groups and ideas have been central to a range of global political developments in recent years, including resistance to the intensification of neoliberalism (Dreher and Smith Reference Dreher and Smith2016); the COVID-19 pandemic (Linke and Jankowski Reference Linke and Jankowski2022); the politics of climate change (Arbuckle Reference Arbuckle2017); and the recent wave of nationalist and/or populist forms of politics (DeHanas and Shterin, Reference DeHanas and Shterin2018). By failing to engage with religious themes in a more comprehensive way, scholars invariably are missing a key force in contemporary politics and limiting their understanding of the wider political world. This does not mean that we need to develop a “political science of religion” (see, e.g., Kettell Reference Kettell2015; Potz Reference Potz2019). However, efforts to promote a greater understanding of the multiform ways in which religious forces act politically in different contexts and parts of the world would go a long way toward further developing the academic study of politics and religion.