WHAT EXPLAINS SEGMENTATION IN PUBLICATION PATTERNS?

Publications are largely a product of choices and resources. Authors choose where to submit manuscripts partly based on professional conventions about which journals are appropriate placements for their research. Authors also might consider their chances of success when choosing where to submit; however, the level of uncertainty of the publication process makes it difficult to do so with confidence, especially in more competitive journals. By “resources,” we are referring to information that is valuable for successfully navigating the publication process at the journal of choice—for example, how to structure the manuscript or highlight a contribution. Those who share these conventions and have acquired these skills and information tend to publish in similar journals.

First, scholars are exposed to norms and information through their graduate training. Thus, we expected the likelihood of publishing in internationally visible journals to be related to where authors obtained their graduate degree. The idea that higher-impact journals are the expected placements for one’s research is practically unquestioned in Global North universities, but it is not a generally accepted norm in the Global South. Thus, our first hypothesis is that:

Second, norms and information travel through professional networks. Research on coauthorship patterns in major political science journals revealed the existence of a large network of authors located worldwide (Metz and Jäckle Reference Metz and Jäckle2017). Whereas ties to Global North scholars can disseminate norms and resources beyond those who attended graduate school abroad, they also can foster publication in global journals through the coauthor’s own professional incentives. Universities in the Global North, in general, are more explicit than in the Global South about which publication outlets are needed for promotion. Therefore, a Global North coauthor might be less interested in publishing in less-visible regional venues. Thus, our second hypothesis is that:

Third, we considered whether norms and information also are shared within university departments. Shared norms can generate convergence through emulation if individuals expect reputational costs for not publishing similarly to colleagues in the same department. Convergence also could be a product of information sharing about an inherently uncertain process, such as publishing a paper. Scholars should be receptive to information that would allow them to shorten the process. A colleague’s positive experience with a journal provides information that the possibility of publishing in that journal is relatively good. Similarly, a colleague’s rejection in another journal might discourage a scholar from submitting a paper for publication there. We do not argue that only one of these two mechanisms—emulation or information sharing—explains the relationship between individuals’ publication records and their departmental colleagues’ records. Rather, we view them as complementary and pointing to the same expectation—that is, convergence within a department regarding publication outcomes. Thus, our third hypothesis is that:

ALTERNATIVE EXPLANATIONS

There recently has been an increased interest in gender differences in political science. Some of the findings from this research are that prominent journals publish fewer articles by women than by men (Teele and Thelen Reference Teele and Thelen2017) and that men submit more articles than women (Brown and Samuels Reference Brown and Samuels2018; Djupe, Smith, and Sokhey Reference Djupe, Smith and Sokhey2019). These findings lead us to expect similar dynamics in the Argentine scientific community resulting in women authors being less likely to publish in global journals.

A second possibility is that attributes of the articles, and not the authors, explain where they are published. For several reasons (e.g., language, distance, and resources), Argentine scholars might be in a better position to access higher-quality evidence on their own country and the Latin American region than elsewhere and, as a result, be expected by outsiders to speak about their own region with more authority than about other parts of the world. Even the belief that this expectation exists could result in authors self-selecting into certain journals based on the subject matter, expecting more favorable results in global journals with work on their country or region.

A third possibility is that, given the greater prominence of quantitative methods in higher-visibility journals (Mahoney Reference Mahoney2007), scholars will be more likely to publish in global journals when they use quantitative methods. We recognize that IR has a long tradition of qualitative research, but we focused on publishing patterns in the past decade. This coincides with a trend in the field that gives increasing importance to quantitative research.

A fourth factor that influences placement may be the literature with which an article is in conversation. Reviewers may react more positively to manuscripts that reference literature they recognize. Reviewers in global journals are more likely to be Global North–based scholars, who usually are less familiar with academic work from the Global South than regional academics who generally review manuscripts in regional and domestic journals. Conversely, reviewers from regional journals might suggest referencing other Global South–based work. Because we observed only published versions, we cannot know whether authors preemptively adapt their reference list to what they expect reviewers will like, or whether reviewers require the inclusion of more Global North or Global South references, or both.

Fifth, material incentives in the form of promotion requirements–which are fairly common in the Global North—also may explain publication patterns of Global South scholars. If this were the case, then it would be difficult to disentangle material incentives from the previous argument based on the dissemination of convention and information. However, as mentioned previously, whereas explicit publication standards for promotion are rare in Argentine universities, they do exist for scholars affiliated with the Argentine Research Council, commonly known as CONICET and akin to the American National Science Foundation. CONICET researchers are ascribed to universities and receive monetary compensation that supplements their university salary with the expectation that they will devote their time to research. CONICET has more formalized standards for evaluating its researchers’ publications—international journals have greater value—and uses them as metrics for promotion. These are extremely sensitive processes because negative reviews can result in researchers losing their appointments or being delayed in a particular category, thereby missing out on salary increases. Therefore, we expect articles authored by CONICET researchers to be more likely to appear in global journals. More important, because scholars with CONICET-appointment work in a university department but not all faculty in that department will have a similar appointment, we can observe which affiliation—and, therefore, which mechanism—has more explanatory power.

DATA AND METHODS

To test these expectations, we collected data on publications by IR scholars based in Argentina from 2008 to 2020. “Argentine scholars” refers to individuals based in and publishing from Argentina, independent of their nationality. Our dependent variable consisted of two indicators for global and nonglobal journals. The first was whether journals were included in the Scopus database. Journals are entered in Scopus when they reach certain levels for their H-index, SCImago Journal Rank, and other metrics. We believe that this is a good indicator of whether journals represent the global discipline of IR because they are more visible and widely read. Among those, journals that are published mainly in English are even more likely to be read widely and globally. Thus, our second indicator identified Scopus-indexed journals that publish mainly in English. To confirm the validity of this indicator, we consulted the most recent Teaching, Research, and International Policy (TRIP) survey of faculty in 30 countries that asks which IR journals publish articles with the greatest influence. All identified journals are indexed in ScopusFootnote ⁴ Footnote ⁵ (Maliniak et al. Reference Maliniak, Peterson, Powers and Tierny2017). (See table 1 in the online appendix for the complete list.)

Data on articles published by Argentine scholars in journals indexed by Scopus are from the Scopus website. The data on non-Scopus publications were collected manually. We began by collecting articles in journals indexed in the regional database SciELO, a bibliographic database of open-access journals created to serve scientific communities of Latin American countries as well as Portugal and Spain. We supplemented those with other regional and local journals that are not indexed in these databases but that qualitative evidence and informal interviews suggested are common journals in which IR Argentine scholars publish. The result is a dataset of 791 articles. Figure 1 shows the distribution of our dependent variable over time. The local IR community has become more productive but most of this output generally appears in nonglobal journals.

Figure 1 Articles by Type of Journal

To test our first hypothesis, we calculated for each article the ratio of authors who obtained their PhD from a university in the Global North. However, those who perform an academic role in Argentina—teaching at the university level, researching, and publishing—do not necessarily pursue a doctoral degree, which results in a significant proportion of articles in our dataset not being authored by PhDs. Thus, to distinguish the impact of a doctoral degree from the impact of connections to the Global North, we included a second variable: the ratio of authors in an article who obtained their PhD from a university in the Global South. For our second hypothesis, we created a dummy variable that takes the value of 1 if an article has any coauthors based in the Global North. For our third hypothesis about emulation and information sharing within departments, we identified each author’s departmental affiliation and then calculated the share of papers that were published in global journals in the previous three years by all other scholars in the same department.Footnote ⁶

To control for alternative explanations, we calculated the ratio of authors who are women. To determine whether Argentine scholars are more likely to publish on certain topics using certain methods in either type of journal, we classified all articles into five topics: Democracy and Human Rights, Foreign Policy Analysis, International Political Economy (IPE), International Security, and Theory. (See figure 1 in the online appendix for the distribution of articles by topic.) We coded whether each article focuses on Latin America, which means that the article analyzed the region as a whole or a specific country and whether it used quantitative methods. To control for whether the referenced literature tends to vary with the type of journal, we identified all references cited in at least three of the 721 papers in our dataset. We then coded whether each uniquely referenced author was based in the Global South or Global North and calculated the ratio of referenced authors from the Global North. Finally, to control the potential effect of material incentives through promotion requirements, we calculated the ratio of authors with a CONICET appointment. (See table 2 in the online appendix for descriptive statistics of all independent variables.)

RESULTS AND DISCUSSION

We ran a set of logistic regressions predicting the likelihood of a paper being published in a journal indexed in Scopus (model 1) and in a journal indexed in Scopus and published in English (model 2). The results, presented in table 1, provide support to our three hypotheses. The probability of an article appearing in a global journal increased with the proportion of authors with a PhD from a school in the Global North. This suggests that doctoral training in Global North countries makes authors more likely to submit to and/or be successful at global journals. Having at least one coauthor in the Global North also increased the probability of publication in a global journal. The impact of coauthoring can be explained purely by the preoccupation of Global North colleagues with their tenure requirements, but it also may indicate shared understandings of what are desirable venues for publications and how to arrive at them.Footnote ⁷

Table 1 Logistic Regressions

Notes: *p<0.1; **p<0.05; ***p<0.01

Authorship by PhDs from Global South universities also is positively related to publication in global journals. This suggests that, unlike our expectations, doctorate programs in Argentina are socializing students around similar norms defining what is professionally desirable and providing comparable resources about how to obtain it. If this is the case, arguments about lack of diversity in political science journals may simply be reflecting that scientific communities are smaller. However, the positive and statistically significant effect of both variables on the likelihood of publication in global journals does not mean that their substantial effect is equal. Figure 2 shows the predicted probabilities as each independent variable changes values. In both plots, the change in probability of publication in global journals from having no authors trained in the developed Global North to having only authors trained in the developed Global North is considerably larger than when going from none to all authors trained in the developing Global South: from 0.14 to 0.48 and from 0.12 to 0.24, respectively.

Figure 2 Predicted Probabilities

Departments appear to generate convergence among their faculty’s output because articles are more likely to appear in a global journal when their author is from a department with a higher proportion of placements in global journals among recent publications. Because promotion requirements are not formalized in Argentine universities—unlike in the United States and the United Kingdom, for example—we interpreted departmental convergence to be a product of prevailing norms that impact scholars’ sense of their own reputation more than their salary goals. Moreover, CONICET appointments appear to have no effect on the probability of publishing in global journals, suggesting that promotion requirements are a less effective incentive to publish in higher-impact journals. In summary, our results suggest that the Argentine contribution to the globalized discipline is made by scholars with ties to the Global North and by those working in a department where there already is an inclination to make that contribution.

Regarding other predictors, some article characteristics appear to make it more or less likely for Argentine scholars to publish in global journals. Democracy and Human Rights and IPE articles have positive and statistically significant coefficients in models 1 and 2, respectively, whereas Foreign Policy Analysis and Theory have opposite results. In general, none of the topics appears to be statistically significant across both models. Choices of methods do appear to affect the likelihood of publication in global journals. Although quantitative methods are rare in Argentine IR, they appear to correlate more strongly with global journals, and the relationship appears to be stronger for English-language journals. A higher presence of Global North authors among the referenced literature apparently has no statistically significant effect on the dependent variable. On a positive note, results do not seem to vary when articles are authored by a higher proportion of women.