Realists as Lone Wolves?

Note: A couple of weeks ago I posted a short piece noting that realists seem less inclined to co-author than other IR scholars, and I speculated that this might be due in part to a personal affinity. Specifically, those who see international politics as a “self-help” system are also more inclined to see their own work in similar terms. I offered some anecdotal evidence to illustrate the point, but nothing more than that. In this guest post, Lindsay Hundley of the College of William & Mary offers more substantial empirical evidence. –S.W.

Lindsay Hundley writes:

Stephen Walt recently hypothesized that realists may be more likely than liberals to do solo-authored research, in part as a result of their basic personality traits or worldview. He suggests that the part of a scholar’s personality that causes him or her to embrace realism may be the same quality that leads him or her to do solo-authored work (or, conversely, to co-author for liberals).

Walt’s evidence that realists tend to practice “self-help” when it comes to publishing consists of a list of nine prominent (mostly retired) realist scholars whose research typically appears under a single byline. This is followed by a second list of nine prominent IR scholars who are some combination of liberal and constructivist, but are characterized by Walt as liberals. Since the Teaching, Research, and International Policy (TRIP) Project includes systematically collected evidence that includes a larger number (of both prominent and obscure) realist and liberal publications, we thought it might be interesting to see whether Walt’s empirical premise holds water. Are realists less likely to co-author than liberals? Some simple hypothesis testing seems to be in order!

It turns out that an overwhelming 82 percent of realist articles published in the top 12 peer-reviewed journals between 1980 and 2011 are solo-authored. Liberal articles, on the other hand, were much more likely to be co-authored, with (a comparatively small) 63 percent being solo-authored. Of course, even if realists are less likely to co-author than liberals (or other isms), other factors might explain this pattern. First, we know that co-authorship has become much more prominent in the IR literature over time. Second, we know that realism has become a less popular theoretical framework over time as seen in surveys of IR scholars and as documented in publication patterns of peer-reviewed research of both books and articles. Third, we know that women are more likely to co-author than men and are less likely to self-identify as realists. Finally, we know that liberals are more likely to study different issue areas and use different methods than realists. Some or all of these factors might shape the propensity to co-author.

Yet, even when we control for these factors and some others (as reflected in these more extensive regression results), realism retains a significant and inverse relationship with co-authorship. As Walt suggests, realists do appear to solo-author, in part, because they are realists.

We did not, however, find significant correlations between co-authorship and liberalism (or any of the other major IR paradigms, for that matter). Prior to controlling for methodology, it looks like liberal and non-paradigmatic scholars are more likely to co-author (see Model 1), but once we control for methods, the relationship disappears. This suggests that methodological diversity, not personality, is driving co-authorship among liberal and non-paradigmatic scholars. Below, we chart the predicted probability of co-authoring by paradigm after controlling for methods in articles published since 2000.

It is not surprising that type of methods is more strongly related to co-authorship than realism (or any other paradigm). Quantitative or experimental methods are more likely to appear in co-authored articles, whereas qualitative, descriptive, or analytic nonformal methods appear more often in solo-authored pieces. We would guess that these results reflect diversity in or need for specialized skills in order to answer certain types of questions, publish in certain journals, or get past peer reviewers in certain issue areas, rather than an author’s personality type. Regardless, the relationship between realism and solo-authorship is striking and worthy of further research, speculative blog posts, and conversations at the ISA bar.

On the more interesting question of why different scholars select different theories to help them explain the world of international politics, our rough and ready empirical analysis tells us very little. It does confirm one aspect of Walt’s 1st image hypothesis of theory selection: If a scholar is a realist, he or she is more likely to solo-author. Moreover, the mixed results for the other theoretical approaches suggest that even if personality or worldview is one factor behind the relationship, the exact dynamics are likely complicated as personality interacts with other factors.

Lindsay Hundley is a research associate at the Institute for the Theory and Practice of International Relations at the College of William & Mary. In the fall she will be studying for her Ph.D. in political science at Stanford University.