Book Review: Multimodal Political Networks

The analysis of political networks has grown substantially since David Knoke’s initial examination of the structural perspective on political behavior. Although these political behaviors usually unfold in networks that include multiple kinds of actors, the field has been slow to adopt new multimodal network analytic methods. The authors of Multimodal Political Networks aim to overcome this inertia, by highlighting “additional opportunities multimodal political networks offer… by introducing intermediate and advanced methods for analyzing such networks” (p. 4). The volume therefore takes an important step in this direction by offering numerous examples of the kinds of questions for which multimodal political networks are useful, and thus will be thought provoking for political network scholars.

The authors begin with a conceptual overview of multimodal political networks, presenting a central argument that has two components. First, in general, such networks can be viewed as comprised of individual actors (e.g. people) and collective actors (e.g. organizations) that have agency and therefore can interact with each other, as well as with entities that lack agency (e.g. events). Second, these networks coordinate political behavior by organizing the actors into communities by shared location or interest. The conceptual overview is followed by a methodological primer that introduces key concepts in the analysis of 1-mode networks, 2-mode networks, projections of 2-mode networks, and 3-mode networks. The remaining chapters apply these techniques in examples covering a broad range of political phenomena including political leaders, public policy, participation in civil society, collective action, international relations, and political action committees’ legislative influence. The volume concludes with a similarly broad discussion of future substantive applications for multimodal political network analysis on topics such as the environment, healthcare, and public health.

Several strengths make this volume a valuable resource to those interested in political networks, as well as to those more broadly interested in multimodal networks. First, the methodological primer presented in chapter 2 is concise but clear. By starting with the basics of ordinary 1-mode networks, the authors provide a very useful introduction to networks that makes the rest of the volume accessible even for newcomers to political networks. Indeed, because the publisher offers individual chapters and (at the time of writing) chapter 2 is available open access, this primer gives any student of social networks an excellent place to start. Second, the field of political networks is quite diverse, yet the authors have done an outstanding job covering that diversity through the range of contexts from which their examples are drawn. These examples include case studies in the US, UK, Italy, Germany, and at the global scale, and focus on different types of political actors including individuals, elites, organizations, and states. As a result, this volume not only provides a broad overview of political networks, but of issues in comparative politics generally. Further complementing these empirical examples, the authors provide access to seven of the associated datasets. Third, the authors have struck a balance between theory and method, taking care to highlight the close link between the two in network analysis and explaining that “theory and method go hand-in-hand” (p. 197). Perhaps the key strength of this work is that it reads neither as a book about political networks alone, nor as a book about multimodal network analysis alone, but as a book about using multimodal network methods to explore theories about political behavior.

This volume usefully connects the theory and method of political networks through a range of examples. However, certain features may limit its utility as a practical guide to actually applying these methods. First, the description of analytic steps often omits key details that will make it difficult for readers to follow along. For example, the authors note that “due to space constraints, we only present a summary of the insights generated from three different [methods]” (p. 116). Potentially exacerbating this challenge, the analyses rely entirely on UCINET, which is a closed-source commercial software available only for Windows-based computers. The authors once mention an R package associated with the book called mi-graph, but never refer to it again. Second, despite the volume’s title, the empirical examples deal with a relatively restricted range of multimodal networks. For example, although the authors claim that “3-mode network analysis offers a way to examine the interaction between structure and agency,” (p. 91) only chapter 6 (and briefly in chapter 4) examines a 3-mode network, with most chapters focusing instead on 1-mode and 2-mode networks. Similarly, although the authors encourage readers to analyze 2-mode networks because “projection is neither helpful nor necessary” (p. xvii), chapter 5 presents an analysis of projections and chapter 7 estimates a model where the dependent variable is a projection. Finally, throughout the empirical examples, the identification of communities in political networks is a central theme. However, these analyses switch between different community detection methods (e.g., CONCOR, Louvain, affiliated graph model) without explaining why, and none of the analyses use community detection methods developed specifically for multimode networks that the authors discuss earlier in the volume. Without understanding how or why to choose a specific community detection method, readers may be left unsure of what to make of the fact that “By applying three community-detection methods to the same restricted 3-mode legislative network, we uncovered three distinct legislative community structures“ (p. 194).

In sum, this volume has much to offer, but will appeal to different readers for different reasons. It will appeal most to those wanting to get a broad overview of the types of questions that can be asked about political networks, and the types of data that are useful for answering them. Indeed, the final chapter sketches several promising avenues for multimodal political network research that will give readers a head start on what to do next. For those interested in network methods, it will especially appeal to those interested in community detection in political networks, which is an analytic goal in five of the six empirical chapters. It will also appeal to those interested in 2-mode analysis, but will be less useful for those seeking a detailed discussion of 3-mode and higher-order multimodal structures. Finally, for those looking for a practical guide to multimode political network analysis, the second chapter offers a useful overview. However, this volume stops short of walking readers through how to perform such analyses.