In the decades following the formative observation in John Mueller’s (1973) War, Presidents, and Public Opinion that popular support for war declined as wartime deaths rose, a large and vibrant literature has struggled to make sense of this pattern. International relations scholars have come to favor a rational calculus approach that interprets collective-level opinion dynamics as a function of perceived benefits, costs, and chances of success for a given military action (e.g., Christopher Gelpi, Peter Feaver, Jason Reifler [2009], Paying the Human Costs of War). In contrast, a rival approach informed by public opinion research (e.g., Adam Berinsky [2009], In Time of War) and political communication scholarship (e.g., Matthew Baum and Timothy Groeling [2010], War Stories) favors an elite signaling model in which types and levels of dissensus among political leaders serve as a primary driver of mass opinion change during wartime. Scott Sigmund Gartner and Gary M. Segura’s Costly Calculations draws upon the American experience of war since 1950 to move the terms of debate onto firmer theoretical ground by developing a “general framework for understanding war initiation, war policy, and war termination in democratic politics, and the role that citizens and their deaths through conflict play in those policy choices” (p. 2).

In contrast to earlier versions of the rational calculus approach, Gartner and Segura move casualty information into the center of the theoretical story. Their “price theory of war” posits that anticipated losses serve as a key component for weighing the value of a war’s potential benefits, while already-incurred losses serve an informative heuristic for assessing the war’s likely outcome. Against the elite signaling approach, the authors offer evidence that position taking by elected leaders is structured by expectations of future losses, as well as by the losses occurring within their constituencies. But while Gartner and Segura’s account places casualty information as the primary mover of war support for both elites and masses, theirs is no reductionistic story. Gartner and Segura’s nuanced theoretical framework entails a sophisticated array of variables and conditioning factors that together yield a satisfyingly complex theoretical account.

The authors’ theoretical model hinges on the observation that wartime deaths serve as information that is both readily available and highly credible for assessing how a conflict is going. Both citizens and leaders are posited to be sensitive to the Expected Total Costs (ETC) of a conflict: “an individual’s estimate of the total costs that will be required to achieve the understood aims of the war effort” (p. 40). ETC can only ever be estimated, so individuals are posited to infer ETC dynamically from three forms of casualty information (pp. 40–52): their perceptions of the cumulative number of casualties that have occurred since the start of the conflict (which tell them how the war has gone so far), their perceptions of recent casualty numbers (which tell them how the war is going right now), and their perceptions of whether recent casualty levels are rising or falling (which tell them how the war is likely to go in the future). Individuals are also posited to have a Reservation Point (RP), which “is the number of estimated—that is, perceived—casualties the observer is willing to expend to accomplish a specific goal” (p. 55). These two pieces of information offer a straightforward heuristic for assessing the merit of a conflict: “When these Expected Total Costs (ETC) of a war exceed the value of the goal itself (i.e., the price the citizen is willing to pay), the citizen has passed his or her Reservation Point (RP), and will no longer support paying those costs” (p. 40). The authors expect ETC to vary over time and across individuals as casualty information and other news about a war differentially reaches citizens. Individual-level variation in ETC comes from several factors, including local experiences of war casualties, but should become increasingly more accurate the longer a war lasts. In contrast, RP is conceived as a fairly stable element with potential to shift only if the war’s goals and tactics change over time (p. 55).

In terms of organization, the book begins by laying out this new “price theory of war” that positions information about military casualties as a central component for assessments about a war’s potential value and expected costs (chapt. 2). The argument then turns to explicating how citizens assess the merits of war before the onset of conflict and as the conflict progresses, detailing a variety of component elements relevant for calculating the model’s RP term (chapt. 3). Several expectations from the theoretical model are then tested, yielding evidence that casualty sensitivity among citizens operates in ways that align with the premises of the theory in experiments (chapt. 4) and in observational data (chapt. 5). After considering the impact of localized attention to war casualties as well as observational data on social network connections to military losses (chapt. 6), the book’s empirical section concludes with case studies exploring how casualties affect the position-taking behavior of US Senate candidates as well as how those positions have electoral consequences related to constituent experiences of localized losses (chapt. 7).

Marshalling an impressive array of experimental and observational data, Gartner and Segura succeed in establishing the plausibility of their theoretical model. Nonetheless, the authors acknowledge that because their primary aim is to develop a conceptual framework that can generalize across conflicts, they test only some of the propositions that their theoretical model entails. This leaves ample ground for other researchers to explore, refine, and dispute elements of Gartner and Segura’s innovative model. Three areas seem especially ripe for further investigation. First, leader dynamics are less thoroughly tested here than mass opinion dynamics, and because the book’s “price theory of war” seems especially well-suited to elite decision making there is an important opportunity for other scholars to apply this approach more broadly for understanding elite position taking and conflict-initiating behavior. Second, the book’s theoretical story currently lacks a well-grounded psychological model, as the complex weighing of expected versus perceived costs runs against the grain of current thinking within political psychology about how judgmental processes operate. The authors acknowledge that the RP component of their model is unobserved and can only be assumed (p. 226), even though “the shape of the RP distribution is the critical component in determining how casualties manifest themselves in domestic politics through opposition to ongoing conflict” (pp. 65–66). A theoretically compelling individual-level account of psychological mechanisms underlying ETC and RP calculations would help to solidify the case for the book’s theoretical model. Third, although the authors provide ample evidence that mass casualty sensitivity follows the general contours of their theoretical model, the book offers no head-to-head test of their model against alternative explanations such as those offered by the elite signaling approach or by previous formulations of the rational calculus approach. This opens a good opportunity for future research to assess the relative strengths of competing claims across these different theoretical perspectives.

Costly Calculations serves as a capstone synthesis that cumulates insights from the authors’ nearly quarter-century collaborative effort to understand how wartime casualties influence opinion dynamics. It presents the most thoroughgoing, ambitious, and nuanced version of the rational calculus approach to date. Gartner and Segura’s volume is equally valuable as an invitation for more scholarship to test its predictions, fill its gaps, and refine its intuitions about the underlying mechanisms that structure how domestic politics is affected by the conduct of war. While offering a more forceful challenge to earlier formulations of the rational calculus approach than to the elite signaling approach, the theoretical clarity of the book’s conceptual model is a signal achievement that has much to recommend it to both camps.