Why language matters: Shaping public risk tolerance during deterrence crises

ABSTRACT

Despite the obvious the risks involved in deterrence bargaining, we know surprisingly little about how the public evaluates risk during a crisis. A limited deterrence scholarship considers domestic preferences and tends to assume that the public’s risk tolerance remains stable throughout a deterrence episode. Yet, robust findings in cognitive psychology suggest that people’s risk tolerance can shift dramatically based on how the potential outcomes of military options are framed. We evaluate these competing views through a series of decision experiments grounded in prospect theory. Across an array of potential extended deterrence scenarios, we demonstrate that simply reframing the language used to describe the possible outcome of military options affects both the public’s willingness to accept risks and, therefore, their willingness to escalate the crisis. These findings condition long-standing concerns about the public’s intransigent risk disposition, which underpin research on costly signaling, deterrence traps, and nuclear self-deterrence.

KEYWORDS:

• Prospect theory
• foreign policy analysis
• deterrence
• United States
• experiments

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Jeffrey D. Berejikian is a Josiah Meigs Distinguished Teaching Professor at the University of Georgia, and Senior Fellow at the Center for Trade and Security. His research on cognitive science and foreign policy decision making appears in the top journals in the discipline of political science, including the American Political Science Review, Political Psychology, and the Journal of Conflict Resolution. Recent work (published and forthcoming) undertakes the statistical analysis of risky conflict behavior by incorporating prospect theory into traditional tools of foreign policy analysis. His research on prospect theory and deterrence is referenced in the Deterrence Operations Joint Operating Concept. His book on prospect theory and international politics (SUNY Press) and has been translated into Korean.

Zachary Zwald is in the Department of Political Science at the University of Houston, where he serves as the director of the Center for International and Comparative Studies. His research focuses on the psychological aspects of decision-making on national security issues. Recent work examines nuclear deterrence policy preference formation, risk disposition in multidomain deterrence crises, and the cognitive drivers of human trust in autonomous weapon systems. His forthcoming book advances the first systematic analysis of the information processing dynamics that lead to judgments and decisions during military technology innovation.

Notes

1 For a more detailed discussion of prospect theory, (see Kahneman, Slovic, and Tversky 1982; Levy 1992; Tversky and Kahneman 1991).

2 One of the benefits of experimentation is that it permits investigators to circumvent an inherent challenge to applying prospect theory to explain foreign policy behavior, namely that prospect theory does not provide a “theory of framing.” As Kahneman and Tversky (1982, pp. 171–172) observe, in the real world, an actor’s “reference point is the state to which one has become adapted” and that there are many cases in which “the reference point is determined by events that are only imagined.” However, in an experimental setting, the reference point is set by the terms of the experiment.

3 See Svyantek and DeShon (1991) for an alternative view of the relationship between framing and catastrophic consequences.

4 While we disagree with the characterization that prospect theory explains “errors” in decision making, we agree with the broader point that prospect theory identifies empirically robust patterns of non-maximizing behavior.

5 2002 National Strategy to Combat WMD, p. 3

6 2006 Deterrence Operations-Joint Operating Concept, p. 21.

7 The dependent variable is coded as deterrence = 0 and escalation = 1.

8 We also recruited our sample for the follow-on analysis from Mechanical Turk. We used Mturk worker IDs and IP addresses to screen subjects and assure participants in the first panel did not participate in the second.

9 The dependent variable in this experiment is reverse-coded: 1 = deterrence, 0 = escalation.

10 Models implemented in Stata 16. Results reported with standard (non-clustered) errors.

Additional information

Funding

This research was funded by a research grant from the Center for Public Policy at the University of Houston's Hobby School of Public Affairs.