Book Project:

Seeing is Disbelieving: The Depths and Limits of Fake News and False Beliefs in War

Abstract: From Syria to South Sudan to Ukraine, “fake news” and misinformation are an endemic feature of modern warfare. But when is such false information actually believed by the people who live in conflict zones, and when is it not? This question is important, as the spread of false beliefs can provoke violence, trigger refugee flows, and impede efforts to find peace. In this book, I advance a new argument about the formation of factual beliefs in war. In contrast to intuitive ideas about people believing whatever they want in war, I argue that the accuracy of people’s wartime beliefs hinges critically on their exposure and proximity to the events in question. While war is rife with lies, those who are close enough to the action have the means and motives to sift through them. In short, when it comes to wartime misinformation, seeing is disbelieving.

I investigate these dynamics by examining original surveys and experiments fielded in Pakistan and Iraq, a range of violent event data from both disputes, and the pairing of the two in new and powerful ways that reveal how factual beliefs change with exposure to the fighting. Additionally, my analyses are also informed by fieldwork in Pakistan and Jordan, and supported by a wealth of anecdotal material from a wide variety of conflicts. Ultimately, the book shows that – as in peace – lies, propaganda, and fake news are pervasive in war, but they can be punctured by high stakes and personal exposure. In so doing, it offers key contributions to both our understanding of the dynamics of modern armed conflict and the role of facts in contemporary politics.


What Shapes Civilian Beliefs about Violent Events? Experimental Evidence from Pakistan. Journal of Conflict Resolution, Advance online publication. PDF. Online Appendix. Replication Materials.

Abstract: Why do civilians in warzones often hold widely divergent beliefs about what is happening in the fighting? While there is a burgeoning literature on the micro-dynamics of armed conflict, variation in civilians’ factual beliefs has received scant attention. Yet such beliefs are critical, as they form the basis for wartime opinion and action. I argue that – particularly for civilians outside the direct “line of fire” – this variation comes not chiefly from an event’s empirical nature, but from civilians’ prior political orientations in the dispute. In order to investigate these dynamics, I fielded a survey experiment in Pakistan in which I manipulated the features of a reported counterinsurgent air strike and then measured civilians’ ensuing beliefs about it. The results show that these beliefs are most driven by the perpetrator’s identity and civilians’ own preexisting attachments. While actual casualty levels matter too, these findings suggest that civilians’ beliefs about conflict events are often deeply biased in nature.

Foreign Policy Attitudes towards Islamic Actors: An Experimental Approach (with Mujtaba Ali Isani). Political Research Quarterly 69(3): 571-582. PDF. Online Appendix. Replication Materials.

Abstract: This article examines how Western foreign policy opinion reacts to the perceived Islamic character of foreign actors. Studies show that the target actor’s dominant religion is a key ingredient in foreign policy opinion: Western audiences react more hostilely to “Muslim” than “Christian” targets. Yet actors differ not only in which world religion they belong to, but also how that religion is politicized both by themselves and by others. We argue that Islam can be politicized in three major ways – via Islamic rhetoric, policies, and labels – that shape foreign policy attitudes. To examine our claims, we field a survey experiment in which we attach common Islamic rhetoric (“Allahu Akbar”), policies (“Shari’a law”), and/or labels (“Islamist”) to a foreign actor in the context of the Syrian civil conflict. We find that these cues strongly harm attitudes towards the actor, and the results vary widely by type. Indeed, the Shari’a policy cue does the most damage to attitudes, emotions, and preferences towards the actor. Moreover, the Islamic cues reinforce each other in fueling these fearful reactions, and are particularly potent on conservative citizens. These results paint a richer picture of how outgroup religious cues influence foreign policy attitudes.

Ideas, Interests, and Information: The Sources of Iraqi Support for ISIS (with Munqith Dagher and Karl Kaltenthaler). Studies in Conflict & Terrorism, Advance online publication. PDF.

Abstract: This article explores the amount and sources of support for the Islamic State among Iraqis. We argue that, in addition to shared identity and ideology, a neglected factor in debates about support for Islamist militancy is the messaging and information that individuals receive about a given group. We test these arguments using regression analysis on public opinion data collected in Iraq in April 2015. The analyses largely support our contentions, showing that exposure to news coverage of the Islamic State in Iraq and the Levant substantially reduces support for the group, even among alienated Sunnis or ideological Islamists.

Working Papers:

Seeing is Disbelieving: The Depths and Limits of Factual Misperception in Modern War (with Munqith Dagher and Karl Kaltenthaler). Under Review. PDF.

Abstract: When do civilians develop accurate – or biased – factual beliefs about what is happening in war? While there is ample literature on the micro-dynamics of conflict, there is little work on peoples’ factual beliefs in war. Yet such beliefs are crucial, as they form the basis of wartime opinion and action. In this study, we build an original theory of factual beliefs in conflict which distinguishes between two types of civilian communities based on their exposure to the fighting. We test these claims with a unique survey in Iraq which asked a series of factual questions about the Coalition anti-ISIL airstrikes in Iraq. We find that factual misperceptions about the strikes are widespread, but civilians with more exposure to the campaign are far less likely to embrace these falsehoods. This reveals that lies, misperceptions, and “fake news” are endemic in war, but can be punctured by high stakes and direct exposure.

Public Fears of Terrorism, Partisan Rhetoric, and the Foundations of American Interventionism (with Christopher Gelpi and Daniel Kent). PDF.

Abstract: An American’s yearly chance of being killed by a terrorist attack sits at about 1 in 3.5 million. Yet over 40% of the public have been consistently found to believe they or a family member are likely or very likely to be the victim of a terror attack. Can these heightened estimates of the risks posed by terrorism be brought closer to reality? With trillions of dollars spent on the War on Terror, this question is not just theoretically but practically important. To investigate, we use an experimental approach, testing whether participants update beliefs when presented with factual information about the risks posed by terrorism. In a pilot experiment conducted on Mechanical Turk in late 2017, we found that people did update their risk estimates to more accurate levels when given factual information, but that the information only had a strong and enduring effect when it was reinforced by a cue from a co-partisan political elite. In this paper, we first present these pilot results and then lay out a design for a revised and expanded study which will build on them. In particular, the revised design will: (1) conduct the survey experiment on a nationally representative sample, and (2) test the impact of a richer set of elite cues as treatments. Overall, this project helps show how the American public’s overblown fear of terrorism – a particularly costly and sticky misperception in contemporary world politics – can best be mitigated and reduced to more realistic levels.

Rebel Leader Attributes in Civil Wars: Introducing the ROLE Database (with Benjamin Acosta and Reyko Huang). PDF.

Abstract: Existing literature on civil wars relies predominantly on national and group-level factors to explain their dynamics and outcomes. In contrast, we propose in this paper that individual rebel leaders’ personal attributes help explain the behavior of the organizations they lead. Just as IR scholarship has long highlighted the importance of state leaders’ biographical attributes in explaining interstate war and diplomacy, so we build on existing studies in political science and psychology to suggest that the characteristics of rebel leaders affect their groups’ decisions and actions in civil wars. In order to do so, we introduce original data from the Resistance Organization Leaders (ROLE) Database, which contains a range of biographical information on all rebel, insurgent, and terrorist leaders in civil wars between 1980 and 2011. We first explain the construction and content of the database, and then use it to present a number of descriptive statistics on rebel leaders as well as to replicate and extend a recent study by Wood and Thomas (2017). Ultimately, our paper urges a new research agenda that goes beyond rebel groups or campaigns as the units of analysis to examine the importance of the people who organize and lead them in conflict studies.

Personal Control, Political Control, and Belief in Conspiracy Theories: Evidence from the Islamic World (with Lars Berger). PDF.

Abstract: In recent years, there has been a surge of scholarly interest in public belief in conspiracy theories. In addition to motivated bias against the perpetrator, one of the key explanations for their broad appeal is perceived powerlessness and personal lack of control. In this article, we argue that the perception that an individual lacks personal control is different from, and less relevant than, the perception that a potential conspirator has power and control. In many Middle Eastern countries, for example, the notion that the U.S. has vast control over world events is widespread, helping to fuel the appeal of anti-American conspiracy theories. We test our argument using existing public opinion data about several American-centric conspiracy theories in four major Islamic countries – Egypt, Morocco, Pakistan, and Indonesia. The results confirm that perceptions of U.S. control have a significant impact independent of views of the U.S. or perceived personal lack of control. These results show that conspiracy theories thrive when perceived motives and perceived means align, and hint at the costs of actors overinflating their strength in world politics.

Help from Any Quarter? Sectarianism, Nationalism, and Support for External Intervention in Iraq (with Munqith Dagher and Karl Kaltenthaler). Under review.

Abstract: What drives citizens’ attitudes toward external military intervention in a society experiencing an armed conflict? Conventional wisdom suggests that nationalism is a crucial source of opposition and resistance to such intervention. In contrast, we argue that the impact of nationalism on views of outside intervention hinges on the strategic context facing the target nation. When the primary threat to the country comes from the intervener itself, nationalism will indeed reduce support for outside intervention. But when that threat comes from elsewhere, nationalism will actually boost support for external intervention in order to defeat it. In order to explore these dynamics, we use public opinion data from a unique survey fielded across Iraq in late 2016 that includes questions about the outside intervention against the Islamic State by the U.S.-led Coalition as well as Iran. The results are broadly consistent with our argument, showing that unlike other key factors such as sectarianism, nationalism pushes Iraqis to seek foreign military help from any quarter when it is deemed necessary for national integrity or survival.

Too Late to Apologize? Collateral Damage, Post-Harm Compensation, and Insurgent Violence in Iraq. Revise & Resubmit.

Abstract: A key piece of conventional wisdom among scholars of modern armed conflict is that collateral damage is often strategically quite costly in war. Yet most combatants already know this and take actions after mistakes – most prominently, the distribution of “condolence payments” to civilian victims – in order to mitigate these costs. Do these payments work? This question is important not only for policymakers but also for deeper theoretical debates about how civilians respond to combatant signals in war. In order to examine these issues, I use micro-level conflict event data on 4,046 condolence payments made by Coalition forces to civilian victims during the Iraq War from 2004 to 2008, matching it with corresponding data on collateral damage and insurgent violence. The results of this analysis reveal that post-harm compensation does significantly diminish local rates of insurgent violence, and that this is true across different types of payments and different levels of prior trust in the perpetrator. Ultimately, these patterns can be best explained by a rationalist mechanism in which civilians update their beliefs about violent events based on new information about combatants’ wartime intentions. The results thus provide a compelling strategic rationale for combatants to compensate their victims in war, and suggest that civilians are not blinded to new information about conflict dynamics by their preexisting biases.

Drone Strikes and “Hearts and Minds”: A Quasi-Experimental Analysis in Pakistan. PDF.

Abstract: How does counterinsurgent violence influence the “hearts and minds” of civilian populations? Despite extensive theoretical debate, we have little systematic evidence on this matter. I add to this literature by analyzing how one increasingly important form of violence in modern asymmetric conflict – targeted killing across state borders – influences “hearts and minds” in the target society. In particular, I exploit a natural experiment in the timing of U.S. drone strikes and public opinion surveys in Pakistan that allows me to estimate the effects of these operations on national Pakistani opinion. The results of this analysis show that the strikes do have substantial anti-American, anti-incumbent, and pro-militant effects on the Pakistani population. Yet closer analysis also reveals important nuances in these responses. Indeed, while the strikes legitimize some target organizations (e.g., the Taliban), they do not fuel support for other prominent militant groups (e.g., Al Qaeda). Moreover, they actually boost Pakistani support for U.S. financial and humanitarian aid to fight extremism, creating demand for policy change rather than simply disengagement. Ultimately, the results show that cross-border targeted killing does alienate and radicalize “hearts and minds” across the targeted society, while also highlighting the limitations and potential opportunities for mitigation of these effects as well.

The Only Legitimate Game in Town? Attitudes towards Democracy and the Caliphate in the Islamic World (with Mujtaba Ali Isani).

Abstract: In recent years, essentialist claims about the incompatibility of democracy and Islam have been largely swept away by a wealth of public opinion research showing that democracy garners vast support in the Islamic world. However, while this literature has effectively demonstrated the popularity of democracy over authoritarianism, we contend that it misses an essential piece of the puzzle by not examining Muslim support for an alternative model of government: the Caliphate system. After surveying some of the most influential elite visions of the Caliphate throughout Islamic history, we analyze its appeal and meaning today with existing and original surveys across four major Muslim countries (Egypt, Pakistan, Morocco, and Indonesia). Our results show that, like democracy, the Caliphate is understood in largely instrumental terms, as a vehicle for broad social welfare and justice rather than a specific institutional configuration.

Beyond “To Shoot or Not to Shoot”: Rethinking Military Responses to Mass Uprisings after the Arab Spring.

Abstract: Authoritarian regimes facing mass uprisings often rely on their militaries to crush them in order to survive. Yet, historically, the responses of these institutions to revolutionary protest have varied widely, from Tahrir to Tiananmen and beyond. While we have studied why some militaries “shoot” while others choose “not to shoot,” we have neglected critical variation in these two categories that impacts regime change as well as civil war outcomes. In this article, I introduce a richer framework to capture this variation, and use it to generate new hypotheses about the role of patronage and coercion in these processes. I demonstrate the value added of the new framework and hypotheses through a comparison of Tunisian, Egyptian, Libyan, and Syrian military responses to the 2011 “Arab Spring” uprisings.