Explaining Support for ISIS in Iraq: Experimental Evidence from a National Survey

• In Event: Ideology and Support for Violence

Abstract

This paper presents unique, new evidence on popular support for ISIS from a large national survey we fielded in Iraq at the height of the insurgency in 2015. While most public opinion research in conflict theaters relies on direct question techniques, which likely yield biased responses to sensitive questions about views on warring factions, we administered the first survey experiments in Iraq in order to attain reliable measures of civilian attitudes toward ISIS, the Iraqi government, and U.S. airstrikes. This research design and data enable us to demonstrate asymmetric dynamics in support for civil-war parties, which run counter to existing public opinion research in rebellious environments.

A near-consensus exists among practitioners of counterinsurgency that the outcome of counterinsurgency campaigns critically depends on the relative success of each side to win the support of civilians in the conflict theater (Thompson, 1966; U.S. Army, 2007). The harm warring factions inflict upon civilians is a key determinant of civilian attitudes and behavior (Shaver and Shapiro, 2015; Weidmann and Zürcher, 2013). The recent literature argues that violence against civilians shifts the victims' allegiance away from the perpetrator toward its opponent (Lyall, Blair, and Imai, 2013; Condra and Shapiro, 2012). In this paper, we argue that an external military intervention that favors the government changes this dynamic. While we expect that victimization of civilians by an insurgency deteriorates public attitudes about the insurgency and rallies support for the counterinsurgency, violence against civilians at the hands of the government does not necessarily increase support for the insurgency. Instead, we expect that government violence enhances support for the external intervention – the other party in the counterinsurgency – as long as the foreign intervention is viewed as independent from the government that harms its citizens. The experience of victimization by the government thus may not turn civilians into sympathizers of the insurgency when there is an external third party that also fights the insurgency and that causes less harm to civilians.

In line with our expectation, we find that Iraqi civilians do not engage in a binary choice between supporting the insurgency or the government. Our list experiments show that victimization by ISIS is associated with a lower probability of sympathy for ISIS and stronger support for the Iraqi government. At the same time, victimization at the hands of government forces does not shift support away from the government and toward the insurgency. Instead, it is associated with stronger support for U.S. airstrikes against ISIS – but only among respondents who view the airstrikes as independent of the government. When these strikes are tied to the Iraqi government, victimization by government forces is associated with lower support for U.S. airstrikes. We experimentally manipulate perceptions of the government’s endorsement of the U.S. airstrikes through a randomized cue embedded in the survey experiment.

This paper makes three contributions to research and public policy. First, it tests a novel argument on popular support for civil-war parties. Second, new data from an original survey in Iraq and indirect question techniques yield the first robust evidence on Iraqi public attitudes toward ISIS and its opponents. Third, our findings imply that being viewed by the public as independent alternatives to the insurgency can help counterinsurgent governments and their foreign allies maintain popular support for the counterinsurgency even if one counterinsurgent inflicts collateral damage on the civilian population.

Authors

• Christoph Mikulaschek, Harvard University

• Saurabh Pant, Fondation JJL-IAST