Testing the “Proximate Casualties Hypothesis”: Local Troop Loss, Attention to News, and Support for Military Intervention

Abstract

In academic and policy circles, it is believed that the American public is casualties-averse when sending its troops into war and that as casualties mount the public will increasingly call for withdrawal of troops from foreign military interventions. This study tests a variant of this “casualties hypothesis” by examining whether the public conceptualizes the human costs of war in local as well as national terms. Using a massive quantity of public opinion data combined with information on deaths to U.S. soldiers in the Iraq war theater between 2003 and 2006, we show that recent deaths to soldiers from a person's home state prompt an increase in the likelihood of the person supporting withdrawal of troops, independent of cumulative national casualties and recent casualties to soldiers from elsewhere in the country. These results are explained in terms of journalism norms that make recent, local casualties salient to consumers of news, who then use this salient information when their opinions about the progress of a war are probed.

Notes

∗The Pew Research Center for the People and the Press bears no responsibility for the analyses or interpretations of the data presented here.

¹By this choice, we are not suggesting that we believe those who say they do not know and those who say the military should stay in Iraq have the same attitude. Our focus in the analysis we report here is on modeling the likelihood a person reporting the United States should withdraw rather than saying something else, and whether this likelihood increases as a function of proximate casualties. We saw no need for the purpose of this analysis to model attitudinal differences between those who report something other than a withdraw response.

Note. All predictors are grand mean centered, with the exception of the interactions, which are based on products of grand mean centered variables. Results are from the unit-specific model with robust standard errors. n = 30,666 respondents (level-1 unit) distributed across 48 states and the District of Columbia (level-2 unit). Attn = attention to Iraq news; Prox. = proximate.

^a t = 16

^b t = 23.

∗p < .05. ∗∗p < .10. ∗∗∗p < .001.

²We ran an identical set of analyses after discarding participants who said they didn't know or refused to answer the question about withdrawing troops. The results were largely the same. The only notable difference was a slightly weaker interaction between proximate casualties at some values of t and attention (with the interaction changing from significant to marginally significant in a few instances).

Additional information

Notes on contributors

Andrew F. Hayes

Andrew F. Hayes (Ph.D., Cornell University, 1996) is an associate professor in the School of Communication at The Ohio State University. His research interests include statistical methodology and public opinion.

Teresa A. Myers

Teresa A. Myers (M.A., The Ohio State University) is a doctoral candidate in the School of Communication at The Ohio State University. Her research interests focus on the intersection of media, public opinion, and foreign policy.