Secondary Casualty Information: Casualty Uncertainty, Female Casualties, and Wartime Support

Abstract

I develop a theory of wartime public opinion that identifies different types of casualty information and the varied roles they play. These roles include Primary Casualty Information (e.g., monthly casualties), which directly influences opinion; Contextual Casualty Information (e.g., enemy casualties), which mitigates the effect of primary casualties; and two new concepts, Casualty Uncertainty, the inability of individuals to infer casualty patterns, and Secondary Casualty Information, which represents alternative cost information. I theorize that when casualty uncertainty is low, and casualty patterns are clear, individuals rely on primary casualty information for evaluating a conflict. As uncertainty increases, however, and the picture painted by primary casualty data becomes increasingly opaque, individuals employ secondary casualty information to evaluate costs. I use the reporting of female casualties to represent an example of secondary casualty information. If casualty uncertainty is low, I anticipate that the existence of female casualties should have little influence on public attitudes. If casualty uncertainty is high, however, the reporting of female casualties will exert a negative effect on public support of the conflict. An experimental study finds strong support for these arguments. I examine how the results contribute to our understanding of wartime public support, women in combat, and the Iraq War.

Keywords:

• casualties
• public opinion
• uncertainty
• war
• women

Notes

¹“Insurgents have taken over the country of Fiji, an island in the Pacific, and have threatened to kill the 1,100 Americans living and visiting there and to destroy the $87 million worth of US property located in the country. The US has begun a military intervention, deploying the 101st Air Mobile and 82nd Airborne divisions—totaling approximately 32,000 troops. Their objective is to defeat the insurgency and restore the previous government.”

²Given that each casualty pattern adds up to the same 1380, I am unable to include cumulative casualties (which with monthly casualties would perfectly predict duration).

³The large number of ties leads to inefficient hazard ratio estimates (Box-Steffensmeier et al., 2004). I employ the Cox estimator Exactp (using Stata™), which provides more efficient estimates. Analyses with uncorrected Cox estimators are similar (not shown). Those who support intervention after the last round are censored.

∗ p < 0.1;

∗∗ p < 0.05;

∗∗∗p < 0.01.

⁴ Monthly Casualties cannot be included since they are unique to each pattern-specific time period.

⁵ Female Casualties includes the reports of both female and male monthly casualties.

∗ p < 0.1;

∗∗ p < 0.05;

∗∗∗p < 0.01.

∗∗ p < 0.05;

∗∗∗p < 0.01.

⁶Inclusion of the variables Medium Uncertainty and Medium Uncertainty ∗ Female Casualties in either model has no effect, and is dropped from the analysis for ease of presentation (not shown).

⁷The question reads: “In general how supportive or concerned are you about the integration of women into the U.S. Armed Forces?”