A recent poll by the Pew Research Center/USA TODAY is being described by commentators as revealing a high level of US public support for airstrikes in Iraq. But how much comfort can the White House take in these poll results?

The survey found that 54 percent of Americans approve of the airstrikes and 31 percent disapprove.

I. Limited to protection of Iraqi civilians?

The Pew Research Center provides a long and detailed summary of the findings, but the summary does not mention that the poll question specifically asked about support for airstrikes “in response to violence against civilians.”

Indeed, the word “civilian” does not appear anywhere in the Pew summary (you have to go to the full report). Likewise, commentary on the poll findings similarly overlooks this feature of the survey question.

How should this affect our interpretation of the results?

On the one hand, it may suggest the poll underestimates public support. That is, people were asked about support of airstrikes for the humanitarian objective outlined by President Obama (e.g., protection of the Yezidis). One might expect even greater (or additional) support for the other stated objective—protection of US personnel inside Iraq including at the US consulate. And one might expect even greater support still if the question involved the national security threat that ISIS could pose to the United States and our allies (let’s call that the third, unofficial objective).

On the other hand, it is remarkable how low American support is for the airstrikes given that the second and third objectives are a large part of the public discourse. The Pew Research Center, for example, also provided its poll results from the NATO intervention in Kosovo in 1999 as a historical comparison. In that case, American public approval for airstrikes ranged 53-62% during March-June 1999. But that campaign was much closer to a purely humanitarian mission.

Commentators, such as James M. Lindsay at the Council on Foreign Relations, have seized on the fact that the support for the Iraq airstrikes is “comparable to the support the public showed for the 1999 U.S. airstrikes against Serbia in the[sic] Kosovo.” Susan Page at USA Today (reminder: it is a Pew Research Center/USA Today poll) also notes that support for the Iraq airstrikes is just “a bit lower than public support for NATO airstrikes in Kosovo in 1999.”

But shouldn’t we expect much greater support for the airstrikes in Iraq which also serve more direct US national security interests? It is not as though Milosevic was threatening to attack the United States.

II. Civilian collateral damage from airstrikes

In an earlier post at Just Security, I described social science studies which suggest that American public support for airstrikes declines greatly in the event of civilian casualties. (Indeed, one of those studies involves a critique of polls like Pew/USA Today for failing to include the element of civilian casualties in the survey questions.)

There are two lessons here for the Pew/USA today findings on Iraq:

First, the American public support may be relatively shallow—and will drop if US airstrikes involve more complicated environments in which civilian casualties are likely to occur.

Second, if the social science research on civilian collateral damage is correct, American public support for the Iraq airstrikes is potentially far weaker than American public support for the Kosovo air campaign. Why is that? Because there have not been major incidents of civilian casualties from US strikes against ISIL in Iraq. In contrast, during the Pew surveys of public attitudes on Kosovo, NATO airstrikes resulted in multiple disastrous instances of civilian casualties. Some of the most salient and widely reported incidents of civilian casualties included: (1) an airstrike on a civilian passenger train on April 12, 1999; (2) an airstrike on a convoy of Albanian refugees on April 4, 1999; (3) an airstrike on Serbian Radio and TV Station in Belgrade on April 23, 1999; (4) an airstrike on the Chinese Embassy on May 7, 1999; and (5) an airstrike on Korisa Village on May 13, 1999.

It is, therefore, a bit of a false comparison to look at public support for current airstrikes in Iraq with no such incidents of civilian casualties and public support for Kosovo airstrikes with repeated civilian casualties. Imagine one campaign involved no loss of American life (with 54% public support) and another campaign involved a significant loss of American life (with around 54% public support)—it would be correct to describe the latter public support as much stronger.