In the event that US troops are killed in the course of a peacekeeping operation is the majority of Americans likely to respond by wanting to withdraw US troops?

Summary:  In the event of US troop fatalities in the course of a peacekeeping operation, the majority of Americans is unlikely to favor an immediate withdrawal of US troops.  The majority is more likely to favor an assertive response such as reinforcing US troops or striking back at the attackers.  Only a minority of the public wanted to withdraw immediately in response to the troop fatalities in Somalia in 1993.  Although a majority of Americans mistakenly believes that a substantial number of US troops have been killed in the peacekeeping operation in Bosnia, the majority continues to support US participation.  Presented with a variety of hypothetical scenarios involving substantial fatalities in peacekeeping operations, only a small minority says it would respond by wanting to withdraw US troops.

Whether such fatalities would affect support for the operation itself appears to be primarily a function of the perceived success of the operation. 

Somalia

            Shortly after eighteen American soldiers were killed in a Somalia firefight in October 1993, television networks broadcast graphic pictures of dead GIs being dragged through the streets of Mogadishu. The very evening following the news reports of the deaths, with these images still fresh in the minds of the American public, polls taken by ABC and CNN/USA Today. Only 37% and 43% of respondents, respectively, said they wanted US troops to withdraw immediately. Three other polls taken over the next week produced similar results.

            Moreover, it appeared that a majority of Americans favored increased involvement after the firefight, at least in the short run. In polls by CNN/USA Today, ABC, and NBC, respectively, 55%, 56%, and 61% supported sending more US troops. Seventy-five percent favored retaliating against Somali warlord Mohammed Farad Aidid with a “major military attack” if American prisoners could not be released in a timely manner through negotiations, ABC found.

            Several polls showed that a majority did want to withdraw at some point and did not want to stay in until the country was fully stabilized.  A PIPA poll taken October 15-18 found that only 28% wanted to withdraw immediately, 43% supported the President’s plan of withdrawing in six months and only 27% did not want to withdraw “until we have stabilized the country, even if it takes longer than six months.”  Some polls found as many as 65% saying that the US should withdraw from Somalia, though when asked specifically whether this meant that the US should withdraw now, less than a majority felt such urgency.

            However, it does not appear that this desire to eventually withdraw was prompted primarily by the fatalities.  Earlier polls show that it was already in place before the fatalities occurred.  In September 1993, 57% favored stopping US involvement in combat with Somali warlords (CNN/USA Today).  Similarly, when Time/CNN asked, "Do you think the US troops in Somalia should be responsible for disarming the rival warlords there, or should the U.S. troops only be responsible for making sure that food is delivered to the areas affected by the famine?" only 22% said troops should be responsible for disarming rival warlords, while 69% said they should only protect famine relief.  Thus it is not clear that the fatalities had the effect of changing attitudes, though they may have consolidated them.

            Furthermore, there is evidence that after the fatalities occurred, the reason for wanting to withdraw eventually was not necessarily prompted by an unwillingness to suffer casualties, but rather by a belief that most Somalis wanted the UN and the United States to leave—not a surprising assumption after seeing the television images described above.  In PIPA’s October 1993 poll, 58% believed that most Somalis wanted the UN and the United States to leave.  When asked whether the US should leave if a substantial majority of Somalis wanted this, a resounding 88% of respondents said yes (only 8% said no). When asked how they would feel if most Somalis wanted the US to stay, 40% still wanted to leave, but the majority, 54%, said the US should remain.

            Despite many Americans’ doubts about the wisdom of getting involved in the civil war, and despite the fatalities, fairly strong majorities continued to support the mission. In two polls taken by CBS in October 1993 and one the following December, 64%, 67% and 62% respectively said the US “did the right thing” by going into Somalia.  In PIPA’s April 1995 poll, only 43% retrospectively supported the effort to resolve the civil war, but 82% affirmed the humanitarian operation. 

Perceived Fatalities in Bosnia

Although no American troops have been killed due to hostile fire in Bosnia, it appears that the majority of the American public believes they have.  This offers a unique opportunity to see the effect of these perceptions on attitudes about the operation. 

In February-March 1998 PIPA asked respondents “Is it your impression that American soldiers have or have not been killed by hostile fire in Bosnia over the last year?” almost a two-thirds majority—63%—said that Americans have been killed.  In reality, as of this writing, no Americans have been killed by hostile fire.  Only 22% of respondents knew this, while 15% could not answer the question.  Among those who said American soldiers have been killed, when asked to estimate how many have been killed over the last year, the median estimate was 25 deaths.  A substantial number of respondents also gave estimates in the hundreds, so that the average estimate of US fatalities was 172.

Nonetheless, these perceptions did not lead to a desire to withdraw US troops.  In, the same poll 65% said that they supported US participation in the operation.  Fifty-seven percent supported extending the missions beyond the original deadline for withdrawal.

In May 1999 PIPA reasked the question about fatalities in the previous year.  Once again a majority, in this case 56%, believed that US troops had been killed.   The median estimate for the number of dead was 20. Nonetheless, though most Americans perceived that the US had suffered yet another year of substantial troop fatalities support for the operation was statistically unchanged from the previous year—63%.

Presumably, if Americans are highly sensitive to casualties, the misperception that fatalities have taken place would be an important factor in shaping their other attitudes about the Bosnia mission, and their attitudes would differ from those who know that there have been no fatalities.  This follows logically from the original premise that the American public cannot withstand fatalities in an ongoing military operation.  For this reason, PIPA did analyses to see if differences existed between the two groups.

What is perhaps most striking is that, in both years, there was no relationship between the perception of US fatalities and support for US participation in the Bosnia operation.  Among those who believed fatalities had occurred support was just as high as for those who believed there had not been.  Also there was no relation between the number of perceived fatalities and support.  Even among those who believed that there had been more than 1,000 US fatalities support was no lower than for the general sample.

                  The 1998 study included a question which posed a battle scenario involving US fatalities and asked respondents what course of action they would support among four offered.  According to the conventional wisdom, one might assume that those who think combat deaths have already occurred in the operation might be more inclined to cut further losses than would those who know no combat deaths have occurred.  Virtually no reliable differences existed between the preferences of those who thought fatalities have really occurred, and those who knew they had not.

                  It could be assumed that those who believe that the Bosnia operation has already involved significant costs in American lives would be more wary of undertaking extra risks than those who know that no Americans have been killed.  Respondents were asked in 1998:

There is a controversy about whether the NATO force in Bosnia should seek out and arrest the two Bosnian Serb leaders who have been charged with war crimes and turn them over to the World Court.  Some say that the NATO force should arrest these leaders because they are responsible for the systematic killing of thousands of civilians.  Others say that such an effort might lead to armed conflict as in Somalia, and some American troops might be killed.  Do you favor or oppose having the NATO force carry out these arrests?

Again, no effect for perceived casualties was discernible.  Those who believed that American troops have died in Bosnia under hostile fire were just as likely to favor the operations described in the question as were those who knew that US troops have not been killed in Bosnia.

Hypothetical Scenarios

            In a series of poll questions presented by PIPA over the last six years respondents have been are asked to imagine how they would respond to a substantial number of fatalities in the course of a military operation.  In every case only a small number—no more than 25%--have said they would want the US to withdraw its troops. A majority has backed a vigorous response--either bringing in reinforcements or striking back at the attackers.

At the time of the Kosovo war in May 1999, when the option of invading with ground troops was widely discussed, PIPA asked respondents to imagine that “in the course of carrying out a ground war, 50 Americans were killed in a battle.” and to imagine that they had seen pictures of the soldiers’ dead bodies on television.  They were then asked whether they would then want to withdraw all American troops, strike back hard at the attackers, bring in reinforcements so that future attacks can be met with overwhelming force, or simply stay the course.  Only 20% said they would want to immediately withdraw US troops.  The majority backed one of the two assertive responses. Bring[ing] in reinforcements so that future attacks could be met with overwhelming force” was selected by 33-35%, and “strike back hard at the attackers” was chosen by 19%, while 19-21% elected to “not react in any of these ways but to simply stay the course.” 

            The situation in Bosnia has elicited similar responses.  Using the same type of question an April 1995 PIPA poll presented respondents a scenario for a UN operation there in which 200 UN troops were killed, 100 of them American.  Just 24% said they would want to withdraw.  In late winter 1998, when US troops were in Bosnia as part of the NATO operation, respondents were asked to consider a scenario with 20 US fatalities and just 15% favored withdrawal.

In a PIPA poll conducted in July 1994, respondents were asked to imagine that 25 to 100 American troops were killed in a UN peacekeeping operation in Haiti and to imagine that they had seen pictures of the soldiers’ dead bodies on television.   Only 21% opted for withdrawal.  The majority favored more assertive responses; bringing in reinforcements was chosen by 34%, whereas 24%) said they would strike back hard.

            Also in July 1994, PIPA posed comparable scenarios for hypothetical UN operations in the civil war then in progress in Rwanda.  The responses were almost exactly the same as for Haiti.  No more than 21% wanted to withdraw, and a fairly strong majority favored active responses.

Even if most Americans would not favor withdrawing troops in the event of fatalities it is highly significant how they would feel about the original choice to undertake the operation if fatalities did ultimately occur.  A number of poll questions have tried to get at this somewhat complex issue.  Sometimes by asking respondents whether they would be willing to sacrifice soldiers to achieve an end.  Basically it seems that most Americans are uncomfortable affirmatively answering a poll question where they are effectively being asked to make the choice for troops to die.  Also if Americans are presented a scenario in which it is spelled out that troops will die but no other outcome is defined this is not an attractive proposition.  However if Americans are asked to assume that the operation succeeds, then Americans will endorse the original decisions even with a surprisingly high number of fatalities.

As a general rule, many Americans resist making the explicit choice to sacrifice the lives of soldiers, even when doing so is implicit in other positions they may take. On three occasions a modest majority (averaging 53%) said it would not “be worth the loss of some American soldiers’ lives to help bring peace to Kosovo” (ABC). When Time/CNN on March 25 asked, “How many American lives would you be willing to sacrifice to achieve US goals in Kosovo?”  74% said none.

These responses should not however be read to mean that most Americans would only support using ground troops in Kosovo if they were sure that no troops would be killed.  In five polls conducted by four polling organizations over April-May (Pew, NBC/Wall Street Journal, ABC/Washington Post, and Newsweek), majorities ranging from 65% to 71% said they believed air strikes would not be enough to achieve NATO objectives and that it would be necessary to intervene with ground troops.  A CNN/USA Today poll also found an overwhelming 78% predicting that US ground troops would ultimately be used.  Asked how many soldiers would likely be killed the median estimate was 15-24.  Nonetheless, a substantial majority—73% in a March MSNBC poll—said they would support going into Kosovo with ground troops if it was the only way to stop the Serbs.  Thus it appears that the answer to questions about sacrificing lives is largely an artifact of the form of the question. 

If the question limits the possible range of fatalities, or speaks of risking rather than sacrificing lives, this can shift the balance toward a plurality or a slight majority in favor of acceptance.  Asked by Gallup on April 6 whether achieving NATO’s goals in Kosovo “is worth having a few American casualties in a limited military action,” 50% said that it was, while 42% said that it was not.  In an April 8, 1999 Louis Harris poll, 53% disagreed with the statement, “It’s not worth risking American lives to bring peace in Kosovo” (41% agreed).  In late March, 54% said it was “worth risking the lives of American soldiers in order to demonstrate that Serbia should not get away with killing and forcing people from their homes” (Los Angeles Times).

When respondents are asked to evaluate a scenario in which it is spelled out that American troops would die in an operation but no other information is given about the outcome majorities will tend to disapprove of it.  In April 1999 NPR/Kaiser/Harvard asked those who favored intervening with ground troops in Kosovo, “Would you still favor sending ground troops if 100/500/1,000 American soldiers were killed?”  Only a small minority in all cases said they would favor doing so. When no other information is given the value of the lives of American troops is given precedence.  Also providing information only about fatalities and not about the outcome implies that the operation is not going well.

However, if the operation is explicitly portrayed as succeeding, then the respondent is being asked to weigh the value of American lives and the value inherent in the goals of the mission.  In this case majorities tend to opt in favor of the value inherent in the success of the operation.  In May 1999 PIPA asked respondents to “Imagine that over the course of the ground war, 250 Americans were killed but the operation succeeded in driving Serb forces out of Kosovo so that ethnic Albanian refugees could return.  Do you imagine that you would think that NATO did or did not do the right thing by going into Kosovo with ground troops?”  In this context, a substantial majority (60%) endorsed the choice to pursue the mission, despite the 250 fatalities posed by the question (not right thing: 33%).[1]