A multivariate analysis

In Tables 6.8-6.11 we look at the results of a regression analysis of the impact of the various situational variables under the condition of keeping the impact of other variables constant. The first analysis (Table 6.8.) compares the four selected cases, the second (Table 6.9) compares the scores for each of the three groups on all four cases taken together. Finally, in Tables 6.10. and 6.11 we look at the results for the four cases from the perspective of the US and the EU respectively.

To start with the impact of casualties, the available evidence is rather clear and confirms the conclusions drawn above. With a few exceptions the expectation that the fear or actual occurrence of casualties depresses support considerably (keeping other factors constant) is confirmed in general. The scores are all in the expected direction and most of them are highly significant in all cases and with respect to military as well as civilian casualties. Looking at the data for the US and the EU separately, one is struck by the differences between the two. In the US only in the case of Kosovo for civilian casualties, the expectation is not confirmed, although the relationships are not always strongly significant. In the EU, however, the evidence is mixed. There is a strong (and mostly very significant) negative impact of the incidence of civilian casualties in three of the four cases, but the impact of military casualties is much smaller and not significant across the board. In the case of Iraq it is not even negative.

The importance of the factor of legitimacy (positive, in the form of a UN mandate or action, a NATO decision and the support of allies; or negative, in the form of the absence of such support or mandate) is much less than we assumed. The evidence is contradictory. In the US, only the absence of a UN mandate in the case of Iraq has the expected

Table 6.8 The impact of situational variables on support for military force. Regressions by cases

Table 6.9 The impact of situational variables on support for military force. Regressions by countries

negative (and indeed significant) impact. In the case of Kosovo, the fact that the intervention was controversial in international legal terms but defensible as a humanitarian action, is reflected in the limited impact of considerations of legitimacy. The element of support of NATO even played a (limited but significant) role in reducing rather than increasing support. This may have been a reflection of the unilateralist sentiment in the US ('we shall call you when we need you'), which was so visible in the reactions to 9/11 (see the negative impact of international support in the cases of Terrorism and Afghanistan in the US case). In Europe, the

Table 6.10 The impact of situational variables on support for military force. Regression for the US by cases

Table 6.11 The impact of situational variables on support for military force. Regression for EU-15 by cases

picture is somewhat clearer. The absence of international support (negative legitimacy) has the expected (and highly significant) impact in the cases of Kosovo and Iraq and the same is true for these same cases when positive legitimacy is mentioned in the relevant poll questions. In the cases of Terrorism and Afghanistan the evidence is less easy to interpret and ambiguous.

One of the factors frequently mentioned in the literature and specified in the model developed in Chapter 2 is the role of the goals and purposes for which military action is taken or contemplated. Our data present considerable confirmation for the importance of this factor but also some puzzling outcomes. Let us first look at the role of direct national interests. Taking all cases together, the pursuit of national goals has a positive impact for all three groups of countries but it is not significant for the EU countries. In the US, national goals have a positive impact in the cases of Kosovo and Iraq, but not in the other two. One could argue that only a few people believed that a national purpose was being served by the military response to the threat of terrorism, but this is contradicted by the huge support for the military action to remove the Taliban regime in Afghanistan in 2001 (detailed in Chapter 5). In Europe, the appeal to national goals does not carry much weight. The case of Iraq is an exception here, where the impact is significant.

The other category of goals to which a number of questions refer can be labeled international goals. These include humanitarian actions, action to remove dictatorships or to keep the peace after a conflict. The results are remarkable. Case by case we see that the impact of this factor was positive and (strongly) significant in the case of Kosovo and negative (and significant) in the case of Iraq. The arguments that the intervention in Kosovo served a humanitarian purpose was clearly widely appreciated but equally widely rejected in the case of Iraq. The outcomes in the other two cases were inconclusive (negative but not significant). Interestingly, overall, internationalist goals do not seem to carry much weight in the US and the EU. In the group of other countries they form the strongest argument in support of military force.

Looking at the individual cases, we note that in the US internationalist goals played only a (strongly) significant, but negative role in the case of Iraq. Apparently, Americans did not accept the argument that the removal of the Iraqi regime was a humanitarian purpose worth going to war for. In the EU, the same pattern obtains with the exception that the internationalist argument did also carry some weight (but now positive) in the case of Kosovo.

Finally, some other considerations that tend to play a role in the discourse around the use of military force deserve our attention. One is that because war is seen by many people in principle as a measure of last resort the argument that the time for war has not yet come and that more time should be given to diplomacy can be expected to play a considerable role, especially in issues where there is no immediate threat to national security. Our data present some confirmation for the strength of this argument. In general, the argument has a negative impact for the cases on which data are available, but that impact is significant only for the cases of Terrorism and Iraq, in both the US and EU. One remembers that these were cases where this question played a major role in the debate over whether military action was indeed unavoidable now. In Europe, in the case of Kosovo, the argument also played a role but its impact was reversed. Apparently, Europeans felt that action to protect the Kosovars was overdue.

A second consideration concerns the risks involved in military action, the fear of the consequences of having 'troops on the ground' and of a war of long duration. These factors do have a considerable negative impact on support, which is particularly notable with respect to the troops-on-the-ground argument. This particular influence is strongly significant in three out of the four cases, and especially in the US and in Other countries. In the EU it only plays a significant role in the case of Kosovo.

The 'long duration' argument in general is significantly influential in the US (where it is strongly significant for the cases of Afghanistan as well as Iraq), as well as the EU. In the latter case this is an outcome which is hard to explain.