Legality, Deaths and International Responsibility in the Netherlands: A public opinion research on the Dutch attitude towards Responsibility to Protect missions.

(1)

Master thesis

Crisis and Security management

Legality, Deaths and International Responsibility in the Netherlands

A public opinion research on the Dutch attitude towards Responsibility to Protect missions.

Name:

Iris Kok

Student number:

s1685244

Supervisors:

Dr. E.T. Aloyo

Dr. H. Mazepus

(2)

2

Chapter 1: Introduction & Research Question

With the end of the Cold War a feeling of peace and safety started to develop in Europe. After 80 years of bloody wars, tensions, and division Europe had to rethink their position in the world. The same thing was happening for the military forces, if not to defend the home country, what were they going to be used for? From 1991 onward western states started to intervene in other countries on humanitarian grounds. This was one way of using the existing military forces and keeping them relevant. Some of these missions can be labelled successful but others had some known failures. The most well-known mission in the Netherlands is the protection of Muslims in Srebrenica. This mission ended cabinet-Kok in 2002 and raised the question on what was morally right and how much we, as a western society, are prepared to give up to achieve this ideal world?

Problem outline

This research is rooted into several problems and trends. The first trend is the change from military force used for the survival of the state to force used for other purposes, specifically force used for humanitarian intervention abroad.

Secondly, humanitarian military intervention comes with a confusing system of international law, especially for responsibility to protect missions (R2P). The concept of R2P is, according to international law, applicable in four defined situations: genocide, crimes against humanity, war crimes and ethnic cleansing (United Nations, 2005). Of course there are other forms in which civilians can be harmed that do not fall under any of these four categories, these cases are excluded from the R2P concept. The literature on this concept can be labelled as confusing because specific definitions on, for example, genocide or mass killings, do not necessarily correspond to the situation as experienced by civilians and military personnel on the ground. This gap between law and reality brings in the need for interpretation and re-evaluation of the principles this concept is based on.

The third phenomenon is a democratic deficiency on the topic of foreign policy and military missions in the Netherlands. Civilians have little to say about whether the Netherlands participates in an international mission or not. Even though citizen participation can lead to better decision-making according to democratic theory. Empirical democratic theory values citizen participation for several reasons: the multitude of opinions and creative thinking, the broad knowledge citizens can bring into decision-making and, maybe most importantly, the critical assessment of moral ideas that can be used as a justification for policy (Christiano, 2006). On a more practical level, public support is needed for a democratic government to stay in power. When there is a large dissatisfaction among citizens, representatives will not be re-elected during the next elections (Ringmose & Børgesen, 2011, p.508). When the democratic deficiency decreases the legitimacy of the government will increase.

(5)

5 For these three reasons it is interesting to research what factors influence public opinion on military missions, more specifically the humanitarian intervention missions. To get a better understanding of the legality of a missions versus the humanitarian need, I have decided to research what influence legality has on public opinion. Another factor, that was much discussed after the Srebrenica missions, is the effect of military casualties on public support. Is the safety of military personnel a reason not to start a humanitarian intervention? What is worth sacrificing for groups in need outside of the Netherlands?

Academic relevance

Academically this research can help to clarify what factors do and do not influence public support. The research on this topic is extensive though not conclusive in any way. The use of an experimental design to research this topic will give the chance to either reject or confirm a causal relationship. The expected deaths variable was chosen because of how it fits the post-heroic age. The legality variable was chosen because this has not been researched yet in the Dutch context and because of the interesting relation between R2P as a political concept and international law. The research is focused on possible missions in the near future instead of the public opinion during a missions like most research on military missions. In order to get a better understanding of the public opinion in the decision-making phase before it is influenced by factors occurring during the mission. This perspective has rarely been used. More information can be found in chapter 2, the literature review and chapter 3, the methods.

Societal relevance

According to some (Everts, 2000; Van der Meulen & Soeters, 2005ab; Burk, 1999) there is a trend of increasing risk aversion among civilians since the beginning of the Second World War in comparison to the period before the world wars. Are Dutch civilians prepared to make human sacrifices when no direct interests are involved? Answering this question will help decision makers to increase democratic legitimacy. It will bring a broader understanding about what is seen by civilians as a just war and what they are prepared to sacrifice for a humanitarian cause. The difference between perceived attitudes by politicians and actual attitudes of civilians can be a problem in democratic decision-making. Obviously decisions are made on perceived attitudes by politicians (Everts, 2000, p.102). When politicians perceive a high level of fear for military deaths amongst the public it is possible this can become a self-fulfilling prophecy. When politicians think there is a high level of fear amongst the public they will emphasise this in public debate. As a consequence more people come into touch with this idea and might adopt this fear as well. To counter this it is therefore desirable that there is no gap between perceived attitudes by politicians and actual attitudes.

(6)

6

Research question and hypothesis

The goal of the research is to see if there is a measurable effect of either the amount of expected deaths or the legality of the mission on public support for a military R2P mission. The research question is therefore as follows:

Does the international legality of armed humanitarian interveners or the number of expected Dutch military deaths influence public support for an armed responsibility to protect mission?

To be able to measure this the research question is divided into three hypotheses.

H1: As the number of expected Deaths increases, public support for military R2P missions will decrease. H2: Military R2P missions with UN Security Council authorisation will lead to higher public support compared to missions without Security Council support.

H3: A combination of UN Security Council authorisation and low expected deaths among Dutch military personnel will lead to high public support.

More information on these will be given in the methods section of the thesis.

Reading guide

This thesis is structured in the following manner. First the theoretic framework will be discussed in a literature review in chapter 2. Decision-making in foreign policy, prior research on risk aversion and humanitarian military intervention will be analysed in this chapter. After that, in the third chapter, the method of the research will be discussed as well as the limitations of the research. Chapter 4 will contain the results of the experiment. The thesis will be concluded with a discussion and suggestions for further research.

(7)

7

Chapter 2: literature review

To place this research in the existing body of knowledge there are a few subjects that need to be addressed. The first is the historic context of civilians in foreign policy decision making and some case relevant information about the decision-making process in the Netherlands. The second part discusses humanitarian intervention, more specifically the responsibility to protect concept and its usage and flaws. Thirdly some risk aversion theories will be discussed with an emphasis on casualty aversion. Following this some alternative factors for changing public support will be discussed. Finally, there will be some case specific information about military missions the Netherlands has participated in since the end of the Cold War. This chapter will identify some gaps in the existing literature where this thesis is based on.

Civilians in foreign policy

Before the Second World War there was an active theoretical debate between the liberal school, who was enthusiastic of including public opinion in foreign policy decision making, and the realist school that opposes that statement (Holsti, 1992, p.439-441). In the years after the Second World War, research on public opinion increased. The overall consensus became that public opinion was no stable foundation to base policy decisions on (Holsti, 1992). The main arguments for this were that public opinion is highly volatile, and neither coherent nor influential. This consensus has been called the Almond-Lippmann consensus after the two most well-known scholars who were working on this theory during this time (Lippmann, 1955; Almond, 1956, p.376-378). Realist scholars claim that the ambition of a country in its foreign policy is mainly influenced by their place in the international system and their relative material power capabilities (Rose, 1998). This leaves no room for the public opinion to make a difference. These are all still existing arguments against a more liberal approach. However, the Almond-Lippmann consensus was believed to be inaccurate after the Vietnam War (Ringmose & Børgesen, 2011, p.510).

During the period before the Vietnam War little research on foreign policy included public opinion. It was indicated that The State Departments in this period in the United States saw the public as a group to educate and lead instead of a group to listen to or follow (Holsti, 1992, p.444). The Vietnam War has also been called the first television war. After the Second World War the number of televisions and increased rapidly which caused the public to be more easily and better informed (Kratz, 2018). After the Vietnam War it became clear that not considering public opinion was doing harm to politicians personal careers and political trust in general. After this discovery the influence of public attitudes on foreign policy on politics and decision-making became acknowledged (Ringmose & Børgesen, 2011, p.510). Page and Shapiro (1988) established that changes in public opinion were not random but reasonable and event driven (Holsti, 1992) This had become the new consensus after the Vietnam War. An example of research in this new tradition is Mueller’s (1971) research on the effect of casualties on public opinion.

(8)

8 In this same tradition, there is an increased interest in the influence politicians have on public opinion in wartime. This branch states that the relationship between politicians and civilians is a two way street. The public opinion influences the foreign policy through politicians but the politicians can also have an influence on public opinion (Ringmose & Børgesen, 2011; Van der Meulen & Soeters, 2005b, p.537-539).

In the last decades scholars arrived on a new consensus, It is now believed that that multiple factors can influence public opinion in general and public opinion on foreign policy more specifically. It is also believed, in western democracies, that public opinion on foreign policy matters from a democratic perspective. To provide more information on which factors influence public opinion on foreign policy this thesis will research the influence of two possible factors, legality and expected casualties, in a specific context: military missions in the Netherlands.

Decision-making in the Netherlands

In the Netherlands decisions about participating in peace missions are made by the government (Article 97.2, Constitution). What this mission will look like in terms of practical decisions and resource management is decided by the ministry of defence. When there is a proposal for participating in or prolonging a military mission, the government is obligated to inform the Parliament about their plans by sending an article 100 letter (Article 100, Constitution). This structure divides the tasks of the state, citizens and the military. This division is in line with Huntington’s idea of objective citizen control, where citizens define the important security policies and the military is free in how they will achieve these goals (Burk, 2002, p.9-10; Feaver, 2009 p.215). However the actual influence the public has on specific policies is limited in this case. The article 100 letter is the only connection between a policy decision and the public representatives in Parliament. The function of the letter is to inform the Parliament and it gives them no executive power to stop, alter or approve a future mission. Their only power is to speak their arguments and hope they will be heard. This is different than for a declaration of war (article 96, constitution).

In the beginning of the Cold War the Dutch Parliament was barely involved in decision-making on military missions. However in the 1970s this changed and the informal influence of Parliament increased. In the first half of the 1990s the process of formalisation was started (Van der Ven, 2011, p.15). The article 100 letter is a part of the Dutch constitution since 2000. The groundwork for this was laid in 1995 by Ministers Van Mierlo and Voorhoeve (Constitution, article 100). The addition of the article in the constitution was part of a trend, but gained priority as a direct consequence of the mission in Srebrenica.

An important part of the article 100 letter is the term acceptable risk, which is used since the early 1990s in military-civil discourse. What exactly is and isn’t an acceptable risk is based on a cost-benefit analysis of a specific mission by the government (Van der Meulen & Soeters, 2005b). It is therefore a political

(9)

9 concept rather than an objective calculation. The number of deaths is never specified in these letters but casualty avoidance is a prominent factor when considering deployment in all western countries today (Van der Meulen & Soeters, 2005ab). The Dutch historically look at themselves as amilitary, unheroic

and nonmartial (Van der Meulen & Soeters, 2005b, p.538). In a survey done in 2008 on the mission in

Uruzgan only 63% of the Dutch civilians responded to be aware of the mission and 35% responded to actually be interested in the subject (Ministry of Defence, 2008). These are not very high numbers that could confirm the self-image. However, it must be noted that historians today have a more nuanced view According to them, Dutch civilians have often been in favour of the use of force at crucial moments, an example is the violent break-up of Yugoslavia (Van der Meulen & Soeters, 2005b, p.543). This conflict between the Dutch identity and the historic reality is part of what is researched in this thesis. A better understanding of what constitutes a “crucial moment” in the eyes of the population and what does not, can be useful in decision-making. This research will contribute to a more clear understanding on how Dutch citizens evaluate future military missions and what motivates these decisions.

Responsibility to protect

This research is focused on armed humanitarian interventions, more specifically military Responsibility to Protect (R2P) missions. The goal of this research is to measure the public support for military missions so the gap between the ideas of decision makers and the public can be narrowed. Wars are fought for many different reasons but historically most have either economic or strategic motives. This touches the state as a whole but does rarely touches the citizens as urgent or necessary for moral reasons. A humanitarian mission has as a goal to do something good for people in the international community that are less fortunate. “It touches on the idea that you would also like to be helped if it was you in this less

fortunate situation.” As a respondent pointed out in the comments of the survey. Situations described

as reasons for a R2P mission claim to be universally seen as bad. Therefore it is more closely related to what citizens experience as doing the wrong or right thing. Reality is more nuanced than a dichotomy between state interests and personal morality. However, this difference makes public opinion on R2P mission an interesting subject to study.

Just war and humanitarian military intervention

The just war theory (Jus Bellum Justum) is the leading is the most used method of evaluating military interventions. The theory consists of three parts: right to go to war (Jus ad bellum), law in war (Jus in

Bello) and, the latest addition, law after war (Jus post bellum). For this thesis the first, jus ad bellum, is

important. This part consists of six criteria which are contrary to what one might think, not equal in importance. Each criteria comes with its own ideas, advantages and disadvantages. This thesis will focus on one of the six criteria specifically, the right authority principle, but all principles will be important to understand how decision-making on humanitarian military intervention works (Seybolt, 2007, p.25-26). An important fact to always keep in mind when taking about the decision to go start a humanitarian

(10)

10 military intervention is the fact that it is always a political choice (Seybolt, 2007 p.25-27). This can be stated about every war that is not about survival of the own state.

The six criteria for starting a (humanitarian) military intervention according to the just war theory are as follows: a just cause, the right authority, the right intention, military action should be the last resort, it must be proportionate and there should be a reasonable prospect of success (Seybolt, 2007,p.25-26; Chinkin & kaldor, 2017). The criteria of just war are used by politicians and decision makers but can also be used by the public to check the decisions of the government. This is possible when the instrument used to measure whether a war is either just or unjust does measure what it is supposed to measure. The principle of just war claims to be correct on what is in general thought of as right and wrong. If the instrument is not a reflection on public opinion this can be a problem for democracies who make decisions accordingly.

The Responsibility to Protect principal

The process of globalisation brings many changes to world order. The increased interconnectedness in different fields and a rising awareness of a single human community opens the discussion for change in a still mainly state centric legal framework for the use of force. Chinkin and Kaldor (2017) discuss five different conceptions of world order and the corresponding ideas on sovereignty. The most important to our research being the Responsibility to Protect (R2P) model. R2P is based on three pillars: Within this model sovereignty is conditional, the state has a responsibility to take care of their citizens and protect them from extreme harm. When a government is not able or willing to protect its citizens the international community should offer help, by assisting with resources or diplomatic help. When this second pillar fails the last resort is that the international community is able to militarily step in. However, the norm of non-intervention can only be overridden under certain predefined circumstances: genocide, ethnic cleansing, crimes against humanity and war crimes, and with permission of the United Nations (UN) Security Council (Chinkin & Kaldor, 2017, p.78-86; p.82-83 p.195). With accepting the R2P concept the UN moves from being state centred towards a stronger focus on human rights (Seybolt, 2007, p.190). With this focus on the security of citizens instead of the rights of states, the concept uses well institutionalised measures in a new way.

(11)

11 The concept responsibility to protect started out as an exclusively humanitarian aid provision, however, it is in practice not possible to make a clear divide between political, humanitarian and military action within a conflict zone (Chinkin & Kaldor, 2017, p.179-180). Seybolt (2007) has created a typology to help categorize humanitarian military interventions into 4 categories. The first variable is whether the approach is focused on the victims or on the perpetrators. The second variable is whether the problem is privation or violence. Privation can be defined as a shortage or the absence of basic necessities like food, water or healthcare. Violence is the absence of security. When these two are crossed different objectives for missions appear as can be seen in the table 1 (Seybolt, 2007, p.38-40). The category useful for this thesis on military interventions in R2P missions is where the focus is on the victims and they are saved from violence. What type of

mission and strategy are chosen is decided by politicians. For a successful mission, in the case of humanitarian intervention that is measured in saved lives (for more information Seybolt, 2007), the conflict, the type of mission and the strategy should fit together. The last factor needed for success is more practical: a sufficient amount of resources (Seybolt, 2007). The R2P concept fits into a broader idea of humanitarian

intervention and consists of military as well as non-military intervention and of combating privation as well as violence (Chinkin & Kaldor, 2017). However, for this thesis the focus will be on military missions to combat violence against civilians.

Problems with R2P

There are several problems that occur alongside the R2P principle. The first one being a question of definition. How to establish a definition of a certain situation on which action can be based? Definitions are important to assess whether a mission is legal or illegal. This research can be done by an independent fact-finding body but this is costly and causes delay of action. During this time the situation can possibly escalate while still no action is being taken. It also gives a false sense that something is being done while on the ground there is no sign of help (Chinkin & Kaldor, 2017, p.212). There is strong prove that the success of a humanitarian mission depends on a timely military intervention (Seybolt, 2007, p.177). There is a conflict between timely military intervention and achieving legal ground for a mission. Both are important and this is not a dilemma solved easily.

As Seybolt (2007) recognizes is the decision to start a humanitarian intervention is highly politicised. The second critique touches on this. Politicians are sometimes accused by other politicians or researchers for selectively applying the R2P principle (Seybolt, 2007, p.220-221). The high level of complication

Table 1: typology for military humanitarian intervention (Seybolt, 2007, p.40)

(12)

12 in some cases undercuts the transparency of the decision. There have been several atrocities that did not get any international help but would technically have been able to get it on the basis of earlier R2P cases. The ICISS supported a proposal to let decision making in the Security Council on humanitarian need become less obstructed. But no such approach has been materialized yet (Chinkin & Kaldor, 2017, p.215-216). Transparency and consistency on this subject is extremely valuable for the democratic legitimacy of an international organisation like the UN.

A problem that occurs after authorisation is given is more practical. It is often unclear which measures will or will not be effective and efficient in a humanitarian crisis. Therefore when the mandate is given it is usually not determined what should be done on the ground (Chinkin & Kaldor, 2017, p.225). This can possibly create confusion and make the mission ineffective. A second problem that occurs on the ground is more of a moral dilemma. Described well in a quote from a UN medical doctor in Srebrenica:

‘But for me there’s only one lesson ... If blue-helmeted UN peacekeepers show up in your town or village and offer to protect you, run. Or else get weapons. Your lives are worth so much less than theirs.’

(quoted in Cain, Postlewait & Thomson, 2006, p.253)

There are two ways to protect civilians: point protection and area protection. Both of these methods are used to keep civilians save and both have their own advantages and disadvantages. Two factors are important besides conflict specific factors: the environment and the willingness of soldiers to fight and even maybe accept deaths to protect these points or areas. “The willingness” sounds like it is a matter of personal willingness but this willingness is a political choice (Seybolt, 2007, p.184-189). As the quote suggests the intervening countries that are there to provide help may not be willing to make these necessary sacrifices for a successful mission. There has been proof that during the humanitarian mission in Srebrenica the avoidance of casualties among Dutch military personnel was top priority (Van der Meulen & Soeters, 2005b, p.545). Also in other humanitarian missions with other western countries this played an important part in decision-making. An example of such a choice is air pilots who were advised to fly higher so they were less likely to be shot. Even though this meant that they could shoot less accurate and would hit more civilians on the ground (Seybolt, 2007, p.86). This willingness to sacrifice possibly the lives of military personnel to protect people in conflict areas, especially in the Dutch case, is what is being researched in this thesis.

Legal versus illegal

R2P is a political concept more than a legal concept. The concept is largely shaped by the times it is or is not applied. An interesting part of this process is the authorisation by the UN Security Council. In the original theory as thought of in the 1990s this was a clear condition for the use of force. However, this has changed with recent missions. Western states do generally wait for authorisation by the UN Security Council before the use of armed forces. However in the case this isn’t possible states seek support of other states and civil society to justify their actions against atrocities (Chinkin & Kaldor, 2017,

(13)

p.175-13 176). A good example of a mission without authorisation is the mission in Kosovo in 1999, which intensified the debate with the UN (Seybolt, 2007, p.14-15). When intervention takes place without Security Council authorisation the international court of justice (ICJ) has to decide whether the mission was legal or at least legitimate. This has to do with terms as proportionality and whether the atrocity can be defined as one of the four situations in which R2P is appropriate. It must be said that all R2P cases have some kind of legitimacy basis. Different actors will assess legitimacy differently and therefore also whether this has been sufficient and or proportionate. R2P intervention without UN Security Council authorisation is often happening in a grey area between legal and illegal. Whether the mission will be seen as legitimate depends on a committee who will take a look at the situation ad hoc (Chinkin & Kaldor, 2017, p.222-224). This process is complicated and can take a long time. To summarise, a mission is legal when the Security Council has found the situation in need for an intervention and has given the states permission to intervene. A mission is illegal when there is no authorisation from the Security Council but the non-intervention principle is being violated regardless. These missions will be reviewed by the ICJ and can be labelled ex post facto as legitimate.

Risk aversion and deaths

There are multiple societal and historical factors that affect the relationship between citizens and military missions. At the core of this debate stands the fact that most western democracies do not fight for the survival of their own state anymore. As mentioned before, historically this has been the main argument to use military force. This shift from self-defence to peace or humanitarian missions is one of the main arguments for higher casualty aversion (Smith, 2005, p.492-494). Another argument is that war is now associated with a zero-defect mentality as a part of the post-heroic age (Van der Meulen & Soeters, 2005b, p.537 & p.552; Smith, 2005, p.505-507). In the post-heroic age, developed after the Second World War, military deaths are not perceived as heroic but rather as failures of the system (Smith, 2005, p.505-507). Reasons for this attitude are among other things a great trust in modern technology and strategic warfare (Smith, 2005). This post heroic age goes hand in hand with a growth in risk aversion (Everts, 2000, p.93). Besides trends like the shift to peace and humanitarian mission and the rise of the post-heroic age there are several other phenomenon that contribute to the growth in risk aversion. The first phenomenon is that citizens accept different costs depending on what the character or goal of the use of force is. For example, people are more likely to accept high risks when the goal of the mission is to protect the homeland than they are for mission with low national interest (Smith, 2005; Mueller, 1971). This phenomenon is closely related to how a particular conflict is perceived by citizens. The media are important actors in the civil-military relationship (Everts, 2000, p.94). The past few decades the most well-read papers in the Netherlands have not been supportive on participating in peacekeeping missions (Van der Meulen, 2014, p.5). This could possibly have a negative effect on the level of risk Dutch citizens are willing to take.

(14)

14 Another phenomenon that can possibly explain rise in risk aversion is based on historical research. The aversion to casualties and death in general is by some seen as a characteristic of modern times. Death is now less a part of everyday life than it was in any other point in history, especially compared to, for example, the medieval period (Smith, 2005, p.502, Foucault, 2010). As a consequence people can be negatively surprised by death when confronted by it. This argument may explain the crisis in Rwanda (1994) where 10 Belgian soldiers were killed during a humanitarian mission and the government almost immediately decided to stop participating, against UN advice (Seybolt, 2007, p.182). The mission in Somalia (1993) and Bosnia-Herzegovina (1994) had also disappointing results for respectively the Canada and the Netherlands. After these mission western countries get more sensitive about deaths and more sceptical about saving oppressed civilians with military force (Seybolt, 2007, p.86; Klep, 2008). This argument is, however, debatable, a relevant counter argument would be that media now brings death and war close to home in the form of news, games and movies. Still it is an argument worth taking into account. Mainly because this issue relates to how civilians value deaths and lives.

Body bag syndrome & casualties hypothesis

There are two theories that specifically focus on the relation between public opinion and military casualties instead of risk in general. The first one is the casualty hypothesis by Mueller. This theory states that the initial support will decline rapidly when military personnel starts to die (Mueller, 1971; Everts, 2000). This is a comparison between the initial support for a missions and the support after the first deaths. This hypothesis was supported in 1990 in the Netherlands using hypothetical situations (Everts, 2000, 95-96). It can be argued that this theory emphasises an emotional response to military deaths from civilians. The label emotional or rational is hard to put on this decision to not support the mission any further. Later research suggests the opposite, namely, that the response is not emotional but rather rational, including a cost-benefit analysis by citizens (Ringmose & Børgesen, 2011, p.511). In recent years the distinction between emotions and the rational has become smaller. It is now believed by many that emotions can have a rational value (Scarantino & Sousa, 2018).

The second theory is the body bag syndrome. It is essentially built on the same principle as the casualty hypothesis. The difference is that instead of comparing the support during the mission, this theory seeks to explain the effect of expected deaths on future missions (Smith, 2001; Raes, Du bois & Buts, 2019). The main hypothesis being: citizens are supportive of future military missions as long as there is a low chance for casualties (Everts, 2000, p.93). This definition comes with a word of warning, some authors use the body bag syndrome and the casualty hypothesis as interchangeable. A study in the case of Bosnia showed that a significant part of Dutch civilians (about 2/3) were not willing to support a mission with more than 10 perceived Dutch military casualties (Everts, 2000, p.96). While almost half of the respondents were in favour of the mission before this question was asked (Everts, 2000, p.96). This shows some support for the body bag syndrome in the Netherlands. What should be noted is that it is not clear in this case whether likelihood of success or perceived deaths was the variable to cause the

(15)

15 effect (Everts, 2000, p.97-98). Because of this internal validity issue, no hard conclusions can come from this research. The hope is that this connection will be clearer in a controlled experimental setting as is the idea for this thesis. Another example of the body bag syndrome in action is Kull. Kull (1995) found some evidence for the body bag theory as well as the casualty hypothesis when polling Americans about UN peacekeeping operations. In a US survey the support for a peacekeeping mission dropped 15 percentage point, when the question made respondents aware of the chance of casualties (Kull, 1995, p.57). However for the specific case of the Gulf War no such evidence was found. Kull (1995) concludes that public support does decrease when casualties are brought up but that it cannot be confirmed that it is decisive for an overall change in public opinion.

Ethics of agent-neutrality

How to value life and especially different lives has been a part of ethics since at least the 18th_century.

According to utilitarianism, a form of consequentialism, the action that will produce the largest amount of good is the morally right choice. The theory is known for its impartiality among some philosophers. To them everyone’s happiness has the same value (Driver, 2014). In this case that would mean that sacrificing several military lives for a large amount of possible victims is the morally right choice. Of course this is not the only philosophical theory that is applicable to this dilemma, as utilitarianism is also highly criticised. However it is interesting to look at this dilemma with more philosophical lens, in further research.

Philosophical and psychological research has found that personal relationships can have an influence on someone’s personal decision-making. It is assumed that humans are essentially social beings that see the relationship with others as a part of their identity. According to Scheffler (1999) that would indicate that these personal relationship bring responsibilities to the group members and can serve as a basis of preferential treatment (Nortvedt & Nordhaug, 2008). In this case the personal relationship and shared identity is the nationality shared with the military personnel, and the lack of a relationship to the victims. This would increase the possible effect of the manipulation.

Other factors

It must be clear that the amount of expected casualties/deaths is not the only factor that influences public opinion in military missions. Many more factors are considered in research. The most heard are closely related to each other: the likelihood of success and the duration of the mission (Gelpi, Feaver & Reifler, 2009, p.53-54). This again points out the lack of internal validity in previous research. Others, like Burk (1999) claims that whether the mission is perceived as meaningful and effective is decisive in calculating support. Everts (2000) as well as Burk (1999) and Ringsmose & Børgesen (2011) emphasise the role of the media and politicians in creating a certain public sentiment within a democracy. However, Kull (1995) concludes that whenever congress and the president (dis)agree on the mission public support is all but certain, only 13 to 36 percent of the respondents is willing to change their opinion to match the

(16)

16 decision of congress (Kull, 1995). These arguments are in agreement with the current consensus that politicians influence public opinion as well as the other way around.

Dutch military missions

This thesis is focused on the Netherlands. This section will provide background information about military missions in the Netherlands has participated in since the end of the Cold War. Dutch military missions in this period have not been very lethal. 15 missions have resulted in one or more deaths among military personnel since the end of the Cold War. The average amount of deaths in these missions is 3.9. However this number is highly influenced by the 25 deaths during the International Security Assistance Force (ISAF) mission in Afghanistan (2002-2014). Eight missions only reported one military death (Staten-Generaal, 2019). Not all military missions since the end of the Cold War have known casualties. There have been 88 international missions that were still operational or started after the end of the Cold War that had Dutch military personnel involved. Most of these missions were not or barely discussed in the media. Some of these missions were strictly humanitarian, like help after a flood, and some only involved a few Dutch participants in supporting functions, like medical personnel (Ministry of Defence, 2017). The diversity of the missions makes them hard to categorise and compare.

The mission that had the most impact on the Netherlands in the 1990s would be the fall of Srebrenica. In 1993 the UN Security Council declared Srebrenica a UN safe area. The Dutch UN mission on the ground had the task of protecting somewhere around 40.000 Muslim citizens in the area. The Dutch soldiers were not able to prevent the fall of the enclave and around 8.000 Muslim men were killed. In 2001 the International Criminal Tribunal for former Yugoslavia (ICTY) declared this mass killing a genocide (Klep, 2008, p.79). After this event many things have changed in how Dutch politicians and citizens view armed humanitarian interventions. During this mission there were seven Dutch military deaths. This raised the question in the years after: Could/should we have done more?

In February 2008 a public opinion poll was done in the Netherlands concerning the ISAF mission. At that point 13 Dutch had died in this mission. The survey started with knowledge of the mission, 63% was aware of the mission in Uruzgan. More interesting numbers are: 29% of the respondents thought of the risks for military personnel as acceptable and 36% responded that military deaths would be a reason to immediately stop the mission. The results show a low interest (35% said to be interested) in the mission and little trust in the success of the mission (Ministry of Defence, 2008). The results show a low national interest, low chance of success and a high perceived risk for military personnel. In this case the Dutch participant responded in line with the literature in risk aversion.

(17)

17

Chapter 3: Methods and Limitations

In this chapter the methods used for the data collection and the limitations of the research will be discussed.

Hypotheses

H1: As the number of expected Deaths increases, public support for military R2P missions will

decrease.

The first hypothesis is based on the common believe that if something involves a greater cost

the willingness to do so is less than at an average or low cost. This research does however not

measure the actual costs but the expected costs and therefore the risk. The choice is between

participating in a mission and having possible costs and benefits or not participating and having

neither costs nor benefits. The benefit in this scenario is helping a population in need and

participation in the process of international peace.

H2: Armed humanitarian interventions with UN Security Council authorisation will lead to

higher public support for military R2P missions compared with armed humanitarian

interventions without Security Council support.

The second hypothesis is based on the right authority principle that can be found in the just war

theory. The UN Security Council can overrule the principle of non-intervention in certain cases

as discussed at the 2005 UN world summit (United Nations, 2005). This hypothesis is about

trust in the international system and law but also on at what costs do you want to help others.

The costs in this hypothesis are disobedience towards international law and possible punishment

in court.

H3: A combination of Security Council authorisation and a low chance of Dutch casualties will

lead to high public support.

The third hypothesis is a combination of the previous two and is made possible because of the

2x3 design this experiment will have. It measures whether a specific combination of the two

dependent variables has a significant effect on the public support.

(18)

18

Dutch case

This research is only done with Dutch citizens and the results are therefore specific to the Netherlands. Research suggests that sensitivity of civilians on casualties can differ between cultures and countries (Everts, 2000, p.101-102). Human right abuses are differently assessed in different situations and in different countries. Lupu and Wallace (2019) researched the effect of international law on public opinion as well as whether the government was violent towards an opposition. They found that the effect of breaking with international law was country specific and depended on multiple factors like history and political culture (p.416). Because of these findings this research will focus on one country and is not easily applicable to other countries.

The choice for the Netherlands is partly a convenience choice, for I am Dutch and have therefore a network and a good understanding of the language. Dutch civilians shall be defined as everybody with a Dutch nationality above the age of 18. The age limit is 18 because that is the legal age to vote, the most common way to participate in the political process. I am aware that the majority of research on R2P is done from a western lens (Chinkin & Kaldor, 2017). This perspective accentuates the fact that western state intervene and that this responsibility is on them while it situates non-western states as the place where intervention happens. The choice to intervene is in the hands of the western states (Chinkin & Kaldor, 2017, p.218). Those reasons combined are the reason the Netherlands has been chose in this research.

Experimental design

To answer this research question I have chosen to use an

experimental design in which participants will read a

vignette and respond to a survey. There will be 6 different

treatments in a 2x3 design. Two possible options for the

legality variable and three different options for the

expected deaths variable, see table 1. In every vignette the

same atrocity is described, the varying variables will be

UN support and the mention of expected military deaths.

After the vignette all respondents will react to the same

questions measuring public support for military R2P

missions.

Table 2: The six vignettes

Legality variable Expected deaths variable N in vignette Legal mission Below average 27 Average 24 Above average 30 Illegal mission Below average 24 Average 19 Above average 24

(19)

19

Explanation of the vignettes

The survey used in this experiment consists of several blocks: the introduction, the vignette

with the manipulations, the questions of the dependent variable, control questions,

demographics and a possibility to leave comments. The randomizer in the survey program

Qualtrics was used to randomly assign respondents to a certain manipulation. The two options

for the legality variable were Legal (54.7%, N=81) and Illegal (45.3%, N=67). The three options

for the expected deaths variable were below average (34.8%, N=51), average (29%, N=43) and

above average (36.5%, N=54). Considering this data it can be concluded that the randomiser

sufficiently placed the respondents into the different treatments.

The introduction

The introduction includes some information on the research in two ways: first, some

information on the topic of mass atrocities and military missions and second, on the survey: the

expected duration and what respondents will be reading. After this there is an option to continue

to the research or to stop. Nobody has chosen the stop option by clicking the stop-button. There

are however some people who stopped the survey here by closing the window on their computer

screen, their responses are deleted in Qualtrics and do therefore not appear in the dataset.

Background information on the conflict

The usage of words in a vignette can have an impact on the respondent. There is an

acknowledged gap between what words mean to an academic or court and what it means to the

general public. This power of certain words and phrases is calls framing or issue framing in

political sciences. Framing is a way of organizing information connecting facts to pre-existing

knowledge the reader assumingly has. what frame is chosen is to affect the attitude of

respondent (Valentino & Weinberg, 2017). In this vignette words are used as they most often

appear in the media. All words, with exception of the manipulation of course, are exactly the

same. Therefore, all the vignettes use the same frame.

The survey has 6 manipulations in a 3x2 design. The manipulations are randomly assigned

using the randomiser in Qualtrics. All respondents read the same background information on

the conflict in the made up country “Oshor”. A name made up by an AI country name generator.

The background information includes the following elements: the government as a perpetrator,

a specific group that is being discriminated, examples of crimes against humanity and literally

(20)

20

the term crimes against humanity, 1000 people who are killed, the prediction that these crimes

will not stop on their own or soon, and last but not least the fact that the Netherlands has no

strong economic or cultural ties with this country.

The government has been chosen as a perpetrator because it is historically the most likely

institution to kill civilians in large groups (Rummel, 1994). The state has power over its citizens

and a monopoly of violence, this brings them into an asymmetrical power relation. There are

examples where weak states lose their monopoly on violence without a coup, but those are rare.

There is no reason, even with the trend towards more extra-state entities in conflict to deviate

from taking the state as a perpetrator. The victim group is specified as “a community in the east

of the country”. Specifying the group further would make the research detailed on a point that

was not a manipulation/part of the research question. For length reasons only the neutral factor

of where they lived was presented.

Crimes against humanity is in and of itself a very complicated legal concept to understand. It is

barely used in journalism or everyday speech. The Dutch translation is complicated because

crimes can be translated in a variety of ways, as well as the word humanity. The Dutch and

Belgian authorities have for legal reasons a different translation. In this survey the words

“misdaden tegen de menselijkheid” is used. Misdaden is an informal translation of crimes often

used in the media for illegal behaviour. The formal and official option in Dutch is “misdrijven”

for high category criminal behaviour but this term is little used in everyday language. Humanity

is translated as “menselijkheid” which means human dignity instead of the possible translation

“mensheid” which means all of mankind. Both terms are used in journalism and everyday use

but the one I chose is preferred in the Netherlands (as opposed to Belgium) (Taalunie, 2009).

What falls under the concept crimes against humanity and what not is more of a legal discussion.

Article 7 of the Rome Statute of the International Criminal Court (UN General Assembly, 1998)

states the following:

“For the purpose of this Statute, "crime against humanity" means any of the following acts

when committed as part of a widespread or systematic attack directed against any civilian

population, with knowledge of the attack”

After this it lists crimes from the first one “murder” to the last one “Other inhumane acts of a

similar character intentionally causing great suffering, or serious injury to body or to mental

or physical health.”

(21)

21

The crimes listed are well known to civilians. In this research I have chosen 3 out of this list:

persecution, murder, and torture. There are no examples in the text of what these crimes would

further entail, it is beyond the scope of this research to do so.

The number of people that have already died in this conflict is literally “over a thousand”. I

have chosen this number because it is a lot but not an unbelievable amount. I did wat to use a

specific number instead of “a lot” so I would have some control over this factor. The additional

information that these crimes will not stop soon on their own is to create a sense of urgency

with the reader.

The information that the Netherlands have no strong cultural or economical connection with

the country in the survey is added because the attitude towards different countries could play a

role in how people evaluate a conflict or deaths. Examples of cultural connections are former

colonies, and European Union member countries. Examples of economic connections are a little

more complicated and less known to the public but because trade is one of the most important

parts of the Dutch culture and identity this part is added as well.

The Legality variable

After the background information on the conflict, respondent read the manipulation on legality

and UN Security Council authorisation. One manipulation (A) states that the missions was

legal. This is explained by saying that this is because the UN Security Council has agreed upon

starting a mission to protect victims against violence. It does not elaborate more on how this

mechanism works. Although it does state in the manipulation that the UN calls upon its

members whether they are willing to send troops, to give some explanation of the consequences

of this agreement for the Netherlands. The manipulation literally states that the mission in this

case is legal.

The other manipulation (B) states that the mission in Oshor is illegal. Again explaining that this

is because there was no UN Security Council authorisation. It stated that the members of the

UN Security Council could not agree on the necessity of starting a mission to protect the victims

against the violence. It literally states entering another country without UN Security Council

permission is illegal. This is a complicated statement in and of itself. The chance that The

Netherlands would enter into an illegal war is highly unlikely. When a mission has no UNSC

authorisation but a country does want to interfere it often does so in a coalition of countries

either with an already existing organisation like NATO or an ad hoc coalition. In the past this

(22)

22

has been done in, for example, Kosovo in 1999. The complicated aspect is that these

interventions are reviewed after the fact by the international court. In almost all cases the

intervention was legitimized in court. This means there was enough ground to start a mission

without being prosecuted. There is therefore a grey area between legal and illegal that for

audience and time reasons was not included in this research.

The expected deaths variable

The last part of each treatment was the expected deaths manipulation to measure the influence

of possible deaths among Dutch military personnel. Information on the risk in a certain area is

given by the Ministry of Defence (MoD). This organisation is chosen because it has no ties to

a certain political preference. Besides this the MoD is one of the most trusted governmental

organisations in the Netherlands (Schmeets, 2018). It is not uncommon for the MoD to give

advice about the risk of future missions, though there is little chance they would do this publicly

without political interference. For the sake of clarity it is added in this hypothetical situation.

Control questions

Both control questions, one for each manipulation, were measured on a likert scale from 1-5

(Bryman, 2016, p.166). To make sure the manipulations were understood by the respondents I

used an independent analysis of variance (ANOVA) to compare the manipulation of expected

deaths with the answers to the control question. This resulted in a significant difference of

means, F(2, 153) = 145.4, p < .001. For the legality variable the mean for legal was 4.77 while

the mean for illegal was 2.03, to test whether this difference was significant I used a t-test that

resulted in the following significant results, t (92, 6) = 14.73, p < .001. It can be concluded from

these tests that both manipulations were noticed by the respondents.

Explanation of the questions

The dependent variable will be support for participating in a military R2P mission. To measure

this several questions will be asked that focus on the amount of responsibility respondents

acknowledge, their general support for these kind of missions and under what circumstances

they would be willing to participate in a mission and some more specific questions on decisions

that should be made if the Netherlands decided to start military action. Together these questions

should give a good indication of the support for military R2P missions. All the support questions

will be done on a scale of 1-5, 1 being negative and 5 being positive, this is also the

measurement that is used for public support in this research.

(23)

23

The survey consisted of 10 questions, excluding the demographics questions and control

questions. All these questions have something to do with public support for participation in

military R2P missions. Two questions directly ask the respondent about what influence they

think the independent variables should have on the decision of participating in a military

mission. One question was excluded from the analysis because of vague wording, a quite similar

question was included so there was no gap in the information. To get a better understanding on

what the collected information consisted of a factorial analysis was conducted.

Factor analysis

A principal axis factor analysis was conducted on the 9 items with oblique rotation (direct

oblimin). The Kaiser– Meyer–Olkin measure verified the sampling adequacy for the analysis,

KMO = .74, and all KMO values for individual items were greater than .53, which is above the

acceptable limit of .5 (Field, 2013). An initial analysis was run to obtain eigenvalues for each

factor in the data. Three factors had eigenvalues over Kaiser’s criterion of 1 and in combination

explained 38.63% of the variance. The scree plot did not show inflexions that would justify

making the number of factors smaller. Table 3 shows the factor loadings after rotation. The

items that cluster on the same factor suggest that factor 1 represents support for R2P missions,

factor 2 represents Willingness to send military personnel and factor 3 involvement of

international factors. The factor Support for R2P missions has an acceptable reliability

Cronbach’s α = .73. The factor: willingness to send soldiers had a relatively low reliability,

Cronbach’s α = .53. The factor involvement of international factors showed an unacceptably

low reliability, Cronbach’s α = .22

1

_.

1_{This last two results are not surprising for the Cronbach’s alpha is sensitive to the amount of variables in a factor.}

Both factors have only two variables with a value above 0.3 that were taken into account in this reliability analysis. Therefore I think it is not so much the low correlation between the question as it is the low quantity of relevant questions in the factor.

(24)

24

Table 3: Factor analysis for the dependent variable questions

Rotated Factor Loadings

Questions Support for R2P missions Willingness to send military personnel Involvement of international factors Dutch responsibility to protect the group in the vignette. ,765 ,012 -,150 Dutch responsibility to protect people when their own government is

not able to do so.

,758 -,147 ,041

Priority of participating in R2P missions in the Netherlands. ,619 ,111 ,015 The responsibility the international community has to protect people

from mass atrocities.

,508 -,040 ,384

Dutch responsibility to protect people when their own government refuses to do so.

,477 -,096 -,017

The safety of Dutch military personnel should be the main consideration when deciding on R2P missions.

,340 ,156 -,025

The Dutch government should send a military mission to the area instead of trying to find a diplomatic solution.

-,123 ,690 ,068

The Dutch government should send military personnel to protect the group in the vignette.

,357 ,533 -,165

Authorisation of the UN Security Council should be the main consideration when deciding on R2P missions.

-,024 ,016 ,347

Eigenvalues 2.86 1.40 1.05

% of variance 31.80 15.59 11.64

(25)

25

Creating dummy variables

With the results of the factor analysis in mind, I created dummy variables to use for the

statistical analysis. I excluded the questions that literally ask the respondent for their opinion

on the independent variables so I can use them as a comparison to the results of the analysis.

This leaves the factor international involvement with only one variable left. I have made the

decision to not test for this factor for both this reason as well as the low reliability. The question

was also included in the factor support for R2P missions, so will not be excluded for the

analysis. This leaves us with 5 questions in the factor support for R2P missions and 2 questions

in the willingness to send soldiers factor. Two new variables were created with the means of

these factors. Both these variables were normally distributed, see figure 1 and figure 2. Together

these variables give an adequate picture of the dependent variable, support for a military R2P

mission.

Figure 1: Histogram of normal distribution R2P support

variable

Figure 2: Histogram of normal distribution Willingness to send military personnel

(26)

26

Data collection

Data collection is done by a self-administered online questionnaire. This method of data

collection is relatively cheap, and can easily be done in large quantities. The method is also

convenient to the respondent, for they can do the questionnaire in their own time at their own

speed (Bryman, 2016, p.222). The downside of this method of data collection, in comparison

to a structured interview for example, is that the researcher is less able to make sure the

respondent has understood the question and has answered truthfully. For the sake of this

research it is assumed that the respondent has understood the questions and has answered

truthfully. Another difficulty is respondent fatigue, this is the most realistic problem for this

research (Bryman, 2016, p.223-224). This research needs some background information for it

works with a made up scenario that needs to be explained. There is a dilemma between

explaining the casus properly and keeping it interesting for the respondent to answer the whole

questionnaire. The questionnaire will be distributed by a snowballing effect on social media

platforms and by the researchers own network. This convenience sample is further discussed in

the demographics section of this chapter.

Demographics

Some limitations in this research occur because of the way in

which the survey was distributed. This survey was a convenience

population (N=148), recruited from within my own network.

Among other findings this can explain that the majority of the

respondents (70.7%) was young adults between 18 and 30 years

old, see figure 3. A different explanation is that the survey was

only distributed online and young adults are more familiar with

using the internet in a social manner. The gender variable was

sufficiently equally distributed, male (44.9%) and female

(55.1%), see figure 4. There was no option in the survey for a

different gender. One person objected to identifying as either

male or female in the comments. This person is therefore

excluded from this particular analysis.

Some of the respondents did not fall into the original target group.

The target group was Dutch citizens over the age of 17. This

group was chosen is because the goal of the research, to give a

Figure 3: Age in categories

(27)

27

better idea about the ideas of the Dutch electorate. This way a possible difference between how

politicians think civilians judge a certain mission and how they actually judge can be brought

to light. The respondents that were excluded from the analysis were 7 people under the age of

18 and 2 people without a Dutch nationality. There is a total of 5 people who did not finish the

survey but did read the vignette. These people will be included or excluded per analysis,

depending on whether they answered the relevant questions.

A serious limitation in this research is the education level of

participants. The group of people with higher and academic

education levels is 70.7%, see figure 5. according to research of

the Central Dutch institute of statistics this is in reality only 19%

of the population (CBS, 2018). This overrepresentation can be

explained by how the survey was distributed. I am aware that

the majority of my network consists out of this demographic

group. Recent studies have shown that educational level and

political trust are positively related (Hooghe, Marien, & de

Vroome, 2012). This can partly explain the high level of

political trust in the current government in making decisions on

military missions (58.1% trusts the government oppose to 21.6% who does not). These

demographics are an important consideration when reading the results of this research. To

overcome these obstacles in a further study a different method for recruiting the respondents.

It took respondents on average less than 5 minutes (median: 4min27sec) to finish the survey.

There were no respondents who chose the option to stop the survey when given the chance after

reading the introduction. There were some instances where people closed the survey in this

stage but the data for this is not available in the dataset because it was already deleted in

Qualtrics to avoid confusion. As stated before, five people continued the survey after the

introduction but did not fully complete the survey.

Validity

Experiments are high on internal validity (Bryman, 2016, p.44). The use of a control group

rules out the possibility of an alternative explanation for a certain phenomenon. In this research

the respondent will not be aware that they are filling out one of six versions of the survey.

Therefore it is not possible for the respondent to be aware of the objective of the research and

Figure 5: Bar chart of the educational level of the respondents

(28)

28

Legality, Deaths and International Responsibility in the Netherlands: A public opinion research on the Dutch attitude towards Responsibility to Protect missions.

Master thesis

Crisis and Security management