Public Attitudes towards Use of Force Abroad: An Experimental Analysis for the Microfoundational

Transcription

1 Public Attitudes towards Use of Force Abroad: An Experimental Analysis for the Microfoundational roots of external intervention 1 Ali Çarkoğlu acarkoglu@ku.edu.tr Belgin San-Akca bakca@ku.edu.tr Department of International Relations Koç University Istanbul, Turkey * Draft, do not cite without permission. Earlier versions of this paper were presented at the Annual Meeting of International Studies Association, March 15-19, 2016, Atlanta, GA, USA and the Annual Conference of the Network of European Peace Scientists (NEPS), Catholic University of the Sacred Heart, June 20-22, 2016, Milan, Italy. 1 The research in this paper was funded by a Koç University Seed Fund, Open Society of Turkey and by Turkish Academy of Science (TUBA).

2 1 Abstract What are the conditions under which public supports military intervention abroad? The research on the use of force abroad has extensively focused on studying public opinion formation in US. The underlying assumption was that in democratic regimes, leaders care about public opinion vis-à-vis foreign policy decisions they make. This body of research explored factors, such as public perception of foreign policy objectives or context of the use of force, individual predispositions with respect to isolationism vs. internationalism, threat and power perception, and the utility of use of force. We build and expand on this body of research in two ways: 1- we move the debate to a non-us setting, and 2- we explore the influence of ideology and identity on people s support of intervention abroad by defining a new context of intervention that has not been explored so far; i.e. intervention to support protestors with political claims in a neighboring state. Considering the recent rise of populism across the world, it is essential to identify the codes and signals on which populist leaders can mobilize support for their external undertakings. We conduct three nationally representative on-field survey experiments in a one-year period to explore the influence of identity, democratic dispositions and the type of actor against which intervention is directed on public opinion towards the use of force abroad.

3 2 The ousting of Ukrainian President, Viktor Yanukovych, in February 2014 as a result of Euromaidan protests and the subsequent Russian aggression including the annexation of Crimea, situating troops in Eastern Ukraine, and supporting the pro- Russian separatists in Donetsk have revived the scholarly debate about the use of force in international politics. While some scholars argued that Putin was acting irrationally and had a plan to revitalize the Russian Empire (Braun 2014), some others stated that Russia was not acting irrational. In contrast, the Russian acts were driven by a perceived threat from a Europeanized or Westernized Ukraine located right at its border (Mearsheimer 2014). The more interesting issue, though, was the Russian public support for the intervention in Ukraine (Kolonitskii 2014). Similarly, when Turkey started Operation Euphrates Shield in Syria in August 2016, the public had almost unanimously agreed that it was required to protect Turkey s interests in Syria. Indeed, to some people, intervention in Syria was a long-delayed response to what was going on in Turkey s neighborhood. In both countries, there are highly popular and strong leaders, who can mobilize their people behind use of force abroad. By the election of Trump as the new president of US, scholars of Political Science started examining the recent rise of populism around the world (Inglehart and Norris 2016). The interest in populism is not only driven by its rise in Europe and North America, but also possible ramifications this might have for world peace and stability. Many now believe that the current situation of world politics has many similarities to the pre-wwii period mostly because populist leaders or political parties came to power everywhere. Therefore, it is essential to understand the codes and signals such leaders use in justifying the use of force abroad and whether individual citizens buy into these codes and signals and how. In this paper, we build on extensive research on the

4 3 internationalization of ethnic conflict, outside intervention in civil wars and social identity theory to build the theoretical framework and conduct the experimental tests. Considering existing research, we address the following questions with respect to public support of intervention in an internal conflict in a neighboring country: 1. What are the conditions under which public agrees with use of force abroad? 2. Do the most popular causes expressed by state leaders, such as protecting ethnic kin or spreading democracy are accepted as legitimate causes of external intervention by domestic audience? 3. How do predispositions about democratic values and ethnic identity influence individual citizens attitude towards the support of protestors or government side of an internal conflict? Considering that the recent conflicts in Syria and Ukraine have been exasperated by the intervention of multiple states, it is essential to understand how the public makes its decision to support foreign policy involving use of force abroad. In the rest of the paper, we explore the existing research on the conditions under which public supports external intervention as a foreign policy instrument and develop the theoretical framework to generalize from US public opinion to another setting. Next, we present the experimental design and empirical findings. I. Existing Research on Public Opinion and Foreign Policy The views that public cannot make informed decisions about foreign policy and thus, leaders should not design foreign policy in response to public opinion have been countered now by extant research. We learnt that public has consistent views on foreign policy issues and leaders, regardless of whether they are in liberal or illiberal regimes. This issue has become more significant by the rise of strong leaders across

5 4 the world, who have the potential to bypass audience cost or compensate for any foreign policy failures. It does not mean that such leaders are less attentive to public. In contrast, staying in power requires them to work hard to convince their public about a potential foreign policy undertaking. There is research on how various actors, such as politicians and media engage in framing to influence public opinion on certain issues (Jacobs and Page 2005, Druckman, Peterson et al. 2013). For example, Putin consistently made the case that Russia was acting to protect ethnic Russians in Ukraine to justify Russian intervention in Donetsk. Yet it is beyond the scope of this paper to track how strong leaders influence public opinion through issue framing, charisma and/or some other way. Even if that is the case, it does not change the fact that leaders work hard to respond to public demands not only about domestic politics but also foreign policy. Then, how does public make decisions about whether or not to use of force abroad? The study of public opinion about foreign policy issues has been driven primarily with an interest to study public s support of ongoing wars of US. Most of this work flourished during times of war, such as Korean and Vietnam Wars and the ongoing US involvement in Iraq, Afghanistan and Syria. We refer to two main traditions in the existing research as utility-based vs. value-based explanations. Utility-based explanations mostly refer to cost and benefit perceived by individual citizens for a specific course of foreign policy action. Scholars, who had this approach, emphasize the goal of a military intervention, the cost to be incurred, namely casualties, and perceived power of their country to carry out the military intervention with success as the major factors shaping public perception of utility from intervention (Mueller 1973, Jentleson 1992, Gartner and Segura 2000, Gelpi, Feaver et al. 2005/2006). Jentleson (1992) argues that the approval of use of force

6 5 abroad is determined according to citizens perception of objectives of a military mission. He finds that when the use of force is about prevention of aggression, people are more likely to support it in contrast to when it involves nation-building or humanitarian purpose. Value-based explanations build on individual predispositions shaped by specific values, such as conservatism, morality, and universalism and views related to international isolationism, nationalism and human rights (Kull, Ramsay et al. 2003/2004, Federico, Golec et al. 2005, Rathbun, Kertzer et al. 2016). These two distinct views of public attitudes towards foreign policy can also be referred as macro vs. micro-level explanation of public attitude. The utility-based explanations build on individual perceptions of cost and benefits attached to specific foreign policy pursuits whereas the value-based explanations investigate how these perceptions are influenced by micro-level factors in the first place, such as psychological predispositions and moral standards and norms held by citizens. The increasing attention to war by proxies and external intervention in internal conflicts require a detailed assessment of how public in the outside interveners respond to the policies of their government. Answering this question is even more significant if we consider that almost all outside interveners are either advanced democracies, such as the USA, France and UK or illiberal democracies, such as Russia. The attitudes of public in these countries will determine whether we will witness a major militarized dispute in the Middle East instigated by major powers listed above. In this paper, we develop and test a theory about public attitudes towards the use of force abroad and improve the external validity of the existing research by conducting a survey experiment in a non-european and illiberal democratic context.

7 6 2. A Micro-Level Theory of Intervention There is ample amount of work on how and why public opinion matters in foreign policy making and how domestic politics, in general, influence external behaviors of states. In general, the scholarly debate about the interaction between domestic and international politics can be divided into two camps. A group of scholars focus on domestic institutional and rather macro level attributes to explore domestic effects on external behavior of states (Maoz and Russett 1993, Bueno de Mesquita, Smith et al. 2003). 2 The other camp engages in micro-level factors and individual behavioral attributes to understand domestic public s view on foreign policy decisions (Page and Shapiro 1982, Page and Shapiro 1983, Page 1994, Page and Barabas 2000). Micro-level analyses gained prominence recently with the protracted campaigns in Afghanistan and Iraq and scholarly efforts to understand public support of military intervention in these two countries (Kertzer and McGraw 2012, Kertzer, Powers et al. 2014, Reifler, Clarke et al. 2014, Mader August 31, 2015). When building cases of military intervention abroad, states leaders frequently make reference to existing norms or ideational ties with the actors, which are the subject of intervention. Democratic norms and ethnic ties are extensively used by US and Russia respectively in the past couple decades. Not surprisingly, the 2003 invasion of Iraq was named Operation Enduring Freedom with a claim to democratize Iraq. By the same token, Russia sees it pretty much her natural right to interfere wherever there is a group of ethnic Russians (Öniş and Yılmaz 2015). In each case, 2 The literature on domestic politics and their influence on international behavior of states is immense. Only the most representative works on democratic peace, which is the most developed theory under this research field, have been cited. In addition, Fearon s work on domestic audience cost can be considered as a micro level theory since he indeed focuses on democratic public s responses, yet his empirical tests do remain limited to institutional characteristics. Only recently, there is a renewed interest on this issue. Researchers started putting more time and effort into capturing individual attitudes towards foreign policy choices of their governments.

8 7 interveners make a reference to the ideas and norms that hold their society together. US policymakers mostly use rhetoric of democracy-related norms and liberating the other societies. On the other hand, Russian policymakers build their case on the basis of transnational ethnic affinity. Are these norms and ideas reverberate in public opinion? Are citizens moved by democratic norms and ideas and/or transnational ethnic ties in forming their opinion about intervening in a conflict within the borders of another state? Extant research on external intervention in civil wars and ethnic strife establishes the following factors as major drives behind the decision of outside states to intervene on the government or opposition side of an internal conflict: interstate rivalry (between targets and supporters of armed opposition groups) (Saideman 1997, Salehyan 2010, Maoz and Akca 2012), transborder ethnic kin (Carment and James 2000, Saideman 2001, Saideman 2002, Cederman, Girardin et al. 2009, Cederman, Gleditsch et al. 2013), and relative strength of target states vis-à-vis intervening states (Saideman and Jenne 1992, Saideman 1997, Bapat 2007, Salehyan, Gleditsch et al. 2011). Building on these empirical findings, it is anticipated that leaders use rhetoric to justify a decision to use force abroad along these lines. How do leaders benefit from a rhetoric relying on transborder ethnic kin or promoting democracy as justification to use force abroad? Building on Social Identity Theory (SIT), we suggest that leaders use such rhetoric to elevate the status of their support base at home so that they can mobilize material and human resources to use force abroad. SIT is a micro-level social-psychological approach, which explains the process of individual identity formation, forming ties with individuals from similar ideational and cultural background, and inter-group comparison. It was formed to explain the competition between groups in a society initially (Tajfel, Billing et al.

9 8 1971, Tajfel and Turner 1986). It starts with the individual as the major agent in search for solidarity with other individuals, who feel similarly about a specific issue of contention. In addressing their grievances, individuals either join low-status groups or seek collective action to elevate the status of in-group at the expense of out-group. Or they join a high-status group through upward mobility and drive utility from being a member to this high-status group. Furthermore, whether individuals support intervention to promote democracy abroad or protect their ethnic kin should be a function of their predispositions vis-àvis democratic values and their own ethnic identity. In this paper, building on two main bodies of research, namely democratic peace and transnational ethnic ties, in explaining inter-state conflict, we propose that citizens are positively influenced when there are democracy-related claims in an internal conflict and when they share ethnic times with the parties involved in a given conflict. Democratic peace argument would imply that democratic norms and institutions should be spread across the world to achieve permanent peace (Doyle 1983, Maoz and Russett 1993). Adhering to this idea, we have recently witnessed intervention by various states, such as the US and France, to support rebels trying to establish a democratic regime in places, such as Libya, Iraq and Syria. We conduct an experimental test to see how individuals respond to intervention in the neighboring states especially in a state, which is in the middle of a conflict-region. The fact that these recent militarily superior interveners, such as US, UK and France, are also advanced democracies is puzzling for the relationship between domestic politics and foreign policy. With respect to the mechanisms through which domestic politics effect is observed, scholars frequently mention institutional design, whether democratic or autocratic (Bueno de Mesquita, Smith et al. 2003) and the

10 9 effect of various structural factors on the public perception of leaders and the leaders response to their public with respect to foreign policy failures (Chiozza and Goemans 2011). Especially with increasing number of democracies in the world, the scholarly community shifted from focusing on military might as a sign of credible signaling and threat in interstate environment to domestic mechanisms through which leaders are held accountable for their foreign policy actions and decisions. States credibility in international arena is measured by how resilient the leaders are to domestic pressures if they pursue a faulty foreign policy strategy or renege on their words related to foreign policy matters. For a while, the scholarly community had ignored the public s view about foreign policy based on the assumption that laymen are rarely knowledgeable about foreign policy matters (Sartori 1969, Carr [1945]2001, Morgenthau [1948]1993). As a result, foreign policy choices should have little reflection of public demands. Yet these views have long been challenged. Leaders are kept accountable for their foreign policy decisions as much as domestic political decisions. More importantly, politics is a two-level game (Putnam 1988). Leaders try pleasing their domestic audiences with their foreign policy choices and decisions as well. Therefore, even the recent militarily superior interveners frame their actions around democratic norms and values to get support of their public. Though the Euromaidan protests mostly supported by US and European states proved successful in toppling a pro-russian government and replacing it with a pro-european one, Ukraine ended up paying a very high price when Russia decided to annex Crimea and intervene in the Eastern regions of Ukraine through its support of pro-russian militants. The international community remained inactive while Russia was expanding its territory. The US and European states did not

11 10 go beyond diplomatic engagement and economic sanctions to deter Russia from interfering in Ukrainian domestic affairs. The use of force is frequently a very challenging option especially for democratic states, which are directly accountable to public for any policy failure. Furthermore, a possible direct armed confrontation with Russia was also not among the options for the US and European policymakers. Yet states that have democracy promotion in their agenda and support bottom-up Arab Spring type democratization protests need to make more informed decisions when it comes to making sure that these bottom-up movements succeed. Part of the reason why various Arab Spring revolutions failed to bring democracy is attributed to the lack of decisive international support from major powers, such as UK, France, and US. The research on democracies and interstate conflict is ample, yet the causal mechanism linking democracy to peace at the international level is still debated. Given the direct democratic accountability for their actions in office, political leaders are argued to be attentive to the expectations and preferences of their mass constituencies in domestic as well as foreign policy areas. Since they face a risk of loosing office in case of a failure to meet mass preferences, leaders of democracies are hesitant to get involved in war against other states to begin with, particularly against democracies (Bueno de Mesquita et al. 2003; Maoz & Russett 1993). Taking the debate further, researchers have then been exploring what makes publics in democracies and autocracies to oppose war against other states (Tomz 2007, Weeks 2008). Studies that examine individual attitudes towards war have yielded significant empirical findings about how public opinion towards war is shaped. Gartner (2008) conducts an experimental study about the US public opinion for supporting the war

12 11 against Iraq and finds out that as casualties increase, public support declines. When it comes to public approval of war, the most frequently explored area is the causal link between casualties and public support (Mueller 1973, Gartner and Segura 2000, Gelpi, Feaver et al. 2005/2006). Yet this research remains limited to the US. By the same token, Jentleson (1992) examined the role of the identified objective of a military intervention. He finds out that if the objective is identified as nation-building, people are less likely to support the intervention. We would like to retest this argument by clarifying that the goal of the intervention is to support protestors with democratic claims. H1: Individual citizens are more likely to support intervention if the objective is to support protestors with democratic claims. Scholars, who study intervention in civil war, extensively examine the role of transnational ethnic ties in the decision to intervene and on whose side to intervene (Saideman 1997, Saideman 2001, Cederman, Girardin et al. 2009, Cederman, Gleditsch et al. 2013). We know that states whose majority ethnic group sharing similar ties with an ethnic rebel group in the borders of another state is more likely to support the group in question. Therefore, we test this finding in the context of ethnically contiguous protestors within the borders of other states. We argue that; H2: If protestors are from the same ethnic background as the individuals, the latter is more likely to support foreign policy to abet and empower the protestors. Research on public opinion s effect on foreign policy has yielded three main theoretical arguments: 1- Rationality: individual citizens follow the basics of rational actor assumption by building their decisions on a cost-benefit calculus (Roussenau

13 , Kertzer and McGraw 2012) 3 ; 2- Leaders Charisma: citizens decide in line with whether they like their leaders or not (Reifler et al. 2014); 3- Morality: individual citizens follow certain norms and principle codes of conduct in making their decisions about which foreign policy act is a proper one (Kertzer et al. 2014). The theoretical argument building on morality borrows from the Moral Foundations Theory (MFT) (Haidt 2012) and examines the role of moral attributes, namely, harm/care, fairness/reciprocity, authority/respect, ingroup/loyalty, and purity/sanctity in shaping public opinion in cooperative internationalism and militant internationalism. The first two attributes constitute the individualizing foundations of morality, then the last three belong to binding foundations of morality (Keltzer et al. 2014). The findings suggest that individuals making their decisions on the basis of realist ideology usually associate with binding moral values, which make individuals to value obedience and respect for authority and role fulfillment. Drawing parallel with the suggestions of realism, we test the implications of power, threat and utility perception for citizens attitudes towards several foreign policy options including the direct use of force. We initially test typical hypotheses: H3: If a citizen has a high perception of its country s military capacity, he/she is more likely to support a masculine foreign policy. H4: If a citizen is worried about a future threat or aggression against his/her country, he/she is more likely to support a masculine foreign policy. H5: Citizens will choose the option with the highest level of utility according to their perception. Another relevant body of research examines public attitudes towards war in more direct manner by employing surveys and relying on rational choice theory. 3 Kertzer and McGraw does not specifically talk about rationality or cost-benefit analysis as the main drive behind public opinion. Yet their work builds on realism and liberalism being a character of individuals as well as states. Realism would make the assumptions with respect to power and threat perceptions, which constitute the major elements of rational actor calculus in international relations.

14 13 Publics in democracies support wars if they anticipate economic benefits from war as rational actors (Caverly 2014). Furthermore, more recent work emphasized the democratic targets and tried examining the micro-level factors in seeking support for democratic peace. Tomz and Week (2013) talk about some mediator variables, such as threat and morality in seeking evidence for democratic peace in individual attitudes. They found that individuals in democracies oppose war against other democracies because they do not perceive threat from other democracies. In addition, war against other democracies is perceived to be immoral by democratic publics. Yet we all know that democracies fight wars, despite that these wars are more frequently against autocracies. One argument we make about the democracy-causality nexus is related to indirect war. If democracies are worried about casualties in promoting democracy, protestors who make democratic demands from their government present a good opportunity to promote the democratic agenda without risking their own soldiers. In other words, the long-standing divide and rule strategy is also present at the microfoundational level. Rather than sending troops to a foreign territory, democracies take on the ongoing conflict between a rebel group and target state. This kind of international aggression has been recently observed in the case of Syria and various other Middle Eastern states going through internal turmoil. Free Syrian Army received support from various Western states including the USA in its fight against the Assad regime. People who oppose war will prefer indirect war if the need arises. In this paper, the analysis is simply limited to protestors, who are not specified whether they use violence or not. H6: Individuals are more likely to support protestors with democratic claims from their government rather than protestors with no such claims.

15 14 3. Empirical Analysis We conducted a survey/experiment on a nationwide representative sample of 4709 participants/respondents throughout December 2014 and December 2015 across three periods: December 2014-January 2015 (N=2334), September-October 2015 (1086) and November December 2015 (N=1289). Since Turkey is geographically contiguous to the most recent conflict zones in the world including Ukraine, Syria, and Iraq, it is not unimaginable for Turkish public to think about a scenario where Turkey gets militarily involved in a conflict. In addition, right prior to conducting the third survey, the Russian jet incident happened, thus external threat environment was a natural treatment. The pre-treatment vignette describes the situations in Syria, Ukraine and Iraq (without providing a specific country name) and implies the recent refugee flow from Syria. Subsequently, respondents were presented with three policy options available to government and asked to specify the utility of each option. Below is the vignette each respondent received in the first and second surveys: Imagine that there are demonstrations in a neighboring country wherein demands for a better democratic regime are expressed. The leadership of the country uses oppressive measures leading to clashes between the security forces and demonstrators, which result in, several killings from both sides. As a result significant groups of people start fleeing into Turkey. Perceiving all of these developments as a threat to Turkey s security the government has three alternative policy options: 1. Turkish government should close its borders with the neighboring country and start applying economic sanctions (close border/economic sanctions). 2. Turkish government should start supporting the protestors (support protestors). 3. Turkish government could declare that all possible military precautions will be taken including all necessary uses of military force against the neighboring country s regime. (use force) In the third survey, the following vignette was received by each respondent; Imagine that there are demonstrations in a neighboring country wherein some demands from the government are expressed. The leadership of the country uses oppressive measures leading to clashes between the security forces and

16 15 demonstrators, which result in, several killings from both sides. As a result significant groups of people start fleeing into Turkey. Perceiving all of these developments as a threat to Turkey s security the government has three alternative policy options: 1. Turkish government should close its borders with the neighboring country and start applying economic sanctions (close border/economic sanctions). 2. Turkish government should start supporting the protestors (support protestors). 3. Turkish government could declare that all possible military precautions will be taken including all necessary uses of military force against the neighboring country s regime. (use force) Throughout the period we conducted the surveys, two major incidents influenced the domestic and international landscape of Turkey: in October 13, 2015, two bombs were detonated during a rally against the ongoing violent confrontations between the government and PKK (Kurdistan Workers Party) in Ankara leading to the killing of 95 people. The attacks were connected with an individual who had connections to ISIS (Islamic State of Iraq and Syria). During the period prior to the attacks, there was increasing security alert within the country due to the operations Turkish governments launched against PKK, specifically in some eastern towns to demolish the group s strongholds. The August-September 2015 surveys were conducted during a period where security alerts were gradually on the rise. On November 24, 2015, Turkey downed a Russian jet, which violated the Turkish airspace. The third survey was conducted right around that period, when Russian president was issuing threats against Turkey. Though, unplanned, these events turned out to be a natural treatment giving us an opportunity to really single out the effect of threat as a confounding variable with a natural experiment. Figure 1 displays the average level of threat perception ranging from no threat (0) to high threat (10) among the respondents across three surveys. The threat perception question was asked prior to the vignette.

17 16 (Figure 1 about here) In the first stage of the experiment, respondents are asked to first answer the following questions. Question 1. (Utility) To what extent do you think that each one of these policy options would be beneficial for Turkey? Express your evaluation on a scale ranging from 0 as the least beneficial to 10 the most beneficial for each alternative. (Table 1 about here) Table 1 presents the utility preferences clustered around low, moderate and high levels across three surveys. Across all three surveys, respondents attribute the highest level of utility to the option of closing border and applying economic sanctions. Indeed, from a descriptive perspective, the figures on Table I show that almost 59% of respondents attributed the high level of utility to use of force option, while it was 44% and around 42% in the preceding two surveys, respectively. Examining the figures with respect to support of protestors, though there seems to be a decline in the second survey, there is an immediate increase subsequently in the third survey. Considering the nature of the threat environment, it can be argued that the recurrence of conflict between the government and PKK divert attention entirely to the domestic political arena. There is a significant increase in the utility attributed to the close border/apply economic sanctions option (a shift from 52% to almost 61%) whereas there is a decline in the utility attributed both to the support protestors and use force options approximately around 4%). In other words, when domestic threat perception increase, individuals are less likely to support active engagement or militarist options in the domestic politics of other states. When it comes to the increasing external threat environment, there is a drastic increase in the utility attributed to both options. Almost 6% of respondents attributed moderate or high level

18 17 of utility to support protestors option. With respect to use of force, the shift seems to be more drastic, almost 64% of respondents attributed moderate and high level of utility to this option after the downing of the Russian jet. While around 42% of respondents attributed a high level of utility to this option, after the incident, this figure rose up to 59%. Multivariate analysis will help controlling for the confounding effect of threat and power perception as well as other variables, such as economic evaluation and the experimental treatments with respect to the nature of protestors in the neighboring country and their ethnic identity. When asked about the most preferred outcome, though, across all three periods, the isolationist option seems to be the most preferred one with more than 50% of respondents choosing this option. Whereas there is a decline in the option of support the protestors and use force across three surveys. As mentioned above, the first and second surveys specified that the protestors were democratic-oriented whereas the third survey used a neutral language. We see the highest level of support for support protestors in the first survey while a decline in this figure in the second survey despite the fact that the same specification was done with respect to the nature of the protestors. As mentioned earlier, this could be attributed to the internal threat environment of Turkey. The multivariate analysis will show the effect of democracy treatment by controlling for all these confounding effects. Question 2. (Policy preference) Among the above three alternatives, which one would you support the most? (Table 2 about here) We next asked our respondents to provide to us an assessment as to the degree to which they feel confident that Turkey has the necessary capacity to succeed in this policy line that they support the most. Again we used a 10 point scale ranging from

19 18 0 representing no capacity to 10 representing full capacity to succeed in this policy. Figure 2 shows that our respondents are predominantly confident that the government has the power to pursue the chosen policy. Nevertheless we observe that about 15% appear unconfident that the government has the capacity to pursue their most preferred policy option. (Figure 2) In the second stage of the experiment, we asked respondents their most preferred option after specifying that the protestors speak the same language as themselves. Table 3 shows the figures after receiving the ethnic treatment. (Table 3 about here) Comparing the figures across three identified options before and after treatment, we can see the shifts across each option after the ethnic kin treatment. The highest shift seems to be within the close border-apply economic sanctions and support protestors. Among those who chose the close border option before the ethnic treatment, 16.5% shifted to the support protestors option and 8% to use force option after hearing that these protestors are from the same ethnic background like them. Among those who preferred support protestors option prior to the treatment, 58% continued to say so, but 17.5% shifted to use force option and 13.3% shifted to close border option. By the same token, among those who preferred use force option before the ethnic kin treatment, almost 65% continued to say so, while around 14% shifted to support protestors option. The descriptive analysis still demonstrates that the highest persistence is within the close border option. 68% of respondents continue supporting the same option even after the ethnic treatment. Among those who did not state any option prior to the treatment, we see that 8.2% shifted to use force, 5.2% shifted to support protestors, and 6.2% shifted to close border option. This is a sign that people

20 19 get moved by their ethnic kin suffering in the neighboring countries from state repression. Multivariate Analyses We estimated two sets of models. Table 4 displays the findings of multinomial logit analysis with base category selected to be close border/apply sanctions for both democracy and ethnic treatment. The dependent variable is the corresponding category of most supported policy option. Table 5 estimates binary logit models for each polity option after controlling for the utility variables. The purpose is to control for a potential confounding effect of utility attributed to each policy option. Model 1 shows the effects of the key variables on the likelihood of preferring a government policy devised to help the protestors against the government of the neighboring state. In another paper, we examine the effect of violence, secessionism and terrorism by using these as treatments to identify the protestors. In this paper, protestors are not specified to use violent means neither we made an implication to emphasize that they were not using violence. The language was neutral; simply implying that the neighboring country s government was trying to repress the protestors. The estimates show that when it is specified that protestors are rallying for more democratic reforms and rights, the respondents are more likely to prefer the support protestors option whereas there is no significant effect on use of force option. This implies that Turkish citizens do not want to meddle in the domestic affairs of a neighboring state to impose regime change by force (Model 2). They are ok with supporting the protestors, but not with a direct war against the neighboring country s regime. This also has implications for Turkish public attitude towards proxy war or indirect war. But for now, we rather want to stay within the confines of supporting the protestors, avoiding stretching the implications of our findings. In future research, this

21 20 is another venue we would like to expand. When respondent is from a minority background, in Turkey s case Kurdish, it has a positive effect on the likelihood of a policy geared towards supporting the protestors. This is reasonable considering the situation of the Kurds in Syria though we did not specify country names. Surprisingly, this variable is also significant for the use force option (model 2). Again, this is expected if one examines the events in Syria. Model 3 shows that ethnic kin treatment has a positive effect on the likelihood of preferring an option involving support of the protestors. It matters whether these protestors are from the same ethnic kin, in this case speak the same language, as the respondent. Across all models, threat and power perception have a negative effect on the most preferred outcome. As threat perception increases, people are less likely to prefer militarist options involving support of the protestors and use of force against the neighboring country. This is also in line with the descriptive findings mentioned above. The only time power perception has a positive effect is when after ethnic treatment, people preferred the use of force option. This can be attributed to the Russian attacks against Turkmen in Syria. Turkmens are from the same ethnic background like Turks and this might have an influence for those who wanted to use force to protect them. Further analysis is required to check the effect of being a Turk in comparison to a Kurd to see if divergent patterns are present. This will have implications for transnational ethnic kin research, which frequently investigates which transnational ethnic kin matters: the majority ruling the country or the minority. With respect to economic evaluations, it seems to have a positive effect across all models except the use of force after ethnic treatment (Model 4). In other words, people are more likely to support militarist options if they perceive their economic standing better off. Yet it does not seem to matter if they choose the

22 21 use force option if their ethnic kin is being treated badly by their host government. They will run to rescue regardless of their economic situation. Table 5 presents the findings for three preferred options after controlling for the utility attributed to each option by the respondents. Findings with respect to democracy treatment are reconfirmed after controlling for the utility of each option, though they are a little puzzling. Whereas in multinomial logit (table 4), democracy treatment does not have a significant effect on the use of force option, this time it seems to have a significant positive effect both on support protestors and use force options. In other words, people s attitudes with respect to foreign policy options are determined both by the assessed utility and ideational factors associated with the preferred policy option. Use of force to help with democratization efforts within a neighboring country is only preferred if it is also perceived to be useful (model 7). Furthermore, while satisfaction with current situation of democracy in Turkey does not have any significant effect in the multinomial logit estimates, after controlling for utility, it seems that if a respondent is satisfied with the democratic situation in Turkey, he or she is more likely to be isolationist. On the other hand, if a respondent is not satisfied with the level of democracy in Turkey, they are less likely to prefer use force option. Another finding providing further empirical evidence for the ethnic kin effect, when a respondent is from Kurdish background, it is less likely to support the isolationist option (model 5), though it does not seem to have a significant effect on support protestors and use force option. This could be attributed to the lack of confidence by Kurdish constituents in Turkey towards a possible policy by which Turkish government will help the protestors from Kurdish ethnic background in a neighboring country or use force to protect them.

23 22 4. Conclusion Our initial hypotheses about threat and power perception seem to be not confirmed. Yet the Turkish case presents some puzzles and possible venues to think about the threat perception and use of force relationship in more detail and offer alternative theories. Our hypothesis with respect to conventional effect of utility (H3) is confirmed. Findings from table 5 show that citizens have a tendency to prefer the option with the highest expected utility. Both of our treatments seem to influence the preference of people in the expected direction. People are more likely to support protestors if they pursue democratic claims from their target governments and are from the same ethnic background like themselves, yet they do not want to go as further as intervening through direct military actions (models 2 and 4). Indeed, they prefer use of force directly against the target government of the protestors if the find it useful; i.e. they assess the possible costs and benefits. There seems to be some findings about the micro-foundational sources of democratic militarism and transnational ethnic kin. Further analysis will explore these linkages in more detail by considering possible interactions with the threat environment and the objectives of the protestors and means to pursue them.

24 23 References Bapat, N. A. (2007). "The Internationalization of Terrorist Campaigns." Conflict Management and Peace Science 24(4): Braun, A. (2014). Putin's dangerous misadvanture: a 'greater Russia' or a 'great Russia?'. The Globe and Mail. Bueno de Mesquita, B., et al. (2003). The Logic of Political Survival. Cambridge, The MIT Press. Bueno de Mesquita, B., et al. (2003). The Logic of Political Survival. Cambridge, Massachusetts Institute of Technology. Carment, D. and P. James (2000). "Explaining third-party intervention in ethnic conflict: theory and evidence." Nations and Nationalism 6(2): Carr, E. H. ([1945]2001). The Twenty Years' Crisis. Basingstoke, Palgrave. Caverly, J. (2014). Democratic Militarism: Voting, Wealth and War. Cambridge, Cambridge University Press. Cederman, L.-E., et al. (2009). "Ethnonationalist Triads: Assessing the Influence of Kin Groups on Civil Wars." World Politics 61(3): Cederman, L.-E., et al. (2013). "Transborder Ethnic Kin and Civil War." International Organization 67(2): Chiozza, G. and H. E. Goemans (2011). Leaders and international conflict. Cambridge, Cambridge University Press. Doyle, M. (1983). "Kant, liberal legacies, and foreign affairs." Philosophy and Public Affairs 12(3): Druckman, J. N., et al. (2013). "How Elite Partisan Polarization Affects Public Opinion Formation." American Political Science Review 107(1): Federico, C. M., et al. (2005). "The Relationship Between the Need for Closure and Support for Military Actions Against Iraq." Personality and Social Psychology Bulletin 31(5): Gartner, S. S. (2008). "The Multiple Effects of Causalities on Public Support for War: An Experimental Approach." American Political Science Review 102(1): Gartner, S. S. and G. M. Segura (2000). "Race, Casualties, and Opinions in the Vietnam War." Journal of Politics 62(1):

25 24 Gelpi, C., et al. (2005/2006). "Success Matters: Casualty Sensitivity and the War in Iraq." International Security 30(3): Haidt, J. (2012). The Righteous Mind: Why Good People are Divided by Politics and Religion. New York, Pantheon. Inglehart, R. F. and P. Norris (2016). Trump, Brexit, and the Rise of Populism: Economic Have-Nots and Cultural Backlash. Harvard Kennedy School Faculty Research Working Paper Series. Jacobs, L. R. and B. I. Page (2005). "awho Influences U.S. Foreign Policy?." American Political Science Review 99: Jentleson, B. (1992). "The Pretty Prudent Public." International Studies Quarterly 36(1): Kertzer, J. D. and K. M. McGraw (2012). "Folk Realism: Testing the Microfoundations of Realism in Ordinary Citizens." International Studies Quarterly 56(2): Kertzer, J. D., et al. (2014). "Moral Support: How Moral Values Shape Foreign Policy Attitudes." The Journal of Politics 76(3): Kolonitskii, B. (2014). Why Russians Back Putin on Ukraine. New York Times. Kull, S., et al. (2003/2004). "Misperceptions, the Media, and the Iraq War." Political Science Quarterly 118(4): Mader, M. (August 31, 2015). "Citizens' Perceptions of Policy Objectives and Support for Military Action: Looking for Prudence in Germany." Journal of Conflict Resolution Early view: Maoz, Z. and B. S. Akca (2012). "Rivalry and State Support for Non-state Armed Groups (NAGs)." International Studies Quarterly 56(4): Maoz, Z. and B. Russett (1993). "Normative and Structural Causes of Democratic Peace, " American Political Science Review 87(3): Mearsheimer, J. (2014). "Why Ukraine Crisis is the West's Fault." Foreign Affairs 93(5). Morgenthau, H. J. ([1948]1993). Politics among Nations: The Struggle for Power and Peace. Boston, McGraw Hill. Mueller, J. E. (1973). War, Presidents, and Public Opinion. Lanham, MD, University Press of America.

26 25 Öniş, Z. and Ş. Yılmaz (2015). "Turkey and Russia in a shifting global order: cooperation, conflict and asymmetric interdependence in a turbulent region." Third World Quarterly Online First. Page, B. I. (1994). "Democratic Responsiveness? Untangling the Links between Public Opinion and Policy." PS: Political Science and Politics 27(1): Page, B. I. and J. Barabas (2000). "Foreign Policy Gaps between Citizens and Leaders." International Studies Quarterly 44(3): Page, B. I. and R. Y. Shapiro (1982). "Changes in Americans' Policy Preferences, " Public Opinion Quarterly 46(1): Page, B. I. and R. Y. Shapiro (1983). "Effects of Public Opinion on Policy." American Political Science Review 77(1): Putnam, R. D. (1988). "Diplomacy and Domestic Politics: The Logic of Two-Level Games." International Organization 42(3): Rathbun, B. C., et al. (2016). "Taking Foreign Policy Personally: Personal Values and Foreign Policy Attitudes." International Studies Quarterly Online First: Reifler, J., et al. (2014). "Prudence, Principle and Minimal Heuristics: British Public Opinion Toward the Use of Military Force in Afghanistan and Libya." British Journal of Politics and International Relations 16(1): Roussenau, D. L. (2007). "Identity, Power, and Threat Perception: A Cross-National Experimental Study." Journal of Conflict Resolution 51(5): Saideman, S. M. (1997). "Explaining the International Relations of Secessionist Conflicts: Vulnerability vs. Ethnic Ties." International Organization 51(4): Saideman, S. M. (2001). The Ties That Divide: Ethnic Politics, Foreign Policy, and International Conflict. New York, Colombia University Press. Saideman, S. M. (2002). "Discrimination in International Relations: Analyzing External Support for Ethnic Groups." Journal of Peace Research 39(1): Saideman, S. M. and E. K. Jenne (1992). The International Relations of Ethnic Cofnlict. The Internationalization of Communal Strife. M. I. Midlarsky. London, Routledge: Salehyan, I. (2010). "The Delegation of War to Rebel Organizations." Journal of Conflict Resolution 54(3): Salehyan, I., et al. (2011). "Explaining External Support for Insurgent Groups." International Organization 65(4):

27 26 Sartori, G. (1969). "Politics, Ideology, and Belief Systems." American Political Science Review 63(2): Tajfel, H., et al. (1971). "Social Categorization and Intergroup Behavior." European Journal of Social Psychology 1(2): Tajfel, H. and J. Turner (1986). The Social Identity Theory of Intergroup Behavior. Psychology of Intergroup Relations. S. Worchel and W. Austin. Chicago, Nelson-Hall: Tomz, M. (2007). "Domestic Audience Costs in International Relations: An Experimental Approach." International Organization 61(4): Tomz, M. and J. Weeks (2013). "Public Opinion and the Democratic Peace." American Political Science Review 107(4): Weeks, J. L. (2008). "Autocratic Audience Costs: Regime Type and Signaling Resolve." International Organization 62(1):

28 Figure 1. A Natural Stimulus for Threat 27

29 28 Table 1. Utility Preferences across Policy Options (% of Respondents) Close Border/Economic Sanctions Support protestors Use Force Utlility Level Protestors with democratic claims Protestors with some claims Survey 1 Survey 2 Survey 3 December 2014-January 2015 August- September 2015 November- December 2015 Low Moderate High Low Moderate High Low Moderate High

30 29 Table 2. Most Preferred Policy Option Protestors Protestors with democratic claims with some claims Survey 1 Survey 2 Survey 3 December 2014-January 2015 August- September 2015 November- December 2015 Close Border / Economic Sanctions Support Protestors Use Force N/A

31 30 Figure 2. Power Perception Response to Question 3. (Power perception) Please also provide to us an assessment as to the degree to which you feel confident that Turkey has the necessary capacity to succeed in this policy line that you support the most. Again use a 10 point scale ranging from 0 representing no capacity to 10 representing full capacity to succeed in this policy.