Emotional reactions to immigration and support for EU cooperation on immigration and terrorism

ABSTRACT What explains variation in European citizens’ support for common EU immigration and counter-terrorism policies? We advance extant literature that focuses on the utility versus identity debate by focusing on individuals’ emotional reactions. Drawing on theories of affect, we show that citizens’ emotions about immigration are integral to their preferences for EU cooperation on the dual questions of immigration and terrorism. We hypothesise that while anger about immigration is associated with opposition to cooperation on both policies, fear about immigration is associated with support for a common EU counter-terrorism strategy. Using a large-N cross-sectional survey conducted in Germany and the Netherlands, our analyses confirm our hypotheses. Our findings have implications for the progress of European integration and the scope of public approval of EU common policies.


Introduction
The influx of more than 1.5 million refugees, asylum seekers and migrants into Europe, especially since the war in Syria, has coincided with immigration and terrorism becoming the two most important issues for European Union (EU) citizens, replacing economics (Eurobarometer 2018). Although the vast majority of refugees are hosted in non-EU member states and immigration also occurs within the EU (Vasilopoulou and Talving 2019), the European immigration crisis has attracted significant attention for at least two reasons. First, host-country nationals may perceive this new flow of people as a potential threat to their existing social and demographic fabric. Second, the origin and religion of the majority of refugees has allowed far right political actors not only to politicise identity, religious and value-based differences in their campaigns, but also to link immigration to terrorism.
To address immigration and security-related challenges, European governments are faced with two options. The first possibility is to address these issues domestically. Some EU member states, such as Italy, Hungary and Poland, have taken unilateral action against the EU's official line. The second option is to deal with these issues through cooperation at the EU level, which might allow member states to share the burden of immigration, protect refugees and collectively fight terrorism. The challenge here is to build a strong multilateral framework while at the same time ensuring public approval of all these initiatives.
Against this background, this article examines the impact of citizens' emotional reactions to immigration towards their preferences for common EU immigration and counter-terrorism policies. Emotions influence the ways in which individuals perceive and process threats and how they form their opinions (Brader and Marcus 2013;Redlawsk and Pierce 2017). Previous work has examined the effects of emotions across policy domains, including terrorism (e.g., Keltner 2000, 2001;Huddy et al., 2005Huddy et al., , 2007, immigration (e.g., Brader et al. 2008;Erisen and Kentmen-Cin 2017), climate change (Davydova et al. 2018), and economic crisis (Magni 2017). Following the same scholarly strand, we draw on theories of affect and posit that emotional reactions triggered by immigration influence the extent to which citizens prefer domestic over international solutions to the dual questions of immigration and terrorism.
Using evidence from two online surveys conducted in Germany and the Netherlands in 2015-2016, our study makes a three-fold contribution. First, we fill a gap in the literature on Euroscepticism by focusing on preferences for international cooperation on immigration and terrorism. Existing research has focused on preferences for renegotiation (Vasilopoulou and Wagner 2017), widening versus deepening (Hobolt 2014), European economic governance (Kuhn and Stoeckel 2014), EU freedom of movement (Vasilopoulou and Talving 2019) and trust in EU institutions (Armingeon and Ceka 2014). However, we still do not know how Europeans view EU cooperation on the questions of immigration and terrorism, whether they view the two policies as related, and what explains variation in their sentiment towards these two highly salient issues.
Second, we examine the emotional underpinnings of preference formation specifically in the context of EU public opinion. By and large, political science has drawn upon political economy and cultural/identity approaches in order to explain both attitudes towards immigration and public support for European integration (Hobolt and De Vries 2016;Hainmueller and Hopkins 2014). In doing so, it has ignored the impact of the psychological processes that may underpin attitude formation, and are often better at explaining short-term changes (Erisen 2017;Magni 2017;Vasilopoulos 2018). In addition, research exploring the role of emotions in understanding citizen concerns about immigration primarily focuses on the American context (Brader et al. 2008). The role of emotional reactions in attitudes towards the EU is limited (see Garry 2014;Vasilopoulou and Wagner 2017). No contribution has explicitly addressed the specific impact of immigration-related emotions on EU cooperation in this field, despite the fact that immigration is an emotionally loaded political issue at the forefront of domestic and EU-level debates.
Third, by incorporating theories of affect into the study of EU policy cooperation, we show that distinct emotions have divergent effects on people's preferences. Anger about immigration leads individuals to oppose EU level cooperation on both immigration and terrorism, essentially preferring that these policies are exclusively managed at the nation-state level. Anxious individuals, on the other hand, support national sovereignty on the question of immigration, but EU decision-making in the field of terrorism. Our results are robust as they hold controlling for key variables, including trust, European identity and hostility against immigration. Our findings have implications for the progress of European integration and the scope of public approval of EU common policies.

Emotional reactions to immigration and attitudes towards EU policy cooperation
Research in political psychology suggests that a combination of conscious and preconscious appraisals generate discrete emotions, including anger, fear and enthusiasm (Berkowitz and Harmon-Jones 2004). Anger and fear on the one hand tend to be activated by situations where people feel under threat, e.g., terrorist threats (Huddy et al. 2005;Iyer et al. 2014). Although these emotions can be experienced simultaneously, they tend to derive from different factors. A threat stimulus activates anger when the individual perceives a high degree of normative violation perpetrated by external actors (MacKuen et al. 2010). Anxiety, on the other hand, is activated by the novelty or uncertainty of a situation, or a sense of lack of resources to control that situation (Huddy et al. 2005(Huddy et al. , 2007Albertson and Gadarian 2015). Enthusiasm is linked to individuals' exposure to positive and reassuring evidence of prior preferences (Brader and Marcus 2013). These emotions, in turn, initiate distinct, but associated, attitudes and behavioural outcomes.
The affective intelligence theory suggests that anger activates the disposition system (MacKuen et al. 2010;Brader and Marcus 2013). Angry individuals are more likely to rely on habitual routines and seek to remove threats through, often, aggressive and risk-taking preferences. They are less likely to look out for information and prone to resist information that undermines their existing beliefs. The moral violation aspect of anger is important as it primarily provides the motivation for action. Blameworthy actions or events that cause undesirable outcomes tend to trigger anger (Smith and Kirby 2004). These situations include events where people tend to get angry due to a perceived moral failure arising from a situation or a perceived violation of moral norms. Anger tends to curtail democratic deliberation and fuel biased reactions to political arguments (Suhay and Erisen 2018).
Fear, on the other hand, activates the surveillance system (Brader and Marcus 2013). It functions as an interruption mechanism, which heightens an individual's perception of risk and tends to be associated with heightened vigilance, risk-averse behaviour and compromise (Lazarus 1991). It, thus, increases an individual's propensity to seek out for new information about the threatening stimulus, deep and effortful processing of knowledge, and decreases reliance on long-standing political beliefs and partisanship (Marcus and MacKuen 1993). Enthusiasm is different to both anger and fear in that it has a positive valence. It is similar to anger, however, because it also activates the disposition system (Marcus et al. 2000). Enthusiasm tends to strengthen reliance to prior convictions and habitual routines.
The question of immigration is likely to trigger different emotional responses across the public (Brader et al. 2008;. Immigration is a transnational and controversial issue. It is related to multiple policy issues including the economy, cultural integration, criteria of entitlement and demographic change. Some members of the public may respond to immigration with enthusiasm. These individuals tend to view immigration as a way of enriching the national culture and bringing vital skills and human capital to the host country. Others may perceive immigration as a threat. Such conscious or preconscious appraisals of migration may elicit anger and fear. Anger will be triggered when immigration is interpreted as a normative violation, such as viewing immigrants as a threat to the existing social and moral order. This emotional state is less related to the actual dangers and risks arising from immigration. Rather, immigration is perceived as an unfair disruption to a person's desired condition . Fear, on the other hand, will be activated when immigration is perceived as something new with unpredictable consequences. Uncertainty is more likely to trigger fearful reactions, turning individuals to breaking from their habitual routines and engaging in information-seeking (Valentino et al. 2011;Erisen 2018).
One possible policy approach to address public concerns about immigrants is to develop a common cross-country strategy dealing with immigration and terrorism at the EU level. However, supranational policies in these areas also have the potential to influence immigrant volume, rights and duties of residents, crime rates, security concerns, social identities, and national and individual economic considerations (Erisen and Kentmen-Cin 2017;Luedtke 2011). We posit that emotional dispositions towards immigrants influence attitudes towards first, common EU immigration policy; second, common EU counter-terrorism strategy; and third, combinations of the two policies.
Starting with immigration policy, we argue that those who are enthusiastic about immigrants would support a common EU policy on immigration. The key EU competence in this sphere is to set the conditions for immigrants' entry and legal residence in EU member states. EU cooperation in the field of immigration is intended to provide international protection for those in needalthough often the content of EU policies has been criticised (Trauner 2016;Scipioni 2018). Common international standards aim to improve the status of immigrants, and are often perceived as encouraging more people to migrate to Europe. If enthusiasm 'increases people's preferences for variety and broaden their arrays of acceptable behavioural options' (Fredrickson et al. 2003: 368;Kahn and Isen 1993), then, those who are enthusiastic about immigrants would be more willing to receive immigration flows and to provide immigrants with better legal, political and economic standards.
Those who feel angry about immigration, on the other hand, would oppose EU cooperation on immigration. Such individuals view immigration as a threat violating existing moral values and social standards and would seek to remove this threat through risky and often confrontational strategies (MacKuen et al. 2010;Erisen 2018;Vasilopoulos et al. 2018). The EU is an external actor that, by its actions, may be seen as challenging national power and control over immigration policy. For example, angry individuals may view EU initiatives, such as the Commission's proposals on asylum-seeker relocation and resettlement 1 as having the potential to attract more people into Europe, thus further contributing to the perceived threat of immigration. By definition, anger evokes attribution thought towards external actors (Lazarus 1991, Berkowitz andHarmon-Jones 2004), and as such, people angry about immigration are likely to place blame on the EU for their immigration concerns (Smith and Kirby 2004). The EU would be seen as the external illegitimate actor that does not allow the nation-state to take drastic unilateral action to address the immigrant threat.
Angry individuals would thus seek to take aggressive action to remove the 'immigrant threat' through opposition to EU cooperation on this policy issue and support for national control of immigration. Given that such individuals tend to rely on heuristics, for them, a protective immigration policy that is designed in the habitual place of the nation-state is more likely to be directly responsive to national sociotropic needs. An aggressive-disruptive position against immigration cooperation at the EU level would enable retaining perceived control over this policy issue; or at least it would contribute to a feeling of maintaining control of this policy closer at home.
Hypothesis 1: Enthusiasm about immigration increases support for EU cooperation on immigration policy, whilst anger decreases it (1a); Anger about immigration increases support for national control of immigration policy (1b).
Our second set of hypotheses relates to the impact of emotions on attitudes towards the creation of a common EU counter-terrorism strategy. Terrorism is one of the two top issues in contemporary European politics (Eurobarometer 2018). Yet, just as on immigration, a common EU approach on terrorism, has not yet been formalised. Borders are at the heart of this discussion as they serve to define nation-states and delineate the limits of their sovereignty. International terrorism may be seen as posing a great threat to society. Random ISIS attacks in a number of EU member states (e.g., Germany, Spain, France and the UK) pose a great danger to the daily lives of EU citizens, in addition to presenting serious challenges to national and EU policymakers. One solution would be to develop a common strategy dealing with terrorism across Europe. By its nature, i.e., a strategy rather than a policy, EU cooperation on terrorism would not replace policy-making at the national level, but rather complement it. Nevertheless, public support for such a policy goal is essential, given that terrorism has the potential to overshadow every aspect of economic, cultural, social and political life.
Although both fear and anger are negatively valenced, they result in different reactions to terrorist threats Keltner 2000, 2001;Huddy et al. 2005;Erisen 2013Erisen , 2018Vasilopoulos et al. 2018). As mentioned above, threat stimuli may be perceived differently leading to different emotional reactions, which in turn promote distinct proclivities. Those individuals who are angry about immigration tend to perceive immigrants as an external threat that constitutes a normative violation to the existing social and moral order. Equally, immigrants are perceived to increase the potential of physical harm through crime and violence, which in turn leads to heightened personal threat, growing prejudice, and the feeling of insecurity. Angry individuals about immigration are also likely to perceive the EU's inability to protect its external border as a prime factor contributing to terrorist attacks in Europe. Such individuals will cope with the threat by relying on previously learned routines and by seeking to maintain control full control over the threat often with more confidence and risk-taking approaches (MacKuen et al. 2010;Suhay and Erisen 2018). We thus suggest that for angry citizens a domestic policy on terrorism is preferred for three reasons. First, it prioritises decision-making at the nation-state level, which constitutes the 'habitual' place for policy design, approval and implementation. Second, the ability of citizens to directly hold their own government accountable further promotes a sense of heightened control over the policy. Third, people angry about immigration could shift the blame to the EU for their concerns on security and terrorism, and as such, a common EU counter-terrorism strategy would be unacceptable. The EU's foreign and distant centre of authority lacks familiarity and does not allow angry individuals to maintain control (Lerner and Keltner 2000). From a similar standpoint, in the perceptions of those who feel anger when thinking of the issue of immigration, a domestic policy on terrorism is more likely to be considered as restrictive, and thus completely remove the threat by 'eliminating the enemy'.
Anxious individuals, on the other hand, are likely to perceive immigration as unpredictable and associate it with uncertainty. They tend to perceive high risks related to immigration. Unlike angry individuals, they prefer eliminating such threats through risk-averse behaviour, caution and support for less risky policies (Huddy et al. 2005(Huddy et al. , 2007. Anxious individuals perceive that they have a low degree of control and strive for more security and safety. In this quest they are more open to compromise (Brader et al. 2008;Albertson and Gadarian 2015). Therefore, we posit that, unlike angry individuals, they are likely to view international cooperation as an additional shield vis-à-vis terrorist threats, especially given that EU cooperation on terrorism does not negate domestic policy in this sphere. Such individuals tend to break from the habitual attachments (Marcus et al. 2000). In the case of terrorism, this 'break' will be translated as a departure from the nation-state being the exclusive and habitual locus of decision-making and the support for compromise within the framework of EU cooperation instead. For them, a common counter-terrorism policy will provide consistent standards across the EU that would likely reduce terrorist threats and increase national security (Huddy et al. 2005). Enthusiastic citizens on immigration may not necessarily link terrorism to immigration as, often, terrorist attacks are carried out by EU nationals. We thus do not expect that enthusiasm would have an effect on preferences for a common EU counter-terrorism strategy.
Hypothesis 2: Those who feel angry about immigration are less likely to support a common EU counter-terrorism strategy (2a) whereas those who feel anxious about immigration are more likely to support it (2b).
Finally, whereas some individuals may have clear preferences on both policies and view them as complementary, i.e., support or oppose both, we also want to consider the possibility that some individuals may support one policy but reject the other (see also Hobolt 2014;Vasilopoulou and Keith 2019). EU cooperation on immigration policy may not be seen as complementary to EU cooperation in the field of terrorism. This is because these two policies relate to and address threat differently. The intuition here is the following: whereas for some individuals EU cooperation on immigration may be perceived as a potential risk; a common EU counter-terrorism strategy might alleviate some of the problems related to extra-EU immigration. In other words, even though a common EU counter-terrorism strategy does not fully eliminate the immigrant threat, it may serve to contain it. The EU becomes an additional layer of protection vis-à-vis terrorism. In combination with an immigration policy where the nation-state is the central authority, it might be the most cautious policy option. Given that anxiety relates to risk-aversion, prevention and protection (Lerner and Keltner 2000;Huddy et al. 2005;Redlawsk et al. 2010), we posit that the 'domestic-level immigration and EU-level counter-terrorism' policy nexus would be particularly appealing to anxious individuals.
Hypothesis 3: Those who feel anxious about immigration are more likely to support a common EU counter-terrorism strategy and at the same time oppose a common EU immigration policy.

Data and measures
We used a marketing research firm, PanelClix, able to draw a representative sample of the populations of both Germany and the Netherlands. We conducted the study online in the late months of 2015 and early 2016. The German sample included 694 participants and the Dutch sample included 659 participants. An average German participant was male, 56 years of age, with 15 years of schooling and an average monthly income of €3000-3499. An average Dutch participant was male, 47 years of age, with 15 years of schooling and an average monthly income of €3000-3499. Ideologically, the average person in both samples was moderate (Germany: M = 5.11; SD = 1.74; the Netherlands: M = 5.74; SD = 1.87) on the left-right scale, ranging from 1 to 10 (see Appendix Table A1 for descriptive statistics).
The timing of our survey coincided with the beginning of the refugee crisis in 2015 and the November and January Paris attacks. These events made immigration particularly salient in party, public and media debates across Europe. Germany and the Netherlands provide the ideal environment to test our propositions regarding the effect of emotional reactions to immigration on international cooperation. These two countries host a significantly different number of immigrants and refugees giving us a powerful comparison. According to EuroStat (2018), Germany not only hosts a large body of foreign-born immigrant population (approximately 15% of the population) but also has received approximately one million Syrian refugees in recent years. In the Netherlands, approximately 11% of the population is foreignborn and the country hosts approximately 100.000 refugees. Despite variation in levels of immigration, however, both countries have experienced rise in farright party support. The leaders of both the Alternative for Germany party (AfD-Alternativ für Deutschland) and the Dutch Party for Freedom (PVV-Partij voor de Vrijheid) have voiced their rejection of immigrants, refugees, and the EU. Both parties have been unfavourable towards immigration and have criticised domestic immigration and integration policies and the EU's response to the refugee crisis. Equally important, the PEGIDA (Patriotische Europäer gegen die Islamisierungdes Abendlandes) movement has held various protests in both countries against immigration, often linking immigrants to terrorism.

Dependent variables
In line with our hypotheses, we formulate dependent variables that measure the level of support and opposition towards EU cooperation on immigration and terrorism. To capture the level of support for a common immigration policy (H1) we asked 'How much do you support the creation of a common EU immigration policy?' on a scale ranging from Strongly Oppose (1) to Strongly Support (5). 2 In addition, to understand the perceived assessment of the effect of a common EU immigration policy on the flow of the immigrants entering the EU system, we used the following item: 'Do you think that a common immigration policy will decrease (0) or increase (1) the flow of immigrants to the EU?' To detect any potential differences of authority on the subject we asked which of the two institutions (the EU or the national government) should set the conditions of entry and residence for immigrants.
To assess the level of support for a common EU counter-terrorism strategy (H2), we asked 'Do you support the creation of a common counter-terrorism strategy?' on a scale ranging from Strongly Oppose (1) to Strongly Support (5). Finally, we test the potentially conflicting nature of the two types of attitudes towards these two policies (H3). We construct a variable where 1 denotes individuals who oppose EU cooperation on immigration but support an EU counter-terrorism strategy, and 0 those who view these policies as complementary, e.g., support or oppose both processes (see also Hobolt 2014; Vasilopoulou and Keith 2019). Given that we are interested in those individuals who actually have a clear positive or negative preference on these two policies, we exclude those individuals who have neutral opinions. We also exclude those individuals who have the opposite view, i.e., support EU cooperation on immigration but not on terrorism, as the group is very small (n = 7) (also see Appendix Table A2).

Measures of affective reactions to immigration
We measure emotional reactions toward immigrants through a battery of items. Each item scored from 1 (Not al all) to 4 (Very) and asked participants how anxious, proud, angry, hopeful, worried or enthusiastic they felt about immigration. We combine the positive emotions (proud, hopeful and enthusiastic) into a single variable (α = 0.80) and anxious and worried into another variable (α = 0.85) while keeping the anger measure separate (MacKuen et al. 2010). Our measures allow us to capture the specific emotion elicited by immigration as opposed to general emotions. Overall, we find that there was moderate-to-high amount of anxiety (M = 2.56; SD = 0.92) and anger (M = 2.12; SD = 1.06) as opposed to lower degree of enthusiasm (M = 1.59; SD = 0.68) on immigration.

Control variables
Our prime aim is to assess the effects of emotions on attitudes towards EU cooperation in two fields, i.e., immigration and terrorism. To isolate the effect of emotions, our models include a number of confounders. First, we include a battery to capture the level of perceived hostility toward immigrants, which is one of the key antecedents of preferences towards immigration. The specific item asked respondents to rate immigrants on a list of adjectives (Honest-Dishonest; Trustworthy-Untrustworthy; Safe-Dangerous; Non-violent-Violent; Good-Bad; Democratic-Undemocratic) ranging from 1 to 5. We combined the responses into a single measure (α = 0.85), where higher scores indicate greater hostility. This measure captures both substantively and contextually relevant variance regarding immigration. Second, we control for trust and identity, which are important drivers of EU attitudes (Hobolt and De Vries 2016;Armingeon and Ceka 2014;McLaren 2002). We separately capture trust in the national government and trust in the EU with two items asking whether the person tends to trust (=1) or not to trust (=0) the respective institution. We include a measure of EU identity by asking how much the respondent is proud of being a EU citizen on a scale of 1 (Not at all) to 4 (Very much). This measure primarily differentiates between Eurosceptics and others. Third, we control for political ideology, age, gender, and education. We measure political ideology by two items ranging from 1 (Left/Liberal) to 10 (Right/Conservative). Education was captured by years of schooling. Finally, we included a dummy variable for country to detect any potential differences between Germany and the Netherlands.

Models
In line with our hypotheses and measures as described above, we estimate the probability of change (from one category to the other) through ordered logit in the models on support for EU cooperation on immigration and terrorism (H1 and H2). We employ a logit model in order to predict preferences for the combined policy measure (H3). Table 1 reports our findings related to our first set of hypotheses. In line with our expectations (H1a), we find that anger felt as a result of immigration significantly decreases the probability to support a common EU immigration policy. Enthusiasm about immigration, on the other hand, promotes a collective policy response to immigration. Importantly, we also find that both anger and enthusiasm increase the expectations that a common EU immigration policy will increase the flow of immigrants coming to Europe (two middle columns in Table 1). This finding supports our argument that although both citizen groups agree about the potential consequences of EU cooperation, their different emotional reactions to immigration result in different preferences for where the authority for this policy should lie. The last two columns of Table 1 show that anger towards immigrants significantly increases support for the national government to set the conditions of entry and residence for immigrants. As hypothesised (H1b), anger about immigration motivates people to prefer their own government as opposed to the EU as the central decision-making authority.

Common EU immigration policy
Taken together, the effect of emotions across an array of related domains on immigration policies suggest that anger about immigration promotes refusal of collective policies driven by the EU. In contrast, enthusiasm encourages support for collective decision-making at the EU level. When focused on the consistent and robust effects, we see that anger and enthusiasm differentiate between the two cohorts that respectively oppose and support a common EU immigration policy. We find that anxiety is not a key driver of such attitudes.  1288 Robust standard errors in parentheses; + p < .10, *p < .05, **p < .01, ***p < .001.
Our models include a number of variables controlling for the main confounders of the relationship between affect and EU immigration policy support, including trust in the EU, EU pride, hostility towards immigrants and key socio-demographics. Appendix Figure A1 reports the marginal effects of all these variables. EU pride is positively associated with a sustained preference in favour of EU authority in policy-making. Hostility towards immigrants dampens the level of support (Hainmueller and Hopkins 2014) whereas trust in the national government marginally increases support for a common EU immigration policy. Older male individuals are more likely to support a common EU immigration policy. Finally, we find a major difference between the Germans and the Dutch. Germans are significantly more likely to support the common EU immigration policy compared to the Dutch, although this effect disappears when controlling for additional independent variables.

Common EU counter-terrorism strategy
Our second set of hypotheses related to the effect of emotions about immigration on support for a common EU counter-terrorism strategy. We focused on the different effects of anger and fear suggesting that whereas those who feel angry about immigration are less likely to support a common EU counter-terrorism strategy, anxious individuals are more likely to support it. The first two columns in Table 2 demonstrate that there is a major difference between anger and anxiety on the issue of a common EU counter-terrorism strategy. In line with our hypotheses H2a and H2b, those who felt angry about immigration are significantly more likely to oppose the policy in contrast to those who felt anxious. We argue that this is because of a major difference in threat perceptions (MacKuen et al. 2010;Erisen 2018;Vasilopoulous et al. 2018). Angry individuals prefer the nationstate level as the habitual place of decision-making that would allow them to take immediate action against immigration. Domestic decisions would correct the normative violations perceived to be posed by immigration. Angry citizens may also blame the EU of failing to provide security against the threats of immigration and terrorism. Anxious individuals, on the other hand, perceive immigration as something new with unpredictable consequences. For them, international cooperation constitutes an additional protection vis-à-vis terrorism. As presented in Figure 1, anger decreases the propensity to support a common EU counter-terrorism strategy whereas anxiety increases it, while holding other variables at their mean values.
A few substantively related findings require further elaboration (marginal effects are presented in Appendix Figure A2). First, trust in national institutions is an important predictor of one's position on forming a common strategy dealing with the threats of terrorism. Second, we see that older people are more likely to support this policy in contrast to females who oppose it.
Finally, unlike the previous model on a common EU immigration policy, there are no country differences.
Preferences for EU cooperation on terrorism, but not on immigration We assess the potentially conflicting nature of preferences regarding EU cooperation on immigration and terrorism. Specifically, we are interested in the individuals who oppose a common EU immigration policy, but support the creation of an EU counter-terrorism strategy. Our third hypothesis suggests that anxiety is the key driver of such a policy combination. This is because anxious individuals tackle the immigration threat in a distinct manner, i.e., they prefer the most risk-averse option. We argue that the 'domestic-level immigration and EU-level counter-terrorism' policy nexus is the most cautious policy option.
The logit models reported in the third and fourth columns of Table 2 confirm our expectations. 3 Anxious individuals support a common EU counter-terrorism strategy but oppose a common immigration policy. This effect is maintained Table 2. Support for the common counter terrorism strategy and combined policies. Robust standard errors in parentheses; + p < .10, *p < .05, **p < .01, ***p < .001.
(even at the marginal level of significance) when controlling for all substantively relevant indicators. Interestingly, whereas anger has no effect in the full model, enthusiastic individuals are less likely to support this policy combination. Figure  2 plots the predicted probabilities of the emotion effects. As we have seen from our previous models, angry individuals have very congruent preferences. They oppose EU cooperation on the dual questions of immigration and terrorism. Enthusiastic citizens support EU cooperation on immigration, but not in the sphere of terrorism. This is possibly because they do not associate immigration with terrorism. As expected, hostility draws significant negative weight when examining the results on the combined policy measure. Those who are hostile towards immigration are more likely to oppose this policy combination (see appendix Figure A3). These individuals tend to express reluctance towards EU cooperation. Even with the robust effect of hostility in the models, however, the direct effects of emotions still hold.

Discussion
In this article we have put forward and empirically tested an affective model for understanding citizen preferences towards EU cooperation in the spheres of immigration and counter-terrorism. Based on an original survey carried out in Germany and the Netherlands and controlling for major confounders, we have shown that citizens' emotional reactions to immigration are integral to their policy assessments. First, whereas enthusiasm about immigration increases individuals' likelihood of supporting a common EU-wide immigration policy, anger increases the likelihood of opposition to the same policy. Anger about immigration increases reliance on domestic rather than supranational decision-making authorities on this specific policy question. Second, anger and anxiety promote distinct proclivities regarding support for a common EU counter-terrorism strategy. Anger is associated with opposition to such cooperation whereas anxiety results in support. For angry individuals who seek to fully eliminate the immigrant threat, state unilateral action is a confident and habitual response. Anxious citizens, on the other hand, have a heighted perception of risk, which is more likely to be alleviated through cooperation. Besides, a cautious approach of EU cooperation on terrorism does not rule out domestic control. Third, anxiety increases citizens' desire to support a policy combination where immigration is resolved exclusively at the nation-state level but terrorism is also addressed at the EU level. We argue that this is the most risk-averse policy preference. Our findings have implications for future research on European integration. First, they demonstrate the emotional underpinnings of EU preferences. Research has shown, for example, that Brexit was a response to economic and cultural insecurities stemming from immigration (Hobolt 2016). By providing evidence that anger towards immigration creates strong preferences for domestic policy, we highlight the need to explore the role of psychological processes that motivate voters (see Vasilopoulou and Wagner 2017).
Second, previous studies have suggested that anxious individuals would be more likely to feel depressed, lonely, and helpless. This is because they tend to overestimate risks and perceive threats as uncontainable (Huddy et al. 2005). However, our findings indicate that anxious individuals can also prefer cooperation, supporting international solutions in the form of a common EU counter-terrorism strategy. These findings provide new insights for those studying foreign terrorist attacks, such as Charlie Hebdo or 9/11. Third, our findings have implications for the Europeanization of other policy areas related to terrorism, such as crime prevention and control. With the Lisbon Treaty entering into force, the EU can influence national policies and projects regarding crime prevention. However, the question about whether the public is supportive of the EU intervening in crime matters remains open. We demonstrate that emotions can engender a substantial impact on how EU citizens may react to EU's new roles.
We should underscore that we looked at the emotionality of immigrationnot only as an attitude but also as an affective state. Our design has allowed us to test our argument on two countries, Germany and the Netherlands, where the question of immigration is highly politicised by far-right parties (AfD and PVV). Future research should shed light on the relationship between emotions and attitudes towards EU cooperation in the spheres of immigration and terrorism in countries where far right parties do not operate. Overall, research that considers the role of emotions on policy preferences opens up new avenues for understanding the scope for international cooperation. Notes 1. http://europa.eu/rapid/press-release_IP-16-829_en.htm, accessed on 1 February 2019. 2. The EU's common immigration policy concerns non-EU migrants. However, due to the EU's freedom of movement principle, immigration also occurs within the EU. Given that the correlation between attitudes towards EU freedom of movement and attitudes towards immigration from outside the EU is very low (Vasilopoulou and Talving 2019), we can reasonably expect that in most cases survey respondents can differentiate between the two policies. 3. We maintain our findings when those who selected the neutral policy position are included in our estimation. We also conducted the same models for each group separately (Appendix Table A3).