When Catholics turn right: the effects of the Islamic terrorism threat on the fragmented Catholic Italian voters

Transcription

1 Journal of Ethnic and Migration Studies ISSN: X (Print) (Online) Journal homepage: When Catholics turn right: the effects of the Islamic terrorism threat on the fragmented Catholic Italian voters Matteo Vergani & Enrico Maria Tacchi To cite this article: Matteo Vergani & Enrico Maria Tacchi (2015): When Catholics turn right: the effects of the Islamic terrorism threat on the fragmented Catholic Italian voters, Journal of Ethnic and Migration Studies, DOI: / X To link to this article: Published online: 27 Nov Submit your article to this journal View related articles View Crossmark data Full Terms & Conditions of access and use can be found at Download by: [Monash University Library] Date: 27 November 2015, At: 14:47

2 JOURNAL OF ETHNIC AND MIGRATION STUDIES, When Catholics turn right: the effects of the Islamic terrorism threat on the fragmented Catholic Italian voters Matteo Vergani a,b and Enrico Maria Tacchi c a Alfred Deakin Institute for Citizenship and Globalisation, Deakin University, Burwood, VIC, Australia; b School of Social Sciences, Department of Politics, Monash University, Melbourne, VIC, Australia; c Dipartimento di Sociologia, Università Cattolica del Sacro Cuore, Milano, Italy ABSTRACT Italy, as well as most European countries, has been hit by a wave of anxiety arising from groups such as ISIS and Al Qaeda, whose effects on political attitudes are still under-examined. This article investigates the effect of the perceived threat of Islamic terrorism as a potential driver for a right turn in the Catholic Italian electorate with open-ended interviews and an Internet-based experiment in which voters were randomly assigned to a terrorism threat manipulation and to a control condition (N = 138). The results show that the Islamic terroristic threat significantly increased the support for centre-right leaders who promoted ingroup identity and out-group hostility towards Muslims. Implications for the debate about the effects of perceived threat on political opinions and the relevance of the findings beyond the Italian case are discussed at the end of the article. ARTICLE HISTORY Received 27 April 2015 Revised 2 September 2015 Accepted 21 October 2015 KEYWORDS Threat; terrorism; Italy Europe has been recently hit by a wave of anxiety arising from the Islamic terrorism threat posed by groups such as ISIS and Al-Qaeda, and by the extensive media coverage of events such as the Charlie Hebdo incident. As demonstrated by empirical research in the US, not only the actual attacks but also media narratives reinforce the idea that a deadly event is likely to happen in the near future and increase the perceptions of threat through the highlighting of threat-messages, terror alerts and assessments by government officials and opinion leaders (Nacos, Yaeli, and Shapiro 2007). This is also likely to be the case of Europe, where media emphasise threatening information and increase the perception of a great risk of a future terrorist attack. The threat of terrorism particularly Islamic terrorism has often been associated in Europe with Islam and migration discourses, especially in the narratives of right-wing populist parties that often merge anti-system, anti-elitist and anti-immigration motives (De Witte and Klandermans 2000; Bleich 2009; Schmuck and Matthes 2014). Yet sometimes anti-immigration narratives also win over moderate politicians and voters, who then back restrictive immigration policies on immigrants: an example is the so-called Bossi Fini bill, which was approved in Italy in July 2002 by a centre-right government supported by MPs affiliated with Catholic moderate parties and movements (Colombo, CONTACT Matteo Vergani 2015 Taylor & Francis matteo.vergani@monash.edu

3 2 M. VERGANI AND E. M. TACCHI Sciortino, and Craveri 2002; Zincone 2006). Why do moderate voters turn right? What processes can explain the increase in support for right-wing leaders and policies in a Christian electorate who should identify with an ecumenical worldview of making strangers into friends (Hanley 1994, 8)? This article investigates these issues in Italy, where the Catholic electorate has been fragmented since the end of the so-called First Republic, when the Christian Democracy (Democrazia Cristiana) dissolved after a corruption scandal that hit the whole political system. This article tests the explanatory power of the notion of perceived threat within the context of Social Identity Theory. This theory provides an effective framework to explain the interaction between media discourses and political attitudes through the concept of activation (Tajfel and Turner 1979). 1 We believe that media discourses about Islamic terrorism can generate in the Catholic electorate perceived threat, which can provoke a shift in political attitudes towards in-group bias and out-group avoidance and aggression. As identities are fluid constructs, we also argue that Islamic terrorism threats can activate different in-group identities and different forms of in-group versus out-group thinking. In this article, we investigate whether the threat of Islamic terrorism primes religious, national or other multiple identities among Italian Catholic voters. Previous research explored the impact of the perceived Islamic terrorism threat on political attitudes: recent examples are Bail s work on the impact of fringe anti-muslim organisations on public opinion (Bail 2014), Gadarian s work on the effect of terrorism evocative imagery on candidate evaluations (Gadarian 2014b) and Iyer s work on the effects of terrorism on emotional and political responses (Iyer, Hornsey, et al. 2014; Iyer, Webster, et al. 2014). This article expands this literature by investigating the impact of Islamic terrorism threat in the Italian context, where political identities are distinct from the US and other EU countries because of many factors 2 among which the peculiar role and history of Catholic religion in shaping political attitudes. This article provides an original contribution to the literature for several reasons. First, it adds to the debate about what factors motivate opposition to immigration in Europe by focusing on the effects of the Islamic terrorism threat, which is an under-examined factor of anti-immigration attitudes in the current debate on group threat (Hjerm 2007; Pettigrew, Wagner, and Christ 2010; DeWaard 2014; Helbling 2014; Van Assche et al. 2014). This article also contributes to the understanding of the relations between European parties and the religious other : if the use of anti-immigration motives is more common in right-wing European parties (Hagelund 2003; Van Heerden et al. 2014), the complex relations between Catholic moderate electorate, Christian democracies and Muslims in Europe are under-examined. Second, while most of the current empirical research on the issue is focused on US public opinion, this article focuses on a European country, Italy, which is often beyond the discipline s radar. Last but not least, this work contributes to expanding the literature on Christian Democratic parties, which have been an underresearched political phenomena that have, however, attracted renewed interest in recent years (Bale and Szczerbiak 2008; van Kersbergen 2008; Grzymala-Busse 2013; Warner 2013). This article contributes to these debates providing original empirical evidence that high perceived threat can increase the support for centre-right leaders who promote in-group identity and out-group aggression. Before presenting the results of the field research, the next section will introduce and discuss the literature on threat and will contextualise the research in the Italian historical and political context.

4 JOURNAL OF ETHNIC AND MIGRATION STUDIES 3 Threat, in-group identity and out-group aggression Threat can be defined as any risk of future harm either material or symbolic (Valentino et al. 2008, 157). The impact of threat on political attitudes and opinions has been investigated by several empirical studies. Scholars argued that perceived threat especially terroristic threat can increase support for a specific type of political worldview that is aggressive and punitive in nature (using an American terminology, a hawkish political worldview). For example, threat has been found to increase ethnocentrism and xenophobia, intolerance and willingness to forego civil liberties, as well as the support for taking action in response to the threat, including military action and restriction of civil liberties such as increased surveillance and immigration restrictions (Huddy et al. 2002, 2005). Threat can also increase trust in government and leadership and it can be associated with rally effects (Sinclair and LoCicero 2010). Merolla and Zechmeister conducted three experiments in the US between 2004 and 2008, and found that the leadership qualities become more consequential for voters after exposure to a terrorist threat (Merolla and Zechmeister 2009). Perceptions of threat can also make people cling to their own worldviews, which can be ideologies or religious beliefs (Abrams 2011). However, previous research also found that not all the individuals primed with the same threatening message react in the same way. For example, Gadarian found that American Democrats reacted to terrorism news in a politicised context by delegitimising the source, because they considered the communication as manipulative and threatening their freedom of act (Gadarian 2014a). These findings are all consistent with Social Identity Theory, which posits that group membership and group identity become more relevant during intergroup conflict because of the increased awareness of shared grievances and threats (Tajfel and Turner 1979). Yet the effects of threat on political attitudes should not be intended as a manipulation causing voters to completely switch allegiances: as Gadarian pointed out the power of terrorism threat, especially in a context where terrorism is not imminent, may be more limited by who gives the cue and who receives it (Gadarian 2014b, 297). It is known that some factors can moderate responses to threat, such as emotions (anxiety, anger or fear predict different reactions to threat) and political orientation: for example, conservatives tend to focus on aggression against an out-group when threatened (Huddy et al. 2005; De Zavala, Cislak, and Wesolowska 2010). Skitka et al. (2006) found that anger, associated with appraisals of certainty and personal control and lower risk assessment, predicted support for violent counter-terrorism policies whereas fear, which is associated with appraisals of uncertainty and higher risk assessments, predicted the support for the deportation of groups perceived to be associated with the terrorists. Similarly, Iyer and colleagues found that different types of images of terrorist attacks provoked different emotions, which predicted different reactions to the threat. More specifically, images that depicted the victims aroused sympathy, whereas images of terrorists aroused both fear (if the public appraised the terrorists as dangerous) and anger (if the public focused on the attack as an illegitimate and unjust action). These different emotional states predicted different reactions to the threat: sympathy predicted support for policies to help the victims, fear predicted support for the government s efforts to negotiate with terrorists, anger predicted support for violent counter-terrorism policies (Iyer, Webster, et al. 2014).

5 4 M. VERGANI AND E. M. TACCHI We use Social Identity Theory as the theoretical framework of this article because it explains the interaction between media discourses and political attitudes through the concept of activation. We expect Islamic terrorism threat to activate in-group identity, making classify others in groups and thinking in terms of in-group versus out-groups. We explore which identity is activated by such threat in the sample of Italian Catholic voters, and we expect the activation of religious identity: the Islamic terroristic threat could make salient the Muslim out-group to the Italian Catholic voters, increasing the in-group sense of belonging and the hostility towards Muslims. Historically religious in-group versus out-group thinking has been endemic (Azrout, Van Spanje, and De Vreese 2013, 6), especially in Europe where empirical research has found that perceived religious threat is a major predictor of opposition to Turkey being accepted into the EU. Right-wing political opinion leaders in Europe have been exploiting these identity cleavages by characterising Muslim civilisation as evil and as threatening Europe with terrorism, thus contributing to the creation of a context of Islamophobia and prejudice (Field 2007; Meer and Modood 2009). However, we also acknowledge that this activation is not to be intended in a determinist manner: the reactions to the Charlie Hebdo attacks showed, for example, that the entire French society mobilised against terrorism, not just the religious groups. This would suggest a second possibility, which would be the activation of national identity (i.e. Italian nativity and citizenship) following the Islamic terrorism threat. This would increase the sense of belonging with the national community and hostility towards foreigners irrespective of the religion of belonging. Another possibility would be the simultaneous activation of both identities. To sum up, in this article we want to test if (and to what extent) Italian Catholic voters react to the Islamic terrorism threat by changing their political attitudes. Building on previous research (Huddy et al. 2005; Gadarian 2010, 2014b; Bail 2014), we expect Catholic voters to increase support for hawkish worldviews following threat, especially when expressed by a centre-right leader because as Gadarian pointed out terrorism threat enhances candidate evaluations even more when the policy position matches the party s usual position (Gadarian 2014b, 296). Moreover, we expect threat to activate in-group identity, and we explore which identity is activated in the sample of Catholic Italian voters. The next section contextualises the research question and the hypotheses in the Italian context. This discussion will be the basis for generating the measures used in the experiment and the interpretation of the results in the Italian context. The nature and identity of the Italian Catholic vote To understand the relevance of religion and the role of the Catholic Church and the Vatican in explaining the Italian Catholic vote, it is necessary to look at the history of Catholic political participation in Italy. During the Restoration, numerous Catholic voices opposed the idea of an Italian state under the Vatican. Pius IX did not want Catholics to participate in Italian politics, issuing in 1868 the non expedit bull (which sanctioned the motto neither elector nor elected ). Mistrust between the Church and the Italian state became even greater in 1870, when Italian troops entered the Papal States to make Rome, a year later, the new capital of Italy (Pollard 2008). This annexation opened the so-called Roman Question (i.e. the issue of the Church s status in Italy), which led to a Church state controversy

6 JOURNAL OF ETHNIC AND MIGRATION STUDIES 5 that brought about difficult relations between the emerging Catholic political participation, the Vatican and the Italian state (Warner 2013). The non expedit ruling was gradually transgressed and various strands of Catholicism developed in Italian politics. After the first two attempts to constitute a Catholic political organisation in Italy (the Opera dei congressi in 1874 and the Partito Popolare in 1919), De Gasperi s Christian Democracy became a mass party with a dual structure made by political and religious opinion leaders (Allum 1997). Religious belonging and religious commitment, successfully combined with widespread anxiety about the totalitarian threat of the Communist Party coming to power, gave the Christian Democracy the relative majority across the whole Cold War period, from 1948 to 1992, despite a clear declining trend (Figure 1). 3 Extreme diversity of political views has always been the defining characteristic of the Christian Democracy in Italy, combined with a diverse structuration of the party in different areas of the country (Allum 1997; Baccetti 2007). Sartori s famous definition of the Italian system as a polarised pluralism with an anti-system party at each end of the political spectrum and the Christian Democracy in the centre perfectly describes the position of the party in the political system (Sartori 1976). Christian Democracy attracted the vote of the absolute majority of the Catholic Italian voters, despite their ideological diversity (Bellucci and Segatti 2010). Despite the clear and stable position at the centre of the Italian political system, Christian Democracy was not a homogeneous party: it included numerous different strands and political attitudes within the party itself, which resonated with a diverse electorate (Marradi 1982; Donovan 1989). In 1992, the Mani Pulite investigation uncovered corrupt practices across the entire Italian political spectrum. After years of mounting scandals, Christian Democracy was disbanded in 1994 and Catholic political elites and electorates were spread across the political spectrum (Newell and Bull 1997; Zincone 2006). All major political parties, from left and right, attracted a significant proportion of Catholics within their electorate: between 1996 and 2001 Forza Italia, the main centre-right party, had about 30% of practising Catholics in the electorate, Margherita about 12% and Democratici di Sinistra (DS) about 20% (Donovan 2003, 108). Electorate, members and MPs grouped around the various political identities that already existed in Christian Democracy: the leftist strand gave birth to the Popular Party, which later become Margherita and then merged with former Communist Party in 2007 giving birth to the centre-left Democratic Party (PD). Others, more aligned with centre-right ideological stances, participated in Berlusconi s Forza Italia. The former Christian Democracy members who gave birth to centrist parties (which have dissolved Figure 1. Percentage of vote obtained by Christian democracy in national elections.

7 6 M. VERGANI AND E. M. TACCHI and reassembled with several denominations, the most long-lasting of which has been the Union of the Centre, UDC) have generally been allied with Berlusconi s centre-right coalition (Gundle and Parker 1996). Similarly, the Catholic electorate did not concentrate the vote on a single party, yet it remained fragmented across the political spectrum: for example, surveys conducted in 2011 showed that among the voters who attributed high importance to religion, about 30% voted PD, 20% PDL (Popolo della Libertà, founded by Berlusconi in 2007), 10% Lega Nord (a right-wing populist and especially in the early 1990s separatist party with a strong anti-immigration focus), 7% UDC (Cataldi 2012). Since 2008 the UDC has decided to remain in opposition to Berlusconi s party and participated in Monti s government between 2011 and In 2014 UDC formed a unified parliamentary group with New Centre Right party (NCD), composed of MPs who defected from Berlusconi s party and participated in the government guided by PD leader Matteo Renzi. Is it possible to establish a core set of values that characterise the Catholic political positions in Italy despite the great diversity of experiences and political paths described so far? Scholars investigated the distinct ideological and moral stances that distinguish Christian Democracies from other parties (Van Kersbergen 1994; Bale and Szczerbiak 2008). An example is the solidarism or social personalism, which corresponds to attempts to eliminate tensions between different social classes and to synthetise collectivist and liberal views of society: not surprisingly the Italian Christian Democracy has often been described as an all-class party (interclassista) (Allum 1997). In addition, social capitalism has been described as a core element of the socio-economic policy of Christian Democracy: yet as already pointed out not all the strands of the party agreed on this view, as demonstrated by the choice of many former members to join centre-right parties and coalitions with liberal views. More distinctive values are surely religious identification and religiosity, Christian ethics and the view of the traditional family as the key means to achieve societal equilibrium (Diamanti 2007). These values are shared by all political parties that compose the fragmented Catholic political spectrum in Italy, and by their electorates. As for Christian ethics, the defence of life from conception to natural end is probably the most important value, as John Paul II outlined in his Evangelium Vitae. 4 This value stands for an inflexible opposition to the death penalty, abortion, embryo eugenics and euthanasia. Other core principles are the support for traditional familial structures regulated by marriage and composed of male and female and the freedom of private property, education and religion, which was symbolised by the motto of the Italian Christian Democracy: Libertas, which means freedom. The vast majority of Catholics in Italy support the idea of private schools and education, consequently, this has been a distinctive policy of Catholic political parties. The core Catholic values in the Italian political sphere were used to create the dependent variables that measured the in-group values and identity. Data A convenience sample of 138 Italian Catholic voters was collected between October and December The sample was selected on the basis of identification with Christian Catholic religion, and it was collected through advertising via ing lists and Facebook pages of Catholic organisations and political parties with a large Catholic base. 5 We also

8 JOURNAL OF ETHNIC AND MIGRATION STUDIES 7 asked respondents to forward the questionnaire link to each Catholic person they knew above 18 years of age, regardless of the political affiliation. This method allowed recruiting a sample of Catholic voters from the entire political spectrum. As we used a convenience sample, we collected baseline measures to ensure that the sample was demographically and politically diverse to include the diversity of worldviews that can be found in the Catholic electorate. We made sure that the sample was composed of self-identified Catholic believers, and we included in the sample only respondents who declared to be Catholic. Moreover, we included a measure of religiosity asking how important is religion in your daily life (1 was not at all and 10 very much ) and we controlled for this measure in the data analysis: the mean score was 8.2 (standard deviation 2.3), the median 9 and the mode 10. We collected age and gender, and we found that the sample was not very different from the profile of voters of centrist parties in Italy as emerging from ISPOS 2009 and 2014 survey. 6 The age range of our sample was from 19 to 86 years old (mean 54.5, standard deviation 14.1), 81.2% were men. In order to ensure that the sample was distributed across the political spectrum from left to right, we collected baseline measures of party identification and political orientation. As for the party identification 36.8% of the sample wrote none, 14% Partito Democratico, 14% Nuovo Centro Destra, 8.1% Unione Di Centro, 3.7% Forza Italia, 3.7% Lega Nord and the rest were fragmented in smaller percentages with other parties. The mean political orientation of the sample (where 1 is very progressive and 10 is very conservative ) was 4.9, the median 5 and the mode 5. We controlled for this measure of political orientation in the data analyses. The general picture is of a centrist moderate sample without a strong party identification and fragmented into many different electoral preferences. As 36.8% of respondents did not identify with any political party, we wanted to explore whether this group was skewed towards one end of the political spectrum, and we found that the group that declared no identification had a mean political orientation of 4.39 (standard deviation 1.85), as opposed to the 5.24 (standard deviation 2.49) of the rest of the sample. The difference is less than 1 point in a 10 points scale, but independent samples t-test yielded a significant p-value (p =.04). This tendency towards conservativism of the respondents who declared no affiliation could reflect the conservative respondents feelings of being a minority, which would be consistent with the presence of a popular Catholic left-wing Prime Minister in Italy at the time of the data collection: Matteo Renzi. 7 As for the political opinions, we collected attitudes about the political system (two-party system versus multi-party system), about the immigrants and about the need of a religious Catholic party in the Italian political system. On average, the sample slightly agreed with the idea that a two-party system obliterates the democratic richness of political pluralism (mean 6.0, SD 3.0 where 1 is completely disagree and 10 completely agree ). The sample slightly disagreed with the idea that there is more popularism in centre-left than in centreright (mean 4.6, SD 2.7) and slightly agreed with the idea that a centrist political pole can save Italy from the damages of a two-party system (mean 5.6, SD 3.4). Clearer opinions emerged about immigration and multiculturalism: on average, the sample agreed that the coexistence of different cultures is a positive value of modern society (mean 7.4, SD 2.5), and that migrants are not just a problem but also a benefit to Italian society (mean 6.6, SD 2.7). This is consistent with what Bale and Szczerbiak called the emphasis on transnational, as well as domestic, reconciliation and with the universalistic claims of the Roman Catholic Church of mutual understanding and reciprocity

9 8 M. VERGANI AND E. M. TACCHI (Bale and Szczerbiak 2008, 482). Finally, on average the sample agreed that it is important that a Catholic party governs Italy (mean 6.8, SD 3.0) and felt only slightly more represented by Catholic candidates (mean 5.5, SD 3.2). The general picture is of a sample with very diverse political opinions about the party system, institutional asset and the idea of society and the role of religion in the political life. This is consistent with the diverse political worldviews and backgrounds of Catholic political participation, as discussed in the previous section. We also collected information about the political participation of the sample. 63.0% had been a member of a political party in his/her life, but only 28.3% were still members at the time of the research. 47.1% had been a candidate in the past, but only 7.2% were political representatives in a local institution, 18.8% had a role in a party and 19.6% defined him/ herself as a political activist. This picture describes a sample of politically active individuals, many of whom were engaged in political participation in the past. Consistent with the general decline in political engagement, the participants reported being less active than previously. Yet, the sample was still actively engaged in voting: 93.4% voted at the last elections and the same percentage reported willingness to vote at the next political elections. The participants reported high levels of political attentiveness and necessity to evaluate : 73.9% identified as having more or much more opinions about politics as compared to friends and acquaintances, 55.8% identified as discussing more about politics and 65.2% as looking more for information about politics than others. Yet, only 39.1% identified as participating more in political activities than others. We collected those measures because, as we recruited our sample through Catholic organisations and political parties, we predicted that the sample would have been more active than the average Italian population. Moreover, we suspected a self-selection of respondents motivated to fill this questionnaire without any additional incentive as being more engaged with politics than the general population. We reported these data for completeness of information but we do not foresee any interaction between the manipulations used in this experiment and the measures of political participation. We decided to conduct an experiment following previous research that tried to establish clear causation between terrorism threat and political attitudes (Gadarian 2014b; Iyer, Hornsey, et al. 2014; Iyer, Webster, et al. 2014). The experimental design aims at assessing if and to what extent, when Islamic terrorism threat is experimentally heightened, participants: (1) increase in-group membership and out-group hostility (towards Muslims) and (2) follow centre-left, centre-right or centrist party leaders who advocate Catholic identity, who call for out-group aggression and who promote core values of the Catholic political tradition. The experimental design is a 2 (threat versus non-threat) 3 (centre-left versus centre versus centre-right leader). Results of the experiment We randomly assigned 48.6% of the sample to read a newspaper article about the ISIS threat on the Vatican City, and 51.4% to a control manipulation (an article about the Independence referendum in Scotland). After presenting the two articles, we asked respondents to explain with two open-ended questions what would be the best answer of an Italian political party to the problem described in the article and if the issue represents a political challenge for Italy and Europe. We asked those two questions for two

10 JOURNAL OF ETHNIC AND MIGRATION STUDIES 9 reasons: first, to make sure that they read carefully the newspaper article, and second, to investigate the identities activated by the threat with the open-ended answers to the article about ISIS. We used as manipulation check an adapted version of Huddy Huddy et al. s (2005) measures of perceived threat. As Table 1 shows, the manipulation significantly increased all the measures of threat in the sample. This manipulation check assessed the effectiveness of the manipulation, which significantly increased the perceived threat of the people who read the newspaper article about the Islamic terrorism threat, as opposed to the control group. No interaction effects of the manipulation appear with factors such as political orientation, religiosity, gender or age (not shown). After the manipulation check, the sample was randomly assigned to one of the three following texts: Text 1: Think if in the next 5 years a new and young charismatic leader emerges, who belongs to a CENTRE-RIGHT party and who wants to re-unify the Catholic moderate electorate in Italy with a party affiliated with the European People s Party and competitor to the Party of European Socialists. Text 2: Think if in the next 5 years a new and young charismatic leader emerges, who belongs to a CENTRIST party and who wants to re-unify the Catholic moderate electorate in Italy with a party affiliated with the European People s Party and competitor to the Party of European Socialists. Text 3: Think if in the next 5 years a new and young charismatic leader emerges, who belongs to a CENTRE-LEFT party and who wants to re-unify the Catholic moderate electorate in Italy with a party affiliated with the Party of European Socialists and competitor to the European People s Party. Thereafter, participants were asked to rate their agreement with several statements made by the above-mentioned leader during an electoral campaign. The leader s statements were the dependent variables of the experiments: three statements promoted core values of the Christian political identity (preservation of human life, freedom of education and defence of traditional marriage), two promoted Christian identity (in-group identity) and one attacked Muslims (out-group aggression). Figure 2 reports the statements and the mean level of support found in the whole sample. The figure shows that on average there is a high support for the core values of the Christian political identity, Table 1. The manipulation check. The higher the mean score, the higher is the agreement with the statement. Threat Mean (SE) Non-threat Mean (SE) F-value p-value This article raises a problem that could affect Italy soon 7.70 (.23) 5.14 (.35) This article raises a problem that could affect myself, 6.07 (.30) 3.84 (.37) my friends or my family This article makes me feel anxious 4.03 (.32) 2.61 (.23) fearful 3.52 (.30) 2.38 (.22) worried 6.11 (.34) 3.46 (.27)

11 10 M. VERGANI AND E. M. TACCHI Figure 2. The dependent variables. The higher the score, the higher is the support for the statement (scale from 1 to 10). Identity statements are in black, core values in white. while the agreement with other generic identity statements valuing the in-group and attacking the Islamic out-group is lower. One-way analysis of variance (ANOVA) revealed no significant difference between the high-threat and low-threat groups on the support for the dependent variables (p values >.1). Yet, generalised linear model ANOVA (ANCOVA) revealed a significant interaction of the threat condition with the centre-right leader frame for the identity statements. The interaction remained statistically significant for each identity statement even after controlling for gender, age, religiosity and political orientation (Table 2). Under low threat the participants tended to support more centre-left leader, but under high-threat the participants tended to support more a centrist and centre-right leader promoting in-group identity and out-group aggression (Figures 3 5). As for the dependent variables about the core Christian values, no effects of the threat manipulation (and no interaction effects with the leader s frame or any other potential moderators) emerged on the support for the statements promoting core Christian values. In other words, irrespective of the leader s ideological frame and of the political orientation and demographic characteristics, the participants showed a high support for these statements. Table 2. F-values and p-values of the generalised linear model ANOVA (ANCOVA). In Italy there is no space for mosques F-value (p-value) Christians in politics should unite F-value (p-value) There is the need for a new Christian F-value (p-value) Gender 0.73 (.40) 0.04 (.85) 0.10 (.75) Age 0.30 (.59) 0.88 (.35) 1.93 (.17) Political (.00) 6.42 (.01) 6.35 (.01) orientation Religiosity 1.11 (.29) 4.25 (.04) 4.21 (.04) Threat 1.81 (.18) 2.67 (.11) 2.90 (.09) Leader frame 0.98 (.38) 0.71 (.49) 0.34 (.71) Threat * leader frame 4.82 (.01) 5.86 (.00) 3.04 (.05)

12 JOURNAL OF ETHNIC AND MIGRATION STUDIES 11 Figure 3. Support for the statement In Italy there is no space for Mosques. Higher values in the y-line mean higher support for the statement. The open-ended interviews Half of the participants (only the group in the high threat experimental manipulation, N = 67) provided their opinions about the best answer of an Italian political party to the terrorist threat and if terrorism is an important political challenge for Italy and Europe. The answers provided insights on what group identity the Islamic terrorism threat activated in the sample. As already discussed, the sample was politically diverse, therefore, the answers were very different: they went from the promotion of dovish worldviews (promoting cooperative foreign policies, international and domestic aid, cultural dialogue) to hawkish worldviews (promoting aggressive and punitive measures, defence spending, armed intervention). Most of the people who promoted more hawkish solutions called for increasing funds for law enforcement agencies and secret services, and called for more security and control of the territory. Most of the people who promoted more dovish worldviews called for prevention, integration, resolution of social and economic causes of terrorism and coordination with other states at a European level. We conducted thematic analysis on the open-ended answers to identify the identities activated by the Islamic terrorism threat. Numerically, religious identity was the most Figure 4. Support for the statement Christians in politics should unite not argue against each other. Higher values in the y-line mean higher support for the statement.

13 12 M. VERGANI AND E. M. TACCHI Figure 5. Support for the statement There is the need for a new Christian Democracy to lead Italy in the future. Higher values in the y-line mean higher support for the statement. prevalent. Religious was indicated by references to the in-group (Christians, the Vatican) and the out-group (Islam and Muslims). None explicitly mentioned Italy or Italian culture, but many wrote about strengthening State functions such as security (i.e. police, information services) and foreign policy (i.e. diplomacy, coordination with other countries). Although these could be interpreted as indirect indicators of national identity, we argue that the most important identity emerging from this analysis was the religious one because, as the next paragraphs show, many respondents explicitly referred to it. Among the people who suggested hawkish measures as a response to the terrorist threat, it is interesting to notice that some expressed concern about Muslims, associating the terrorist threat to the whole Muslim community. For example, some respondents wrote: To increase the level of alert with Muslims and friends To increase control over Islamic organizations Similarly, it is interesting to notice that some of the more dovish respondents, although expressing a very different political approach, reacted to the terrorist threat by activating a religious identity. This means that even some of the people who preferred a dovish solution to terrorism thought in terms of religious belonging when primed with the Islamic terrorist threat. For example, one respondent wrote: To demand the institutions to engage in dialogue with the representatives of other religions Interestingly, a respondent reacted to the Islamic terrorist threat refusing the association between religion and terrorism, but sill citing it as if they recognised that this association is a widespread phenomenon. Unfortunately people confuse the word terrorism with religion. It is necessary to teach people (since childhood) that Muslims are not terrorists This type of reaction is similar to what Gadarian identified as reactance, which is the backlash effect of presenting information about terrorism to Democrats who recognised it an attempt to manipulate them and to push them towards more hawkish worldviews (Gadarian 2014a).

14 JOURNAL OF ETHNIC AND MIGRATION STUDIES 13 Many expressed their concerns about immigration, particularly the uncontrolled inflow of immigrants. Some respondents directly associated the terrorist threat with the immigration phenomenon. For example, respondents wrote: Constant surveillance and a thorough control of the people who enter in our country A commitment to security and illegal immigration Favouring a controlled migration of people who aim at integration, punishing violations Immigrants are not directly associable with a uniform religious other : for example, Italy experienced a large Christian immigration from areas like Eastern Europe (e.g. Romania and Ukraine), Latin America and Asia (e.g. the Philippines). 8 Yet, the public perception of immigrants in Italy (especially the clandestine immigrants) is commonly linked to the Muslim other because of the numerous arrivals of refugees from Northern Africa and the Middle East on Italian shores and because of media narratives about immigration, which also are the most relevant driver of discrimination as perceived by the immigrants themselves (IDOS 2014). Moreover, the narrative associating migration with terrorism risks is recurrent in the Italian media. 9 Therefore, it is not surprising that many respondents suggested the same association. Last but not least, the Islamic terrorist threat activated a strong in-group identity, both with generic statements and with more specific ones that explicitly recalled the Catholic religion. For example, respondents wrote: To acknowledge and defend the Christian identity of the Italian history and culture To protect security and safeguard the Vatican State One respondent recalled generic Western culture as the identity that is threatened by Islamic terrorism, and others called for safeguarding generic roots. The threat comes from Islamic extremism, therefore the issue is important for the survival of Western culture Yes we have to safeguard our roots We do not have to degenerate our identity These answers again confirm that the Islamic terrorism threat activated religious identity in many of the respondents, in terms of in-group identification and out-group hostility. Discussion This article provides some evidence, the first in Italy, about the effects of perceived threat (and more specifically the Islamic terrorist threat) on the political attitudes of Catholic voters. The results show that the Islamic terrorist threat activated religious identity in the Catholic voters who participated to the experiment. This activation, following Social Identity Theory, increased in-group affiliation and out-group aggression for most of the respondents, as revealed by the open-ended answers. Being religious identity salient, ingroup identity mostly emerged as support for the Christian values and defence of the Vatican threatened by terrorism, and out-group aggression emerged as a generic anti- Muslim and anti-immigration sentiment. These results are consistent with the literature that found threat to increase ethnocentrism, xenophobia and out-group hostility (Huddy et al. 2002).

15 14 M. VERGANI AND E. M. TACCHI Yet, threat was not found to directly change the support for any of the dependent variables in the experiment. The effect of the experimental manipulation reached statistical significance only when the source of the statement was a centre-right Catholic leader, and to a lesser extent a centrist leader. This means that perceived threat (from Islamic terrorism) is a factor that can potentially play a role for the reunification of the Catholic electorate, especially when the political leader who advocates in-group identity and out-group aggression comes from a centre-right party. Under high-threat conditions, the Catholic participants in this experiment increased their support for a centre and centre-right leader rather than a centre-left leader. 10 The issue about Catholics turning right is of special relevance because of the recent unprecedent levels of electoral volatility of the Italian electorate (Verney and Bosco 2013). The results of this study could also be explained by what previous research (Pettigrew and Meerteens 1995; Zanier 2002; Zincone 2006) have identified as a subtle and blatant prejudice against immigrant among centre-left voters, which may be activated under conditions of threat. The effect of perceived threat was present even after controlling for political orientation, gender, age and religiosity. This suggests that the effect was consistent across partisan and ideological lines: in other words, under high-threat conditions not only those who already identified with a right-wing and centre-right party increased their support for a centreright leader, but also those who identified with centrist and centre-left parties. This may be explained by the fact that right-wing leaders are generally recognised as more committed to the defence of traditional values and beliefs and with hostility towards threatening others, while left-wing leaders are recognised as more likely to adopt narratives of multiculturalism and integration. Moreover, the centrist leader in the current political Italian context may be perceived as less effective than the centre-right leader because the most important parties since 1994 have been either centre-left or centre-right. In short under high-threat conditions, the participants, consistent with the existing literature, preferred a right-wing leadership, perceived by the respondents as stronger, more effective, more defensive of in-group identity and more hostile towards the threatening out-group (Huddy et al. 2002; Merolla and Zechmeister 2009; Gadarian 2010). This finding appears to be consistent with the studies conducted by Gadarian in the US. Gadarian found that individuals exposed to terrorism news stories were more likely to adopt hawkish foreign policy views, especially when the news stories were emotionally charged with fear-inducing clues (Gadarian 2010). Also, Gadarian found that graphic news images of terrorist attacks increased approval for Bush s policies across partisans lines and for a leader s foreign policies, but only among the same party members (Gadarian 2014b). Interestingly the manipulation had a significant effect on out-group aggression and generic statements about in-group identity (such as Christians should unite and there is the need for a new Christian Democracy ). Yet threat had no effect on the defence of core values such as Christian ethics, traditional family structure and freedom of education. Participants in all conditions scored very high on the support for those values, suggesting that they perceive the values as part of their identity irrespective of the political orientation or party affiliation of the leader promoting them, and also irrespective of the perceived threat from an out-group. Those values appear to be a non-negotiable and inflexible identity marker for the participants. Yet, it is also important to notice that the emphasis on universalistic reconciliation and the worldview of making strangers into friends (which was identified as a core value of Christian Democracy by already cited scholars such as

16 JOURNAL OF ETHNIC AND MIGRATION STUDIES 15 Hanley, Bale and Szczerbiak) is more susceptible to change after the threat manipulation. Although we did not use as dependent variable any consistent scale of openness or attitudes towards immigrants, the increased out-group aggression towards Muslims that emerged from both the experiment and the open-ended interviews suggests this pattern. As much comparative research suggests, the nature of European Christian parties is highly diverse, even in countries where Christian Democracies enjoyed success similar to the Italian case (at least until the early 1990s) such as Belgium, the Netherlands and Germany (van Kersbergen 2008; Grzymala-Busse 2013). Therefore, it is difficult to predict if the trends suggested in this article may be replicated in other EU countries. Yet, this article is consistent with previous research findings that showed how, in time of crisis of identity and salience of the Islamic threat, Christian Democracies turned right: an example is the case of the Christian Democratic response to the rise of Pim Fortuyn s xenophobic campaign in The Netherlands, which consisted of a formulation of a critique of multiculturalism that was meant to profit from xenophobic anxiety in the Christian moderate electorate (van Kersbergen 2008, 270). This article adds another piece of evidence to the argument that associations between Islam and violence have underpinned many of the concerns about the place of Muslims in Europe over the past decades (Bleich 2009, 362). As scholars successfully pointed out, this violence-centric perspective makes Muslims more likely to be marginalised and also that future terrorist acts will increase repression and discrimination against Muslims, as in the reactive co-radicalisation paradigm (Pratt 2013), and as already happened after 9/11 in European countries (Sander 2006). As mentioned at the beginning of the article, one of the important geopolitical factors that united the Catholic vote under the banner of the Christian democracy during the Cold War was the Communist threat. An appropriate question would be: to what extent is the Islamic terrorist threat comparable to the Communist threat for the effects that they can have on the political attitudes of the Italian Christian electorate? The Communist threat pushed Christian Democracy to adopt a narrative based on a scelta di civiltà (civilisation choice), between totalitarianism and democracy (Galli 1972; Allum 1997). Although the narrative of the clash of civilisation between the West and Islam is somewhat similar to the Cold War one at a global level, there are a number important differences such as the different nature (and magnitude) of the players and the conflict, the different geopolitical context, the different cultural, social and political issues that define the alleged cleavages between the two so-called civilisations and so on (Huntington 1993; Inglehart and Norris 2003). As Minkenberg pointed out, there is nothing deterministic in the relations between religions and democracy, and other factors (such as patriarchal orders and geopolitical factors) explain secularisation and democratisation of societies better than religion do (Minkenberg 2007). This article contributes to this debate providing evidence about how the Islamic threat can be a contextual factor that is able to drive Italian Catholic political attitudes (but only to a smaller extent the Catholic values) towards more hawkish and right-wing worldviews. This result is consistent with studies on anti-immigration attitudes and perceived threat in EU, which suggested fear of conflict over values and culture to be a major predictor of cross-national differences in anti-immigration attitudes (Silke 2008). Yet, the reactance and backlash effects of the Islamic threat need further attention, especially in the centre-left Catholic electorate. Another extension that could be useful for investigating some general mechanisms that might apply to other electorates would be to