Classical and Modern Prejudice toward Asylum Seekers: The Mediating Role of Intergroup Anxiety in a Sample of Italians

: (1) Background: Increasing waves of immigration have not only changed the demographic features of European societies but have also had important implications for intergroup relationships. Based on the theoretical model of intergroup threat theory, this study examined whether and to what extent intergroup anxiety mediates the relationship between antecedents and attitudes toward asylum seekers. (2) Method: In a sample of 470 Italians (mean age 30.21), using a survey-based study, we assessed the antecedent of intergroup anxiety (national identiﬁcation) and its consequences (classical and modern prejudice). (3) Results: The results supported most of the expected predictions. While intergroup anxiety was positively associated with prejudice toward asylum seekers, national identity was positively related to intergroup anxiety and prejudice. Moreover, the relationships between antecedents and consequences were mediated by intergroup anxiety. (4) Conclusion: This study has identiﬁed antecedents that can potentially increase or decrease intergroup anxiety and proposes certain strategies for improving social inclusion policies and relationships between asylum seekers and settled communities. The implications for future studies are discussed.


Introduction
Over the last decade and on a global scale, the rise in the incidence of forced migration has resulted in an increase in the number of asylum seekers in European countries. Asylum seekers are individuals whose request for international protection is yet to be processed, and statistics show that, every year, around one million people seek asylum (UNHCR 2019). Refugees, on the other hand, are those who have been forced to flee their countries by reason of war, violence, or persecution. While not every asylum seeker is recognised as a refugee, every refugee is initially an asylum seeker. Since the beginning of the civil war in Syria, the number of refugees has rapidly increased. Accord to the latest report published by the United Nations High Commissioner for Refugees (UNHCR 2019), over 5.5 million people have fled Syria, and around six million have been internally displaced, with almost half of them living in besieged areas. Among the Syrians who have fled their country, more than 5 million live in neighbouring countries in the Middle East region (e.g., Lebanon, Jordan, Iraq, and Egypt). However, since these refugees experience very poor living conditions in these countries, they often flee further to Europe to seek safety, protection, and stability to become asylum seekers (Koc and Anderson 2018). While around 1 million displaced Syrians have moved to Europe as asylum seekers or refugees since the conflict in Syria began, aid agencies, however, have reported incidents of brutality and violence toward them. Late last year, a report by the Danish Refugee Council (2018) explained that vulnerable asylum seekers who arrived in Italy found difficulty accessing accommodation, risked destitution, and underwent substandard reception conditions. As in the past, the Eastern Mediterranean and the central Mediterranean routes-mainly located near Southern Italy owing to its geographic position in the Mediterranean Sea-have represented the main routes of access for irregular foreigners to the European Union, accounting for over 40% of the total number of entries. Southern Italy, therefore, is one of the chief destinations for this international migration. Although each country that accepts asylum seekers tries to offer social support for the newcomers, asylum seekers continue to face widespread rejection as demonstrated by the number of hate crimes committed against them, often with the support of nationalist movements and parties (Kotzur et al. 2019).
While the presence of immigrants brings increased social diversity that can bring opportunities for economic growth and cultural enrichment, this presence is also often perceived as a threat by both majority and minority groups. The perception of threat is then used to justify the application of restrictive territorial policies oriented at closing European borders. These policies divert attention away from refugees who may be in real need of protection from violence in their homeland. In recent years, the number of refugees arriving on the borders of the European Union have fallen, but some Italian newspapers continue to focus their attention on the immigration issue. This societal debate on immigration contributes to the impact of what many Italians see as uncontrolled and illegal immigration, offering discussions on population flows and suggestions on how the crisis could be resolved (Bottura and Mancini 2016).
In the psychosocial literature, there is evidence that feelings of threat are associated with prejudice against minority groups (Caricati et al. 2017;Stephan and Stephan 2017;Visintin et al. 2017). Thus, the present research used the intergroup threat theory (Stephan and Stephan 2000) as a theoretical framework to investigate classical and modern prejudice toward asylum seekers within a sample of Italians. We examined the antecedent factors of intergroup threat and its consequences, which in previous studies had been shown to predict negative outgroup attitudes either directly or indirectly through intergroup anxiety (Anderson 2018;Brylka et al. 2015;Koc and Anderson 2018;Montgomery and Zhang 2018). Intergroup anxiety, which involves the anticipation of negative outcomes from intergroup interaction, can be considered as a subtype of threat centring on apprehensions about interacting with outgroup members (Stephan and Stephan 2000). Therefore, understanding the social psychological factors that Italians show toward asylum seekers can be important to promote inclusiveness and acceptance and prevent social exclusion risks.

Prejudice toward Asylum Seekers
Prejudice is defined by Allport (1954) as a feeling of "antipathy based upon a faulty and inflexible generalization" (p. 9). For majority groups, prejudice intensifies the relationship between perceptions of threats and ideological beliefs toward minority groups (Mancini et al. 2018;McConahay et al. 1981). Policies on immigration and the reception of immigrants and asylum seekers are controversial issues since they generate unease and uncertainty given that these policies often consider the immigrant from an economic and not necessarily humanitarian and human perspectives (Louis et al. 2013).
These negative attitudes toward minority groups are partly influenced by the process of social categorization (Brown 2013;Tajfel and Turner 2001). When a person is aware of belonging to a social category, the person tends to consider his group differently from the external group, both with respect to the emotional and human dimension. Perceiving differences between social groups can have effects on the relationship between attitudes and behaviour (Hartley and Pedersen 2015;Kaiser and Wilkins 2010;Pedersen and Walker 1997). Additionally, in intergroup relationships, the belonging of a person to a group defines the significant aspects of the self, playing a crucial role in the way a person perceives the self.
Results from previous studies, mainly conducted in Australia, showed evidence of negative and prejudicial attitudes toward asylum seekers (Anderson 2018;Pedersen et al. 2005;Pedersen and Thomas 2013). The main concerns associated with negative attitudes are about terrorism and fear of Muslims (Pedersen et al. 2012). Moreover, several other factors have also been connected to prejudice toward asylum seekers. For example, asylum seekers' values are not aligned with the social values of the majority group (Esses et al. 2008). Other factors related to prejudice toward asylum seekers include demographic variables such as low levels of education, being male, political conservatism, and old age (Anderson 2018) as well as high levels of national identification (Nickerson and Louis 2008).
Recently, Anderson (2018) proposed a new scale to measure the prejudice against asylum seekers (PAAS) which includes two factors: classical prejudice and conditional prejudice as a form of modern prejudice. Classical prejudice is related to the blatant or explicit form of prejudice. According to Pettigrew and Meertens (1995), blatant prejudice is described as "hot, close, and direct" and they suggest that it consists of two components: (1) perceived threat and rejection of the outgroup and (2) avoidance of intimacy, especially sex and marriage. By contrast, conditional prejudice is a "form of modern prejudice defined as negative attitudes that are socially acceptable to express because they are endorsed concurrently with a qualifying statement" (Anderson 2018, p. 695).
In Italy, few studies have investigated prejudice toward asylum seekers. The results of a recent study carried out by Mancini et al. (2018) indicated that the perception of asylum seekers as a threat for the ingroup resources or values strongly and directly affected prejudice against this group. Based on the dual process model of ideology (Duckitt and Sibley 2010), the authors found that higher scores of realistic and symbolic ingroup threats were associated with increased prejudice. Finally, the results of a review of the last five years of Italian researches on asylum seekers and refugees indicated that only 12 studies investigated this topic (Tessitore and Margherita 2017). Thus, this result highlights the importance of intensifying new investigations in this field by researchers with the aim of understanding and improving the quality of intergroup relationships.

Intergroup Anxiety
According to Stephan (2014), "intergroup anxiety is a type of anxiety that people experience when anticipating or engaging in intergroup interaction" (p. 240). Ingroup members may feel intergroup anxiety either because they could be rejected or exploited by outgroup members, or because of a history of antagonism between the two groups. Therefore, anxiety plays a crucial role in intergroup relations, showing a detrimental effect. Recent advances in the socio-psychological literature suggest that fear and anxiety about "the other" make dominant group members avoid intergroup contact, hence perpetuating informal group segregation, and inhibiting the benefits of intergroup contact (Paolini et al. 2016, for a review). Thus, an increase in intergroup anxiety can lead people to express prejudice toward outgroup members. Intergroup anxiety has been shown in previous studies to affect negative outgroup attitudes (Brylka et al. 2015;Montgomery and Zhang 2018;Voci and Hewstone 2003) as well as unfavourable prosocial support intentions (Hutchison et al. 2018). While results from prior studies, also conducted in the Italian context, have confirmed the meditating role of intergroup anxiety (Brown and Hewstone 2005;Pettigrew and Tropp 2008;Voci and Hewstone 2003), this issue needs further investigation, in particular when different type of predictors as well as different intergroup perceptions are considered (Visintin et al. 2017).
As suggested by Stephan (2014), intergroup anxiety is an individual-level threat that can include both realistic and symbolic threats to the self. Intergroup anxiety increases the desire to avoid contact with outgroup members (Paolini et al. 2018, for a review) and toward asylum seekers (Murray and Marx 2013;Turoy-Smith et al. 2013). In their meta-analysis of 95 studies, Riek, Mania, and Gaertner (Riek et al. 2006) found that intergroup anxiety showed the strongest relationship with outgroup attitudes. Another study found that intergroup anxiety was a stronger predictor of social distance (study 1). Participants' projected feelings toward how they might feel while interacting with a refugee predicted their likelihood of avoiding them (Koc and Anderson 2018). By contrast, another study conducted in Australia showed that lower levels of intergroup anxiety toward indigenous Australians and refugees were significantly associated with lower levels of prejudice (Turoy-Smith et al. 2013). The results of the studies just mentioned have confirmed the key role of intergroup anxiety in the expression or inhibition of prejudicial attitudes toward outgroup members. Given the high level of prevalence of prejudice toward asylum seekers and the likely continuation of the global refugee crisis, the antecedents of such prejudice are worth investigating (Pedersen and Thomas 2013).

National Identification
Asylum seekers are often portrayed in the global and national media as a threat to national security and to the nation's economy (e.g., O'Doherty and Augoustinos 2008). Fear of terrorism and economic pressure are also common themes in the Italian media and newspapers. Hence, the perception of certain groups of people-in this case, asylum seekers-as a threat to the nation becomes a powerful basis for experiencing anxiety about these social groups. In addition, the recent "European refugee crisis" which also affected Italy, may be relevant in predicting attitudes toward asylum seekers.
According to Stephan and Stephan (2000) there is a positive association between ingroup identification and perceived threat. Ingroup identification serves as a basis for categorizing other people into ingroups and outgroups (Stephan 2014). People who feel strongly identified with their ingroup are more likely to be more concerned with protecting ingroup interests and preserving social norms, values, and customs than are low identifiers (Hutchison et al. 2018). Previous studies have found that ingroup national identity is positively correlated with intergroup anxiety because ingroup members are worried about losing their distinctiveness through interaction with outgroup members (González et al. 2010;Paolini et al. 2018). Hence, people who belong to a social group are motivated to perceive the ingroup as positively distinct from outgroups because of the increased value that the important social group gives to their self-esteem. According to Social Identity Theory (Tajfel and Turner 2001) the members of a group (e.g., a nation) compare themselves with members of an outgroup (e.g., immigrants) to get a favourable view of themselves. To protect their positive distinctiveness, members of the ingroup may try to limit the outgroup's resources via denigration and hostility and may perceive the outgroup members as a social menace (Steinmayr 2016).
As suggested by Sides and Citrin (2007), increasing numbers of asylum seekers from the Middle East may have unleashed perceived cultural threats in potentially receptive European countries. Asylum seekers are perceived as a threat not only to the ingroup's symbolic resources (e.g., national identity) in terms of positive distinctiveness, but also to the economic resources. In a similar way, Jackson, Brown, Brown, and Marks (Jackson et al. 2001) suggest that a positive relationship between national identification and negative attitudes toward immigrants might be expected because people who identify with the nation have a greater concern for the national interest. This means that people perceive immigration as a bad condition for the national interest. Previous studies have also shown that the Australian media and federal government have collectively framed asylum seekers as the "other", whose entry into the country poses a threat to national identity (Anderson and Ferguson 2018;Cowling et al. 2019). By emphasizing differences rather than commonalities, people may be more willing to endorse the exclusionary policies that safeguard the values, norms, and beliefs of many Australians (Mckay et al. 2012). In a recent study, Anderson (2018) found that national identification was strongly associated with conditional prejudice, r = 0.42, p < 0.001, rather than classical prejudice, r = 0.19, p < 0.05, against asylum seekers. A recent literature review suggested that being high in national identification is related to harbouring negative attitudes toward asylum seekers (Anderson and Ferguson 2018). Similar results emerged in another study, where high levels of national identity were more strongly related to negative attitudes than personality-based attitudes, such as self-esteem, toward asylum seekers (Pedersen et al. 2005). Other researches have also shown that national identification predicted the following relations: blatant and subtle forms of prejudice toward foreigners in European countries (Mummendey et al. 2001;Pettigrew and Meertens 1995); negative attitudes toward immigrants in both Canada and Australia through threat perceptions and dehumanizing beliefs (Louis et al. 2013); and more negative or less positive behavioural intentions toward Syrian refugees in Turkey, especially when the feeling of being under threat was high (Yitmen and Verkuyten 2018). In other words, prejudice toward the outgroup tends to manifest itself when the members of the ingroup perceive that their national identification could be symbolically threatened by the presence of values and cultural and religious habits other than theirs.

The Current Study
In recent years, some European governments, including Italy, for politically motivated reasons, have implemented deterrence-based policies that in some cases have been shown to not only cause human suffering, but also fail to identify vulnerable asylum seekers, including unaccompanied minors. Additionally, the mass-media (e.g., newspapers, TV programs) have focused on government immigration policies, drawing attention to this social issue and thereby generating a sense of solidarity in the Italian population, but also often contributing to the feeling that the country was being invaded and was under threat (Mancini et al. 2018).
The present study used intergroup threat theory (Stephan and Stephan 2000) as the theoretical framework in investigating the negative attitudes toward asylum seekers in Italy. Although prior studies have tested the predictors of intergroup threat theory in relation to ethnic and cultural groups (Stephan 2014;Visintin et al. 2017), this study aimed to examine the relationship between classical and modern prejudice toward asylum seekers and intergroup anxiety and national identification in Italy. Thus, the current study explored the relationships between intergroup anxiety and its antecedent and consequences toward asylums seekers in a sample of Italians.
Intergroup anxiety (Stephan 2014) provides an appropriate conceptual model to investigate Italian public attitudes toward asylum seekers. Previous studies found that asylum seekers are perceived and represented as posing a threat in terms of job opportunities as well as an existential threat to social values, customs, and national identity (Cowling et al. 2019;Koc and Anderson 2018). Thus, the present research tested a mediational model in which intergroup anxiety is conceptualized as a mediator of the predicted relationships between antecedent variables (national identification) and attitudes toward asylum seekers as outcomes (classical and modern prejudice).
Based on the results of the above-mentioned studies, it was predicted that national identification would be positively associated with prejudice toward asylum seekers through the mediating role (full model) of intergroup anxiety (see Figure 1a). To ascertain the mediating effects of intergroup anxiety, we tested an alternative model (partial model), in which we also considered a direct link between national identification and prejudice toward asylum seekers (see Figure 1b). Finally, we hypothesized that the model could be invariant between students on one hand, and young adults and adults on the other hand. It was assumed that prejudice against minority groups, such as asylum seekers, could be influenced by the cultural dimension (e.g., national identification) and situational factors in terms of intergroup anxiety, which involve apprehension about interacting with outgroup members.

The Current Study
In recent years, some European governments, including Italy, for politically motivated reasons, have implemented deterrence-based policies that in some cases have been shown to not only cause human suffering, but also fail to identify vulnerable asylum seekers, including unaccompanied minors. Additionally, the mass-media (e.g., newspapers, TV programs) have focused on government immigration policies, drawing attention to this social issue and thereby generating a sense of solidarity in the Italian population, but also often contributing to the feeling that the country was being invaded and was under threat (Mancini et al. 2018).
The present study used intergroup threat theory (Stephan and Stephan 2000) as the theoretical framework in investigating the negative attitudes toward asylum seekers in Italy. Although prior studies have tested the predictors of intergroup threat theory in relation to ethnic and cultural groups (Stephan 2014;Visintin et al. 2017), this study aimed to examine the relationship between classical and modern prejudice toward asylum seekers and intergroup anxiety and national identification in Italy. Thus, the current study explored the relationships between intergroup anxiety and its antecedent and consequences toward asylums seekers in a sample of Italians.
Intergroup anxiety (Stephan 2014) provides an appropriate conceptual model to investigate Italian public attitudes toward asylum seekers. Previous studies found that asylum seekers are perceived and represented as posing a threat in terms of job opportunities as well as an existential threat to social values, customs, and national identity (Cowling et al. 2019;Koc and Anderson 2018).
Thus, the present research tested a mediational model in which intergroup anxiety is conceptualized as a mediator of the predicted relationships between antecedent variables (national identification) and attitudes toward asylum seekers as outcomes (classical and modern prejudice).
Based on the results of the above-mentioned studies, it was predicted that national identification would be positively associated with prejudice toward asylum seekers through the mediating role (full model) of intergroup anxiety (see Figure 1a). To ascertain the mediating effects of intergroup anxiety, we tested an alternative model (partial model), in which we also considered a direct link between national identification and prejudice toward asylum seekers (see Figure 1b). Finally, we hypothesized that the model could be invariant between students on one hand, and young adults and adults on the other hand. It was assumed that prejudice against minority groups, such as asylum seekers, could be influenced by the cultural dimension (e.g., national identification) and situational factors in terms of intergroup anxiety, which involve apprehension about interacting with outgroup members.
Note. PAAS = prejudice against asylum seekers; CL-P = ssical prejucladice; MO-P = modern prejudice. Moreover, a higher score in national identification was expected to predict more intergroup anxiety and in turn classical and modern prejudice toward asylum seekers. Moreover, a higher score in national identification was expected to predict more intergroup anxiety and in turn classical and modern prejudice toward asylum seekers.

Participants and Procedures
A convenience sample of 470 Italian people (227 males and 243 females) answered a questionnaire that was introduced and distributed by a researcher. The mean age of the sample was 30.21 (SD = 12.56, age range 18-67 years). The student participants (n = 303, age range 18-45 years, M age = 23.58, SD = 4.48) were recruited at our university campus during break periods and psychology classes. The rest of the sample were young adults and adults (n = 167, age range 18-67 years, M age = 42.24, SD = 13.55) randomly recruited in shopping centres and public places. Data were collected in the Italian language from March to May 2019.
After all the participants received an informative introduction explaining the aim of the study, they were invited to complete the questionnaire. Those who consented were given a questionnaire, which took around 20 minutes to complete. The researcher emphasized the voluntary nature of participation in the study, so that the participants could stop completing the questionnaire at any time. In addition, the participants were guaranteed complete anonymity and were assured that the data would be used only for research purposes. Participation in the study was voluntary and the participants did not receive any payment or additional university credits. We collected data following the ethical rules of the Italian Psychological Association (AIP).

Intergroup Anxiety
Intergroup anxiety was measured using the Italian version of the scale (Turner et al. 2007) by adopting the procedure suggested by Stephan and Stephan (1985). It was framed as a hypothetical situation in Italy: participants were asked to imagine finding themselves alone with an unknown asylum seeker. Students rated their mood such as cautious, diffident, embarrassed, and relaxed (reversed item) in that situation. Each adjective was rated on a five-point Likert scale (1 = not at all; 5 = very much). The internal reliability of the scale was acceptable, α = 0.73.

Prejudice against Asylum Seeker
Prejudice against asylum seeker was measured using an Italian translated version of the "prejudice against asylum seekers" scale (PAAS; Anderson 2018). First, the author translated the scale from English to Italian using the forward and back translation method (Behling and Law 2000). Then, it was translated again into English by a bilingual researcher who did not know the original version of the scale ( Van de Vijver and Hambleton 1996). The comparison between the two versions of the scale pre-and post-translation allowed the verification of their semantic and conceptual equivalence. The comprehensibility and applicability of the scale was verified through a preliminary pilot study. The PAAS includes 16-items which measure two forms of prejudice: classical prejudice (8 items, e.g., "Asylum seekers are not our problem", α = 0.82) and conditional prejudice as a form of modern prejudice (8 items, e.g., "Asylum seeker might struggle to integrate into our system", α = 0.77). Participants' responses were rated on a seven-point Likert-type scale (1 = strongly disagree to 7 = strongly agree).
In this study, however, we adopted the general expression "modern prejudice" as including the idea of "conditional prejudice" in contradistinction from the term classical prejudice along the lines that intergroup relationships studies posit the existence of two distinct forms of prejudice: classical and modern.
A parallel analysis with oblimin rotation confirmed that the bi-factorial structure of the PAAS scale was adequate for the Italian sample. The Kaiser-Meyer-Olkin measure of sampling adequacy was 0.91, indicating that the current data were suitable for the analysis. Similarly, Bartlett's test of sphericity was significant, χ 2 (91) = 2769.77, p < 0.001, indicating enough correlation between the variables to proceed with the analysis. However, two items (CL-P4 and MO-P5) showed low loadings and were deleted. The results of the confirmatory factor analysis revealed that the model with 14 items was satisfactory,

Demographic Profile
All participants completed a socio-demographic questionnaire collecting information about their gender, age, and political orientation (from 1 = conservatism to 7 = progressivism).

Data Analysis
Descriptive statistics were computed by using the software SPSS version 25.0. Specifically, we computed means, standard deviations, and tests to verify the normality of the data. The internal reliability was obtained by computing the alpha of Cronbach (α). Subsequently, a structural equation model (SEM) and a multigroup path analysis were estimated using the software Mplus version 7.04. In order to obtain bivariate correlations, we estimated a model with all the observed variables of the study as well as all the possible covariances among the latent variables: national identification, intergroup anxiety, and prejudice. Gender, age, and political orientation were also controlled. Next, we separately tested a full and a partial mediating model to verify whether the proposed mediator, intergroup anxiety, fully explained the relationship between national identification and prejudice (classical and modern) toward asylum seekers. The final model was subjected to an analysis of invariance between students vs. young adults and adults, to assess whether the relationships between the two groups were different or not.
The models were estimated with the maximum-likelihood parameter with standard errors and a mean-adjusted chi-square test statistic that were robust to non-normality (MLM). The MLM chi-square test statistic is also referred to as the Satorra-Bentler (S-B) chi-square. Following Hu and Bentler (1999), multiple indices were used to evaluate model fit (adopted cut-offs in brackets): the chi-square (χ 2 ) test value with the associated p value (p > 0.05), comparative fit index (CFI ≥ 0.95), Tucker-Lewis Index (TLI ≥ 0.95), root-mean-squared error of approximation (RMSEA ≤ 0.06), and its 90% confidence interval, and standardized root mean square residual (SRMR < 0.08). In the process of analysing the invariance between the two groups (students vs. young adults and adults), the comparison between the nested models was applied (the less restricted model vs. the more restricted model). The model constraint was evaluated using a chi-square (χ 2 ) difference test, which provides a statistical test of whether the constraints that produce the nested models are justified (Kline 2016). To establish significant differences between models, at least two out of these three criteria had to be satisfied: ∆χ 2 significant at p < 0.05, ∆CFI < −0.005, and ∆RMSEA > 0.010 (Chen 2007).

Descriptive Analyses
Descriptive statistics and bivariate correlations for each variable are shown in Table 1. Almost all the participants (65%) were students. They mostly professed to have a progressive political orientation (56.7%). Only 38% of the participants knew one or more asylum seeker. Note. CL-P = classical prejudice; MO-P = modern prejudice; a 1 = male, 2 = female. ** p < 0.001.
The means and standard deviations ranged from 2.50 to 30.21 and 0.67 to 12.56, respectively. The skewness and kurtosis values were acceptable below |1.35|, respectively (Kline 2016). Multivariate normality was assessed using the Mardia (1970) measure of multivariate kurtosis. The Mardia's coefficient for the data in this study was, 32.38 (p < 0.001) indicating that the data violated the multivariate normality, and then all the models were tested using robust methods. Table 1 shows the results of the bivariate correlations, estimated through an SEM model, which included all the covariance among the latent variables. The model showed an acceptable fit to the data, χ 2 S-B (409) = 583.50, p < 0.001, CFI = 0.936, TLI = 0.927, RMSEA = 0.03, 90% CI [0.03, 0.04], SRMR = 0.05. As expected, national identification, intergroup anxiety, and prejudice toward asylum seekers were positively correlated with one another. Finally, we found no association between gender and age with intergroup anxiety. However, for the next analyses, gender, age, and political orientation were included as control variables.

Mediation Analysis
The full mediated model showed an adequate fit to the data, χ 2 S-B (385) = 568.17, p < 0.001, CFI = 0.935, TLI = 0.927, RMSEA = 0.03, 90% CI [0.03, 0.04], SRMR = 0.06. Subsequently, a partial mediated model was tested by including a direct relationship between national identification and prejudice (classical and modern). After comparing the two models, since we found that they were similar, ∆χ 2 (1) = 3.03, p = 0.082, ∆CFI = 0.000, the partial mediating model was selected as the final one. As shown in Figure 2, national identification is positively associated with intergroup anxiety, and with the prejudice (classical and modern) against asylum seekers.
Soc. Sci. 2020, 9, x FOR PEER REVIEW 8 of 14 Multivariate normality was assessed using the Mardia (1970) measure of multivariate kurtosis. The Mardia's coefficient for the data in this study was, 32.38 (p < 0.001) indicating that the data violated the multivariate normality, and then all the models were tested using robust methods. Table 1 shows the results of the bivariate correlations, estimated through an SEM model, which included all the covariance among the latent variables. The model showed an acceptable fit to the data, χ 2 S-B (409) = 583.50, p < 0.001, CFI = 0.936, TLI = 0.927, RMSEA = 0.03, 90% CI [0.03, 0.04], SRMR = 0.05. As expected, national identification, intergroup anxiety, and prejudice toward asylum seekers were positively correlated with one another. Finally, we found no association between gender and age with intergroup anxiety. However, for the next analyses, gender, age, and political orientation were included as control variables. Note. CL-P = classical prejudice; MO-P = modern prejudice; a 1 = male, 2 = female. ** p < 0.001.

Mediation Analysis
The full mediated model showed an adequate fit to the data, χ 2 S-B (385) = 568.17, p < 0.001, CFI = 0.935, TLI = 0.927, RMSEA = 0.03, 90% CI [0.03, 0.04], SRMR = 0.06. Subsequently, a partial mediated model was tested by including a direct relationship between national identification and prejudice (classical and modern). After comparing the two models, since we found that they were similar, Δχ 2 (1) = 3.03, p = 0.082, ΔCFI = 0.000, the partial mediating model was selected as the final one. As shown in Figure 2, national identification is positively associated with intergroup anxiety, and with the prejudice (classical and modern) against asylum seekers.
Note. PAAS = prejudice against asylum seekers; CL-P = classical prejudice; MO-P = modern prejudice. ** p < 0.01; * p < 0.05. In the present study, intergroup anxiety showed a partial mediating role in the relationship between national identification and prejudice against asylum seekers, β = 0.12, p < 0.001. In the present study, intergroup anxiety showed a partial mediating role in the relationship between national identification and prejudice against asylum seekers, β = 0.12, p < 0.001.
Next, we computed an analysis of invariance on the partial mediating model comparing the group of students vs. young adults and adults. Due to the higher numbers of observed variables, we calculated the factorial scores for each variable: national identification, intergroup anxiety, and classical and modern prejudice. The partial model has been tested by applying a multigroup path analysis. The less restrictive model fit well to the data, χ 2 S-B (3) = 3.33, p = 0.34, CFI = 0.999, TLI = 0.992, RMSEA = 0.02, 90% CI [0.00, 0.10], SRMR = 0.02. On the other hand, the more restrictive model also showed a good fit, χ 2 S-B (12) = 10.73, p = 0.55, CFI = 1.00, TLI = 1.00, RMSEA = 0.00, 90% CI [0.00, 0.06], SRMR = 0.03. The results of the comparison of the less restrictive model versus the more restrictive model showed that both did not differ significantly, ∆χ 2 (9) = 6.97, p = 0.64, ∆CFI = 0.001. The results of the analysis indicated that the model was invariant between the two groups.

Discussion
The present research used the intergroup threat theory as a theoretical framework to investigate the Italians' classical and modern prejudice toward asylum seekers. In particular, this study was focused on intergroup anxiety, a subtype of threat centring on apprehensions about interacting with outgroup members, which has already been used to examine negative attitudes toward different ethnic and cultural minority groups in the Italian context also (Cowling et al. 2019;Visintin et al. 2017;Voci and Hewstone 2003).
The results of our study reveal that Italians express negative attitudes toward asylum seekers, which are consistent with a previous Italian study (Mancini et al. 2018). In the present study, we explored the role of national identification as an antecedent variable, which has been identified as an indicator of intergroup anxiety and negative outgroup attitudes (O'Doherty and Augoustinos 2008;Paolini et al. 2018). Correlations between the key variables were in line with the general predictions of the study. Consistent with the study of Anderson (2018), our results confirmed the prediction that national identification was positively correlated with prejudice (classical and modern) toward asylum seekers, but not with gender and age. A positive correlation emerged also between intergroup anxiety and national identification. Ingroup identification, such as national identity, serves as a basis for categorizing other people into ingroup and outgroup. This categorization process can produce an ingroup/outgroup bias that can generate intergroup anxiety (Stephan 2014). Similarly, the distinctiveness of an ingroup has been found to be positively correlated with intergroup anxiety, perhaps because ingroup members are worried that their distinctiveness could be negatively affected by interaction with outgroup members (González et al. 2010).
Furthermore, national identification was indirectly related to the outcome variable through intergroup anxiety. Anxiety about interacting with asylum seekers has been found to predict fewer positive attitudes toward them (Anderson 2018;Anderson and Ferguson 2018;Visintin et al. 2017). This result can be explained by the context in which the present study measured intergroup anxiety. It could be suggested that intergroup anxiety is the result of false beliefs derived from media influence that often reinforces the perception of the threat that immigrants pose to national identity security (Cowling et al. 2019). A previous study, for example, found that Australians viewed asylum seekers and refugees as a threat to Australia because of their supposed relationship with Islam and terrorism (Mckay et al. 2012).
Additionally, we tested two alternative mediating models (full and partial) of the effects of intergroup anxiety and national identification on classical and modern prejudice toward asylum seekers. The results revealed that the partial mediating model was better than the full mediated model, in which the direct path from national identification to prejudice was not included. We are aware that more complex paths could have better results. However, the current result confirms the mediating role of intergroup anxiety, which is consistent with previous studies (Visintin et al. 2017;Voci and Hewstone 2003). This highlights the importance of the emotional aspects of prejudice and confirms the central role that cultural and situational factors, which elicit negative emotions (i.e., the images conveyed by the media) may have on anti-immigrant attitudes. In this sense, the present research corroborates the idea of prejudice as a reaction to the perception of ingroup threat, which increases the expression of prejudice.
The results of the multigroup invariance showed that there are no differences between students and young adults and adults. According to McConahay et al. (1981) prejudice against asylum seekers could be influenced by the social and cultural climate rather than by the negative relationship with the members of the minority group. We live in a period in which nations are facing dramatic migratory flows, and given the increase of prejudice and discrimination against migrants, it would be crucial to know the people's anti-immigrant feelings (Caricati et al. 2017). From this perspective, the current results suggest that intergroup anxiety represents a key variable in increasing the prejudice of students and young adults and adults against asylum seekers. This result seems to support the urgency of planning intervention programs aimed at improving the social function of mass-media communication. Media should avoid unnecessarily generating feelings of threat in the population, inasmuch as doing so increases the risk, directly and indirectly, of prejudice and negative attitudes toward outgroup members.

Limitations and Future Research
Most of the findings of the current study were consistent with the predictions derived from the intergroup threat theory (Stephan and Stephan 2000). However, the present study shows certain limitations that future study should address. The first limitation concerns the correlational design of the study, which does not allow causal interpretations. Although our predictions were derived from the perspective of intergroup threat theory, and although the models were compared for fit, it is difficult to draw firm conclusions about causality. For example, it could reasonably be argued that it is not so much the case that intergroup anxiety underlies negative attitudes toward asylum seekers, but rather that individuals with more anti-immigration beliefs tend to perceive more threats. Thus, one could continue to discuss the direction of the influence of the variables explored in this study, and future research should be conducted experimentally and longitudinally to find such evidence. Another issue is related to the use of self-reports. These are vulnerable to social desirability biases and strongly depend on the participants' ability to understand textual cues. This problem can be better dealt with by implicit attitude measures (Anderson 2019). Another limitation is related to the sample. Overall, while our sample was diverse with reference to age and educational level, the size, however, was quite limited and specific, considering the properties of the study, and the participants came from a restricted, although interesting area of South Italy, characterized by different political and cultural traditions in its relationship with immigrants. Consequently, this reduces the generalizability of the results. Future studies should recruit participants by using more sophisticated sampling techniques.

Practical Implications
Despite certain limitations, the findings of the current study have potential implications for the designing of strategies aimed at reducing prejudice and improving positive relations between settled and asylum-seeker groups. The first implication derived from the current research concerns ingroup identification in terms of national identification. In the present study, Italian identification was associated with more modern prejudices and intergroup anxiety. We knew that social categorization and ingroup identification with a superordinate category and/or in conjunction with a subordinate category can reduce prejudice (Tajfel and Turner 2001). Furthermore, it has been demonstrated that the perception of outgroup members as sharing the same mutual affiliation with ingroup members (e.g., people who live in the same residential area, village, or town) can reduce the effects of social categorization and, hence, of evaluative bias (Steinmayr 2016). Future studies should explore these lines of research to examine how different categorization models might help identify strategies for improving social relations between asylum seekers and local communities.

Conclusions
The current research shows that intergroup anxiety (Stephan 2014) can shape the public attitude toward asylum seekers. Additionally, this study has identified antecedents that can potentially increase or decrease intergroup anxiety, along with certain strategies for improving social inclusion policies as well as relationships between asylum seekers and settled communities. The current findings highlight the need to design strategies that reduce the perception of asylum seekers as threats to reduce intergroup anxiety and consequently decrease negative attitudes toward them. Peoples' outlook on asylum seekers and refugees should be changed, by revising the current negative social beliefs in order to contribute to rationalizing human suffering.
Funding: This research received no external funding.

Conflicts of Interest:
The author declares no conflict of interest.