Brexit and Polexit: Collective Narcissism Is Associated With Support for Leaving the European Union

In three studies we examine the link between types of national identity and support for leaving the European Union (EU). We found that national collective narcissism (but not national identification without the narcissistic component) was positively associated with a willingness to vote Leave, over and above the effect of political orientation. This pattern was observed in a representative Polish sample (Study 1, n = 635), as well as in samples of Polish youth (Study 2, n = 219), and both Polish (n = 73) and British (n = 60) professionals employed in the field of international relations (Study 3). In Studies 2 and 3 this effect was mediated by biased EU membership perceptions. The role of defensive versus secure forms of in-group identification in shaping support for EU membership is discussed.

of negative attitudes towards out-groups (see Cichocka, 2016 for a review). It also fosters intergroup perceptions of threat and convictions that other groups aim to harm the in-group. These effects are usually observed over and above the effects of other variables typically associated with intergroup outcomes, such as social dominance orientation or right-wing authoritarianism (Cichocka, Dhont, & Makwana, 2017;Golec de Zavala et al., 2009), and also over and above conventional in-group identification (Golec de Zavala, Cichocka, & Bilewicz, 2013).
In fact, in-group identification without the narcissistic component tends to have different consequences than collective narcissism. Both collective narcissism and in-group identification assume a positive evaluation of the in-group and, thus, are positively correlated. The relationships between collective defensiveness and secure in-group identification can be better understood in reference to personal-level processes (Cichocka, 2016;Golec de Zavala, Cichocka, Eidelson, & Jayawickreme, 2009). Narcissism and self-esteem also tend to positively correlate, although they are underlain by different processes (Marchlewska, Castellanos, Lewczuk, Kofta, & Cichocka, 2019;Marchlewska & Cichocka, 2017;Paulhus, Robins, Trzesniewski, & Tracy, 2004). While personal defensiveness operationalized as individual narcissism aims to protect the self against external threats and is thus linked to enhanced vigilance, secure self-esteem (that is, self-esteem once the variance shared with individual narcissism is co-varied out) implies openness to criticism and readiness to disclose negative aspects of the self which are combined with relatively low need to be liked and recognized by others (Kernis, 2003). In a similar vein, secure in-group identification reflects the successful management of personal needs (e.g., the need for personal control, which is usually referred to as the ability to influence the course of one's own life; Kay, Whitson, Gaucher, & Galinsky, 2009), while collective narcissism is a response to frustrated personal needs (Adorno, 1963), e.g., low personal control (see Cichocka et al., 2018) or generally low feelings of self-worth . Thus, secure in-group identification is a positive attachment to the in-group characterized by the relative absence of defensiveness, that is, by being independent of external recognition and resilient to potential in-group threats and criticism. Using typical measures of in-group identification, we can estimate the effects of secure in-group identification, once we co-vary out the shared variance of collective narcissism and ingroup identification. Research demonstrates that while collective narcissism is related to out-group hostility (Cai & Gries, 2013;Golec de Zavala & Cichocka, 2012), in-group identification without the narcissistic component predicts out-group tolerance Golec de Zavala, Cichocka, & Bilewicz, 2013;Marchlewska, Cichocka, Jaworska, Golec de Zavala, & Bilewicz, 2019).
This might explain their low attachment to the wider international community. Indeed, past work has indicated that British collective narcissism was associated with support for Brexit (e.g., Golec de Zavala, Guerra, & Simão, 2017;Marchlewska, Cichocka, Panayiotou, Castellanos, & Batayneh, 2018). Extending this work, we aim to test whether it might apply beyond the context of Brexit. Drawing on this work, we hypothesize that readiness to leave supranational organizations, such as the European Union, should be positively predicted by national collective narcissism, but not in-group identification. Crucially, we differentiate between a secure and a defensive form of in-group identification and hypothesize that only the latter is associated with biased perceptions of membership benefits, and thereby with stronger support for leaving the EU.
In the current research, we hypothesize that collective narcissism serves as an accessible group-based ego-enhancement strategy. We predict that, as such, it would bias perceptions of current political events in the direction that would help manage individual needs. We go beyond the context of Brexit and aim to test the mechanism behind the association of national collective narcissism and support for leaving the EU. We chose the Polish context because Poland as a country has objectively gained a lot by having been an EU member for the last 15 years as evidenced by governmental reports (Cieślak-Wróblewska, 2019). Despite those economic gains, dissatisfaction with EU is currently being voiced in Poland, as manifested by the rise of the Polexit political party or calls to "verify" Poland's membership by the ruling Law and Justice party. In 2017 when we started this research, 63% of Polish citizens were satisfied with membership, while 16% were dissatisfied (Jędras, 2017). We predict that despite the objective benefits, those high in collective narcissism are reluctant to admit that Poland owes its prosperity to the European Union. We expect that they will rather tend to perceive EU membership as resulting in losses, which fits the central motive of the in-group's superiority and unappreciated greatness. In turn, this biased perception results in support for weakening ties with other European countries.

Overview of the Studies
We tested our hypotheses in a series of three studies examining the associations between national collective narcissism versus national identification and Brexit and Polexit attitudes. Study 1 used a representative sample of Polish adults (N = 635). Study 2 was conducted among 219 Polish students. Study 3 was conducted among employees and associates of the British Embassy in Poland (60 British,73 Polish). In all studies, national collective narcissism and national identification were measured as predictors of voting in the presumed Brexit and Polexit referenda. In Studies 2 and 3, a biased perception of EU membership (i.e., seeing more costs than benefits) was measured as a mediator of the effect of national collective narcissism on voting. As political ideology may be predictive of support for policies leading to further European integration versus disintegration and past work found that right-wing political attitudes were associated with national Collective Narcissism and Support for Leaving the EU collective narcissism Golec de Zavala et al., 2009), in all studies we controlled for political orientation.

Study 1 Method Participants and Procedure
In Study 1 we recruited a sample representative for Polish internet users. With the help of the Ariadna Research Panel (Ogólnopolski Panel Badawczy Ariadna, 2019) we recruited 635 participants, 333 women and 302 men, aged between 18 and 79 (M = 43.22; SD = 15.12). We measured national collective narcissism and national identification as predictors of voting in the presumed Polexit referendum 1 . The order of predictors was counterbalanced. We also measured basic demographics and participants' political orientation.

Measures
Collective narcissism -was measured with the short 5-item version of the Collective Narcissism Scale (Golec de Zavala et al., 2009). Participants indicated to what extent they agreed with the statements (e.g., "I will never be satisfied until Poles get all they deserve", "It really makes me angry when others criticize Poles", "If Poles had a major say in the world, the world would be a much better place") on scales from 1 (definitely not) to 7 (definitely yes), α = .92, M = 4.45, SD = 1.45.
National identification -was measured with 5 items of the Social Identification Scale (Cameron, 2004). Participants responded to items (e.g. "I have a lot in common with other Poles", "I feel strong ties to other Polish people", "In general, I'm glad to be Polish") on scales from 1 (definitely not) to 7 (definitely yes), α = .88, M = 5.25, SD = 1.26.
Voting preference in a Polexit referendum -was measured with a single item: "If you could vote in an EU referendum today, how would you be inclined to vote?", with 528 participants indicating that Poland should remain in the EU (coded as 0), and 107 indicating that Poland should leave the EU (coded as 1).
Political orientation -was a single-item measure, with a scale from 1= definitely left-wing to 7 = definitely right-wing, M = 4.11, SD = 1.19. 1) Beside the variables reported here, Study 1 also involved measures of support for environmental policies included for the purpose of a different project employing the same predictors (Cislak, Wojcik, & Cichocka, 2018; please contact the first author for details).

Zero-Order Correlations
We first computed correlations between continuous variables. National collective narcissism and national identification were significantly positively correlated to each other and to political orientation (Table 1).

Voting Preference in Polexit Referendum
We then examined the hierarchical logistic regression model to verify the relationship between two forms of national identification and voting preference in a referendum regarding Poland's possible exit from the EU. As presented in Table 2, political orientation significantly positively predicted the vote with right-wingers being more prone to vote Leave. While gender was unrelated to voting, age was a significant predictor of the vote, with young people being more prone to vote Leave. Step 1 Step 2 Importantly, when the overlap between collective narcissism and national identification was adjusted for in Step 2, national identification predicted voting Leave significantly and negatively, while collective narcissism predicted it significantly and positively.

Discussion
Study 1 confirmed that national collective narcissism and national identification were positively correlated, yet associated with the tendency to vote Leave in opposite ways. Past work has shown that the level of national collective narcissism was associated with Brexit attitudes. This study extended previous findings by studying the different national context of a country that has not officially communicated its intention to leave the European Union. Further, we showed the opposite effects of defensive and secure national identification on tendency to vote Leave. This allows for a more nuanced understanding of the differences in policies supported by individuals declaring equally strong attachment to their nation . Furthermore, these effects were demonstrated over and above the effects of left-right political orientation.

Study 2
In Study 2 we aimed to replicate these findings and to test the potential social-psychological mechanisms behind these effects. To this end, we replicated the design of Study 1 and tested the biased perception of membership benefits as a mediator of the effect of national collective narcissism on the tendency to vote Leave.

Method Participants and Procedure
Following the recommendations of Vazire (2016), we aimed for a sample size that would provide 80% power to detect the effect of r of .21-a typical effect in social/personality psychology (Richard, Bond, & Stokes-Zoota, 2003). Using GPower, we estimated the target sample size to be at least 173 or larger. In Study 2 we recruited 219 Polish young adults, 167 women and 50 men (2 persons declined to indicate their gender), aged between 19 and 24 (M = 20.30; SD = 0.82).
Biased membership perception -was measured with eight items capturing viewing EU membership as having more costs that benefits, for example "Overall, EU membership brought Poland more losses than gains", "The EU benefits significantly more from Poland being a member than vice versa", "Poland outside the EU will have a stronger economy and military power" or "As part of the EU, Poland has a greater voice in international affairs" (reversed item). Participants were asked to indicate the extent to which they agree with these items on a scale from 1 = definitely no to 7 = definitely yes. Participants' responses to these items showed a high internal consistency; thus they were averaged to create a single measure of biased EU membership perception, α = .84, M = 3.37, SD = 1.06.
Voting preference in Polexit referendum -was measured in the same way as in Study 1, with 183 participants indicating that Poland should remain in the EU (coded as 0), 29 indicating that Poland should leave the EU (coded as 1), and 7 participants declining to answer.
Political orientation -was also measured in the same way as in Study 1, M = 4.37, SD = 1.13.

Zero-Order Correlations
Similar to Study 1, national collective narcissism and national identification were significantly positively correlated to each other and to political orientation, though the relationship between national collective narcissism and political orientation was only marginally significant. All of these variables were significantly positively related to biased membership perception (Table 3).

Biased Membership Perception
We then examined predictors of the biased perception of Poland's membership benefits in a hierarchical regression model. As presented in Table 4, political orientation entered in Step 1 was a significant predictor of biased membership perceptions meaning that Collective Narcissism and Support for Leaving the EU right-wingers perceived more costs than benefits of being part of the EU. Age was unrelated to this variable, while gender was related significantly negatively, indicating that men were more prone to perceive more costs than benefits.

Variable
Step 1 Step 2 Crucially, when the overlap between collective narcissism and national identification was adjusted for in Step 2, the relationship between national identification and biased membership perception was significant and negative. In contrast, the relationship between collective narcissism and biased membership perception was significant and positive.

Voting Preference in Polexit Referendum
To verify the relationship between two forms of national identification and voting preference in a referendum regarding Poland leaving the EU, we examined the hierarchical logistic regression model. As presented in Table 5, when the overlap between collective narcissism and national identification was adjusted for, national identification predicted voting to leaving the EU significantly and negatively, while collective narcissism predicted it significantly and positively.

Mediation Model
Finally, we examined the mediation model presented in Figure 1 using Hayes's (2012) PROCESS macro with a bootstrapping procedure (with 10,000 resamples) in which biased perception of membership benefits mediated the association between collective narcissism and voting to leaving the EU, Cox & Snell R 2 = .34; p < .001; Nagelkerke R 2 = .62; p < .001. Biased membership perception was a significant mediator of the effect of collective narcissism on voting preferences as evidenced by a significant indirect effect, estimate = 1.13 [0.58, 1.97].

Discussion
Study 2 showed that although national collective narcissism and national identification were positively correlated, they were related to the tendency to vote Leave in opposite ways, thereby replicating the results of Study 1. Again, these effects were demonstrated over and above the effects of left-right political orientation. Extending the results of the previous study, we found that the positive effect of national collective narcissism on readiness to vote Leave was mediated by biased membership perception. Individuals higher in national collective narcissism were more prone to perceive that Poland's costs Collective Narcissism and Support for Leaving the EU are greater than Poland's benefits. As they perceived Poland would be better off outside of the European community than inside it, they were more ready to vote Leave.

Study 3
In Study 3 we aimed to test whether support for such a political decision is driven by similar social-psychological mechanisms across a different national context. To this end, we aimed to replicate the design of Study 2 and recruit comparable samples from Poland and the United Kingdom (UK). UK citizens have already participated in the Brexit referendum, and as a result the UK has initiated the leaving procedure. The reason behind selecting UK citizens as a comparison group is the fact that a referendum concerning EU membership has already taken place in the UK. If we observed a similar pattern, it would make a stronger argument for considering national identification as a crucial factor in shaping citizens' readiness to leave supranational organizations.

Method Participants and Procedure
The study was an online survey distributed among UK and Polish employees and associates of the various departments and / or sections operating within the British Embassy in Poland, e.g., Visas & Immigration Section, Political Section, Consular Section, Department for International Trade, etc. in their native languages (British English and Polish respectively). Respondents who participated in the study potentially represented all offices, grades, and positions as the questionnaire was sent out across the Embassy and filled in anonymously by both Polish and British employees. Nevertheless, it should be noted that respondents' nationality was not clearly aligned with their higher or lower rank or position. For instance, a local (Polish) administrative employee from one department could have a higher level of education (or even job position) than a British administrative employee holding diplomatic status from another department (who, in this case, would perform a lower graded job). In most cases though, higher positions were filled by British nationals, although it did not necessarily correspond with a better education. Local employees, by contrast, had a university education in most cases as for them it was typically an obligatory job requirement. We aimed to recruit as many of them as possible within the two weeks preceding elections in the UK in 2017. The final sample consisted of 73 Polish and 60 UK participants. Thirty-four Polish participants were women, 36 were men (3 persons declined to indicate their gender), and were aged between 20 and 64 (M = 29.74; SD = 9.20). Twenty-six UK participants were women, 33 were men (1 person declined to indicated gender), and were aged between 20 and 66 (M = 38.66; SD = 11.06).

Zero-Order Correlations
As in Studies 1-2, national collective narcissism and national identification were significantly positively correlated with each other and with political orientation. Again, all of these variables were significantly positively related to biased membership perception (Table 6).

Biased Membership Perceptions
Similar to Study 2, we then examined predictors of biased membership perception of Poland's and UK's membership using hierarchical regression models. Taking into account the relatively small sample we managed to recruit in that study, we were not able to test measurement invariance, required for cross-cultural comparisons. Thus, we analysed support for leaving the EU separately for those two samples 2 . Similar to Study 2, political orientation entered in Step 1 was a significant predictor of biased membership perceptions with right-wingers perceiving more costs than benefits of being part of the EU. Both age and gender were unrelated to biased membership perception (Table 7).

Summary of Hierarchical Regression Analysis for Relative Disadvantage of the UK's and Poland's Membership as Criteria (Study 3)
Variable UK Poland Step 1 Step 2 Step 1 Step 2 When the overlap between collective narcissism and national identification was adjusted for in Step 2, for both countries biased membership perceptions were significantly and positively predicted by collective narcissism, though not by national identification (among UK citizens national identification predicted biased membership perception negatively, but not significantly).

Voting Preferences in Referenda
Again, as presented in Table 8, the results of logistic regression showed that for both countries collective narcissism, but not national identification, positively predicted voting for leaving the EU.
2) We also tested these effects within one model treating nationality as a moderator and did not find any significant moderation effects. Step 1 Step 2 Step 1 Step 2

General Discussion
The results of all the studies converged to show that national collective narcissism was a strong positive predictor of a preference to leave the EU. We found this pattern in a representative Polish sample (Study 1), a sample of Polish youth (Study 2), and a sample of employees professionally associated with international relations representing two different countries (Study 3). In Study 3, we observed a similar pattern both among citizens of a country that is leaving the EU (the UK) and a country that has not yet officially announced its intention to do so (Poland). The effects of national collective narcissism on Polexit and Brexit attitudes were mediated by perceiving EU membership as having more costs than benefits (Studies 2 and 3).
Crucially, we showed the opposite effects of national collective narcissism and national identification on readiness to vote Leave (Studies 1 and 2). Moreover, across all of the studies, we have found a significant effect of political orientation on voting with right-wingers being more prone to vote Leave, both in Poland and in the UK. Importantly, however, in all three studies the positive effects of national collective narcissism were significant over and above the effect of political orientation. Notably, the negative effect of secure national identification did not replicate in Study 3. This might be attributed to the smaller samples of British and Polish employees of the British Embassy in Poland that we managed to recruit in Study 3 than in Study 1 (involving a representative sample for Polish Internet users) and Study 2 (involving a sample of Polish youth). Thus, Study 3 might have been underpowered, and consequently we did not observe this relatively weaker effect (see Studies 1 and 2) as significant.
Our results are in line with past work on national identification and Brexit attitudes (Golec de Zavala et al., 2017;Marchlewska et al., 2018). We extend past work by verifying the relationship between different forms of national identification and willingness to leave the EU in different contexts (the UK and Poland), using a representative (Polish) sample, and proposing that it is the biased perception of membership benefits that stands behind the effect of narcissistic national identification and support for leaving the EU. The current findings clarify the role that national identity plays in predicting support for membership in multi-national organizations. Previous work has shown that  collective narcissism, but not national identification without the narcissistic component, is associated with hostile intergroup attitudes (Golec de Zavala, Cichocka, & Bilewicz, 2013). Extending this line of research, we have found that collective narcissism, but not national identification, is associated with support for policies undermining close relations with other countries, thereby showing potential ingroup-relevant consequences in the domain of international affairs (see also Cichocka, 2016).

Implications and Future Directions
These results may be interpreted within the broader framework of defensive versus secure ingroup identification. Defensive forms of identification such as collective narcissism are underlain by doubts about whether the in-group is highly regarded and respected by others (Golec de Zavala et al., 2009). We found that the defensive form of national identification, in contrast to the secure one, is associated with the biased perception of the member benefits of one's own nation within multi-national organizations, and thereby with a greater readiness to leave them. Thus, defensiveness is associated with renouncing the benefits that are available to the in-group, because other groups may benefit even more. Although not measured here, similar patterns could be predicted for other conceptualizations of in-group identification, to the extent that they capture an excessive and defensive attachment to the group (such as nationalism or blind patriotism). The conceptualization we used here allows, however, for testing these phenomena across different group contexts and thus may inspire future studies outside the immediate national context (e.g. organizational).
In this work we suggest that collective narcissism is associated with a biased, defensive interpretation of current international politics, which in turn encourages individuals to weaken intergroup ties. Our finding is in line with work showing that collective narcissists might view their group as indispensable to the overarching EU identity (Guerra, Abakoumkin, Golec de Zavala, Sedikides, & Wildschut, 2019). In past research, in-group indispensability has been linked to a negative attitude towards outgroups (Ng Tseung-Wong & Verkuyten, 2010), which suggests that biased membership perceptions might not only weaken willingness to be part of the EU, but also harm relations with other EU members. Future work should explore these interesting questions.
One limitation of the studies presented here is their correlational design. Indeed, one could argue that it is rather the biased perception of benefits stemming from EU membership that increases collective narcissism. In turn, this defensive response results in support for leaving supranational organizations. However, collective narcissism reflects deep-seated frustrations of individual needs (such as low feelings of self-worth; Golec de Zavala et al., 2019 or low feelings of control over one's life; Cichocka et al., 2018). It is thus less likely that collective narcissism develops as a response to current political circumstances. Instead, it helps in managing the psychological needs of the individual and dealing with frustrations (cf., Fritsche et al., 2013), which nevertheless may be rooted in a more generalized sense of relative deprivation . As such, it likely biases perceptions of those current political circumstances, thereby reflecting need management strategy. Future work should experimentally disentangle this mechanism.
These results also show that similar social-psychological processes underpin support for EU integration versus disintegration in different societies. This suggests that the biased membership perception behind citizens' willingness to support policies undermining European integration, stems from defensive national identification. Yet, after controlling for the defensive, narcissistic component of national identification, strong secure identification tended to be negatively related to both dissatisfaction with EU membership and readiness to support an exit. These processes operate over and above political ideology. This work thus implies the ambivalent effects of fostering a strong national or ethnic identification, a phenomenon so frequently observed in today's political landscape. Mobilizing nationalism is the factor implicated for the past fall of multi-national European countries (Hobsbawm, 2012). Our social-psychological research suggests that defensive forms of national identification, such as national collective narcissism, may also be associated with dissatisfaction with multi-national organizations. Our work then suggests that appealing to the unappreciated greatness of one's nation and need for recognition may result in mobilizing dissatisfaction with the effects of European integration and consequently in calling for an exit. Still, mobilizing secure national identification may lead to future integration of European countries. Thus, future work should examine how to differentially foster these two types of strong national commitment.