Understanding identity processes in support for reactionary and progressive social movements among advantaged and disadvantaged groups: The role of collective narcissism and secure ingroup identity

Across five studies (two representative, one pre-registered, total N = 4962), we examined the role of collective narcissism and secure identity in support for reac-tionary (Alt-Right and nationalist) and progressive (Black Lives Matter and Women’s Strike) social movements among advantaged (White Americans and Polish national majority) versus disadvantaged (Black Americans) groups. Among advantaged groups, collective narcissism, more so than secure identity, was related to more support for reactionary movements (Studies 1–4) and less support for progressive movements (Studies 1–3). Studies 3a, 3b and 4 directly compared members of advantaged and disadvantaged ethnic groups in the US and suggested that both collective narcissism and secure identity were positively associated with support for progressive movements among disadvantaged groups. This research contributes to understanding the identity processes involved in reactionary and progressive movements. It highlights the importance of distinguishing secure and defensive (i.e., narcissistic) identities, particularly when investigating support for reactionary movements among advantaged groups.

movements (Muis & Immerzeel, 2017), even leading to the reversal of previously acquired rights (e.g., the near-total ban of abortion in Poland in 2021, BBC News, 2021).Increasing ideological polarisation implies that understanding the identity processes leading to support for, and investment in, reactionary and progressive social movements is of utmost importance.This project examines the ways in which support for far-right reactionary movements and progressive movements in favour of disadvantaged groups is linked to national and ethnic social identity and distinguishes between two forms of identity: defensive (narcissistic) and secure.

1.1
The role of social identity in reactionary and progressive social movements Past work in psychology and sociology emphasises the identity roots of support for social movements and collective action as a means of claiming one's group rights and interests (Tilly, 1976;van Zomeren, 2016).
However, recent work calls for consideration of the ideological roots of support for social movements (see Becker, 2020).From this perspective, support for social movements goes beyond the strict defence of ingroup interests and advocates for the defence of a certain social order.
In this article, we examine the identity correlates of social movements, considering the ideologically grounded classification of progressive and reactionary social movements (see Becker, 2020;Osborne et al., 2019).Progressive movements challenge the status quo for greater equality (Becker, 2020;Osborne et al., 2019).Most of these social movements aim to defend the rights of disadvantaged groups, such as ethnic minorities and women.In contrast, some social movements aim not only to maintain the status quo but also to bring about reactionary social change by, for example, increasing the social hierarchy (Becker, 2020;Osborne et al., 2019).These reactionary social movements ultimately benefit structurally advantaged groups (Osborne et al., 2019).
Several theoretical models have sought to understand support for social movements.The most widely used is the social identity model of collective action (SIMCA, van Zomeren et al., 2008), which positions ingroup identification (typically considered as a group-level self-investment, which encompasses the centrality of one's identity to the self, solidarity with ingroup members, and satisfaction with one's group membership; Leach et al., 2008), perceived injustice and perceived collective efficacy as the three major predictors of collective action support.Among these variables, identification seems to play a prominent role and is both a direct and indirect (by reinforcing perceptions of injustice and collective efficacy) predictor of collective action (van Zomeren et al., 2008; see also Abrams et al., 2020;Sturmer & Simon, 2004).
To date, research on collective action and its determinants has mainly focused on the engagement in progressive movements among disadvantaged groups (see e.g., Thomas et al., 2020).Despite the increasing prevalence of reactionary social movements in today's societies, there is still little research on what motivates individuals to engage in them.A few exceptions do, however, show that national iden-tity might also be a predictor of engagement in reactionary movements.
For example, national identity is a key predictor of engaging in collective action against refugees, especially among those who feel a high level of threat to their social identity (López et al., 2019).Osborne et al. (2019) have also investigated the validity of SIMCA in predicting support for both progressive and reactionary social movements among advantaged and disadvantaged ethnic groups.The authors examined the relationship between system justification, identification, perceived injustice and efficacy on support for progressive (e.g., support for M āori rights in New Zealand and support for Black Lives Matter, i.e., BLM, marches in the US) and reactionary (e.g., support for White rights marches or against the removal of confederate statues in the US) movements.They consistently found that ingroup identification among members of advantaged groups (European New Zealanders and White Americans) correlated positively with support for reactionary social movements but correlated negatively with support for progressive movements.In contrast, findings were mixed for members of minority groups: among New Zealand M āori, identification correlated positively with support for both progressive and reactionary actions, while among American ethnic minorities, identity did not directly predict support for either type of movement.
Another line of research by Thomas et al. (2020) confirmed the central role of identification in collective action among both advantaged and disadvantaged groups.In a longitudinal study, the authors demonstrated a temporal direct and indirect (through increased perceptions of injustice) link between identification and support for future collective action (e.g., participating in demonstrations on behalf of one's ethnic group).Importantly, this temporal link between identification and collective action was found in both disadvantaged and advantaged groups.In other words, ingroup identification plays an important role in predicting support for movements that benefit the ingroup for members of both structurally advantaged and structurally disadvantaged groups, and can thus help explain support for progressive and reactionary social movements.However, some questions remain regarding similarities and differences between groups based on status.
SIMCA is rooted in the social identity tradition (Tajfel & Turner, 1979), where ingroup identification is typically seen as a selfinvestment within the ingroup.Yet, subsequent work has distinguished between different forms of social identity.For example, within the context of national identity, authors have distinguished between patriotism and nationalism (Kosterman & Feshbach, 1989), constructive and blind patriotism (Schatz et al., 1999), or attachment and glorification (Roccas et al., 2008).In this article, we draw on a broader approach which differentiates between ingroup identification and collective narcissism-a belief in ingroup greatness that is contingent upon external recognition of one's group's worth (Golec de Zavala et al., 2009).Collective narcissism and ingroup identification can have distinct-and even opposing-impacts on intra-and intergroup processes (Cichocka, 2016;Cichocka et al., 2021).Distinguishing collective narcissism from ingroup identification might shed further light on the processes underlying support for reactionary and progressive movements and help reconcile past mixed results observed for advantaged and disadvantaged groups (Osborne et al., 2019).

1.2
Collective narcissism, secure ingroup identity and support for social movements Because collective narcissism and ingroup identification typically correlate positively, researchers usually seek to account for their shared variance (Cichocka et al., 2016;Golec de Zavala et al., 2013, 2020;Marchlewska, Cichocka, et al., 2021).Defensive ingroup identity (also sometimes referred to as defensive ingroup positivity or defensive ingroup commitment) is what is achieved from collective narcissism when controlling for the variance it shares with ingroup identification.
Conversely, secure ingroup identity (also sometimes referred to as secure ingroup positivity or secure ingroup commitment) is what is achieved from ingroup identification when controlling for collective narcissism and is characterised by an unpretentious commitment to the group, devoid of the defensive posture and the need for external validation.
Collective narcissism goes hand in hand with the protection and reinforcement of the positive image and status of the group at all costs (e.g., Gronfeldt et al., 2022).In line with this idea, there is preliminary evidence that collective narcissism among advantaged groups is linked to stronger support for pro-ingroup, reactionary social movements and less support for pro-outgroup, progressive ones.Firstly, collective narcissism measured in reference to the national majority group has been associated with several pro-ingroup reactionary movements and actions, including lower willingness to engage in collective action in favour of refugees (Górska et al., 2020), stronger support for Brexit in order to limit immigration (Golec de Zavala et al., 2017;Marchlewska et al., 2018), and stronger support for the war in Iraq and a greater emphasis on military strength (Golec de Zavala et al., 2009).Importantly, these effects were specific to collective narcissism, and not observed for ingroup identification.Secondly, recent studies have investigated the link between collective narcissism, ingroup identification, and support for progressive actions among members of advantaged groups (Marchlewska, Górska, et al., 2021).Collective narcissism (related to gender and sexual orientation) was associated with less support for, and willingness to engage in, progressive collective action (defending women and LGBTQI+ rights, respectively) among heterosexual men.Again, ingroup identification was either unrelated or related to greater support for progressive actions.
In sum, defensive identity promotes pro-ingroup and even antioutgroup actions that maintain the status and positive image of the ingroup (Cichocka, 2016;Golec de Zavala et al., 2009).Therefore, we hypothesise that collective narcissism among members of advantaged groups (e.g., Whites, the national majority) should be related to support for reactionary movements that benefit the ingroup (e.g., defence of White majority rights) and less support for progressive movements in favour of disadvantaged groups (e.g., ethnic minorities' rights, women's rights).In contrast, secure identity reflects a more constructive commitment to the ingroup, coupled with tolerance or even positive attitudes towards outgroups (Cichocka, 2016;Cichocka & Cislak, 2020).Therefore, a secure identification with the ingroup among members of advantaged groups may be unrelated to support for reactionary and progressive movements or may even be related to more support for progressive and less support for reactionary movements.
Although past work distinguishes between the actions supporting advantaged and disadvantaged groups, research has yet to compare the role of defensive and secure identity between the members of advantaged and disadvantaged groups.We mentioned previous studies showing that identification seems to play a similar role in both advantaged and disadvantaged groups with regard to collective actions in favour of their respective groups (Osborne et al., 2019;Thomas et al., 2020).Yet, to date, very few studies examined the role of collective narcissism in disadvantaged groups.Two exceptions are studies by Golec de Zavala et al. (2009), who found that collective narcissism was related to greater support for collective action in favour of the ingroup (boycotting American products in a Mexican population) and to greater anti-White sentiment among Black Americans.However, these studies omitted measures of ingroup identification and thus did not allow for a comparison between the role of defensive and secure forms of identity among members of disadvantaged groups.We aim to fill this lacuna.
As with advantaged groups, defensive identity among disadvantaged groups may also promote attitudes and actions aimed at increasing the status and positive image of the ingroup (Golec de Zavala et al., 2009).Accordingly, we expected collective narcissism to be associated with more support for pro-ingroup, progressive social movements, and less support for pro-outgroup, reactionary social movements among the disadvantaged groups.Thus, we expected defensive identity to be positively linked to support for pro-ingroup movements and negatively linked to support for pro-outgroup movements in both advantaged and disadvantaged groups.However, we expected a different picture to emerge when focusing on secure ingroup identity.As studies on national majority or advantaged groups show, a secure identity encompasses the truly satisfying part of social identity and ingroup investment, but does not seek over-recognition or intergroup dominance and may allow for constructive and egalitarian intergroup relations (see Cichocka, 2016;Cichocka & Cislak, 2020;Tajfel & Turner, 1979).Thus, secure identity among disadvantaged groups might be negatively related to support for reactionary movements (reinforcing the existing hierarchy) and positively related to support for progressive movements (containing pro-equality aspects).Although in the same direction, the positive association between secure identity and progressive movements may nonetheless be stronger among disadvantaged groups than among advantaged groups.Indeed, for disadvantaged groups, progressive movements may also reflect concern for other ingroup members and for the ingroup itself (which is not the case for advantaged groups).
Thus, the positive association between secure identity and support for collective action on behalf of the ingroup may be particularly pronounced among members of disadvantaged groups (see Cichocka et al., 2021).
In sum, while collective narcissism is expected to be associated with support for pro-ingroup and anti-outgroup social movements among advantaged and disadvantaged groups, secure identity might be unrelated or weakly related to lesser support for reactionary movements, and greater support for progressive movements in both types of groups.Thus, distinguishing between forms of social identity would

OVERVIEW
This project seeks to deepen our understanding of the identity processes associated with support for reactionary and progressive social movements.To this end, we examined the differences and similarities in these processes among the advantaged and the disadvantaged groups.Among advantaged (ethnic and national) groups, collective nar- supporting the Charlottesville reactionary social movements among the advantaged group of White Americans.

Measures
Collective narcissism

Ingroup identification
Ethnic ingroup identification was measured with five items from Cameron's (2004) ingroup ties (e.g., "I have a lot in common with people who are the same ethnic group as me.") and ingroup affect (e.g., "Generally, I feel good when I think about being a part of my ethnic group.")subscales.Items were rated on a scale from 1 (Strongly disagree) to 7 (Strongly agree), α = 0.83.

Support for reactionary social movement
Support for reactionary social movement was measured with three items: (1) "How much do you agree that the removal of the statue should have been protested?",with a scale from 1 (Strongly disagree) to 7 (Strongly agree), (2) "To what extent do you support or oppose the "Unite the Right" rally?", with a scale from 1 (Strongly oppose) to 7 (Strongly support), (3) "How likely would it be for you to take part in a rally similar to the one in Charlottesville in the future?", with a scale from 1 (Extremely unlikely) to 7 (Extremely likely), α = 0.79. 2

Zero-order correlations
We computed zero-order correlations between collective narcissism, ingroup identification and support for reactionary social movement (Table 1).All three variables were positively correlated with each other.
2 An additional item "How much do you agree with President Trump's statement about violence on both sides?" was also present.As it does not directly refer to support for reactionary movements, it was not included in the analyses.Including this item did not change the pattern of results.

Discussion of Study 1
In Study 1, we first examined collective narcissism and ingroup identification as predictors of support for reactionary social movement.As expected, when considered together, only the defensive (narcissistic) form of the ethnic social identity was related to support for the reactionary social movement.This first study illustrates the distinct role of defensive and secure social identity in support for reactionary movements among advantaged groups.Study 2 aimed to replicate these findings on a different type of advantaged group (national) using a representative sample, and measuring support for both reactionary and progressive movements.

STUDY 2
Study 2 examined the links between defensive and secure social iden-

Collective narcissism
National collective narcissism was measured with the five-item version of the Collective Narcissism Scale (Golec de Zavala et al., 2009) adapted to the Polish national group (e.g., "I wish other countries would more quickly recognize the authority of Poles.").Items were rated on a scale from 1 (Definitely not) to 7 (Definitely yes), α = 0.91, M = 4.04, SD = 1.39.

Ingroup identification
National ingroup identification was measured with the ten group-level self-investment items (e.g., "I feel a bond with other Poles"; Leach et al., 2008).Items were rated on a scale from 1 (Definitely not) to 7 (Definitely yes), α = 0.97, M = 5.01, SD = 1.5. 3

Support for and participation in social movements
Participants were asked to indicate whether they supported and/or participated in the Independence March (reactionary social movement) and in the National Women's Strike (progressive social movement) by choosing one of four options: "I do not know this initiative" (excluded from analyses: Independence March: 10.4% of the sample; 3 Consistent with the methodology and theorising of previous research on defensive and secure identity (see Marchlewska et al., 2020)

Zero-order analyses
Collective narcissism and identification were positively correlated, r = 0.69, p < .001.We conducted two multinomial logistic regressions with collective narcissism and identification as separate predictors of support for, and participation in, the given social movement.We considered "no support" as the baseline, allowing us to compare "no support" versus "support" on the one hand, and "no support" versus "participation" on the other hand.Collective narcissism and identification were both linked to more support and participation in the reactionary movement, and less support and participation in the progressive movement (all |Bs| > 0.35, ps < .001).

Multinomial logistic regression on support for and participation in reactionary social movement
We then conducted a multinomial logistic regression with collective narcissism and identification as joint predictors of support and participation in the Independence March (with "no support" as the baseline), χ 2 (4) = 152.26,p < .001(Table 2).Both collective narcissism and identification were significantly positively related to support for the Independence March.However, only collective narcissism (and not identification) was linked to actual participation in the march.

Multinomial logistic regression on support for and participation in progressive social movement
We conducted the same analysis on the support and participation in the National Women's Strike, χ 2 (4) = 146.79,p < .001(Table 2).Collec-tive narcissism was negatively related to support and participation in this progressive movement.Ingroup identification was neither related to support, nor to participation in the National Women's Strike.

Discussion of Study 2
As expected, collective narcissism measured in relation to a national majority group was related to more support and participation in a right-wing, reactionary social movement, and less support and participation in a pro-disadvantaged, progressive social movement.The role of secure identity was more complex: it was linked to support for, but not participation in, the reactionary social movement and was not related to support for progressive social movements.This suggests that secure identity may be linked to support for a social movement that is potentially beneficial for one's group, albeit reactionary, but this did not imply lower support for disadvantaged groups (as was the case for collective narcissism).
In the first two studies, we obtained support for our hypothesis that a narcissistic, defensive form of identity among advantaged groups is linked to more support for reactionary movements and less support for progressive movements.In the next studies, we extend this research beyond the perspective of the advantaged group and examine similarities and differences in these processes among disadvantaged groups.Additionally, although the progressive movement in Study 2 was egalitarian, it focused on advancing the rights of a group (women) that cuts across other social categories including ethnicity and nationality.Thus, in the following studies, we focused on an interethnic context along with a progressive movement advocating more rights for ethnic minorities.This allowed us to increase the clarity of the pro-ingroup or pro-outgroup valence of the progressive social movement among the disadvantaged or advantaged groups, respectively.

STUDY 3a
Study 3a investigated the associations ethnic collective narcissism and identification have with support for both reactionary and progressive social movements among advantaged and disadvantaged groups.For the reactionary movements, we focused on the Alt-Right-a nationalist and White supremacist social movement (Forscher & Kteily, 2020).For the progressive movements, we focused on BLM-a movement defending Black people's rights.We further examined this issue by recruiting White (advantaged group) and Black (disadvantaged group) American participants.
We hypothesised that collective narcissism would be related to greater support for pro-ingroup movements but less support for prooutgroup ones.Because reactionary and progressive social movements are likely to benefit advantaged and disadvantaged groups, respectively (Osborne et al., 2019), we expected the relationship between collective narcissism and support for the Alt-Right and for BLM to be moderated by group status.Specifically, collective narcissism should be positively related to support for the Alt-Right, and negatively related to support for BLM among the advantaged group.In contrast, collective narcissism should be negatively related to support for the Alt-Right but positively related to support for BLM among the disadvantaged group.
Conversely, secure identity represents a constructive attachment to the ingroup, which may be compatible with egalitarian motives and actions.Thus, we expected secure identity to be positively related to support for BLM.We did, however, expect this association to be weaker in the advantaged group than in the disadvantaged group, where support for BLM may additionally represent a concern for the ingroup and its members.We tested these hypotheses in Study 3a, using a convenience American sample, and replicated them in Study 3b using a representative American sample.

Collective narcissism
Ethnic collective narcissism was measured with the nine-item Collective Narcissism Scale (Golec de Zavala et al., 2009) in reference to one's ethnic group (e.g., "Not many people seem to fully understand the importance of my group.").Items were rated on a scale from 1 (Strongly agree) to 9 (Strongly disagree) but were reverse-coded so that a higher score corresponds to a higher level of collective narcissism, α Advantaged = 0.87, α Disadvantaged = 0.80.

Ingroup identification
Ethnic ingroup identification was measured with a single-item ("I identify with being [White/Black (or African American)]") from Postmes et al. (2013), rated on a scale ranging from 0 (Strongly disagree) to 100 (Strongly agree).

Support for social movements
Participants were reported their attitudes toward reactionary (the Alt-Right) and progressive (BLM) movements: "How positive or negative do you feel concerning the following (social) movements?"using a scale ranging from 0 (Extremely negative) to 100 (Extremely positive).

Zero-order correlations
Descriptive statistics and correlations can be found in Table 3. Collective narcissism and ingroup identification were positively related among both the advantaged group (Whites) and the disadvantaged group (Blacks).Both collective narcissism and ingroup identification were linked to more support for the Alt-Right, and less support for BLM among the advantaged group.Among the disadvantaged group, collective narcissism and identification were linked to more support for BLM, and were not related to support for the Alt-Right.

Multiple regression analysis on support for reactionary social movement
For the following analyses, collective narcissism and ingroup identification were mean-centred.Group status was coded 0 for the advantaged group (baseline) and 1 for the disadvantaged group.We conducted a multiple regression analysis with ethnic collective narcissism, ingroup identification, group status, group status × collective narcissism, and group status × ingroup identification as predictors of support for the Alt-Right, R 2 = 0.10, F(5, 2502) = 53.21,p < .001(Table 4).The R 2 increase following the inclusion of the two interaction terms was small but significant, ΔR 2 = 0.01, F(2, 2502) = 7.20, p = .001.
The analysis revealed a positive main effect of collective narcissism on support for the Alt-Right.This main effect was, however, qualified by an interaction with group status, with a small increase in R 2 due to the inclusion of this interaction over and above the other variables, ΔR 2 = 0.003.Simple effects indicated that collective narcissism was positively associated with support for Alt-Right among the advantaged group, β = 0.28, p < .001,95% CI [0.23, 0.32], but not among the disadvantaged group, β = 0.00, p = 1.00, 95% CI [-0.18, 0.18].Results also revealed a small, albeit significant, positive relationship between ingroup identification and Alt-Right support.This effect did not differ increase due to the inclusion of ingroup identification × group status in addition to the other variables: ΔR 2 = 0.001).
Results showed a main negative effect of collective narcissism on support for BLM.However, this effect was moderated by group status, with a small R 2 increase, ΔR 2 = 0.004, due to the inclusion of collective narcissism × group status in addition to other variables.More precisely, collective narcissism was related to less support for BLM among the advantaged group, β = −0.23,95% CI [−0.27, -0.19], p < .001,but not among the disadvantaged group, β = 0.06, 95% CI [−0.12, 0.23], p = .545.We also observed a negative main effect of ingroup identification on support for BLM but again, this was moderated by group status.

Discussion of Study 3a
Study 3a partially supported our hypotheses.Indeed, the interaction between collective narcissism and group status was significant for support for both the reactionary and progressive social movements.
Moreover, the interaction between group status and identification was significant when looking at support for BLM: the more the disadvantaged group members (but not the advantaged group members) identified with their ethnic ingroup, the more they supported the progressive social movement.

Measures
Measures were the same as in Study 3a.The internal reliability of collective narcissism was good (α Advantaged = 0.89, α Disadvantaged = 0.83).

Zero-order correlations
The same pattern of correlations identified in Study 3a emerged in the current study (Table 5).Among both groups, collective narcissism and ingroup identification were positively related.Among the advantaged group (Whites), both collective narcissism and ingroup identification were related to more support for the Alt-Right and less support for BLM.Among the disadvantaged group (Blacks), collective narcissism and identification were related to more support for BLM and were not linked to support for the Alt-Right.

Multiple regression analysis on support for reactionary social movement
For the following analyses, collective narcissism and ingroup identification were mean-centred.Group status was coded 0 for the advantaged group (baseline) and 1 for the disadvantaged group.We ran a multiple regression analysis with ethnic collective narcissism, ingroup identification, group status, collective narcissism × group status, and ingroup identification × group status as predictors of support for the Alt-Right movement, R 2 = 0.09, F(5, 879) = 17.96, p < .001(Table 6).Increase in R 2 due to the interactions was small but close to significance, ΔR 2 = 0.01, F(2, 879) = 3.00, p = .050.
Results replicated the pattern observed in Study 3a.The main positive effect of collective narcissism on support for the Alt-Right movement was moderated by group status, ΔR 2 = 0.004.Simple effects revealed that collective narcissism was associated with more support for the Alt-Right among the advantaged group, β = 0.29, 95% CI [0.22, 0.36], p < .001,but not among the disadvantaged group, β = 0.04, 95% CI [-0.20, 0.27], p = .765.Results also highlighted a positive association between ingroup identification and Alt-Right support, which was not moderated by group status, ΔR 2 = 0.001.

Discussion of Study 3b
Study 3b showed similar effects as Study 3a concerning support for reactionary movements, with a positive association between collective Studies 3a and 3b used feelings towards social movements, which are a more subtle measure of support than the ones used in Studies 1 and 2. This measurement difference may explain the simple effect of secure identity on support for the Alt-Right among members of advantaged groups, which may be related to a positive feeling but not to explicit support.In addition, Studies 3a and 3b relied on a single-item measure of identification.Although this measure has been shown to be reliable in previous research (e.g., Postmes et al., 2013), it does not fully account for the multidimensional structure of identification, and particularly the self-investment dimension involved in secure identity (Marchlewska et al., 2020).The inability to capture the complex, multidimensional nature of group identification may have contributed to the inconsistent findings across studies among disadvantaged group members.Study 4 aimed to replicate Studies 3a and 3b while addressing these limitations.

STUDY 4
As in Studies 3a and 3b, Study 4 investigated support for reactionary (Alt-Right) and progressive (BLM) movements among ethnic advantaged and disadvantaged group, that is, White and Black Americans.
In Study 4, we used a measure of actual support for the targeted social movements, as well as a multidimensional measure of identification.
Study 4 was pre-registered (https://aspredicted.org/LTK_M4W) with the same hypotheses as before: we predicted that collective narcissism would be related to stronger support for pro-ingroup social movement and less support for pro-outgroup social movements among members of both advantaged and disadvantaged groups.Furthermore, secure ingroup identity should be related to support for progressive social movement, especially among members of the disadvantaged group.

Participants and procedure
The sample included 608 participants4 recruited via Prolific.We kept participants who reported being White (n = 315) or Black (n = 272).
The final sample consisted of 587 Americans (254 males, 320 females,

Collective narcissism
Ethnic collective narcissism was measured with the same items as in Studies 3a and 3b, on a scale from 1 (Strongly disagree) to 7 (Strongly agree), α Advantaged = 0.87, α Disadvantaged = 0.80.

Ingroup identification
Ethnic ingroup identification was measured with the same ten selfinvestment items from Leach et al. (2008) used in Study 2 (e.g., "I feel a bond with other people of my ethnic group"), rated on a scale from 1 (Definitely not) to 7 (Definitely yes), α Advantaged = 0.94, α Disadvantaged = 0.92.

Support for social movements
Support for social movements was measured with four items.We used the single-item from Studies 3a and 3b and added three items adapted social media?(e.g., Facebook, Twitter, etc.)", rated on a scale from 1 (extremely unlikely), to 7 (extremely likely), with an additional answer "I do not use any social media" (for participants who chose this option the (3) "How likely are you to participate in a [Alt-Right/Black Lives Matter] protest in the future?",rated on a scale from 1 (extremely unlikely) to 7 (extremely likely).The order in which the Alt-Right and BLM items were presented was counterbalanced.Both measures had a good internal reliability: support for the Alt-Right, α Advantaged = 0.98, α Disadvantaged = 0.98; support for BLM, α Advantaged = 0.93, α Disadvantaged = 0.92.

Zero-order correlations
Descriptive statistics and correlations can be found in Table 7.As in previous studies, collective narcissism and identification were positively correlated among both the disadvantaged and the advantaged groups.Among the advantaged group, collective narcissism and identification were positively related to support for the reactionary movement, and negatively related to support for the progressive movement.
Among the disadvantaged group, collective narcissism was positively associated with support for both reactionary and progressive movements, whereas identification was solely related to more support for progressive movements.

Multiple regression analysis on support for reactionary social movement
For the following analyses, collective narcissism and ingroup identification were mean-centred.Group status was coded 0 for the advantaged group (baseline) and 1 for the disadvantaged group.We conducted a multiple regression analysis with collective narcissism, ingroup identification, group status, collective narcissism × group status and ingroup identification × group status as predictors of support for the Alt-Right, R 2 = 0.36, F(5, 453) = 50.12,p < .001(Table 8).Increase in R 2 due to the interactions was small and significant, ΔR 2 = 0.02, F(2, 453) = 5.83, p = .003.
Collective narcissism had a main positive effect on support for the Alt-Right.Despite an increase in R 2 similar to previous studies, ΔR 2 = 0.004, the interaction between collective narcissism and group status was only marginally significant.We observed a positive association between collective narcissism and support for the Alt-Right among the advantaged group, β = 0.81, 95% CI [0.62, 1.00], p < .001,and a positive (although weaker) association among the disadvantaged group, β = 0.49, p = .001,95% CI [0.21,0.78].Results also revealed a non-significant main effect of ingroup identification on support for Alt-Right.As in previous studies, the interaction between group status and ingroup identification was not significant, ΔR 2 = 0.001.

Discussion of Study 4
Consistent with previous studies, Study 4 found a positive link between collective narcissism and support for reactionary movements in the advantaged group.Surprisingly, this positive association was also found, albeit marginally weaker, among members of the disadvantaged group.This suggests that, even among disadvantaged groups, collective narcissism may be linked to conservative and hierarchy-enhancing ideologies (Golec de Zavala et al., 2009;Lantos & Forgas, 2021).
Furthermore, although the interaction patterns were consistent with other studies regarding support for BLM, the simple effects differed.
As hypothesised, collective narcissism and identification were linked to greater support for progressive movements among members of disadvantaged groups, highlighting that both a defensive and a secure form of identity were associated with support for this progressive movement-a movement that is both pro-ingroup and pro-equality.
However, in contrast to previous studies, only secure identity (but not collective narcissism) was related to less support for the progressive social movement among the advantaged group.

INTERNAL META-ANALYSIS
Although congruent with our hypotheses and broadly consistent, some results differed across studies.We conducted an internal metaanalysis to test the consistency of the associations that collective narcissism and ingroup identification had with support for both progressive and reactionary movements in advantaged and disadvantaged groups.

Analysis and data preparation
We used a one-stage Meta-Analytic Structural Equation Modeling (MASEM; Jak, 2015;Jak et al., 2021).This method is a random-effects technique and combines the methods of meta-analysis and SEM to fit a path model from a set of correlation matrices.This analysis also allows for the test of a moderator which, in our case, was group status.
We conducted the analysis in R, using the metaSEM package (Cheung, 2015).The one-stage MASEM was based on eight correlation matrices (one for Studies 1 and 2, and two, one per group, for Studies 3a, 3b and 4).For Study 2, Spearman correlations were calculated for correlations involving support for progressive and reactionary movements, treating these variables as ordinal variables with three increasing modalities (no support, support, participation).

Results
We tested the model positioning collective narcissism and identification as predictors of support for progressive and reactionary movements.The model had a good fit, χ 2 (1) = 0.14, p = .705,RMSEA = 0.00, 95% CI [0.00, 0.03].We then added group status (advantaged groups coded 0 and disadvantaged groups coded 1) as a moderator of all effects.The omnibus test for moderator was significant, χ 2 (5) = 51.83,p < .001.Table 9 reports the between-study variances with and without the inclusion of the moderator.As can be seen, the inclusion of a moderator partly explained the heterogeneity between the studies.The homogeneity of the studies taking into account moderation by group status was satisfactory.

8.2.1
Meta-analytic effects on support for reactionary social movements Group status significantly moderated the association between collective narcissism and support for reactionary movements, β = −0.31,p = .031,such that the overall association was large and positive within the advantaged groups, β = 0. 36, p < .001,but there was no association within the disadvantaged groups, β = 0.06, p = .590.Group status did not moderate the effect of identification on reactionary movements, β = −0.13,p = .335,with a small and marginal positive effect among tive narcissism (but not secure identity) would be linked to greater support for, and involvement in, reactionary movements among the advantaged groups (ethnic and national majority).Our results and the internal meta-analysis largely confirmed this hypothesis.Indeed, collective narcissism in advantaged groups was linked to greater support for the ultra-right in Studies 1, 3a, 3b and 4, and support for a nationalist movement in Study 2. Taken together, the meta-analysis highlighted a large effect of collective narcissism on support for reactionary movements among the advantaged groups.In contrast, secure identity was either unrelated (Studies 1 and 4) or more weakly related than collective narcissism (Studies 2,3a,3b) to support for, and involvement in, these reactionary movements.When distilling these results across studies, the meta-analysis reported no overall significant effect.These Interestingly, our studies showed that collective narcissism in disadvantaged groups was either unrelated (Studies 3a and 3b), or positively linked (Study 4), to support for reactionary movements, leading to an overall non-significant effect in the meta-analysis.The fact that we did not observe a negative effect may be surprising, as reactionary movements can be seen as pro-outgroup for members of disadvantaged groups.Thus, this finding indicates that collective narcissism among disadvantaged groups is not related to less support for pro-outgroup movements, as it is the case among advantaged groups.
This somewhat mixed and overall non-significant association observed among disadvantaged groups may be due to the close linkage between collective narcissism and hierarchy-enhancing ideologies, such as right-wing authoritarianism or social dominance orientation (Cichocka et al., 2017;Golec de Zavala et al., 2009;Górska et al., 2020), and to conservatism (Cichocka et al., 2016;Lantos & Forgas, 2021).In national groups, collective narcissism is linked to reactionary attitudes that go beyond strict group defence and ultimately undermine group functioning, such as support for anti-environmental actions (Cislak et al., 2018) or anti-vaccination movements (Cislak et al., 2021).
Although few studies to date have investigated these links in disadvantaged groups, our results suggest that collective narcissism may have paradoxical effects.On the one hand, it would be linked to more support for one's disadvantaged group.On the other hand, it could also be linked to, or at least not opposed to, support for conservative ideas and movements that may conflict with the interests of the ingroup (see also Jost et al., 2017).This finding supports research showing that people with high levels of collective narcissism are more concerned with their own self-interest than with the interest of the group, possibly leading to attitudes and behaviours that undermine the ingroup and its members (e.g., disloyalty, objectification of ingroup members, Cichocka et al., 2021;Gronfeldt et al., 2022;Marchlewska et al., 2020).
Although collective narcissism among advantaged group members is strongly associated with support for reactionary movements, it could also be used as a brake on opposition to reactionary movements among disadvantaged group members.In this sense, our results provide additional explanations for the work of Osborne et al. (2019) who found that identification among New Zealand M āori was linked to both reactionary and progressive social movements.Our study highlights that this may be due to the defensive, narcissistic part of identity, whereas the secure part is consistently unrelated to support for reactionary social movements.This finding also suggests the importance of distinguishing between hierarchy-enhancing and hierarchy-attenuating motivations that underpin support for social movements beyond the social identity motives (in line with Becker, 2020).

9.2
The role of secure identity in reactionary and progressive social movements Given that a secure identity encompasses a broader egalitarian scope in addition to a constructive ingroup investment, we hypothesised that the relationships secure identity had with support for both reactionary and progressive movements would follow a similar pattern for advantaged and disadvantaged groups.We found results consistent with this hypothesis with regard to support for reactionary movements.Indeed, group status did not moderate the (non-significant) relationship between secure identity and support for reactionary movements.
Concerning progressive movements, we expected a stronger association with secure identity (although in the same positive direction) among disadvantaged than advantaged groups.This means that, among disadvantaged groups, we expected that both defensive and secure identities would predict greater support for progressive movements.Our results and the internal meta-analysis tend to support this hypothesis.Among disadvantaged groups, Studies 3a and 3b highlighted the role of both secure identity (Study 3a) and collective narcissism (Study 3b) in predicting support for progressive movements.
Study 4, which addressed important methodological limitations of Studies 3a and 3b, as well as the internal meta-analysis, demonstrated that collective narcissism and secure identity were both moderately and positively associated with support for progressive movements among disadvantaged groups.
Although collective narcissism may appear to be more in tune with social reality among disadvantaged groups than among advantaged groups, our results highlight that collective narcissism and ingroup identification explain a similar amount of variance in support for progressive movements.In other words, support for progressive movements on the part of disadvantaged group members may be motivated either by ingroup enhancement and a need for power (being related to a defensive identity) or by an investment to truly serve ingroup interests, compatible with intergroup harmony (being related to a secure identity).Further examination of the mediators between collective narcissism, ingroup identification and support for progressive and reactionary social movements could illuminate these processes.For example, examining the self-serving or benevolent motivation to support progressive social movements could help differentiate between the specific roles of defensive and secure identity (see Cichocka et al., 2021, Study 3).
As mentioned before, our results support our hypotheses that a secure identity among advantaged groups is weakly (and overall, not significantly) related to support for pro-ingroup reactionary movements, whereas collective narcissism is largely related to it.However, we did find an (unexpected) small negative effect of secure identity on support for progressive movements among advantaged groups.This suggests that, although egalitarian, these movements may be seen as threatening to advantaged groups, resulting in less support from those who are securely identified with their group (see Depuiset & Butera, 2003, for similar findings on constructive patriotism in situations of intergroup threat).
Taken together, our results identify similarities and differences in the processes underlying disadvantaged and advantaged groups' support for reactionary and progressive collective action.In both groups, support for, and engagement in, a pro-ingroup social movement (reactionary for the advantaged group and progressive for the advantaged group) is associated with a defensive identity, probably reflecting a need to reassert personal or group control (Cichocka et al., 2018;Guerra et al., 2022;Marchlewska et al., 2018).However, both defensive and secure identities correlated with support for the progressive movement among disadvantaged groups, suggesting both self-serving and benevolent motives, although these interpretations warrant further investigation.The effects of secure identity for the reactionary and progressive movements among advantaged groups were, as expected, rather weak.Thus, processes concerning collective narcissism and support for pro-ingroup movements appear to be similar between advantaged and disadvantaged groups, while the specificities concerning the role of secure identity need to be taken into account.

Limitations and perspectives
Although our studies provide important first insights into the origins of support for reactionary and progressive movements, there are some limitations that should be noted and investigated in future work.First, existing models position social identities as precursors to support for social movements and studies have shown this temporal link among advantaged and disadvantaged groups (Thomas et al., 2020).In our studies, we therefore assumed that collective narcissism and secure identity would predict support for reactionary and progressive movements.However, the reported studies are correlational and would require longitudinal data to verify this temporality.
Second, our studies are based on existing social movements, linked to the ultra-right, nationalism, or support for ethnic minorities and women.This choice increases the validity and ecological relevance of these studies but does not allow for a strict comparison between different social movements (see Thomas et al., 2020).For example, Alt-Right and BLM movements may differ in the extent to which they are perceived as pro-ingroup versus pro-outgroup by the advantaged and the disadvantaged groups.While BLM may be perceived as a pro-outgroup movement by White participants, the Alt-Right may be perceived as a conservative but less directly anti-Black by Black people, as suggested by the equal support for the Alt-Right between our White and Black participant samples.If this were the case, it might also partly explain the non-significant association between collective narcissism and support for reactionary movements among members of the disadvantaged groups.Thus, future studies could investigate the perception of these and other social movements among advantaged and disadvantaged groups to ensure comparability.
Third, although we provide evidence that defensive identity in advantaged groups is related to greater support for reactionary movements and less support for progressive movements in various contexts, further studies should be conducted, especially on disadvantaged groups.Replicating our findings in other intergroup and cultural contexts, and on other types of disadvantaged groups (e.g., women), is needed to demonstrate their generalizability beyond the US interethnic context.
Altogether, our studies highlight the diverse roles of collective narcissism and secure identity and demonstrate the need to take this distinction into account when examining differences between advantaged and disadvantaged groups' support for reactionary and progressive movements.Models that place identification as a major predictor of investment in collective movements and actions, particularly SIMCA, could therefore incorporate this distinction in future studies.This would also allow the distinct role of collective narcissism and secure identity to be explored in relation to the other SIMCA variables, namely, perceived injustice and perceived self-efficacy.Indeed, while both identification and collective narcissism are predictors of perceived injustice and deprivation (e.g., Golec de Zavala et al., 2017), studies on collective narcissism highlight that perceived injustice in itself can also be a predictor of collective narcissism (Guerra et al., 2022;Marchlewska et al., 2018).The processes at the heart of SIMCA could then differ between collective narcissism and secure identity.

CONCLUSIONS
Our research found that distinguishing between narcissistic and secure social identities helps us understand support for reactionary and progressive movements among advantaged and disadvantaged groups.
While collective narcissism is at the heart of the identity processes associated with support for reactionary and progressive movements among advantaged groups, both collective narcissism and secure identity are equally associated with support for progressive movements among disadvantaged groups.Our research therefore clearly illustrates the need to take this distinction and group status into account in future work on the identity processes at work in support for social movements.Understanding these processes is necessary to grasp the motives behind movements that can contribute to more egalitarian societies but also threaten the cohesion and even the democracy of countries.
tity and support for reactionary and progressive movements.Contemporary Poland can be considered an ethnically homogenous country, with 97% of the population declaring Polish nationality (Główny Urząd Statystyczny, 2015) and only a small proportion of non-White minorities.We therefore focused on this national majority and measured collective narcissism and ingroup identification relative to Polish identity.As in other countries, the last few years have been marked by a rise of the populist right in Poland.The ultraconservative Law and Justice party assumed power in 2015, after years of more liberal governance.Several reactionary policy decisions have been taken since then, including the almost total abolition of abortion and the creation of anti-LGBTQI+ zones.Reactionary social movements, such as the far-right Independence March, remain vivid in the country (Charlish & Ptak, 2020).These reactionary policies and movements gave rise to the development of large-scale progressive social movements, such as the National Women's Strike in defence of women's rights.Our study draws on this context to examine the role of collective narcissism and secure identity in support for progressive and reactionary movements.1134 adults was recruited by a polling institute in Poland (Instytut Badań Pollster) to take part in a survey on different political topics.The sample was representative in terms of gender, age, settlement size and education, using a quota based on the Central Statistical Office data.We excluded participants who reported not knowing the target social movements (see the Measures section below).The final sample contained 989 participants (515 women, 474 men), aged 18 to 81 (M = 46.75,SD = 15.85).

from
Selvanathan et al. (2018): (1) "To what extent do you support or oppose the [Alt-Right/Black Lives Matter] movement?",rated on a scale from 1 (strongly oppose) to 7 (strongly support), (2) "How likely are you to show your support for [the Alt-Right/Black Lives Matter] through results build upon Thomas et al. (2020) highlighting the role of identification in supporting reactionary movements, and point to the key effects of the defensive (narcissistic) identity.Our work demonstrates that future work must acknowledge the critical distinction between secure and defensive forms of group identification when examining the predictors of social movements and collective action.
10990992, 2022, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejsp.2885 by Test, Wiley Online Library on [17/04/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 10990992, 2022, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejsp.2885 by Test, Wiley Online Library on [17/04/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)onWiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehelp elucidate its role in support for progressive and reactionary social movements, and the similarities and differences among the advantaged and the disadvantaged groups.
Multinomial logistic regression on the reactionary and progressive social movements (Study 2) Leach et al.'s (2008)lf-investment (and not the self-definition) dimension ofLeach et al.'s (2008)ingroup identification scale.10990992, 2022, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejsp.2885 by Test, Wiley Online Library on [17/04/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License TA B L E 2 10990992, 2022, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejsp.2885 by Test, Wiley Online Library on [17/04/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License TA B L E 3 Descriptive statistics and correlations between variables per group (Study 3a) Descriptive statistics and correlations between variables per group(Study 3b) The sample contained 1000 participants recruited by the same company used in Study 3a (CINT).As for Study 3a, participants who failed 10990992, 2022, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ejsp.2885 by Test, Wiley Online Library on [17/04/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License TA B L E 5 Multiple regression analyses on support for reactionary and progressive social movements(Study 3b) TA B L E 6 Descriptive statistics and correlations between variables per group (Study 4) Note.Correlations above the diagonal refer to the advantaged group, correlations below correspond to the disadvantaged group.* p < .05,** p < .01,*** p < .001item was not included in the calculation of support score, n = 19), and TA B L E 7 Multiple regression analyses on support for reactionary and progressive social movements (Study 4) Note.Between-studies variances (τ 2 ) of the model without the group status as moderator are above the diagonal, and τ 2 of the model with group status as moderator are below the diagonal.