When Psychological Contract Is Violated : Revisiting the Rejection-Disidentification Model of Immigrant Integration

In this study, we investigated how perceived ethnic discrimination is related to attitudes towards the national majority group and willingness to confront injustice to promote the social standing of a minority group. We examined this relationship via two mediating factors; national (dis)identification from and out-group (dis)trust of the national majority group. The Rejection-Disidentification Model (RDIM) was refined, first, to account for willingness to confront injustice as a consequence of perceived rejection, and second, intergroup (dis)trust was examined as an additional mediating mechanism that can explain attitudinal and behavioural reactions to perceived rejection simultaneously with national disidentification. The model was tested in a comparative survey data of Russian-speaking minority in Estonia (N = 482), Finland (N = 254), and Norway (N = 219). In all three countries, the more Russian-speakers identified as Russians and the more they perceived ethnic discrimination, the more negative were their attitudes toward the national majority groups and the more willing they were to engage in action to confront group-based injustice. Whereas disidentification from and distrust of national majority group accounted for the discrimination-attitude link to a large extent, both factors had demobilizing effects on willingness to confront injustice, making Russian-speaking immigrants more passive but hostile. The findings are discussed in relation to the risks involved in politicization of immigrants struggling with perceived inequalities.

tegration does not seem possible, immigrants and ethnic minorities resist their group disadvantage to improve their social standing in society (e.g., Klandermans, van der Toorn, & van Stekelenburg, 2008).Significantly less efforts have, however, been made to uncover mechanisms that hinder disadvantaged minority group members' willingness to engage in intergroup confrontation and actions aimed to improve the social standing of their group when perceiving injustice.
This study aims to specify the attitudinal and behavioural intentional consequences of perceived ethnic discrimination among Russian-speakers in three neighbouring countries: Estonia, Finland, and Norway.At a policy level, immigrant integration is understood in all these countries as the possibility of immigrants to live as active, full members of the society (Estonian Ministry of Culture, 2014;Finnish Ministry of Economic Affairs and Employment, 2017;Norwegian Ministry of Children, Equality and Social Inclusion, 2012).In these documents, mutual efforts, a common base for identity, and mutual trust are seen as imperatives to achieve the goal of integration.In this study, we show how critical but also controversial these factors are for understanding intergroup relations from a minority groups' perspective.
At a theoretical level, this study approaches integration as a psychological contract established between minority and majority group members.It shows that perceived ethnic discrimination violates this contract and has attitudinal and behavioural consequences for minority's integration and relations with the national majority.The concept of psychological contract is commonly used in organizational psychological literature on employees' organizational behaviour.A psychological contract (PC) consists of individuals' beliefs regarding the terms and conditions of an exchange agreement between themselves and their organizations (Rousseau, 1989).According to Turnley and Feldman (2000, p. 25), "of critical importance in the establishment of any psychological contract is the belief that a 'promise' (either implicit or explicit) has been made and that a 'consideration' has been offered in exchange for it".Psychological contract violation (PCV), in turn, has been defined as the employees' perception that the organization has failed to fulfil one or more of its obligations as defined by the psychological contract (e.g., Robinson, Kraatz, & Rousseau, 1994;Robinson & Morrison, 1995;Robinson & Rousseau, 1994), and is accompanied by a negative emotional or affective state (Morrison & Robinson, 1997).
In a similar vein, in acculturation research, expectations of positive intercultural contact are central to the understanding of intercultural relations between acculturating groups (Berry, 2013).In societies valuing and promoting multicultural ideology, like in the contexts of this study, these expectations are often voiced in demands for mutual respect and equality.We see such expectations as a contract for integration.Immigrants will integrate if they receive the equality and respect they have been promised.If they face unjust treatment such as ethnic discrimination, in contrast, the contract is broken, preventing integration and leading to intergroup confrontation and ethnic minority activism.Ethnic minority activism aims at modifying norms or practices established by a majority group, thus challenging unequal treatment and discrimination based on the ethnic categorisation of individuals (Pereira, Green, & Visintin, 2017).
In our theoretical model, we combine social and organizational psychological research on the consequences of perceived social disadvantage and psychological contract violation.We aim to complement and extend the Rejection-Disidentification Model (RDIM; Jasinskaja-Lahti, Liebkind & Solheim, 2009; see also Verkuyten & Yildiz, 2007), which suggests that immigrants react to perceived discrimination with negative attitudes towards the discriminating majority, and that national disidentification mediates this association.Particularly, we add an additional mediating variable related to the psychological contract literature: (dis)trust of the national majority.We suggest that perceived ethnic discrimination compromises immigrants' attitudes toward the majority group members and along with strong ethnic identification mobilises immigrants to resist discriminatory treatment and to improve the social position of their minority in-group.We further suggest that disidentification from and distrust of the national majority group may explain why immigrants are often less than willing to confront perceived injustice.

Theoretical Background
Perceived Rejection, Minority Identification, and Intergroup Outcomes Social psychological research on contact hypothesis (Allport, 1954) has recently disentangled the effects of positive and negative intergroup contact on minority groups' attitudes toward the majority and their collective efforts to combat inequality.It has shown a "sedating" effect of positive intergroup contact (Pereira et al., 2017).Positive intergroup contact improves disadvantaged group members' attitudes toward an advantaged group and decreases their efforts to mobilize for social change (e.g., Dixon, Tropp, Durrheim, & Tredoux, 2010;Tausch, Saguy, & Bryson, 2015;Tropp, Hawi, Van Laar, & Levin, 2012).Negative contact, along with perceived disadvantage and discrimination, results, in turn, in negative attitudes toward the advantaged majority group (e.g., Jasinskaja-Lahti et al., 2009) and also leads to more support of and engagement in collective action aimed to improve the in-group's status and situation (e.g., Tausch et al., 2015).The recent study by Hayward, Tropp, Hornsey, and Barlow (2018) found evidence for these two effects of contact on collective action.It showed that while negative contact with the majority and perceived ethnic discrimination are potential drivers, positive contact is a potential inhibitor of social change among ethnic minorities.
Moreover, research informed by the social identity approach (Tajfel & Turner, 1979, 1986;Turner, Hogg, Oakes, Reicher, & Wetherell, 1987) to intergroup processes has shown that perceptions of in-group disadvantage, group based injustice, discrimination and reactions to these are intertwined with and motivated by one's identification with an oppressed minority group (Dovidio, Gaertner, Ufkes, Saguy, & Pearson, 2016).According to Pereira et al. (2017), ethnic identification drives ethnic activism implying recognition of one's group's social disadvantage and motivation to improve the status of the in-group.This line of research follows up Tajfel's (1981, p. 244) notion that social movements should be understood as "efforts by large numbers of people, who define themselves and are also often defined by others as a group, to solve collectively a problem they feel they have in common, and which is perceived to arise from their relations with other groups".The insight that intergroup confrontation and social movement participation is motivated by a collective identification has recently been incorporated as a central element in the social identity model of collective action (SIMCA) by van Zomeren, Postmes, andSpears (2008, 2012).It also forms the core of the social psychological conceptualisation of politicised collective identity (PCI; Simon & Klandermans, 2001).According to Simon and Klandermans (2001), in order for collective identity to politicise, group members should be aware of their shared group memberships, their common opponent, and the fact that it is the wider societal context in which the struggle must be fought out.
Social identification and the motive to establish and maintain positively valued group identities are fundamental to intergroup behavior (Tajfel & Turner, 1979, 1986).Thus, people attempt to defend the value of an important group membership when it is directly attacked by out-group, and when that happens, high identifiers are more likely to think, feel and act for the group (van Zomeren, Kutlaca, & Turner-Zwinkels, 2018) and strike back at the group they perceive as representing the threat (Branscombe, Ellemers, Spears, & Doosje, 1999).It is thus not surprising that group identification has also been given a central place in SIMCA (van Zomeren et al., 2018).As regards empirical research among minorities, Simon et al. (1998) were amongst the first to show that identification with the older people's and gay movements in Germany and the U.S. respectively predicted the willingness of members of these minority groups to participate in collective action to improve the position of their groups.Similarly in a study by Fleischmann, Phalet, and Klein (2011), Turkish and Moroccan Muslims who perceived more discrimination identified more strongly and were most ready to support political Islam and engage in political action to defend Islamic values.Further, in a more recent study by Ufkes, Dovidio, and Tel (2015), a stronger Kurdish identity positively predicted collective action intentions among Dutch and German Kurds.However, while previous research clearly suggests the association between perceived discrimination and ethnic identification on the one hand and the negative link between perceived discrimination and intergroup confrontation on the other hand, the evidence for a direct association between minority identification and attitudes towards the majority is more mixed: the association could be positive, negative, or zero (Duckitt, Callaghan, & Wagner, 2005).In the present study, we suggest that both perceived ethnic discrimination and ethnic identification are simultaneous predictors of willingness to confront injustice among minority group members.When it comes to predictors of out-group negativity among immigrants, we, nevertheless, account for a possibility that ethnic identification could motivate to confront injustice without directly compromising attitudes towards the majority.

Perceived Rejection, National Identification, and Intergroup Outcomes
It has been suggested that for ethnic minority activism, alongside minority subgroup identification (Dovidio et al., 2016), superordinate or common (in this case, national) identification is also needed (e.g., Klandermans, 2002;Simon & Grabow, 2010).As noted by Simon and Klandermans (2001; see also Wenzel, 2000), only by virtue of their membership in a broader community are minority group members entitled to societal support for their claims.
As national subgroups, ethnic minorities can identify with, that is feel attached to, both ethnic and national groups (Pereira et al., 2017).Feelings of belonging within the common superordinate group could thus empower minorities and facilitate their willingness to take action for change on behalf of the minority group (Simon, Reichert, & Grabow, 2013;Simon & Ruhs, 2008).
For example, Greenaway, Quinn, and Louis (2011) conducted an experimental study showing that emphasizing common humanity helped to promote more positive intergroup attitudes at the same time as reducing victims' intentions to engage in collective action to promote social change.Saguy et al. (2009) also studied the implications of common group identification for intergroup attitudes, perceptions of (in)equality, and support for social change among minority group members in laboratory contexts, as well as among Israeli Arabs and Indian Muslims.They found a positive relationship between common group identification and positive attitudes towards the majority and perceptions of fairness among minority group members.Importantly, they also attested that positive contact was followed by their reduced awareness of inequality and support for social change (for similar effect of positive contact see also Hayward et al., 2018).In a recent longitudinal study that Mähönen and Jasinskaja-Lahti (2015) conducted in Finland, strong Finnish national identification made Ingrian Finnish immigrants perceive intergroup boundaries as more permeable, which made them less inclined to improve their group's position in Finland.Pereira et al. (2017) have recently overviewed the research examining the effect of ethnic and national identification on ethnic activism among minorities and also concluded that both ethnic identification-involving separation and grievances-and national identification-bringing entitlement-reflect the collective identification of ethnic minority members and should thus be simultaneously considered when examining the effects of intergroup relations.
The above line of research suggests that while strong subordinate minority identification seems to pinpoint at unfair disparities and promote collective action, superordinate national identification could, in some contexts, be responsible for making disadvantaged group members more positively oriented toward the majority and accept inequality.National identification can increase minority members' expectations to be treated fairly and help them believe in social change (Chryssochoou & Lyons, 2011), making them prone to accept prevailing inequalities instead of engaging in collective action (Dovidio et al., 2016;Jaśko & Kossowska, 2013;Louis, 2009;Ufkes et al., 2015; for a recent review, see Saguy et al., 2017).Considering the controversy between these two lines of research showing opposite effects of superordinate national identification on willingness to confront injustice among minorities, in this study, we expect its positive effect on attitudes toward the majority, but make only an exploratory hypothesis for its effect on immigrants' support for collective action.We also account for possible contextual differences in the associations studied.As will be described below, the intergroup contexts studied differ in the extent to which minority group members may have (ethnic) claims to be members of the national community.Based on the size of their ethnic community and presence in the country over generations, Russian-speakers in Estonia may perceive themselves as an integral part of Estonian society.Being only first-generation immigrants, Russians cannot pose such strong claims in either Finland or Norway.Therefore, we expect a less pronounced effect of national identification on willingness to combat injustice in Finland and Norway as compared to the more pronounced effect in the Estonian context.
Recent research has, however, largely neglected the interrelatedness between perceived disadvantage and willingness to identify with the superordinate national community in the first place.While perceived disadvantage and high minority subgroup identification often accompany or reinforce each other, perceived discrimination jeopardizes minority group members' identification with national superordinate group.Such an effect is captured by RDIM (Jasinskaja-Lahti et al., 2009) that also bases itself on the premises of SIT and argues that when minority group members face injustice and discrimination from the national majority group, their superordinate national identification decreases (disidentification), leading to more negative attitudes toward the majority group.Similar notions have been made by Klandermans et al. (2008) suggesting that when unequal treatment is experienced, national identification fuels minority members' disappointment.Verkuyten and Yildiz (2007) have also shown that perceived injustice makes minority members more likely to distance themselves from majority members and foster negative feelings toward the mainstream society.This reaction to perceived disadvantage was originally attested in longitudinal studies among Ingrian-Finnish returnees from Russia and from the former Soviet Union (Jasinskaja-Lahti et al., 2009;Jasinskaja-Lahti, Mähönen, & Ketokivi, 2012;Mähönen & Jasinskaja-Lahti, 2012).Since then, the RDIM and closely related theorisations have continued to gain support in studies among Norwegian-Pakistani and German-Turkish Muslims (Kunst, Tajamal, Sam, & Ulleberg, 2012), Sunni Muslim immigrants of Turkish origin in Germany and the Netherlands (Maliepaard & Verkuyten, 2018), refugees in the Netherlands and Spain (Bobowik, Martinovic, Basabe, Barsties, & Wachter, 2017), first-generation Latino immigrants in the United States (Wiley, Lawrence, Figueroa, & Percontino, 2013) and in a laboratory study with an international sample of young Muslims (Schmuck, Matthes, & Paul, 2017), only to list some of the most recent.
In this study, we apply RDIM (Jasinskaja-Lahti et al., 2009) to explain the intergroup confrontation or lack of it among the Russian-speaking minority in three Nordic countries.Specifically, we focus on immigrants' willingness to publicly confront perceived injustice (i.e., discriminatory treatment against their group) and to defend their minority rights.Simon et al. (1998) have argued for the need to focus on concrete experiences and actions of people in order to understand when and how members of disadvantaged groups engage in behaviour improving the situation of their groups.In their studies on the older people's movement (Gray Panthers) in Germany (Study 1) and the gay movement in the United States (Study 2), they operationalized collective action intentions as a willingness to publicly intervene with perceived social injustice and also found that it was significantly related to minority's collective identity.We argue that RDIM could assist us to further understand the possible demobilizing or 'ironic effect' of national identification on collective action intentions among immigrants (i.e., how in the face of perceived discrimination it hinders minority members' willingness to confront injustice).More specifically, we test whether national disidentification following discrimination (1) intensifies both out-group hostility and willingness to confront injustice (mobilizing effect), or (2) intensifies out-group hostility but inhibits intergroup confrontation, making minority group members more hostile towards the majority, but avoidant of confrontation and disarmed (demobilizing effect).
To the best of our knowledge, there is only one previous study conducted by Wiley et al. (2013) that explored the associations between ethnic-based rejection, national disidentification and political engagement.In their study, Latino immigrants in the United States, who perceived greater ethnic-based rejection from the majority Americans identified with Americans significantly less, and this decreased their willingness to engage in a political struggle on behalf of their ethnic group.Ethnic identification had rather an independent mobilizing effect on collective action in their study.Another study that deserves to be mentioned here is the one by Pereira et al. (2017), in which they found national identification to serve as a buffering factor of the demobilization process resulting from positive intergroup contact among Roma minority in Bulgaria.Our study mirrors their study in its attempt to show how national identification potentially compromises the mobilizing effect of discriminatory contact experiences.

Psychological Contract
In this study, we take the assumptions of RDIM even further and argue that national disidentification is not the only mechanism potentially explaining attitudinal reactions and willingness to confront injustice when group-based rejection is experienced.Previous research has emphasized the role of emotional mediators, such as anger (e.g., Hayward et al., 2018;Ufkes et al., 2015), fear (Dumont, Yzerbyt, Wigboldus, & Gordijn, 2003) or empathy (Stephan & Finlay, 1999) in predicting reactions to challenging intergroup encounters and disadvantage.The focus on emotional explanatory mechanisms is, however, challenged by the notion of the socio-functional approach to prejudice (Cottrell & Neuberg, 2005): perceived group threats and emotional reactions to them are group-and context-specific.To take this into account, we now take a closer look at the role of intergroup (dis)trust in explaining how perceived discrimination is linked with out-group negativity and willingness to confront injustice among the same minority group, Russian-speakers, in three different contexts.While trust has been acknowledged in organizational research on psychological contract violations (PCVs), it has not been sufficiently incorporated in previous social psychological models explaining reactions to prejudice and discrimination, despite its integral link to (national) superordinate identification.
By definition, "trust in a collective entity represents the extent of trust that an individual places in a collectivity with which she deals" (McEvily, Weber, Bicchieri, & Ho, 2006, p. 54).Brewer (2009) suggests that optimal group identities are bounded communities of mutual trust and generalized reciprocal cooperation.Trust thus accompanies in-group identification, but also characterises the relationships with out-groups.Out-group trust and generalized trust should not, however, be mixed here as they are two very different concepts: While out-group trust is an attitudinal measure of trustworthiness of a particular out-group, generalized trust is a general disposition of the extent to which one trusts (unfamiliar) others in general (van der Linden, Hooghe, de Vroome, & Van Laar, 2017).
Intergroup contact literature suggests that besides reducing prejudice directly, positive intergroup contact increases out-group trust and that way leads to more positive intergroup relations (Dovidio, Gaertner, & Kawakami, 2003;Kenworthy et al., 2016;Schmid, Ramiah, & Hewstone, 2014;Swart, Hewstone, Christ, & Voci, 2010;Tam, Hewstone, Kenworthy, & Cairns, 2009;Verkuyten, Thijs, & Bekhuis, 2010).Similarly, in organizational psychological research, trust in organization has been associated with enhanced communication, higher level of commitments within a relationship and lower levels of conflict as well as with higher satisfaction and lower likelihood to exit (for a review, see Gargiulo & Ertug, 2006).In contrast, when intergroup encounters are problematic and psychological contract between the minority and majority group on mutual intergroup respect is violated, as is the case in illegitimate treatment attributed to discrimination, it is reasonable to assume that not only identification with but also trust in the national group suffers.
Research on PCV (e.g., Niehoff & Paul, 2001;Pate & Malone, 2000;Pugh, Skarlicki, & Passell, 2003) as well as SIT oriented organizational research (e.g., Lipponen, Wisse, & Perälä, 2011;Fortin, Cojuharenco, Patient, & German, 2016) have attested that perceived unfair treatment and injustice have trust ramifications challenging cooperation in both interpersonal and intergroup interactions and other organization benefiting behaviours (see also Ferrin, Bligh, & Kohles, 2008).Previous social psychological studies on collective action only predicate that distrust of discriminating group mediates the relationship between the perception of injustice and collective action (Aryee, Budhwar, & Chen, 2002; see also Hagedoorn, Buunk, & Van de Vliert, 2002).However, as noted by Wright (2009) this mediating effect has never been assessed.There is only one recent study conducted among Russians in Finland (Brylka et al., 2015) attesting the mediating role of distrust in the relationship between perceived cultural discordance and collective action intentions among Russian-speaking immigrants in Finland.
As regards the hypothesized ramifications of (dis)trust in this study, there is no reason to assume other than positive effects of trust on attitudes towards the majority group.The effects of trust on collective action intentions could, however, have not only mobilizing but also demobilizing effects, just as like in the case of superordinate national identification (Brylka et al., 2015;Wiley et al., 2013).This alternative view on "the dark side of trust" has been noticed in organizational trust research.Excessive trust can lead to "blind faith" and cause a drop in the level of monitoring of the trustee's behaviour, push down the performance threshold for corrective action and lead to taking obligations that go beyond the initial psychological contract (for a review, see Gargiulo & Ertug, 2006).
In this study, we argue that perceived discrimination violates the integration contract between immigrants and the majority group.We also test whether both disidentification from and distrust of the national majority group are responsible for making immigrants more hostile towards the discriminating majority group, but also more disarmed when it comes to intergroup confrontation.

The Current Study
Based on the research presented above, in this study, we aim to refine the RDIM and argue that as a reaction to perceived ethnic discrimination, national disidentification is accompanied by out-group distrust, with both having further attitudinal and behavioural ramifications (i.e., out-group negativity and willingness to confront injustice, respectively).
To sum up the previous research, perceived discrimination accompanied by strong identification with one's ethnocultural subgroup has been linked to willingness to participate in social action promoting the rights and social status of immigrants.Although identification with the larger society has been emphasized by many integration ideologies and models of collective action as a necessary element for engagement in social change, empirical research shows that it has both mobilizing and demobilizing effects on collective action intentions among minorities.
The phenomenon is even more complex when considering that disadvantage and discrimination may be associated with national disidentification among ethnic minority group members.Although the consequences of national disidentification for negative attitudes toward the national majority group have been proven in research based on the RDIM, the model has not been tested using willingness to confront injustice as an outcome.Intensifying negative out-group attitudes, national disidentification may disarm minorities and discourage them from willingness to confront injustice, or, in contrast, mobilize them to look for possibilities for intergroup confrontation aimed at social change.Finally, as shared identities presuppose mutual trust and reciprocal cooperation (Brewer, 2009), it is a reason to assume that the disidentification from the superordinate national group as a reaction to perceived rejection is most likely to be accompanied by distrust of the majority group.So, the potentially similar ironic role of distrust as a simultaneous parallel mediator in the relationship between perceived discrimination and disadvantaged identity on the one hand and intergroup attitudes and collective action intentions on the other hand should be considered when testing the premises of RDIM.
The assumptions of the study are summarized in Figure 1.It should be noted that two alternative reactions (mobilizing and demobilizing) to disidentification and distrust are explored.In addition, there is no particular hypothesis for the relationship between ethnic identification and attitudes towards the majority, as it is shown to be highly group and context specific (Brewer, 1999) and therefore needs to be separately explored in each context.Thus, only the positive association between ethnic identification and perceived discrimination and their parallel positive effects on intergroup confrontation could have been clearly hypothesized.
The refined RDIM was tested in three culturally close but historically and politically different countries sharing the same diasporic community: Russian-speakers in Estonia, Finland, and Norway.All three countries share borders with Russia, and all have, to some extent, been influenced culturally, demographically, economically, and politically by Russia.As described below, the degree of Russian influence may be a result of the size of Russianspeaking population in the different countries and the degree of current and previous political relationships with Russia.
The major peculiarity of Estonia, as compared to Finland and Norway, is that as many as one third of the current population is of Russian origin.During the Soviet occupation  the percentage of Russian-speakers in Estonia grew from 26,000 to 602,000 (Vetik, 1993).Regaining independence in 1991 has brought fundamental social and political changes having a major impact on the interethnic issue and integration prospects of Russianspeakers.These changes are related to the new nation-building processes along the lines of pre-Soviet Estonian Republic.For example, the Estonian Parliament adopted the citizenship law, which establishes legal continuity with the Estonian Republic of 1918-1940.The law was exclusive in the sense that automatic citizenship was granted only to those residents and their descendants who were citizens of Estonia before Soviet occupation, which has resulted in the majority of Russians in Estonia becoming stateless.This has a negative impact on their status in the labor market and in the society (Vetik & Helemäe, 2011).As a result, the Russian-language minority confronts an adverse national majority group and tends to counter it through strong defensive identities (Nimmerfeldt, 2009;Vetik & Helemäe, 2011).
In Finland, in turn, until the collapse of the Soviet Union in the 1990s, the Russian-speaking population was quite small.Russian-speakers currently represent the biggest immigrant group with around 22% of those speaking a foreign language as their mother tongue.Nevertheless, Finland is still characterized by a relatively small immigrant population (6%) (Official Statistics of Finland, 2016).Largely due to historical conflicts between Finland and the former Soviet Union, intergroup relations between the Finnish majority and Russian-speaking immigrants have been characterized by considerable discrimination towards this minority (Jaakkola, 2009;Larja et al., 2012).
Compared to other immigrant groups, Russians have belonged to the least welcome groups alongside with Arabs and Somali immigrants (Jaakkola, 2009).Despite being highly educated, the employment rate of immigrants from Russia and the former Soviet Union is considerably lower (61%) than that of native Finns (Nieminen, Sutela, & Hannula, 2015).
Norway inhabits the smallest community of Russian immigrants as compared to Estonia and Finland.Prior to 2000, there were less than 2000 Russian-speakers living in Norway.However, a decade after the break of the Soviet-Union the numbers of Russian speakers in Norway started to increase.There are over 20,000 Russians living in Norway, making them the 12th largest national group in the country (as of 2016 -Statistics Norway, 2017).The majority of the Russians in Norway (ca.17,000) are first generation; and over half of them (57%) are women; having arrived in Norway to join their native Norwegian partners as part of the family reunification program.

Method Participants and Procedure
The data from three countries were from the collaborative MIRIPS-FI and DIMA projects, with both being members of a larger international MIRIPS research network.Data from Estonia was collected among Russians living in Estonia in May 2015 and the random sampling was done based on the previous 2011 census data from Estonia.
The majority (81%) of the participants represented second or third generation of Russians.Data from Finland was collected in 2012 and the survey was sent to a representative sample of Russian-speaking immigrants by postal mail.The selection criteria were: Russian as the mother tongue of participant, currently living in Finland but born in the Soviet Union/the Russian Federation and moved to Finland before 2008.The Norwegian data was collected between November 2015 and January 2016.The participants were sampled from the Norwegian data registry that contains the names and addresses of all legal residents in Norway.Individuals who identified themselves as Russians and were between the ages of 18 years and 60 were randomly selected and individually invited by phone to take part in the study.Upon agreeing to participate in the study, an email with a link to the electronic questionnaire was sent to them.
The response rate was 61% for Estonia, 39% for Finland, and 33% for Norway.In the final sample, only participants who either reported their ethnic background as Russian or reported their mother's and/or father's ethnic background as Russian were included.The number of participants that did not have a Russian background, and hence were excluded, was 19 in Estonia, 60 in Finland, and 33 in Norway.The final sample included 955 participants (N = 482 for Estonia, N = 254 for Finland, and N = 219 for Norway), 65% of them were women (54% in Estonia, 78% in Finland, and 75% in Norway) and the mean age was 45.5 years, SD = 14.03 (M = 48.2,SD = 15.56 for Estonia, M = 43.9,SD = 11.93 for Finland, and M = 40.9,SD = 11.00 for Norway).The gender (χ 2 = 52.33,p < .001)and age distributions (F(2, 952) = 23.72,p < .001)differed between the country samples, with the Estonian sample including more men and older participants than the Finnish and Norwegians samples.More than half (66%) of the participants reported that they had at least 15 years of education (54% for Estonia, 67% for Finland, and 92% for Norway) and 8% of them were unemployed (4% for Estonia, 15% for Finland, and 11% for Norway).Gender, age, and education of the participants were later controlled for when testing the hypotheses of the study.Immigrant generation was not included into the analyses as a separate covariate, but rather accounted for when discussing the results of the study.

Measures
The measures used in this study were partly derived from the survey instrument developed for the international MIRIPS project or specifically developed/adapted from existing sources by Finnish, Estonian and Norwegian collaborators of that project to enable the three country comparisons.
Attitudes towards national majority group were measured with a single item, a so called "feeling thermometer" (Haddock, Zanna, & Esses, 1993; see also Velasco González, Verkuyten, Weesie, & Poppe, 2008), that is considered to be a global measure of in-group and out-group feelings (Verkuyten, 2007).The measure invites participants to rate their feelings towards a majority group: "My feelings towards <MAJORITY> in <COUNTRY> are…" on a scale from 0 = "very cold feelings" to 100 = "very warm feelings".collective action and tapped individual intentions to actively promote the minority rights and publicly stand against the discrimination of immigrants.The importance of so-called "speaking out" when confronting unfairness has also been acknowledged in the behavioural intentions measure of Louis, Duck, Terry, Schuller, and Lalonde (2007).

Willingness to confront injustice
These two items were preferred over those measuring more active forms of collective action behaviour (e.g., participation in demonstrations/campaigns or signing petitions) as in the contexts studied such forms of collective action among immigrants and ethnic minorities in general, and the Russian-speaking group in particular, are rather rare.Participants rated their level of willingness on a 4-point scale ranging from 1 = "No / Totally disagree / Definitely not ready to" to 4 = "Yes / Totally agree / Definitely ready to" in Finland and Estonia.Same measure was rated on a 5-point scale in Norway, including the same response options and "3 = Nor disagree or agree".The response ratings were converted to standardized scores to ensure compatibility across scales and cultures (α = .71overall; α = .65for Finland, α = .60for Estonia, α = .73for Norway).
National and ethnic identification were measured using 4 items each adapted from Mlicki and Ellemers (1996) and from Phinney and Devich-Navarro (1997).The items covered the cognitive ("I feel that I am a member of <HOST> society / Russian-speaking minority") and affective ("I am proud to be a member of <HOST> society / Russian-speaking minority", "I am glad to be a member of <HOST> society / Russian-speaking minority") aspects of national/ethnic identification, as well as commitment to the host society/minority in-group ("It is important for me to be a member of <HOST> society / Russian-speaking minority").Response categories ranged from: 1 = "Strongly disagree" to 5 = "Strongly agree".The reliability coefficients for National identification scale were: α = .89overall; α = .92for Estonia, α = .89for Finland, α = .84for Norway and for Ethnic identification scale: α = .87overall; α = .87for Estonia, α = .85for Finland, and α = .89for Norway.
Trust in the majority group was measured using 3 items adapted from Paolini, Hewstone, and Cairns (2007): "In my opinion most of the <MAJORITY> are trustworthy", "I think that most <MAJORITY> would treat me fairly even if they had a chance to take advantage of me", and, "<MAJORITY> won't take advantage of me if I trust them." Participants responded to the items on a 5-point scale that ranged from 1 = "Strongly disagree" to 5 = "Strongly agree" (α = .85overall; α = .87for Estonia, α = .75for Finland, and α = .87for Norway).
Due to the demographic differences between the three samples and their possible effects on the dependent variables studied, participants' age, gender, and education were included into the analyses as covariates.The survey also included other scales not utilized in this study: satisfaction with life, support for multicultural ideology, self-esteem, use of Russian language at home, appreciation of Russian culture, own and perceived acculturation preferences, intergroup anxiety, and perceived socio-economic status of the in-group (one item).A part of the data (using the Finnish and Estonian samples only) has been previously analysed and reported by Renvik, Brylka, Konttinen, Vetik, and Jasinskaja-Lahti (2018).

The Analysis
In order to test our predictions, we utilised SEM approach with robust maximum likelihood estimator in the Mplus software (Version 7.3).Alongside the Chi-square statistics, the model fit was evaluated by referring to widely used fit indices known as CFI, TLI, RMSEA, and SRMR (Hu & Bentler, 1999).Initially, we performed a confirmatory factor analysis (CFA) by including the items of ethnic and national identification scales, and the items of trust, perceived discrimination, and willingness to confront injustice scales and tested whether the data fits to the fivefactor solution, meaning that each item would reliably fall under the respective latent factor (ethnic identification, national identification, out-group trust, perceived discrimination, and willingness to confront injustice).Later on, we re-did the analysis by forcing the items to load on three factors, in order to see whether they represent our data better than the proposed five-factor solution, and we also did a one-factor model as a baseline model to compare against the other two models.After gaining evidence for the distinctiveness of our five latent constructs, we build our structural model by including two independent variables (perceived discrimination and ethnic identification), two mediators (national identification and trust), and two dependent variables (out-group attitudes and willingness to confront injustice).
Before proceeding with a multi-group analysis, multi-variate measurement invariance tests were performed to see if the five-factor solution model holds across the three countries: Estonia, Finland, and Norway.Finally, in order

The Mediation Model
In our suggested mediation model (Figure 2), national identification and trust in the majority group were predicted by perceived discrimination and ethnic identification.Out-group attitudes and willingness to confront injustice were predicted by national identification and trust, and we expected direct and indirect association between the two independent variables (perceived discrimination and ethnic identification) and the two dependent variables (outgroup attitudes and willingness to confront injustice).The two independent variables (perceived discrimination and ethnic identification), two mediators (national identification and trust) and two dependent variables (out-group attitudes and willingness to confront injustice) were also allowed to covary with each other.The model was tested with and without the control variables (age, gender, and education), but as the majority of the significant results remained with controls included, for the sake of clarity, we only discuss the results of the model without any control variables, and report changes in the significance level of some paths when controls were added.
The model fit was very good, χ 2 (105, N = 955) = 224.62,RMSEA = .04,CFI = .98,TLI = .97,SRMR = .03.As can be seen in Figure 2, and in line with our hypotheses, both perceived discrimination and ethnic identification were negatively associated with national identification and trust in the majority group.In turn, national identification was positively related to both out-group attitudes and willingness to confront injustice.Similarly, trust in the majority group was positively associated with both dependent variables.
To assess the mediating roles of national identification and trust on out-group attitudes and willingness to confront injustice, a bootstrap analysis with 1000 replications and with 95% confidence intervals was conducted.A mediation is considered significant when zero does not fall in-between the confidence interval (Preacher & Hayes, 2008).As shown in Table 2, in line with our hypotheses, the analysis demonstrated that the relationships between perceived discrimination and out-group attitudes and between perceived discrimination and willingness to confront injustice were both partly due to immigrants' national disidentification from and distrust in the majority group.National group disidentification and distrust thus strengthened immigrants' negative attitudinal reactions to perceived discrimination, but demobilized them from confronting injustice.As regards the associations between ethnic identification and dependent variables studied, the results showed very similar patterns as was for perceived discrimination.While the direct effect of ethnic identification on outgroup attitudes was positive, indicating that identifying more with the ethnic group is associated with more positive out-group attitudes, it was inhibited by national disidentification and distrust.In other words, higher levels of ethnic identification were indirectly related to more negative out-group attitudes due to national disidentification and distrust of the majority group.Similarly, the effect of ethnic identification on willingness to confront injustice followed a dual path: while its direct effect was positive, identifying more with the ethnic group was also responsible for less willingness to confront injustice due to its simultaneous association with lower levels of trust in the majority group.
The path between ethnic identification and willingness to confront injustice was the only one not mediated by national disidentification (Table 2).
The model was later on tested controlling for gender, age, and education by including these covariates one at a time.When the model was tested controlling for gender and education, all significant results remained the same.
When age was added as a control to the model, the indirect effect of ethnic identification on out-group attitudes via national identification, and the indirect effect of perceived discrimination on willingness to confront injustice via national identification lost their statistical significance.A further look at the role of age in our model revealed that age was intercorrelated with perceived discrimination (r = .09,p = .01),ethnic identification (r = .13,p < .001),national identification (r = -.13,p < .001),and out-group attitudes (r = .06,p = .05).This means that while age directly associated with one of the dependent variables (out-group attitudes), it could also partly account for the same variance in out-group attitudes and willingness to confront injustice as other variables (except trust) in the model.Moreover, the correlations between the independent variables and mediators were statistically significant with and without age as a covariate.This suggests that national disidentification and distrust were not age-specific reactions to perceived disadvantage.
Lastly, the proposed mediation model was rerun, this time including country as a moderator in order to see if the proposed associations are identifiable across the three countries.The model fit of the multigroup model was appropriate: χ 2 (361, NEstonia = 482, NFinland = 254, NNorway = 219) = 915.38,RMSEA = .07,CFI = .93,TLI = .92,SRMR = .07.The standardized coefficients of each path for each country separately can be seen in Figure 2. The results showed that for most of the paths, the directions were the same across countries even if the coefficients did not always reach statistical significance.There were a couple of notable differences, with none of them being, however, large enough to compromise the findings of the overall model above.Namely, the path between ethnic identification and trust in the majority group was a negative one for Finland, while it was a positive one for Estonia and Norway, but none of the effects were significant.Additionally, the relationship between trust in the majority group and willingness to confront injustice was negative, but non-significant for Finland, while it was positive and significant for Estonia and Norway.

Discussion
This three-country comparative study of Russian-speaking immigrants in Estonia, Finland, and Norway aimed at testing and further developing the Rejection-Disidentification Model (Jasinskaja-Lahti et al., 2009) to explain the attitudinal and behavioural reactions to perceived discrimination.The main contributions of the study are twofold.
First, it showed that perceived discrimination is reacted to not only with negative attitudes towards the discriminating majority, but also with intentions to confront injustice and publicly defend the rights of a minority group.Second, highlighting the complexity of these relationships and their significance for immigrant integration, we found that national disidentification from and distrust of the national majority have dual effects.They intensify the negative attitudes towards the majority but inhibit immigrants' willingness to confront injustice.In other words, national disidentification and distrust account for the relationship between perceived discrimination and out-group negativity, but demobilize immigrants from taking a more active position to improve their social standing.More specifically, perceived discrimination is reflected in weaker national identification and less trust of the majority group.However, both national identification and trust are as such positively related to out-group attitudes and willingness to confront injustice.Thus, disidentification and distrust as two simultaneous reactions to perceived discrimination hinder minority group mobilization that would challenge the discriminating majority group.These results help in understanding why perceived discrimination is not always reacted to with collective mobilization.They also show how common group identity and trust can work as a double-edged sword.
From the perspective of an individual immigrant, identifying with and trusting the national superordinate group promotes integration and positive intergroup relations, which potentially help in combatting negative intergroup experiences, and motivates ethnic activism.Pereira et al. (2017) have recently examined national identification as a buffering factor of the demobilization process resulting from positive intergroup contact.They showed the existence of a non-conforming style of national identification among Roma minority in Bulgaria: the demobilization effect of positive contact was evident only among individuals with low levels of national identification.However, our study shows that considering social injustice faced by a disadvantaged minority group, identification with and trust of national majority may also hinder the struggle towards social equality as they both are reactive to perceived inequality.
To our knowledge, only Wiley et al. (2013) have previously tested the mediational role of superordinate national identification in the relationship between ethnicity-based rejection and collective action tendencies.Their study did not, however, account for trust in the majority group as a potential mediator.Moreover, they did not find ethnic identification to play a role in immigrants' political engagement.We found positive effects of ethnic identification on out-group attitudes and collective action intentions, inhibited by national disidentification and particularly by distrust.This study therefore also provides the research field with new insights about the dynamics of perceived discrimination, ethnic identification, and immigrant integration.It shows that when ethnic identification is intertwined with perceived ethnic discrimination, it could develop into a pattern of incompatible ethnic and national identities accompanied by distrust of the majority group.This pattern is indicative of oppositional identity among immigrants (see also Benet-Martínez, Leu, Lee, & Morris, 2002;Verkuyten & Yildiz, 2007), and seems to account for their attitudinal and behavioral reactions to perceived rejection.As also pointed out by Louis and his colleagues (Louis, 2009;Louis et al., 2007), it makes immigrants more hostile towards the discriminating majority group, but also more disarmed when it comes to intergroup confrontation.
The results of this study showed a consistent pattern across the three contexts studied.The only statistically quite marginal but theoretically notable differences between countries were observed in the directions of the associations between ethnic identification, trust in the majority group and willingness to confront injustice.Ethnic identification and trust in the majority group were positively related to each other in Estonia and Norway, but negatively related to each other in Finland.Likewise, trust in the majority group and willingness to confront injustice were positively related to each other in Estonia and Norway.First, this suggests that these small country differences are not attributable to the differences in the immigrant status of the participants in the three countries studied.In Estonia, Russian-speakers clearly represented an ethno-cultural minority, but the results for the Estonian sample were identical to those in the Norwegian sample of recent immigrants.Second, these results suggest a possibly contextual role of trust in the majority group in the relationship between minority group identification and collective action intentions: while in some contexts (Estonian and Norwegian in this study) trust in the majority group may assist the minority groups to express their demands for equality, in other (Finland in this study) it could rather be a distrust of the majority group that accompanies high minority group identification and willingness to confront injustice.The lack of trust between the Finnish majority and Russian immigrants in Finland has been seen as a re-grettable characteristic of the intergroup relations between these groups.Lack of trust has been previously found to explain the relationship between perceived cultural discordance (i.e., disagreement between majority and minority groups about the importance of cultural maintenance for minority members) and support for collective action benefiting the minority (Brylka et al., 2015).Though to be better corroborated in the future studies, the results of this study seem to suggest that the intergroup ramifications of minority in-group identification are group-and context-specific (see e.g., Brewer, 1999).Depending on the social and political state of affairs, immigrants and ethnic minority groups members may seek for their ways to confront injustice either by engaging in trust building behaviours to achieve social change or by developing an oppositional approach and questioning the intentions of the majority group.

Limitations
This study is not without limitations, with the most significant relating to the cross-sectional design.As a strong sense of attachment to host society has been found to undermine concern for one's disadvantaged in-group (Kulich, Lorenzi-Cioldi, & Iacoviello, 2015;Mähönen & Jasinskaja-Lahti, 2015;Saguy et al., 2009), it is possible that immigrants with high national identification evaluate the position and treatment of their ethnic in-group more positively than immigrants with weaker ties to the majority group.It is also possible that those displaying stronger feelings of attachment and belongingness to the national community do not only have more positive intergroup contact experiences with and trust in the majority group, but are also considered as more integrated or even assimilated.Further, immigrants holding negative attitudes towards the majority might view its behaviour in a more negative light.Both high ethnic identification and negative out-group attitudes may thus affect perceptions of discrimination.Moreover, collective action could lead to identification as an actor, reinforce further polarization of identities, and facilitate future action (Louis, 2009).There was, however, a strong reason to assume in this study that prejudice affects national identification, as rejection-disidentification link is relatively established by recent studies (Bobowik et al., 2017;Wiley et al., 2013), including both experimental (Matschke & Sassenberg, 2010) and longitudinal studies (Jasinskaja-Lahti et al., 2009;Jasinskaja-Lahti et al., 2012;Mähönen & Jasinskaja-Lahti, 2012).Future experimental and longitudinal research should, however, still keep all causal directions in mind.
It should also be noted that the magnitude of the observed effects was not big.Furthermore, while gender and education had no effect on the associations studied, the indirect effects of ethnic identification on out-group attitudes via national identification, and of perceived discrimination on willingness to confront injustice via national identification were lost once the model was controlled for the age of the participants.The results showed that age correlated with out-group attitudes, with the younger immigrants having slightly more negative attitudes towards the majority group than the older immigrants.Age also intercorrelated with other predictors in the model, and so using it as a covariate clearly diminished the strength of the two indirect paths mentioned above.There was, however, no evidence showing that the suggested mediation model was age-specific.In other words, national disidentification and distrust seem to be reactions to perceived disadvantage regardless the age of immigrants.
As a further limitation, perceived ethnic discrimination was assessed with only three items and willingness to confront injustice with only two items.Both scales were adapted from previous research with more items.Our focus on the willingness to confront injustice instead of a more traditional approach to collective action intentions that typically accounts for more active forms of action such as signing petitions and participation in protests and demonstrations has been justified by the characteristics of the contexts studied.In Nordic countries, Russianspeakers are not a unified group with clear political agenda and collective will to actively demand their rights.Instead, what unifies them is the shared experience of every day discrimination that brings along the need to find a civic courage to publicly confront it.Thus, in the expense of the more contextually appropriate scale is our limited possibility to directly compare the results of this study with other studies on collective action among immigrants in other contexts.In addition, not all measures used in this study have previously been validated in all three contexts studied.Namely, the measures of perceived discrimination, national identification, ethnic identification, and outgroup attitudes have also been previously used in studies on Russian-speaking immigrants in Finland (e.g., Jasinskaja-Lahti et al., 2012) and Estonia (Renvik et al., 2018), but have not been previously applied in the Norwegian context.More extensive measurement of context-specifically validated scales of intergroup encounters is recommended in future studies.
Finally, given the theoretical broadness of the concepts of collective action and intergroup attitudes -the dependent variables of the present study-it is plausible that many other factors outside of those studied here may account for their variations.For example, van Zomeren et al. (2018) recently suggested to complement SIMCA model with a new factor -moral convictions (e.g., about group-based discrimination), which they see as having a particularly strong link with politicised identities.
As regards demographic variables used often in studies on immigration and intergroup relations, we were not able to control for the length of stay or the immigrant generation of our participants.Due to differences in the situations of Russian-speakers in Finland and Norway as compared to Estonia, where Russian-speakers represent nowadays rather linguistic ethnocultural group than immigrants, the immigration status of the participants studied overlapped with their country of residence.As a consequence, we did not have equivalent indicators of the length of stay of the participants in the country of residence and could not include immigration status and country as simultaneous moderators of the relationships investigated.This has implications for the interpretation of the results of this study.Namely, the fact that the theoretical model was robust in all three countries suggests that it could be similarly stable across immigrant generations.

Implications and Directions for Future Research
Besides the above-mentioned limitations related to the current empirical research design, we also have more theoretically oriented suggestions for future research.In addition to national identification and trust in the majority group, other potential mediators should be included for further testing and development of RDIM.As one interesting possibility, the role of collective efficacy as highlighted in SIMCA (van Zomeren et al., 2008) would be worth examining.Other forms of political acculturation among immigrants rather than collective action intentions could be examined.For example, in their recent study, Hindriks, Verkuyten, and Coenders (2015) showed that Dutch majority members had more negative feelings when Muslims were presented as politically advancing the interests of their in-group, while Muslims furthering goals that benefit society as a whole were reacted to with considerably less resistance.These results also inspire to enlarge the scope of collective action research among minorities by including different forms and motivations of their collective action.For instance, perceived disadvantage could emphasize the need to collectively achieve societal equality with the interests of each separate groups being acknowledged and served pending the achievement of this bigger goal.
Regarding the practical and political implications of this study, we would like to point out the need to raise awareness of intergroup inequalities and the importance of the active societal position of immigrants and ethnocultural minority groups.To avoid exacerbating intergroup tensions and to assist societal participation, ethnic and national identities of immigrants would, however, not be seen as opposite to each other as it was the case in the contexts of this study.Moreover, our study shows that perceived discrimination and heightened ethnic identity undermine the potential of the development of superordinate national identities among Russian-speakers to serve as a means to build positive intergroup relations and to engage in collective action to promote social equality in the society.Thus, the positive potential of superordinate national identification and intergroup trust should be harnessed for intergroup dialogue that could give a voice to immigrants facing discrimination and pave a way for political mobilization of ethnic minority group members within the prevailing political system.
was measured with 2 items; "[I would] defend the rights of Russian immigrants/minority in a public debate."and "[I would] intervene verbally in situations in which I notice discrimination of <MINOR-ITY>."The items were adapted from Simon et al.'s (1998) four-item measure of Willingness to participate in (future)

Figure 2 .
Figure 2. Unstandardized regression coefficients for the relationship between independent variables (Perceived Discrimination and Ethnic Identification) and dependent variables (Out-group Attitudes and Willingness to Confront Injustice) mediated by National Identification and Trust in the Majority Group.*p < .05. **p < .01. ***p < .001.

Table 1
Means and Standard Deviations of Items by Latent Constructs, and Correlations Between Latent Variables

Table 2
Total, Direct and Indirect Effects on Dependent Variables