text new page (beta)
English (pdf)
Article in xml format
Article references
Automatic translation
Send this article by e-mail
Cited by SciELO 
Similars in
SciELO 
Permalink
Group belonging and identity is one of the most relevant topics in social, cultural, and political psychology (Tajfel, 1979; Tajfel & Turner, 1986), particularly as a resource in dealing with adversity (Branscombe, Ellemers, Spears, & Doosje, 1999). Much of what we know is derived from social psychological studies sampling participants from prototypical Western contexts (Henrich, Heine, & Norenzayan, 2010; Nielsen, Haun, Kärtner, & Legare, 2017), or immigrants that have recently relocated to such contexts that are dominated by one majority group (e.g., Appel, Weber, & Kronberger, 2015; Nadler & Clark, 2011; McQueen & Klein, 2006). Much less is known about non-Western, plural contexts (for such an example, see Sari, Chasiotis, Van de Vijver, & Bender, 2018; Sari, Van de Vijver, Chasiotis, & Bender, 2018), and how identities are employed in close proximity of different groups. We therefore set out to assess whether people benefit from self-affirmation in dealing with threats to their identity in the plural, non-Western context of Lebanon.
Self-affirmation entails that people reflect on valued elements of the self to boost the self’s resources and thereby render threats to the self psychologically less dire (Sherman & Cohen, 2006; Steele, 1988). We investigate this process in two studies with two specific groups: a sample of mainstream Lebanese (Study 1) and Lebanese-Armenians (Study 2). In doing so, we move beyond generic expectations from ad-hoc designs (e.g., minimal group paradigm; Zagefka, 2009) towards an inclusion of meaningful contextualized identities as determinants of how individuals deal with threat (Branscombe, Ellemers, Spears, & Doosje, 1999; Breakwell, 1986). This is in line with the recommendation of expanding research in psychology beyond the usual participant pool drawn from Western, Industrialized, Educated, Rich, and Democratic (WEIRD) societies (Henrich, Heine, & Norenzayan, 2010). More specifically, we address how group affiliations that are typically found to be relevant in Western nation states and which may presumably be less relevant in Lebanon, moderate identity threat and the effectiveness of self-affirmation techniques. As explained below in more detail, the Lebanese context is characterized by political instability, a history of violent intergroup relations, and sectarian politics (Haddad, 2009; Makdisi, 1996; Traboulsi, 2012) a constellation which appears to be accompanied by identity-relevant behaviors that are both consequential and psychologically relevant on an everyday basis.
What is Social Identity Threat?
Instances where an in-group member might fear that their collective would be negatively represented or evaluated, known as social identity threat, has been elaborated by social identity (Tajfel, 1979) and social categorization theories (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987). The experience of such threat precipitates responses at the cognitive (e.g., rationalization), emotional (e.g., anger) and behavioral (e.g., demonstrations) levels, specifically when in-group identity is salient (Stets & Burke, 2000). For instance, a standardized testing context can be threatening to an African American college student, vis-à-vis the stereotype of underperformance as a function of race. This illustrates a social identity threat in a situation where an individual fears negative evaluation based on a commonly held stereotype of his / her group (i.e., stereotype threat; Nguyen & Ryan, 2008; Schmader & Johns, 2003; Steele & Aronson, 1995). The awareness of such a stereotype may deplete cognitive resources and in turn lower performance. Other mechanisms underlying underperformance have also been proposed. These mechanisms, among others, include priming of the stereotype, and threat to self-integrity and subsequent efforts at its correction (e.g., self-handicapping; Spencer, Logel, & Davies, 2016). Social identity threat can also be triggered based on more subtle cues such as the fewer number of in-group members present in a given context relative to a dominant out-group (representativeness threat; Purdie-Vaughns, Steele, Davies, Ditlmann, & Crosby, 2008), and the extent to which the setting itself structurally excludes individuals based on group membership (glass ceiling effect; Baxter & Wright, 2000; Cook & Glass, 2014). The effects of stereotype threat are particularly pronounced for those who highly identify with their group.
How do Individuals Cope with Social Identity Threat?
When threatened, individuals seek to protect their self, and they often shift their focus to an unthreatened (safe) domain of the self. This can occur via affirming one’s values (self-affirmation; Sherman & Cohen, 2006; Steele, 1988), doubling down on one’s in-group identification (Branscombe et al., 1999), and / or disidentifying with the in-group (Jasinskaja-Lahti, Liebkind & Solheim, 2009). Variations of self-affirmation include reflecting upon important values like religion or one’s relationship with family and friends, one’s skills (Cohen & Sherman, 2006; Steele, 1988), or recalling instances of success and overcoming challenges (Tavitian-Elmadjian, Bender, Van de Vijver, Chasiotis, & Harb, 2019). Self-affirmation interventions have been shown to maintain collective self-esteem (Tavitian-Elmadjian et al., 2019), reduce sexual prejudice towards gay men and lesbians (Lehmiller, Law, & Tormala, 2010), reduce prejudice towards out-groups when applied in concert with increased in-group distinctiveness manipulations (Zárate & Garza, 2002), and increase support for Black programs among White participants exposed to collective guilt over their history of slavery (Harvey & Oswald, 2000). Self-affirmation has also been shown to reduce perceptions of racism among Latinos (Adams, Tormala, & O’Brien, 2006). By providing people the opportunity to engage in a task that decreases defensiveness and protects overall self-worth, individuals appear to put the threat they experience into a larger perspective, which effectively dilutes the impact of threat. This reduces intergroup anxiety and increases behaviors conducive for improved intergroup relations (Badea, Bender, & Korda, 2019; Badea & Sherman, 2019).
Social Identity Threat in a Context of Conflict: Why Lebanon?
While findings from Western samples on how social identity threat operates may generalize, we suggest and demonstrate that they are qualified by the specifics of a context (for a similar point, see studies conducted about Roma people; Dimitrova et al., 2017), thus going beyond previous acknowledgements of a situational context’s importance. Lebanon’s context seems to differ from the typically studied Western contexts of threat management. Lebanon is a setting with multiple in-group / out-group dyads in conflict over their perceived survival, where gains of one group are perceived as losses for the other (zero sum outcome). This creates a volatile political structure with a pronounced fear of intergroup conflict, particularly when compared to prototypical Western immigration contexts. With this backdrop in mind, the manner in which individuals negotiate their sense of belonging is bound to deviate from what is observed in Western contexts dominated by one majority group and immigrant minorities (e.g., Appel et al., 2015; McQueen & Klein, 2006; Nadler & Clark, 2011).
Sociopolitical History of Lebanon. Lebanon’s plural nature lies in the sectarian variation of its population. Sectarianism represents the intersectional (i.e., religious, ethnic, political) nature of Lebanese groups (sects). Lebanon has been described as a mosaic of 18 religious groups that include four Muslim sects, 12 Christian sects, the Druze sect, and Jewish groups (Saseen, 1990). However, this metaphor may create the erroneous impression of an overall whole, constituted by its elements. In fact, Lebanon is typically described as a deeply divided society that has been plagued by civil war, politicized sectarian conflict and division (Bahout, 2016; Traboulsi, 2012).
Lebanese intergroup dynamics can be understood in terms of the integrated threat theory where symbolic threats (e.g., differences in religion, values, belief systems, ideologies, and worldview, among others) and realistic threats (e.g., a group’s welfare and resources) may be implicated in determining psychological and behavioral reactions within intergroup contexts (Stephan & Renfro, 2002). The model for integrated threat details a number of variables related to intergroup relations (e.g., history of intergroup conflict, power of in-group relative to the out-group and group size), cultural values (e.g., power distance, individualism and collectivism, and uncertainty avoidance), situational factors (e.g., the setting within which interaction occurs, how structured the interaction is, and if the interaction is supported by authority), and individual differences (e.g., group identification, contact with the out-group, social dominance orientation, and right wing authoritarianism among others) as antecedents for symbolic and realistic threats. Through the lens of the integrated threat model and its antecedents, particular periods in Lebanon’s history, still relevant today, are described next to demonstrate the volatile nature of Lebanese society and the threats experienced within it.
For instance, the first civil war in Lebanon (May 1958) can be understood in terms of ideological differences between Muslims and Christians about the positioning of Lebanon at the time vis-à-vis the emergence of the United Arab Republic: Lebanon as part of the United Arab Republic versus a sovereign Lebanon assuming a neutral position between Arab nationalism and the West (Sorby, 2000). These opposing views represent symbolic threats pertaining to the very definition of what it meant to be a Lebanese national. The displacement of Palestinian refugees in Lebanon after their expulsion from Palestine (1948 - 1967) and Jordan (1970) can also be interpreted as a perceived threat by Christians who, at the time, felt the demographic composition in the country was shifting in favor of Muslims. When discussing threat, we refer to perceived threat which does not necessarily equate with objective accounts. Feeling outnumbered has been classified as a situational antecedent of realistic threat (Stephan & Renfro, 2002) and Christians feared the risk of becoming a minority relative to a Muslim majority (representativeness threat). With Christians holding more power relative to other groups at the time, their perception of threat may have been quite dire. This is in line with research showing that high power groups respond more strongly to threat as they have much more to lose (Johnson, Terry, & Louis, 2005).
While providing a detailed historical account of the Lebanese civil war (1975-1990) is beyond the scope of the present paper, it is important to note that the conflict cannot be attributed to tensions between Christian conservatives and Palestinians at the time. The Arab-Israeli conflict was a larger regional backdrop, that played out in Lebanon and was reflected in the development of the Lebanese National Movement (LNM), constituted of Arab nationalists and leftists (predominantly Druze and Muslim) and that allied themselves with the Palestinians against the conservative predominantly Christian Phalangist Kataeb Regulatory Forces (KRF). These represented two politically different ideologies (symbolic threat), with the LNM calling for a non-sectarian democratic system of governance and the KRF supporting the status quo and later federalism and political decentralization. There was much more nuance, however, with the two fronts changing positions and demands as the war progressed. The conflict was complicated by external interventions from Israel (e.g., 1982 Israeli invasion) and Syria2 (e.g., 1976 military intervention). As the war progressed, further divisions and intra-sect wars emerged (e.g., intra-Christian Maronite and intra-Muslim Shia), increasing anti-militia sentiments in the general public and a call for a resolution to the conflict (Krayem, 1997). With the signing of the Document of National Understanding in Taif, the Lebanese civil war came to an official end in 1990, and Syria assumed a temporary guardianship over Lebanon to ensure reconciliation (Salem, 2006). Internally, parties in support for and in opposition of the Syrian military presence in Lebanon emerged, intersect conflicts continued, this time particularly between Sunnis and Shias. Against this historical background, it is surprising that a relatively recent large-scale study suggests that the Lebanese national identity is a salient source of belonging for Lebanese youth (aged 18-25) across sects (but mostly Maronite, Sunni, and Shia; Harb, 2010). At the same time, however, the study finds a “blatant bias towards one’s sect” “regardless of gender, confession or region of origin” (Harb, 2010, p.14), which clarifies that sectarian identity is a crucial source of social categorization (Tajfel & Turner, 1986). We single out one specific group, Armenians in Lebanon, for further investigation.
Armenians in Lebanon. Armenians voluntarily settled in the Middle East as early as the 1800s, prior to the declaration of the state of Greater Lebanon in 1920. In the late 19th and early 20th centuries, an estimated one and a half million Armenians were massacred in the Ottoman Empire (Jorgensen, 2003; Lewy, 2005). Surviving refugees settled in camps, such as Bourj-Hammoud, Sanjak, and Anjar, under the then French mandate. These camps later flourishing into cities that today are still predominantly inhabited by Armenians. Armenians in Lebanon are considered a religious / ethnic minority as they are Christian and speak the Indo-European Armenian language (Kazarian & Martin, 2006). They are recognized as one of the main factions of Lebanon and are represented in the government based on the proportionality principle of Lebanese power sharing (Article 24 of the Lebanese constitution) with five parliament seats allotted to Apostolic Armenians and one to Catholic Armenians. Armenians perceive their identity as endangered because they live in exile, away from their homeland (Tölölyan, 2010). A prime concern is the group’s continuity (Smeeks, Verkuyten, & Poppe, 2011), which is manifested in collective angst or anxiety about the group’s future (Jetten & Wohl, 2012).
Faced with uncertainty that comes with being a minority sect in exile (Tölölyan, 2010), Armenians tend to protect their collective identity and rights through sustained institutional efforts, such as Armenian schools, churches, youth and cultural organizations, newspapers, radio stations, and television news broadcasts, and, importantly, participating in Lebanon’s political and public life via three political parties (Migliorino, 2008) . We suggest that these efforts make it likely to find that Lebanese-Armenian’s ethnic/sectarian identity is more highly endorsed than their national identity. A threat to the Armenian collective in Lebanon may thus be especially dire and relevant. In turn, self-affirmation may be specifically helpful in increasing the self’s resources and ameliorating the negative effects of such threat.
The Present Studies
We have elaborated on the historical context of intergroup conflict in Lebanon, which we argue may come with implications for the relative endorsement of national and sectarian identity: We expect that the complex and tense nature of intergroup relations in Lebanon would predispose individuals to endorse their national identity less than their sectarian identity, but there is also data that show that national identity is a relevant source of belonging (Harb, 2010).
Against this unresolved issue, we examined whether self-affirmation through the recall of instances of overcoming adversity (mastery recall) would help boost collective self-esteem under a superordinate national identity threat (Study 1). If the identity threatened is not a valued resource for the self, as may be the case in threatening the Lebanese national identity, then self-affirmation may fail to exert an effect. But Lebanese identity could be relevant (Harb, 2010), especially given that we collected data after the 2015 sanitation crisis that precipitated protests lead mainly by young activists (members of civil society organizations, environmental organizations and human rights groups). The protests started off demanding a resolution to the waste management crisis in the country but transformed into protests against government corruption in what started to look more and more like an uprising (Civil Society Knowledge Center, n.d.). In the largest of the protests, citizens took to the streets waving the Lebanese flag and demanding the fall of the regime (Ensor, 2015). These protests may have strengthened a sense of solidarity and superordinate identity among Lebanese.
We also take a closer look at how an ethnisectarian group experiences ethnic identity threat (Study 2). Given the importance of maintaining ethnic identity continuity among Armenians as highlighted above, threats to the Armenian ethnic identity may be particularly relevant and psychologically dire. We therefore set out to understand how individuals in such a context respond to threat and whether their memories - which can be tools used in self-affirmation (Tavitian-Elmadjian et al., 2019) - play a role in their responses.
Study 1
Given the history of intergroup conflict, the Lebanese context provides a fitting opportunity to evaluate a contextualized experience and management of social identity threat. In this historical and political context, a salient national (i.e., Lebanese) identity may be unlikely. Still, data suggest that national identity is an important source of belonging (Harb, 2010) and the emergence of a civil society movement in protest of the sanitation crisis may have contributed to bolstering national identity further. To test whether national identity threat and mastery recall affirmation will be effective among Lebanese (RQ1), we randomly assigned mainstream Lebanese youth to one of three threat conditions (identity-relevant, identity-irrelevant, and neutral), two affirmation conditions (mastery recall versus routine recall), and recall order conditions (before threat and after threat). We employed collective self-esteem as an outcome measure of identity affirmation.
Participants
Participants were recruited via snowball sampling online through social media platforms and undergraduate university student sample pools. A total of 507 Lebanese adults participated, of which we excluded non-Lebanese nationals and participants with links to Jordan (n = 96; the identity irrelevant vignette targeted Jordanians), participants who quit the survey before completing the dependent measure assessment (n =92), participants whose memories did not adhere to the instruction (n = 27) and participants who opted to withdraw their participation after reading the debriefing letter (n = 4) (Table 1). We determined that a sample size of 270 is needed to detect a medium effect size of f = .25 for a power of .80 (calculated using G*Power 3.1; Faul, Erdfelder, Buchner, & Lang, 2009).
Procedure and Materials
The study received approval from the Ethics Review Board (reference EC-2014.30) and was run on Qualtrics. We used a cover story to not reveal the experimental design (“exploring different attitudes among many different groups of people and many different countries”). We then asked participants two questions to exclude non-Lebanese nationals and participants with links to Jordan.
Next, participants completed the Modern and Old-Fashioned racism scale, α = .73 (McConahay, Hardee, & Batts, 1981), which we assessed as a covariate. Participants also completed the Rosenberg Self-Esteem scale (RSE; α = .83) (Rosenberg, 1965) and a variant of the revised Multigroup Ethnic Identity Measure (MEIM-R; α = .84) (Phinney & Ong, 2007) that assessed national as opposed to ethnic identification. We assessed both self-esteem and national identification as possible moderators. Identification strength has been shown to vary responses to identity threat such that high identifiers are more susceptible compared to low identifiers (Armenta, 2010; McCoy & Major, 2003; Schmader, Block, & Lickel, 2015). Although self-esteem has not consistently emerged as a moderator in identity threat and affirmation research, there is evidence pointing to greater perceived stress and attitude change following self-affirmation interventions among those low on self-esteem (MCQueen & Klein, 2006).
Following the assessment of our covariate and moderators, we randomly assigned participants to one of the conditions in a 2 x 2 x 3 experimental design with recall type (mastery / non-mastery), recall order (before threat / after threat) and threat relevance (national identity relevant / national identity irrelevant / neutral) as independent variables and racism as a control variable. We manipulated recall type by explicitly asking participants to remember and write about their daily morning routine or a time in their life where they faced a difficult experience and how they overcame it. We manipulated threat relevance by constructing three vignettes3 2 featuring a negative stereotypical characteristic ascribed to a group (categorization threat; Branscombe et al., 1999).
As a manipulation check, we then asked participants to rate the author of the presented vignette on his/her attitudes and his/her feelings toward the group portrayed in the vignette (Phinney, Chavira, & Tate, 1993). Ratings were in line with our expectations; the identity relevant and irrelevant threat vignettes were rated as equally negative (p = 1) and significantly more negative than the neutral vignette (p < .001). As the dependent measure, we assessed collective self-esteem using the Collective Self-Esteem Scale (CSE; α = .8) (Luhtanen & Crocker, 1992). Finally, we asked six open ended questions to assess reactivity, but found no evidence for it (Chartrand & Bargh, 1999). We used IBM Statistical Package for the Social Sciences (SPSS) version 21 to conduct our analyses.
Results
Participants in our sample were predominantly university students with an average age of 19.53 (SD = 2.33). On average the Lebanese national identity was moderately endorsed (M = 3.52, SD = .78, z = .00) unlike in a previous study where national identification was quite highly endorsed (Standardized value of the mean = 1.27) (Harb, 2010). Similarly, the average score on collective self-esteem was around the scale midpoint (M = 4.25, SD = .8, z = .00).
Table 2 Means of National Identification, Personal Self-Esteem, Racism and Collective Self-Esteem of Mainstream Lebanese Participants (Standard Deviations in Parentheses) (N =289)
| M | SD | |
|---|---|---|
| National identification a | 3.52 | .78 |
| Personal self-esteem b | 3.02 | .47 |
| Racism c | 3.69 | .73 |
| Collective self-esteem d | 4.25 | .80 |
Note. a National identification assessed on a response range of 1 (Strongly disagree) to 5 (Strongly agree) and a midpoint of 3 (Neither agree not disagree). b Personal self-esteem assessed on a response range of 1 (Strongly disagree) to 4 (Strongly agree). c Racism assessed on a response range of 1 (Strongly agree) to 5 (Strongly disagree). d Collective self-esteem assessed on a response range of 1 (Strongly disagree) to 7 (Strongly agree) and a midpoint of 4 (Neither agree not disagree).
We examined main and interaction effects of threat and recall type on collective self-esteem in a factorial ANCOVA. Effects for threat relevance (F(2, 276) = 1.2, MSE = .62, p = .3, η2 = .01), recall type (F(1, 276) = .78, MSE = .62, p = .38, η2 = .00) and recall order (F(1, 276) = .64, MSE = .62, p = .42, η2 = .00) were non-significant, as was the interaction effect between recall type and threat relevance, F(2, 276) = .57, MSE = .62, p = .56, η2 = .01). The three way interaction was also not significant (F(2, 276) = 1.1, MSE = .62, p = .33, η2 = .00). Moderators (national identification, personal self-esteem) were also non-significant (p > .05).
Discussion
Our study aimed at evaluating whether self-affirmation would alleviate national identity threat in a sample of Lebanese youth. Against a backdrop of intergroup tension and conflict we doubted a high endorsement of the Lebanese identity and hence a threat targeting that level being effective. In a representative sample of Lebanese youth, however, a strong endorsement of national identity was found (Harb, 2010) which is an indicator that a threat at a national level may be quite relevant. In the current study, we do not find that self-affirmation is related to threat when it targets a superordinate national identity. There are a number of explanations to this. First, and as expected, the Lebanese national identity may not be as valued an element of the self. National identification in our sample is moderate, clearly not high as reported in a previous study (Harb, 2010). Our sample scored just above the scale midpoint anchored at “neither agree nor disagree.” Our finding also reflects another possibility, where it is not a low endorsement of a national identity that is the issue but rather the manner in which the identity is conceptualized and defined by each group. Shared attributes used to construct a national identity can be based on civic (e.g., territory, legal political equality, civic and cultural ideology) and ethnic (e.g., common language, religion, ancestry, and customs) conceptions (Kunovich, 2009; Shulman, 2002). Neither is possible in Lebanon, where nation building - fraught with consociationalism and external interventions - rendered a unified understanding national identity difficult to forge and maintain (Gürcan, 2007).
It is also possible that our findings may be due to the relatively blatant nature of the threat. A more subtle threat may have yielded a different outcome. However, our manipulation check indicated that the threat manipulation worked, and the vignettes were evaluated as in tended. In addition, national identification at pre-test reflected a rather neutral evaluation of the Lebanese identity which was also maintained at post-test with an evaluation of collective self-esteem, thus decreasing the likelihood that the nature of the threat is responsible for our findings. One additional explanation may relate to the source of the threat which we did not specify. Had we constructed the threat as coming from a non-Lebanese source (e.g. Israel) findings may have differed.
In Study 2, we shift our focus to a specific ethnosectarian group within Lebanon (Armenians) as a target of identity threat. We reason that while a threat targeting a superordinate national identity may have been ineffective, threatening a salient identity will reveal a different pattern of responses. Given the long history of intergroup conflict in Lebanon, the definition of what it means to be a Lebanese national may vary across groups, creating a rather subjective understanding derived from experiences that are specific to each sect. So, despite an endorsement of the Lebanese identity (Harb, 2010), the heterogeneity of its content may reduce the efficacy of threats to that level. This is likely not to be the case when focusing on Lebanese-Armenians with a well-defined, shared understanding of ethnic identity.
Study 2
Armenian ethnic identity is likely a salient, chronically accessible feature of Lebanese Armenians. With an identity constructed around the idea of survival against all odds including multiple invasions (Turco-Mongol raids in the 11th and 14th centuries), forcible relocations (e.g., Isfahan in the 17th century) culminating in the 1915 Genocide (Migliorino, 2008), a discourse of anxiety and fear of offense from out-groups is common among Armenians. These experiences imply that ethnic identity maintenance and the threat of assimilation are primary concerns (Panossian, 2002), effectively constituting a cultural continuity threat (Jetten & Wohl, 2012; Smeekes & Verkuyten, 2015). The threat of assimilation is one that most immigrants grapple with as they undergo changes in different areas of psychological functioning including identity (Berry, 1980). In Lebanon, Armenians live within a politically unstable and confessional society (Sasseen, 1990) with no (state) policies regulating immigration like most Western contexts (Bourhis et al., 1997).
We expect the Armenian ethnic identity to be a salient, valued element of the self (Kazarian & Boyadjian, 2008). Participants were presented with different threats, comparable to Study 1 (i.e., identity relevant, identity irrelevant and neutral). We asked them to produce open-ended text in response to these threats (i.e., mastery recall versus routine recall), and we varied the source of threat (in-group versus out-group). We assessed collective self-esteem afterwards. We set out to assess whether participants that were exposed to relevant threat and that recall mastery memories have higher collective self-esteem compared to those that do not (H1). We also expect that when threat is identity relevant and the source of threat is coming from the out-group, this would result in a lower tendency to engage with the out-group as supported by the rejection-identification model (H2). Finally, we set out to explore whether the content of the open-ended responses differed in terms of references to their collective identity across conditions.
Participants
A total of 252 Lebanese-Armenian adults participated in the present study. The sample size is slightly below the 270 needed to detect a medium effect size of f = .25 for a power of .80 (Faul et al., 2009, 2007). After reviewing the memory entries for adherence to the study instructions, 212 participants were retained (Table 3). We employed a committee approach to inspect the autobiographical memory entries. We excluded entries that involved evaluative statements related to the experimental vignettes and future plans that did not connect to a recall of a past event. Both online and paper and pencil methods of convenience sampling were used. Non-Lebanese nationals and participants who have Egyptian heritage or have visited / lived in Egypt were excluded (the identity irrelevant vignette targeted Copts in Egypt).
Procedure and Materials
As in Study 1, after consenting to participate, we applied the exclusion criteria and then participants completed a measure of heritage (α = .94) and host (α = .91) culture adaptation as a covariate (adapted from Ait Ouarasse & Van de Vijver, 2004), given that the threat vignette featured the Armenian community in Lebanon. Since previous research has supported self-esteem (McQueen & Klein, 2006) and identification strength (Armenta, 2010; McCoy & Major, 2003; Schmader et al., 2015) as moderators, we administered the Rosenberg Self-Esteem scale (RSE; α = .70) (Rosenberg, 1965) and the revised Multigroup Ethnic Identity Measure (MEIM-R; α = .90) (Phinney & Ong, 2007). Next, we randomly assigned participants to the conditions in a 2 x 2 x 4 factorial design with recall type (mastery / non-mastery), counterbalanced recall order (before threat / after threat), and threat relevance (national identity relevant in-group / national identity relevant out-group / national identity irrelevant / neutral) manipulated4(tiene que ser la nota 4). We constructed a vignette based on the stereotype of Armenians living in a “ghetto”, as a blatant threat since we did not specifically target high identifiers (see Cadinu, et al., 2003; Keller, 2002) (see item two of supplementary material). We manipulated the source of the threat such that it was either presented as coming from an in-group (fellow Armenian) or an out-group source. Next, we developed an identity-irrelevant vignette (Copts in Egypt) and a neutral condition (chairs, see Study 1). We kept word count constant across the four vignettes. Afterwards, we adapted a manipulation check from Phinney and colleagues (1993), finding that the identity threat relevant in-group, out-group, and the identity irrelevant vignettes were rated equally and significantly more negative compared to the neutral condition on author’s attitudes and feelings toward the target group (p < .001 for all). As in Study 1, we assessed collective self-esteem (α = .86) as an outcome (Luhtanen & Crocker, 1992) with an additional measure of intentions and willingness to interact with the out-group (α = .95). As in Study 1, we did not find any participant reactivity (with 6 items adapted from Chartrand & Bargh, 1999). We used IBM Statistical Package for the Social Sciences (SPSS) version 21 to conduct our analyses.
Results and Discussion
We did not observe significant effects for both threat relevance (F(3, 185) = .34, MSE = .75, p = .80, η2 = .005) and recall type (F(1, 185) = 2.26, MSE = .75, p = .13, η2 = .01) or their interaction (F(3, 185) =.86, MSE = .75, p =.46, η2 = .01). Moderators (ethnic identification and personal self-esteem) were also non-significant (p > .05) and there were no differences on collective self-esteem across participants who received an in-group threat versus an out-group threat (p > .05). Overall the sample scored high both on ethnic identification (M = 4.07, SD = .76) and collective self-esteem (M = 5.33, SD = .87) (See Table 4).
Table 4 Means of Ethnic identification, personal self-esteem, collective self-esteem, mainstream and heritage adaptation and outgroup interaction of Armenian Lebanese participants (Standard Deviations in Parentheses) (N = 212).
| M | SD | |
|---|---|---|
| Ethnic Identification a | 4.07 | .76 |
| Personal self-esteem b | 3.09 | .46 |
| Collective self-esteem c | 5.33 | .87 |
| Heritage domain adaptation d | 5.85 | 1.04 |
| Host domain adaptation d | 4.94 | 1.07 |
| Interaction with out-group e | 3.68 | .96 |
Note. a Ethnic identification assessed on a response range of 1 (Strongly disagree) to 5 (Strongly agree). b Personal self-esteem assessed on a response range of 1 (Strongly disagree) to 4 (Strongly agree). c Collective self-esteem assessed on a response range of 1 (Strongly disagree) to 7 (Strongly agree). d Acculturation orientations assessed on response range of 1 (Strongly disagree) to 7 (Strongly agree). e Willingness to interact with the out-group assessed on a response range of 1 (Strongly agree) to 5 (Strongly disagree).
The absence of findings may likely be due to the sample size. However, there is one further explanation, which is grounded in the specific circumstances of the Armenian group in Lebanon. For that, we need to take a closer look at the data collection in the Armenian community, which was lengthy and extremely challenging. We suggest that Armenians may have chosen not to participate because they were apprehensive and defensive about being singled out as Armenians, which are common responses to perceived intergroup threat (Steele, 1988). This may have led to both a reduced sample size, as well as an underrepresentation of highly identifying individuals in our sample (but note that sectarian identity is very high). The observed pattern in our phases of data collection is consistent with the notion of defensiveness and threat playing an important role. In phase 1, the online data collection, initial response rates were low, despite extensive advertising (only 49 participants, i.e., ~19% of the eventual sample size, over a period of 7 months). In phase 2, we carried out the study’s paper and pencil version to provide more personalized contact to build trust and reduce defensiveness, which improved the participation rate to 252 recruited participants. Still, participants were defensive and apprehensive: They were skeptical of the identity and nationality of the principal investigators, and they speculated about the intentions behind data collection. Conspiracism is a common concern in Lebanon (Gray, 2010). In phase 3, we sought to address this by soliciting the approval of a prominent Armenian party and other Armenian organizations to collect further data from their partisans. While this further increased sample size, the defensiveness of participants persisted. In short, the response we received to our recruitment efforts, may have already been a response to a perceived threat - although it was not operationalized as such in our design.
Coding of autobiographical memory entries. We content-coded the open-ended responses of our participants to identify instances in which they responded with references to collective Armenian identity as their shared identity (Ashmore, Deaux, & McLaughlin-Volpe, 2004). For this, the entire sample was retained with the exception of three cases where memories were not provided (N = 249). Two independent coders coded any reference to Armenian identity (e.g., “after waking up from 3-4 alarms, I wash my teeth and face. Then I check the social media on my phone and talk to my Armenian friend to check if he is awake, so that I pick him up on my way to university. I go down the stairs and put the Armenian newspaper at its spot. I pick up my friend and go to the university”). Agreement between coders was high, Cohen’s κ = .96, p < .01 (Viera & Garrett, 2005).
Collective references per threat condition. A total of 32 out of 249 (12.85%) collective references were identified across conditions. In other words, even when required to provide a routine memory, a reference to the Armenian collective identity was made in a large number of cases. This may help explain why our affirmation exercise failed to exert an effect, as it overrides the experimental design by enhancing in-group favoritism (Turner, et al., 1987). These self-categorizations occurred in a context of threat, and classifying the self as Armenian may help participants evade negative effects of threat that stems from differences among people, as a participant highlight:
“I have a shower every other day, get prepared for work, have a healthy breakfast, head to work. There is nothing related to my heritage that I do during the mornings. Actually, I'm an Armenian person, and I am attached to the heritage; however, I'm the type who's open to other cultures as well. I do not treat Armenians better than others, but I cannot reject the fact that being surrounded by Armenians gives me some sort of comfort. It feels like home most of the time. Being open to people different from us is essential. Differences are the main causes of hate in this world, along with human greed. I don't care about religion (I'm not religious) neither about nationality. All I care about is having a good heart and making a positive difference in this world.
25-year-old male
A pattern specific to the Lebanese-Armenians. We then examined whether the pattern is also found within the mainstream Lebanese sample. We applied a similar coding scheme for data gathered from Lebanese participants in Study 1, and, of the 316 memories assessed, only 10 (3.16%) included references to the Lebanese identity (e.g., “Every day in the morning, I wake up from bed almost every day at 6 am….weird songs that are old (some of them are good and that I enjoy listening to them like the ones for Fairuz)…”) as opposed to 32 (12.85%) in the smaller Armenian Lebanese sample from Study 2. These proportions were significantly different (χ2(1) = 16.22, p = .000; phi / Cramer’s V = -.16). This suggests that the finding is specific to the Armenian group in Lebanon, and it points towards their identity being chronically more accessible (Fiske & Taylor, 1991). For instance, an Armenian is quickly recognizable by merely stating their last names (e.g., Hagopian, Melikian…) or their area of residence.
Although endorsement of identity for both mainstream Lebanese and Armenian Lebanese was above the scale midpoint (refer to Table 2 and Table 4), endorsement of the Armenian ethnic identity was higher than Lebanese national identity. To examine whether the observed difference is significant, we compared the average scores on national and ethnic identification from Studies 1 and 2 respectively, as well as collective self-esteem from the two samples using independent samples t-tests. We found a significant difference in national / ethnic identification (t(490) = 14.13, p = .00, d = 1.28) and collective self-esteem (t(498) = 7.92, p = .00, d = .71) when comparing Armenian Lebanese and mainstream Lebanese participants. Armenian Lebanese endorsed their ethnic identity (M = 4.07, SD = .76) and derived a sense of worth from their group belonging (M = 5.33, SD = .87) to a greater extent compared to mainstream Lebanese (M = 3.52, SD = .78 and M = 4.25, SD = .80, respectively).
General Discussion
In the present study we set out to test identity threat management through mastery recall in a non-Western context (Lebanon) with regard to a relatively less salient identity (Lebanese national identity among general Lebanese) and relatively more salient identity (Armenian ethnic identity among Armenians). Our results support this pattern specifically when examining the qualitative component of both studies: Armenian-Lebanese spontaneously referred to their collective ethnic identity as opposed to mainstream Lebanese who did not do so. We also find a significantly higher endorsement of their ethnic identity among Armenians relative to Lebanese endorsing their national identity, which further points to the salience of an ethnosectarian identity relative to a national identity. The notion that self-affirmation via mastery recall protects collective self-esteem when national Lebanese identity (Study 1) and ethnic Armenian identity are threatened (Study 2) was not supported. We reasoned that the historical background of Lebanon would make it less likely that an affirmation intervention would benefit the Lebanese national identity (Makdisi, 1996). We expected, however, that self-affirmation for the Armenian ethnic identity would be effective. While sample sizes were too low for firm conclusions in Study 2, we suggest that this was a direct outcome of the challenges in data collection that we encountered, from unresponsiveness to online data collection to mistrust and threat-relevant defensiveness on the side of our Armenian participants. We find, however, a pattern of spontaneous references to the salient Armenian collective identity - irrespective of threat - which speaks to a chronic affirmation that may have overshadowed our experimental manipulation.
Ongoing Intergroup Conflict Weakens
National Identification
In study 1, we found the experimental vignettes to be effective, but there was no difference in collective self-esteem associated with threat manipulation or self-affirmation of Lebanese participants. Participants’ average score on the measure was at the midpoint (neither agree nor disagree), indicating that the sample did not necessarily derive a pronounced sense of self-worth from their national group belonging. It could also be that the definition of what it means to be Lebanese is different per sect: For some Lebanese, their Arab identity may be in the foreground, but not for others (such as Armenians). This finding is not surprising given the sociopolitical history of Lebanon which has accentuated sectarian identity (Makdisi, 1996), and we specifically observe this accentuation among the Armenian sample with a spontaneous affirmation of their ethnic identity when faced with threat. While people are able to identify with more than one social category which can overlap and be simultaneously activated (Roccas & Brewer, 2002), in a politically unstable context with ongoing intergroup conflict, an emergent common national identity is hindered. This is in contrast to Western nation states where a relatively stable and democratic political climate has been conducive to the emergence of civic over ethnist forms of national identity (Kunovich, 2009), and that could be regarded as sources of positive self-evaluation. The context of ongoing conflict shapes the content of identity, and one of the defining features may be antagonism with the out-group which has been shown to moderate the relationship between identification strength and negative behavioral intentions among students from Northern Ireland (Livingstone & Haslam, 2008). In such an instance, a unified definition and understanding of what it means to belong to a larger group, a superordinate category, may vary across groups.
A Chronically Accessible Identity
According to self-categorization theory, the activation of a social identity category is determined by salience to which accessibility is key (Turner et al., 1987). The accessibility of a social identity - the extent to which information in the environment will be readily perceived and interpreted in terms of the given social category (Bruner, 1957) - may vary based on contextual cues (temporary accessibility), or it may be chronically accessible if considered important to the self-concept and used frequently (Hogg & Terry, 2000). Chronically accessible identities would be categories such as sex and race (Mackie, Hamilton, Susskind, & Rosselli, 1996). We also argue that the Armenian ethnic identity within Lebanon constitutes such a chronically accessible identity. For instance, a 21-year-old female notes that she thinks of her family and then her ethnic group at large every morning when she wakes up “When I wake up in the morning, first, I think about my family members so that they have a successful day. I also think of all Armenians so they have a successful day. All Armenians in Armenia...to be well. Then, I get ready and head to work.” To many individuals in different contexts this would seem an odd thought to have early in the morning. In line with our earlier suggestions, it is possible that this may have - at least partially - been the result of the data collection procedure, as participants filling out the paper and pencil version of the experiment were approached by Armenian data collectors, including the PI. This may have made the participants’ Armenian identity (and the associated threats) more salient.
This notwithstanding, an identity is salient when important and used frequently (Mackie et al., 1996) and within a context characterized by conflict, prejudice towards perceived out-groups is not uncommon (Stephan & Renfro, 2002) and therefore could contribute to frequent references and uses of one’s ethnic identity in every day interactions. For instance, in response to a prompt for recalling a difficult situation and how that situation was mastered an 18-year-old male Armenian recounts: “I was 15 years old when I started working in a restaurant serving Lebanese food. During my years working there the Arabs constantly cursed at Armenians and gave me a hard time for being Armenian. I could not understand why but after a while I started understanding that the reason for the way they were treating me was related to the way Armenians presented themselves. That is, how Armenians constantly say that they have had hard times. What I mean is that people and individuals who are strong and proud of their identity are respected and have a place on this earth. I changed my way of thinking after that event and I understood that I am not the offspring of a people of genocide, but rather the son of freedom fighters and heroes. Just because I live in Lebanon, it does not mean that I have to succumb to the Arabs.” This memory is not only an example of the experience of prejudice, but it also demonstrates how the content of social identity can be shaped by emphasizing out-group-conflict and antagonism as defining elements of an in-group’s identity which would in turn explain behavior towards the out-group (Livingstone & Haslam, 2008). Furthermore, this entry is an example of increased identification as a function of discrimination and prejudice from the out-group as elaborated by the rejection identification model (Branscombe et al., 1999; Schmitt, Spears, & Branscombe, 2003). This mechanism has been supported among European-born Muslims, who - in a context of discrimination and prejudice - increase their ethno-religious identification while disidentifying with their nation states (Fleischmann & Phalet, 2016). The reference to Armenians as heroes and the process through which he reached that realization (i.e., from adversity to success) are a powerful illustration of the pride and sense of collective self-worth that this participant derives from his group belonging, a key concept in the self-affirmation literature (Cohen & Sherman, 2006; Steele, 1988).
Is Constant Affirmation Adaptive?
The examination of the qualitative entries was key in uncovering a pattern of spontaneous affirmation regardless of our manipulations. It seems that the Armenian identity in the Lebanese context is quite accessible and perceived as threatened, and it may therefore be in a state of constant affirmation and self-defense. Perhaps this affirmation is best articulated in an entry where a participant states that a common issue he faces is “…the concept of being treated in a different way, because of being very Armenian...” This is an illustration of a general struggle related to having his behavior interpreted in light of his group membership otherwise known as stereotype threat (Steele, 1998). One of the main benefits of self-affirmation in addressing identity threats is that it places a threat within a larger context, and, while an individual may still be aware of the threat itself, its psychological effects are rendered less dire. This process is rather adaptive relative to the defensive methods of coping as denial and disidentification (Sherman & Cohen, 2006). In our Lebanese- Armenian sample, there are more cases when threat experiences seem to be heightened and affirmation appears chronic compared to our Lebanese sample. Such chronic affirmation irrespective of situational demands seems defensive and rather maladaptive.
Above and beyond the Lebanese contextual specifics (i.e., political instability and intergroup hostility), Armenians in Lebanon grapple with concerns of their group’s continuity (Smeekes & Verkuyten, 2015). This constant threat of assimilating and dissipating into the majority paired with a socio-politically unstable environment seems to have contributed to a hypersensitivity to identity threats. For instance, in response to a prompt at providing a routine morning memory, one participant who had read an identity irrelevant threat wrote: “I wake up in the morning, I feel proud, happy and sad. Sad because I live in a foreign country away from my motherland, but proud and happy because we Armenians could keep our language and our culture…We still have children, adults and old people who demand to have our Armenia back so that we Armenians could be reunited in our country.” This entry not only displays how the participant used his collective identity and specifically tapped into its continuity as a source of pride, but also shows a heightened sensitivity to threats. Note that the identity threat vignette he had been exposed to was about Copts in Egypt, but that was enough to trigger a sense of threat and a need to address that even when the required task was to simply report on a morning routine. It could be that the condition served as an implicit threat, which if so, only reinforces that Armenians experience elevated levels of continuity threats within the Lebanese context. Perhaps such collective self-affirmation could be understood in reference to defensive high self-esteem where some individuals who score high on self-esteem measures are more defensive as a function of a discrepancy between their explicit and implicit evaluations of the self (Jordan, Spencer, Zanna, Hoshino-Browne, & Correll, 2003). At the collective level, the hypersensitivity to threats observed in our sample could be evaluated as a defensive reaction to an underlying unfavorable evaluation of the group but not necessarily in terms of its competence or values more in terms of its survival and perseverance (i.e., continuity threat is high). This interpretation echoes with the Palestinian narrative that is characterized by lamenting the loss of a homeland and the importance of a one-day return (Brand, 1995).
Continuity of the Group as a Paramount Concern
The concern over the group’s continuity is recurrent in the qualitative reports in our sample with a pattern of increased in-group identification while simultaneously downplaying the need to integrate and adapt to the mainstream context: “…Up until last year I never felt the need to know Arabic… Despite the fact that I would have never wanted to have to deal with a non-Armenian, we have to see? What can we do?” Here, having to learn Arabic and speak to non-Armenian Lebanese is portrayed as an inevitable fact of the life of an Armenian in Lebanon. Such a tendency towards maintenance is not only a function of perceived stereotypes and prejudice but may be understood in terms of continuity threat (Smeeks & Verkuyten, 2015). Research has shown that perceived discontinuity in a group’s past can contribute to increased efforts at maintaining identity especially as it undermines or threatens a groups future survival (Jetten & Wohl, 2012). The need for continuity is not specific to Armenians and is a concern common across immigrant groups (see Dasgupta, 1998; Timotijevic & Breakwell, 2000), and it is often paired with the need for distinctiveness (Jaspal, 2013). Distinctiveness, however, may be more relevant in contexts that are home to immigrants from diverse ethno-cultural backgrounds. For example, in a study addressing identity among British-born Sikhs, both continuity and distinctiveness emerged as prime components of the Sikh identity. A need for positive distinctiveness was articulated in terms of a differentiation from Asian migrants, particularly Muslims (Jaspal, 2013). Armenians in Lebanon are not necessarily driven by a need for positive distinctiveness given the little ethno-cultural variation of the country, but a clear tendency towards maintaining continuity is articulated.
Limitations and Directions for Future Research
In both studies, self-affirmation through mastery recall was not conducive to protecting or restoring collective self-worth when the Lebanese national identity and Armenian ethnic identity were threatened respectively. While it is plausible that our Lebanese sample did not identify with their nation state but rather their religious group (Makdisi, 1996), it is also prudent to consider the nature of the threat as a possible limitation: We used a blatant threat-activating cue, where the Lebanese identity was explicitly referred to in negative terms. This may have resulted in reactance in the identity-relevant threat condition and may have triggered defensive coping regardless of our affirmation intervention (Javadian & Zoogah, 2014; Kray, Thompson, & Galinsky, 2001). Another possibility is that the identity-irrelevant threat condition may have constituted implicit threat. As such, it would be useful to replicate the present study with variations at the level of threat type (e.g., the use of implicit threat) and threat target (e.g., sectarian identity threat). National identity threat and the use of self-affirmation should also be evaluated within a community sample of Lebanese. So far, we only tested a predominantly university student sample. It is possible that this sample would typically be in the identity exploration phase which may have dampened the relevance of a national identity threat.
One of the main limitations specifically in Study 2 is the low statistical power. We suggest that the nature of the data collection was relevant for the interpretation of the group’s situation and its defensiveness in Study 2. Still, only through a larger replication would it be possible to corroborate or refute the efficacy of mastery recall as a self-affirmation tool among Armenian Lebanese. In addition, the content analysis helped in understanding how Armenians in Lebanon (compared to Lebanese) experience social identity threat. As a follow up study, it would be useful to qualitatively evaluate narratives on identity development and social identity complexity (Brewer & Pierce, 2005), specifically within this group. Within an acculturative context, perhaps Armenians who conceptualize different aspects of their social identity (e.g., ethnic, national, and religious) as distinct yet overlapping may be less sensitive to threats at the ethnic identity level. If so, then work towards fostering social identity complexity may be important to reduce perceived threats. While we varied the source of threat in Study 2, we did not specify the source of threat in reference to the Lebanese national identity. Given the history of conflict regionally (e.g., Israel, Syria) we may have found a different pattern of responses if the threat was perceived as coming from a source that could be classified as an outgroup relative to Lebanon as a whole.
Conclusion
Our study situates the interpretation of identity threat and threat management within a larger context, and we show that considering the specifics of the threat context is instrumental in understanding identity processes. This is in line with work calling for the expansion of research to non-Western contexts and groups (Henrich et al., 2010; Nielsen et al., 2017), as expectations derived from WEIRD contexts may not always hold cross-culturally. We point to the need for moving beyond the typically evaluated Western contexts in studies of identity threat and affirmation and to include contexts with pronounced inter-group tensions.
In Study 1, we showed that a threat to the Lebanese national identity did not resonate with our Lebanese sample. We argue that, in this context, national identity may hold different meanings across different groups, and perhaps national identification is even overshadowed by sectarian identification (Bahout, 2006). We explained this nonsalient national identity in reference to the long history of intergroup conflict and violence that has characterized the history of Lebanon (Traboulsi, 2012). In Study 2, we observed a hypersalient, chronically accessible collective identity in our Armenian Lebanese sample. Unlike the Lebanese participants, the Armenian ethnic identity is a major source of esteem within this group and is resorted to as a refuge at the slightest hint of threat. We discuss this threat hypersensitivity in reference to continuity threat (Smeekes & Verkuyten, 2015), where the Armenian ethnic identity is perceived as endangered and, in turn, in need of constant affirmation irrespective of contextual demands.
