Group Processes & Intergroup Relations 2016, Vol. 19(6) 732 –752 © The Author(s) 2016 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1368430216641303 gpir.sagepub.com G P I R
Group Processes &
Intergroup Relations
Research on collective action and social change suggests that individuals, especially members of minority groups, act on the basis of their group identity against a collective disadvantage to improve their group’s conditions and bring about social change. Conversely, research on intergroup relations and prejudice argues that improving intergroup harmony via positive intergroup con-tact and formation of new all-inclusive identities are the key to a more equal and just world,
Predicting support for collective action
in the conflict between Turks and Kurds:
Perceived threats as a mediator of
intergroup contact and social identity
Huseyin Çakal,
1Miles Hewstone,
2Meltem Güler,
3and Anthony Heath
4Abstract
Two studies investigated the role of perceived realistic and symbolic threats in predicting collective action tendencies, and in mediating effects of intergroup contact and social identity on collective action in the context of an intractable conflict. Extending earlier research on collective action, integrated threat theory, and intergroup contact theory, we tested whether realistic and symbolic threats would predict collective action tendencies and outgroup attitudes; and mediate the effects of intergroup contact and social identity on collective action tendencies and outgroup attitudes among the advantaged, Turks, and the disadvantaged, Kurds. Findings from both studies (Study 1, N = 289 Turks; Study 2, N = 209 Kurds) supported the predictive and mediating role of threats on collective action tendencies
and outgroup attitudes. Overall findings suggest that advantaged and disadvantaged groups might not always have disparate psychologies regarding collective action and incorporating perceived threats as antecedents of collective action can help to explain collective action tendencies among both groups especially in conflictual contexts.
Keywords
collective action, intergroup contact, perceived threats, social identity Paper received 19 August 2013; revised version accepted 26 February 2016.
1University of Exeter, UK 2University of Oxford, UK 3Çukurova University, Turkey
4 University of Oxford, UK and University of
Manchester, UK
Corresponding author:
Huseyin Çakal, Department of Psychology, College of Life and Environmental Sciences, University of Exeter, Washington Singer Building, Perry Road, Exeter EX4 4QG, UK. Email: h.cakal@exeter.ac.uk
especially for members of majority groups. These divergent lines of research hinge on the assump-tion that social change can be achieved via differ-ent psychological processes, that is, via collective action among the disadvantaged, and via preju-dice reduction among the advantaged. The pre-sent paper challenges this different psychological processes assumption and argues that in
extraordi-nary circumstances, i.e., protracted intergroup conflicts, both groups might be motivated by a strong desire to protect ingroup interests, for example, well-being, goal attainment, or simply survival, especially when they perceive those interests to be threatened by the outgroup. To be more specific, we advance the idea that when groups are locked in a violent conflict, perceived threats might provide the psychological basis for acting in defence of ingroup interests among both the advantaged and disadvantaged groups.
In what follows, we first briefly review the research on collective action (van Zomeren, Leach, & Spears, 2012; van Zomeren, Postmes, & Spears, 2008) and intergroup contact (Brown & Hewstone, 2005) discussing how they relate to perceived threats (Stephan, Ybarra, & Morrison, 2009). We then test (a) the predictive power of perceived threats on collective action tendencies; and (b) whether threats would mediate the ener-gizing role of social identity and the sedative effect of intergroup contact on collective action among the members of a structurally disadvan-taged group, Kurds, and an advandisadvan-taged group, Turks in Turkey, who have long had tumultuous intergroup relations.
Social Identity and Intergroup
Contact
Research on psychological predictors of collec-tive action suggests that identification with the ingroup is central to appraisals of disadvantage and related psychological processes. The social identity perspective (Tajfel & Turner, 1979; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987) argues that individuals would be more willing to engage in competitive behaviour to improve the positive aspects of their group identity when they
cannot leave their groups (impermeability of the group boundaries), when they perceive their neg-ative position as unjust (illegitimacy), and when they perceive the system as unstable (Mummendey, Klink, Mielke, Wenzel, & Blanz, 1999; Tajfel & Turner, 1979). Thus, under these conditions, one possible path to improve the status of the ingroup is to engage in collective action (van Zomeren et al., 2008). Stronger identification with the ingroup therefore motivates individuals to engage in collective action both directly and indirectly by (a) enabling group-based perceptions of injustice; (b) evoking emotions related to this injustice, for example, anger (Mackie, Maitner, & Smith, 2009); and (c) heightening perceptions of group efficacy (Drury & Reicher, 2009; van Zomeren, Spears, Fischer, & Leach, 2004). As such, how individuals identify with the group is crucial to the psycho-logical processes that trigger collective action.
If the group plays a more central role in the high identifiers’ conceptualization of themselves compared to low identifiers (Leach et al., 2008), then it also follows that high identifiers would be more committed to the group (Branscombe, Ellemers, Spears, & Doosje, 1999), would per-ceive greater threats to the ingroup, and their reactions to threats would be stronger than those of low identifiers (Corenblum & Stephan, 2001; Riek, Mania, & Gaertner, 2006). Conversely, in conflict, perceptions of threats might increase group identification. Specifically, threats to social identity might make the ingroup identity more salient (Doosje, Spears, & Ellemers, 2002; Schmid & Muldoon, 2015), and decrease perceived ingroup variability (Doosje, Ellemers, & Spears, 1995). The relation between perceived threats and ingroup identification will therefore be con-text-dependent (Turner, Oakes, Haslam, & McGarty, 1994). However, in certain circum-stances, identification with the group is chroni-cally salient and can preclude the specific intergroup situation (Ashmore, Deaux, & McLaughlin-Volpe, 2004; Leach et al., 2008; van Zomeren et al., 2012) especially so when individ-uals internalize their identification with the group as central to their self-concept as a result of past experiences with the outgroup, that is, situations
where the groups have a history of conflict (Tajfel & Turner, 1979; Turner et al., 1987). Simply put, in order for one to perceive threats at the group level, one first needs to identify with the group.
Research on intergroup relations also argues that under certain conditions positive intergroup contact might have a dampening effect on motiva-tions to engage in collective action, especially among the members of the disadvantaged groups (J. Dixon, Levine, Reicher, & Durrheim, 2012) by improving outgroup attitudes toward the advan-taged group (Reicher, 2007), disempowering the disadvantaged by giving them the illusion of a meritocratic system (Saguy, Tausch, Dovidio, & Pratto, 2009), and by decreasing perceptions of discrimination and group efficacy (Çakal, Hewstone, Schwar, & Heath, 2011; Tropp, Hawi, van Laar, & Levin, 2012). We believe, however, that intergroup contact might also demotivate individuals from engaging in collective action via perceived threats. Previous research has shown that contact reduces threat perceptions (Curşeu, Stoop, & Schalk, 2007; González, Verkuyten, Weesie, & Poppe, 2008; Schlueter & Scheepers, 2010; Schmid, Tausch, Hewstone, Hughes, & Cairns, 2008; Tausch, Tam, Hewstone, Kenworthy, & Cairns, 2007). Thus, contact with members of the outgroup will decrease perceptions of threat, and threats will positively influence willingness to engage in collective action to benefit the ingroup. In what follows, we detail these processes within the intergroup threat perspective.
Integrated Threat Theory
Integrated threat theory (ITT; Stephan & Stephan, 2000; Stephan et al., 2009) distinguishes between realistic and symbolic threats. Realistic threats reflect the subjective representations of competition and conflict over political or economic resources, whereas symbolic threats represent the incompati-bility of values, norms, and beliefs between the ingroup and the outgroup (Riek et al., 2006).
ITT argues that the relative status of the group and the history of intergroup conflict are particu-larly important and influence perceptions of
threats among the advantaged and disadvantaged groups differently. Low-status, disadvantaged groups are more prone to feel threatened as they control less resources and therefore can be over-powered by the advantaged groups. High-status, advantaged groups’ reactions to perceived threats, on the other hand, are stronger as they control more resources and they have more to lose should the attempts by the disadvantaged group to chal-lenge the status quo be successful (Stephan et al., 2002; Stephan et al., 2009). With regard to the relation between group status and the dimensions of perceived threats, however, both the social identity approach (Reicher, Spears, & Haslam, 2010; Tajfel & Turner, 1979) and ITT (Stephan et al., 2009) also suggest that when groups are involved in intractable conflicts members of both groups are particularly likely to perceive them-selves to be threatened irrespective of their group’s status.
According to ITT, perceived threats have a number of specific cognitive, affective, and behavioural outcomes—for example, prejudice (Aberson & Gaffney, 2008; Tausch, Hewstone, Kenworthy, Cairns, & Christ, 2007); negative emotions, e.g., fear, anxiety, disgust, and anger (Cottrell & Neuberg, 2005; Renfro, Duran, Stephan, & Clason, 2006); aggression toward the outgroup (Maass, Cadinu, Guarnieri, & Grasselli, 2003); and opposition to outgroup-oriented poli-cies (O’Brien, Garcia, Crandall, & Kordys, 2010). These consequences depend on (a) the content and the context of threats; and (b) the subjective identification with the group. Research on social identity (Reicher et al., 2010; Tajfel & Turner, 1979) for instance maintains that threats to differ-ent dimensions of social iddiffer-entity—moral values (Täuber & van Zomeren, 2013), positive distinc-tiveness (Wohl, Giguère, Branscombe, & McVicar, 2011), as well as to the overall value of social identity (Branscombe et al., 1999; Jetten, Spears, & Manstead, 1997)—could also elicit group-defensive behaviour aimed at removing the source of threat, especially among high iden-tifiers (Branscombe et al., 1999). Similarly, Grant and Brown (1995) found that female students who strongly identified as female were more
willing to engage in collective action to improve the conditions for females under conditions of high collective relative deprivation and high social identity threat. More recently, Wohl et al. (2011) investigated the combined effects of social iden-tity and distinctiveness threat on ingroup-protec-tive behaviour via concerns for the ingroup’s future existence as a distinct group. They found that Québécois Canadians and Canadians in gen-eral were more likely to engage in ingroup-pro-tecting behaviour when they were threatened with losing their French Canadian and Canadian distinctiveness, respectively. Exploring the conse-quences of threats to social identity, Täuber and van Zomeren (2013) found that threats to the moral status of the Dutch ingroup predicted more outgroup-oriented outrage which, in turn, predicted reactions to defend social identity.
ITT argues that perceived threats are an essen-tial aspect of intergroup relations in contexts where real groups (e.g., ethnic, racial, or religious) are involved in ongoing conflicts. As such, they can be experienced by both advantaged and dis-advantaged groups and therefore motivate both groups to engage in collective action despite a widespread assumption that advantaged and dis-advantaged groups have disparate psychologies regarding social change and collective action (Saguy et al., 2009; Wright & Baray, 2012; Wright & Lubensky, 2009). Thus, we are neither the first nor the only researchers to study the link between perceived threats and collective action. Earlier research on social identity (Branscombe et al., 1999; Tajfel & Turner, 1979), ITT (Stephan et al., 2009), and more recent research on collective action (van Zomeren et al., 2012; van Zomeren et al., 2008) have investigated the relationship between threats and group-serving behaviour. However, we believe that exploring the explicit role of perceived threats as antecedents of collec-tive action among structurally advantaged and disadvantaged groups locked in a conflict might help to extend our understanding of what moti-vates individuals to engage in collective action among different status groups especially in the context of sustained conflict (Abrams & Grant, 2012; van Zomeren et al., 2008). This is especially
important as current accounts of collective action have mostly focused on disadvantaged groups and this exclusive focus is partly dictated by the premise that individuals engage in collective action to redress a group-based inequality.
Intergroup conflict arises when a group’s col-lective interests are challenged by another group’s actions, value systems, or by the sheer size of the other group. For the majority this challenge may mean the loss of “majority status” and surrender-ing the tangible and symbolic resources associ-ated with this status (Craig & Richeson, 2014) whereas for the minority it may simply mean a greater challenge to their well-being and survival. In protracted conflicts, such challenges are trans-formed or perceived as more imminent and real then they actually are, motivating group members to act collectively to protect group interests (Effron & Knowles, 2015; Stephan & Stephan, 2000; Stephan et al., 2009). Thus, focusing on how such threats motivate members of advan-taged and disadvanadvan-taged groups to protect the ingroup’s interests could inform collective action research as most research on collective action focuses on incidental or novel collective disad-vantages as opposed to structural or subjective disadvantages (Abrams & Grant, 2012).
Moreover, such an investigation among adult populations in a conflict-ridden context could also enable us to go beyond the usual student or WEIRD (Western, educated, industrialized, rich, and democratic; Henrich, Heine, & Norenzayan, 2010) samples obtained within relatively less con-flictual contexts. Finally, as our previous discus-sion indicates, perceived threats are positively associated with social identity (Stephan et al., 2009), and negatively associated with positive intergroup contact (Tausch, Tam, et al., 2007). As such, they could mediate the effects of these two important predictors of collective action, also allowing us to contribute to the current debate on the positive effects of social identity on collective action on the one hand, and on the delimiting effects of intergroup contact on collective action, on the other. There is therefore merit in investi-gating the predictive power of perceived threats on collective action among both advantaged and
disadvantaged groups that have historically been engaged in a violent conflict.
Present Research
The present research tests perceived threats in the context of overt conflict among the advan-taged Turks and disadvanadvan-taged Kurds in Turkey. We propose that members of both groups will perceive the outgroup as threatening, albeit for different reasons. Turks, as members of the advantaged group, may feel threatened, both in realistic and symbolic terms, in the face of ongo-ing mobilization of the disadvantaged group, Kurds, which challenges the foundations of the advantaged group’s economic and political supe-riority by advancing their agenda on equal rights (Ayata & Yükseker, 2005; Okamoto & Ebert, 2010; Rohmann, Piontkowski, & van Randenborgh, 2008; Stephan et al., 2009; Tajfel & Turner, 1979; van Zomeren et al., 2004; Yeğen, 2011). As recent research evinces Kurds are per-ceived as uneducated and feudal, and thereby pos-ing a barrier to Turkey’s modernization attempts as well as its EU aspirations. (J. C. Dixon & Ergin, 2010). For the disadvantaged group, Kurds, the social, political, and economic differences between the groups are large and widening dis-parities threaten both the Kurdish way of life and the very existence of Kurds. Depending on the psychological and material resources available to the group, such perceptions might lead to coping-related behaviour to protect their already existing assets, that is, collective action (Branscombe et al., 1999; Greenglass, 2002; Lazarus, 1991; Leach et al., 2008). Alternatively, if the group is too overwhelmed and devoid of these resources, perceived threats could actually inhibit such behaviour. In the present context, given the recent Kurdish advances in the political sphere (Weaver, 2015), the Turkish state’s inability to crush the Kurdish insurgence on one hand (Sirivastava, 2015), and the obvious advantaged position of Turks as the majority on the other , threats could lead to coping-related behaviour, that is, collective action among members of both groups.
Overview, Research Context, and
Hypotheses
To test these ideas, we conducted two studies in the south-eastern city of Mersin where ethnic Turks are the majority group with a sizeable Kurdish minority, thus testing our predictions among both the advantaged (Study 1: Turks) and the disadvantaged (Study 2: Kurds). Although the conflict between the two groups has been going on for decades, the military and violent dimen-sions of the strife have mainly been confined to rural encounters between the members of PKK (Kurdistan Workers’ Party) and Turkish Army (Akkaya & Jongerden, 2011). The social, eco-nomic, and cultural aspects of the conflict have come to play a more prominent role in the last two decades (Yeğen, 2011) when the Turkish state evacuated 3,000 Kurdish villages in the pre-dominantly Kurdish areas and dispersed their inhabitants with the aim of curbing popular and logistic support to the PKK insurgents. This led to the migration of one quarter of the overall Kurdish population—around four million—to urban areas such as Istanbul, Izmir, and Mersin (Ayata & Yükseker, 2005; Yöruk, 2012).
This population shift then has had significant bearings for the conflict. Firstly, it delineated cultural and social differences between the two communities and created competition over resources. With this in perspective, we operationalize threats as attitudes, that is, subjective evaluations of the actual condi-tions rather than as objective existential threats and we predict that perceived threats will be positively associated with collective action tendencies but neg-atively associated with positive outgroup attitudes. We also think that such abrupt mass displacement has made the ingroup identities chronically salient and these salient ingroup identities would predict collective action tendencies both directly and via perceived threats (Stephan et al., 2009).
Secondly, increasing number of Kurds in urban centres also created opportunities for con-tact and possible improvement of intergroup relations between the two groups. Therefore, to the extent that contact opportunities are exploited, frequent contact with the outgroup
would reduce collective action tendencies (“the sedative effect”) and would improve outgroup evaluations by reducing threats.
Study 1
Method
Participants. Two hundred and eighty-nine adults
(178 females and 111 males, Mage = 31.98 and
SD = 10.92) who identified themselves as
Turk-ish participated in the study on a voluntary basis. They were recruited from Mersin, a multiethnic city in southeast Turkey by a Turkish research assistant who visited them at their homes and invited them to participate in the study. None of the contacted participants refused to participate. Once the participants agreed to take part, they were handed in the questionnaire which they filled in and returned to the research assistant. The majority of the participants (61.4 %) had completed secondary education, with 12.4 % having completed primary education only, and 25.2% having completed a degree course in higher education. In terms of religious orienta-tion, 82% self-identified as Sunni, 14% as Alevi, and 4% declined to report any religious affilia-tion. As for political orientation, 41% identified as right, 16% as left, 40% as centre, and the remaining 3% did not report political orientation.
Measures. Variables were measured on 5-point
scales where higher values indicate stronger ingroup identification, more intergroup contact, higher levels of perceived symbolic and realistic threats, stronger collective action tendencies, and more positive outgroup evaluations (for contact items: 1 = never, 5 = very often; for outgroup
evalu-ation items: 1 = not at all, 5 = very much; for other
items: 1 = strongly disagree, 5 = strongly agree). Predictors. To measure ingroup
identifica-tion we adapted three items from Luhtanen and Crocker (1992): “I am proud to be Turkish,” “In general, being Turkish is an important part of my self-image,” and “I am very happy to be Turkish”
(α = .89). Intergroup contact was measured with three items (α = .90): “How often do you talk to your Kurdish friends?” “How often do you spend time with them socially?” and “How often do you visit them at their home?”
Mediators. We used three items to measure
per-ceived realistic threat (α = .72): “Kurds have too much economic power in this country,” “Kurds have more political rights than Turks,” and “Too much money is spent on education policies that favour Kurds.” Symbolic threat was measured with two items (α = .75, r = .60, p < .001) adapted
from Stephan et al. (2002): “Turks and Kurds have different values,” and “Kurds and Turks think differently regarding many aspects of life.” Given that our aim was to investigate the role of perceived threats—realistic and symbolic, as rep-resentations of various differences in access to economic and social resources, and cultural and social differences, for example, linguistic or value-based—we believe that the items we used tap into perceptions of threats.
Outcome variables. Three items (α = .88) were
adapted from Smith, Cronin, and Kessler (2008) to assess collective action tendencies among the Turkish ingroup: “I would be willing to sign a petition to improve the current situation of Turks in Turkey,” “I would be willing to sign up for a neighbourhood project to improve the conditions for Turks in my neighbourhood,” and “I would be willing to participate in a peaceful demonstra-tion to improve the current condidemonstra-tions for Turks in Turkey.” To measure evaluations of the out-group, we used three positive trait adjectives (α = .82) from Duckitt and Mphuthing (1998). Par-ticipants reported how characteristic the terms “kind,” “polite,” and “sincere” were of Kurds.
Results and Discussion
Descriptive statistics of the variables in the model are given in Table 1. We created latent variables and used structural equation modelling with the Mplus software package (Muthén & Muthén, 2008a, 2008b) to test our model. Missing data
were less than 2% so we did not treat them statis-tically. We used robust maximum likelihood esti-mation (MLR; Schermelleh-Engel, 2003) to handle any possible nonnormality in the data. The model fit was assessed using the χ2 test, χ2/df
ratio, RMSEA (root mean square error of approximation), CFI (comparative fit index), and SRMR (standardized root mean square residual) criteria. Cut-off points for these fit indices were: a nonsignificant χ2 value (Barrett, 2007); χ2/df
ratio lower than or equal to 3; .95 or higher for CFI; .06 or lower for RMSEA; and .08 or lower for SRMR (Bentler, 2007; Hu & Bentler, 1999). Confirmatory factor analysis (CFA) revealed that there were satisfactory loadings for all observed items within the range of β = .60−.85 (Hair, Black, Babin, & Anderson, 2010).
ITT argues that realistic and symbolic threats are theoretically distinct. Therefore we tested a model (Model 1, χ2 = 148.65, p = .005, df = 111,
χ2/df = 1.33, CFI = .97, RMSEA = .037, SRMR =
.066) where realistic and symbolic threat items were allowed to load on separate factors. However, empirical research suggests that these two dimensions of perceived threats are often highly correlated (e.g., Kauff & Wagner, 2012) as was the case in the present study (r = .57, p = .001). Therefore, we tested a second model
(Model 2, χ2 = 156.97, p = .003, df = 112, χ2/df =
1.40, CFI = .97, RMSEA = .037, SRMR = .056) where realistic and symbolic threat items were allowed to load on a single factor. Finally, we tested a third model in which we employed a sec-ond-order factor structure to emphasize the relat-edness of the two dimensions as underlined by a
higher order construct (Model 3: χ2 = 120.10,
p = .240, df = 110, χ2/df = 1.09, CFI = .99,
RMSEA = .018, SRMR = .049; Marsh & Hocevar, 1988). We then compared the relative fit of three models using a Satorra-Bentler scaled χ2
differ-ence test which adjusts for the correction factor when using MLR (Kline, 2011; Satorra & Bentler, 2010). The results of the chi-square difference tests supported the model with the second-order factor structure approach to perceived threats, which fit the data significantly better (Model 3 vs. Model 1: Δ χ2(1) = 26.67, p < .001; Model 3 vs.
Model 2: χ2(2) = 36.12, p < .001). We therefore
retained Model 3 which allowed us to emphasize the relatedness of the two dimensions as under-lined by a higher order construct (Chen, Sousa, & West, 2005; Marsh & Hocevar, 1988).
Figure 1 shows the overall results. As we pre-dicted, intergroup contact was significantly and negatively associated with perceived threats (β = −.26, p < .001). Ingroup identification too
was significantly but positively associated with per-ceived threats (β = .54, p < .001) and collective
action tendencies (β = .41, p < .001). Perceived
threats, in turn, were significantly and positively associated with collective action tendencies (β = .18, p < .001). Additionally, contact was
sig-nificantly and positively associated with outgroup evaluations (β = .48, p < .001), and the association
between threats and outgroup evaluations was also significant but negative (β = −.21, p < .001). In line
with earlier research (Çakal et al., 2011), there was no significant direct association between inter-group contact and collective action (β = −.08,
p = .109), and between ingroup identification and
Table 1. Descriptive statistics and correlations between the latent variables in the model (Study 1).
Variable M SD 1 2 3 4 5 6 Intergroup contact 2.80 0.98 1 −.09ns −.34*** −.23** −.05ns .57*** Ingroup identification 3.32 0.94 1 .54*** .45*** .54*** −.24*** Realistic threat 2.80 0.93 1 .57*** .45*** −.44*** Symbolic threat 3.34 0.95 1 .38*** −.26*** Collective action 3.61 1.01 1 −.13* Outgroup evaluations 2.88 0.89 1 **p < .05. ***p < .001.
outgroup evaluations (β = −.02, p = .391). The
model explained 37% and 27% of variance in our criterion variables, outgroup evaluations and col-lective action, respectively, and 37% of variance in our mediator variable, threats.
Although our results are in line with earlier research on collective action, threats, and inter-group contact, one could argue for alternative causal paths between our variables. For instance, intergroup contact might negatively predict ingroup identification, via deprovincialization (Verkuyten, Thijs, & Bekhuis, 2010), and ingroup identification in turn might predict collective action and outgroup tendencies via perceived threats (Riek et al., 2006). Specifically, those who have frequent contact with the outgroup might
distance themselves from the ingroup, perceiving the outgroup as less threatening and demonstrat-ing less willdemonstrat-ingness to engage in collective action for the ingroup. The alternative model, however, in which we tested the impact of intergroup con-tact on our outcome variables via ingroup identifi-cation and perceived threats fit the data significantly worse than our model (Alternative Model 1: χ2 = 160.01, p = .001, df = 109, χ2/df =
1.47, CFI = .96, RMSEA = .040, SRMR = .084; Model 3 vs. Alternative Model 1: Δ χ2(2) = 38.03,
p < .001), with a significant χ2 value. A second
alternative model in which perceived threats pre-dicted collective action and outgroup evaluations via ingroup identification and intergroup contact fit the data slightly better than our first alternative
Figure 1. Saturated model showing contact and social identity predicting collective action and outgroup
evaluations via perceived threats among Turks in Turkey (N = 289). Model fit values: χ2 = 120.10, p = .240,
df = 110, χ2/df = 1.09, CFI = .99, RMSEA = .018, SRMR = .049. Correlations between variables: contact–social identity r = .08, ns; contact–collective action r = −.08, ns; ingroup identification–outgroup evaluations r = −.02, ns;
collective action–outgroup evaluations r = .08, ns. Standardized coefficients; only significant paths are reported.
model (Alternative Model 2: χ2 = 157.07, p = .002,
df = 109, χ2/df = 1.43, CFI = .97, RMSEA = .039,
SRMR = .84), but still worse than our original model (Model 3 vs. Alternative Model 2: Δ χ2(2) =
35.09, p < .001). A final alternative model
(Alternative Model 3) which tested collective action and outgroup evaluations as predictors (van Zomeren et al., 2012), perceived threats as mediators, and intergroup contact and ingroup identification as outcome variables, also fit the data significantly worse than our original model (χ2 = 144.28, p =.018, df = 108, χ2/df = 1.33, CFI
= .98, RMSEA = .032, SRMR = .059; Model 3 vs. Alternative Model 3: Δ χ2(3) = 22.30, p <.001). All
three alternative models were therefore rejected.
Mediation by perceived threats. In line with our
hypotheses, we also investigated whether our pre-dictor variables had any indirect effects on our out-come variables via perceived threats. We used the bias corrected bootstrap command in Mplus (Muthén & Muthén, 2008b) and created confi-dence intervals based on 5,000 resamples to test whether indirect paths were significantly different from zero. The effect sizes are represented by point estimates (PE) and their values are consoli-dated through confidence intervals. The indirect effect of a predictor is significant when confidence intervals (CI) do not include zero (Hayes, 2009; Preacher & Hayes, 2008). We report the significant indirect effects in Table 2. Intergroup contact had an indirect negative effect on collective action (PE β = −.05, 95% CI [−0.06, −0.03]) and a positive
effect on outgroup evaluations (PE β = .06, 95 % CI [0.04, 0.07]) via perceived threats. Ingroup iden-tification on the other hand had a positive indirect effect on collective action (PE β = .10, 99 % CI [0.07, 0.12]) but a negative effect on outgroup eval-uations (PE β = −.11, 99 % CI [−0.14, −0.09]) via perceived threats.
Results revealed that social identity, measured here as ingroup identification, was a significant pre-dictor of collective action tendencies both directly (van Zomeren et al., 2008), and indirectly via per-ceived threats, suggesting that social identity can impact collective action via perceived threats. In addition, we further analysed whether realistic and symbolic threats differ in their association with col-lective action tendencies.1 These findings also
cor-roborate previous research in showing that contact can have a “sedative” effect among the members of the advantaged group too (Çakal et al., 2011; Saguy, Dovidio, & Pratto, 2008). Having tested the impact of threats on collective action among the members of the majority group, we conducted a second study to investigate whether threats would predict collec-tive action tendencies among the members of the disadvantaged group in the same context.
Study 2
Method
Participants. Two hundred and nine adults (79
females and 130 males, Mage = 31.54 and SD =
12.04) who identified themselves as Kurdish,
Table 2. Mediation bootstrap test results (Study 1).
Path Mediator Point estimate (β) 95% CI 99% CI
Intergroup contact–
Collective action Perceived threats −.046 [−0.058, −0.034]
Intergroup contact –
Outgroup evaluations Perceived threats .055 [0.043, 0.066]
Ingroup identification–
Collective action Perceived threats .096 [0.073, 0.119]
Ingroup identification–
Outgroup evaluations Perceived threats −.112 [−0.138, −0.086]
Note. Bootstrap is based on 5,000 resamples. When confidence intervals do not include zero this shows that there is a signifi-cant indirect effect (Preacher & Hayes, 2008; Williams & MacKinnon, 2008). Standardized coefficients are shown.
from Mersin, Turkey, were recruited by a Kurd-ish-speaking research assistant who visited them at their homes. Similar to Study 1, after consent-ing to participate, the participants were handed in the survey containing our measures that they duly completed and returned to the researcher. Almost half of the participants (49%) had completed sec-ondary education, 29.8% had completed primary education, and 19.7% reported having completed tertiary education. Religiously, 74.6% identified as Sunni Muslims, with 17.4% as Alevi, and a fur-ther 9.2% not reporting any religious affiliation. In terms of political orientation, 32.4% identified as right, with further 29.3% as centre, 26.1% as left, and 12.2% declined to answer this question.
Measures. We used the same items, measured on
5-point scales, as in Study 1 for all of our measures.
Predictors. Ingroup identification was
meas-ured by three items “I am proud to be Kurdish,” “In general, being Kurdish is an important part of my self-image,” and “I am very happy to be Kurdish” (α = .83). Quantity of intergroup con-tact was measured by three items “How often do you talk to your Kurdish friends?” “How often do you spend time with them socially?” and “How often do you visit them at their home?” (α = .86).
Mediators. Perceived realistic threats was
meas-ured by three items (α = .77) “Turks have too much economic power in this country,” “Turks have more political rights than Turks,” and “Too much money is spent on education policies that
favour Turks.” Symbolic threat was measured by two items (α = .71, r = .56, p < .001) “Kurds and
Turks have different values,” and “Kurds and Turks think differently regarding many aspects of life.” Despite the fact that the conflict between the two groups is overt and has been marked by occasional violence, we believe that the items we used tap into a dimension of threats we want to measure better than items that focus on more tangible threats. Kurds still perceive that the sys-tem is too tolerant of the outgroup Turks and that their rights are not being protected enough, as Turks continue to enjoy better access to politi-cal, economic, and social rights.
Outcome variables. Collective action for the
Kurdish ingroup was measured by two items (α = .82, r = .69, p < .001) “I would be willing to sign
a petition to improve the current situation of Kurds in Turkey” and “I would be willing to par-ticipate in a peaceful demonstration to improve the current conditions for Kurds.” We measured outgroup evaluation with the same three positive trait adjectives (α = .80) as in Study 1. Participants reported how characteristic the terms “kind,” “polite,” and “sincere” were of Turks.
Results and Discussion
We report the descriptive statistics in Table 3. We followed the same analytic approach as in Study 1, and tested a series of models in which we mod-elled threats as distinct constructs (Model 1: χ2 =
130.68, p = .061, df = 112, χ2/df = 1.16, CFI =
.96, RMSEA = .044, SRMR = .079), as a unitary
Table 3. Descriptive statistics and correlations between the latent variables in the model (Study 2).
Variable M SD 1 2 3 4 5 6 Intergroup contact 2.96 0.92 1 −.12* −.31*** −.20** −.13*** .34*** Ingroup identification 3.79 0.95 1 . 58*** .45*** .50*** −.17* Realistic threat 3.76 0.96 1 .59*** .50*** −.32*** Symbolic threat 3.52 0.99 1 .41*** −.25*** Collective action 4.18 0.87 1 −.10* Outgroup evaluations 2.78 0.083 1 **p < .05. ***p < .001.
construct where all threat items are allowed to load on a single factor (Model 2: χ2 = 138.95,
p = .055, df = 110, χ2/df = 1.26, CFI = .95,
RMSEA = .048, SRMR = .078); and finally a model in which symbolic and realistic threats are modelled using a second-order factor structure (Model 3: χ2 = 112.15, p = .075, df = 111, χ2/df = 1.01,
CFI = .98, RMSEA = .033, SRMR = .059), to model realistic and symbolic threats as distinct dimensions of a unified perceived threat variable. Chi-square difference tests favoured the model with the second-order factor structure (Model 3 vs. Model 1: Δ χ2(1) = 17.14, p < .001; Model 3 vs.
Model 2: χ2(2) = 23.70, p < .001). Results
are shown in Figure 2. As we predicted, inter-group contact was significantly and negatively
associated with perceived threats (β = −.23,
p < .001). Ingroup identification too was
signifi-cantly but positively associated with perceived threats (β = .56, p < .001) and collective action
tendencies (β = .34, p < .001). Perceived threats,
in turn, were significantly and positively associ-ated with collective action tendencies (β = .22,
p < .001), whereas the association between threats
and outgroup evaluations was also significant but negative (β = −.25, p < .001). Additionally,
con-tact was significantly and positively associated with outgroup evaluations (β = .24, p < .05). Similar to
Study 1, there was no significant direct association between intergroup contact and collective action (β = .04, p = .267), or between ingroup
identifica-tion and outgroup evaluaidentifica-tions (β = −.06, p = .298).
Figure 2. Saturated model showing contact and social identity predicting collective action and outgroup evaluations
via perceived threats among Kurds in Turkey (N = 209). Model fit values: χ2 = 112.15, p = .075, df = 111, χ2/df = 1.01, CFI = .98, RMSEA = .033, SRMR = .059. Correlations between variables: contact–social identity r = −.12, ns;
contact–collective action r = .04, ns; ingroup identification–outgroup evaluations r = −.08, ns; collective action–
outgroup evaluations r = −.05, ns. Standardized coefficients; only significant paths are reported.
The model explained 16% and 27% of variance in our criterion variables, outgroup evaluations, and collective action, respectively, and 43% of variance in our mediator variable threats.
The findings confirmed our expectations as social identity directly predicted collective action tendencies, and perceived threats mediated not only this relation but also the effects of contact on collective action. In line with ITT, stronger identification with the ingroup predicted higher levels of perceived threats which, in turn, nega-tively predicted positive outgroup evaluations. Next, we explored alternative models. As in Study 1, we first tested a model with contact as predic-tor, ingroup identification and perceived threats as mediators, and collective action and outgroup attitudes as outcome variables. This model fit the data worse than our proposed model (χ2 =
137.52, p = .001, df = 109, χ2/df = 1.51, CFI =
.95, RMSEA = .050, SRMR = .090; Δ χ2(2) =
8.19, p < .05). We then tested the role of
per-ceived threats on collective action and outgroup evaluations via ingroup identification and inter-group contact. Again this model demonstrated worse fit than our original model (χ2 = 165.23, p
= .000, df = 109, χ2/df = 1.81, CFI = .92, RMSEA
= .063, SRMR = .92; Δ χ2(2) = 14.13, p < .001).
Finally, we tested collective action and outgroup evaluations as predictors, perceived threats as mediators, and intergroup contact and ingroup identification as outcome. This model did not converge. We therefore rejected all three alterna-tive models.
Mediation by perceived threats. We report the indirect
effects in Table 4. Similar to Study 1, contact reduced perceptions of threat, which, in turn, were associated with reduced willingness to engage in collective action and with more favour-able outgroup attitudes. More specifically, inter-group contact had an indirect negative effect on collective action (PE β = −.05, 95 % CI [−0.09, −0.02]) but was positively associated with out-group evaluations (PE β = .06, 95 % CI [0.03, 0.09]) via perceived threats. Ingroup identifica-tion had an indirect positive effect on collective action (PE β = .12, 99 % CI [0.08, 0.17]) and an indirect negative effect on outgroup evaluations (PE β = −.14, 99 % CI [−0.21, −0.07]).
The indirect effects showed that perceived threats as perceptions of the prevailing condi-tions of intergroup relacondi-tions mediated the effects of both intergroup contact and social identity on outgroup evaluations and collective action, repli-cating our model among members of the disad-vantaged group.2
General Discussion
Across two studies we provided evidence for the predictive power of perceived threats on collec-tive action tendencies. We found that perceived threats could motivate mobilization among mem-bers of both advantaged and disadvantaged groups. Moreover, perceived threats also medi-ated the effects of social identity and intergroup contact on outgroup attitudes and collective
Table 4. Mediation bootstrap test results (Study 2).
Path Mediator Point estimate (β) 95% CI 99% CI
Intergroup contact–
Collective action Perceived threats −.054 [−0.085, −0.023]
Intergroup contact–
Outgroup evaluations Perceived threats .060 [0.030, 0.091]
Ingroup identification–
Collective action Perceived threats .124 [0.076, 0.173]
Ingroup identification–
Outgroup evaluations Perceived threats −.138 [−0.207, −0.070]
Note. Bootstrap is based on 5,000 resamples (Preacher & Hayes, 2008; Williams & MacKinnon, 2008). Standardized coef-ficients are shown.
action tendencies. Although we did not investi-gate perceived threats vis-à-vis other possible predictors of collective action such as group norms, anger, group efficacy, or relative depriva-tion, our model explained a decent amount of variance in our outcome variables. We discuss the implications of these findings in what follows.
Implications for Collective Action and
Intergroup Contact Literature
A number of implications follow from the two studies reported in this article. First and foremost, the findings lend support to the predictive power of threats on collective action. This influence seems to be independent of the structural posi-tion of the group in quesposi-tion, as demonstrated by findings from both studies, and challenges the assumption that advantaged and disadvantaged groups are subject to different psychological pro-cesses at least with regard to perceived threats. As Pettigrew (2006) argues, the role of context in determining these psychological processes is important. In contexts where the status and power imbalance is stable, for example, Israeli Jews ver-sus Israeli Arabs (Saguy et al., 2008), with little hope of change, groups might have different per-ceptions of the situation and therefore might have different motivations. In other contexts, for exam-ple, Turks versus Kurds, where the power and sta-tus imbalance is less stable, similar psychological processes might be at work among members of both groups. These observations therefore invite more research to test the “different psychologies” assumption. Research comparing the impact of intergroup contact on other predictors of collec-tive action among advantaged versus disadvan-taged group members across nonviolent and violent contexts is particularly welcome. The empirical similarity of perceived realistic and sym-bolic threats, as supported by our data, also con-tradicts the assumption that factors such as group power, prior conflict, and relative group size are more related to realistic threats than symbolic threats (Stephan et al., 2009). Most of the sup-porting evidence to this assumption comes from research conducted in western settings among
host and immigrant populations with no apparent violent conflict or real time competition over tan-gible resources, or threats to physical well-being (Crisp, Stone, & Hall, 2006; Riek et al., 2006). As our findings suggest however, majority–minority distinction in terms of perceived threats seem to disappear when groups are involved in a violent conflict.
Second, the indirect impact of social identity on collective action is not limited to already estab-lished predictors of collective action, for exam-ple, anger, relative deprivation, or group efficacy. Perceived threats, as subjective evaluations of the actual intergroup context, can also mediate this relation as our findings imply. The question of whether perceived threats can motivate individu-als over and above other predictors of collective action, for example, anger, relative deprivation, or group efficacy, remains to be answered in future research. As discussed earlier, threats could make group identities more salient, motivating the ingroup members to close ranks and to identify with their group more (Ethier & Deaux, 1994; Schmid & Muldoon, 2015). This in return could predict willingness to engage in collective action or increased outgroup bias both directly or via anger (DeSteno, Dasgupta, Bartlett, & Cajdric, 2004) or perceptions of deprivation (Pettigrew et al., 2008). Alternatively, perceiving the ingroup as relatively deprived could also threaten a posi-tive group identity as well as perceiving the mate-rial sources of the ingroup as threatened by the outgroup (Grant & Brown, 1995; Smith, Pettigrew, Pippin, & Bialosiewicz, 2012). In the absence of these variables, it is difficult to con-clude how empirically different threats are from these constructs. Future research could simulta-neously test the impact of perceived threats and these processes in more complex models. We believe that empirical differentiation of chronic collective disadvantage and perceived threats might be a particularly fruitful avenue to pursue. Chronic collective disadvantage can also be per-ceived as threatening, and this too predicts antici-patory coping, a response in the form of collective action, in reaction to future harm or threat (Lazarus, 1991).
We also need more research on threats vis-à-vis group efficacy. In situations where the system is perceived as unstable and the group as effica-cious we expect threats to motivate individuals to engage in collective action to protect or improve the ingroup’s conditions. If, however, the system is perceived as stable and the group is perceived as incapable and having no resources to mobilize to challenge the system, then threats could dampen motivations to engage in collective action. Finally, it would seem intuitive to think that these conditions would apply more to the disadvantaged group, Kurds, than the advantaged group, Turks, in the present research. Recently escalating violence between Kurds and Turks both in rural areas and urban centres such as Istanbul and Ankara however shows that threats are real for both groups.
Third, as discussed elsewhere, identification with the ingroup is associated with perceptions of threats, which, in turn, as our findings indicate, are associated with collective action tendencies. The relation between identification with a group and collective action via threats might, however, also depend on the level of conformity to group norms (Louis, Taylor, & Douglas, 2005). Extensive research suggests that a salient social identity leads individuals to think and act in line with the norms of that social identity (Blackwood & Louis, 2012). Norms then influence the costs or benefits, at both individual and group levels, of engaging in collective action (Louis et al., 2005). Because acting to remove a threat that challenges the goals or the well-being of the in-group (Effron & Knowles, 2015) is a process that involves costs to the individual, it is safe to assume that stronger adherence to ingroup norms that prescribe the protection and maintenance of group interests will positively influence the link between identification with the group, perceived threats, and willingness to engage in collective action. Future research could investigate this moderating effect of norms on the social iden-tity–perceived threats–collective action path by manipulating the salience of group norms or in more complex models based on correlational data.
Fourth, our data imply that contact tends to reduce collective action tendencies indirectly (J. Dixon et al., 2012), and this effect works among the members of the advantaged as well as mem-bers of the disadvantaged group. However, as discussed elsewhere, the impact of contact on collective action is a rather complex and multidi-mensional one (Pettigrew, 2010). Contact reduces collective action tendencies among the advan-taged, thus contributing towards a weakening of the advantaged group’s determination to main-tain existing conditions. As for the disadvantaged group, findings imply that frequent contact with members of the advantaged group results in psy-chological consequences which might prevent the disadvantaged from entering into direct competi-tion with the advantaged. An optimal solucompeti-tion to this paradox would be to examine conditions under which contact contributes toward building solidarity with members of the disadvantaged group while at the same time allowing members of the disadvantaged group to preserve their will-ingness to mobilize.
Fifth, the indirect effects of intergroup con-tact on collective action tendencies also call for more research on other mediators of intergroup contact, for example, perspective taking, anxiety, or even trust (Islam & Hewstone, 1993; Tam, Hewstone, Kenworthy, & Cairns, 2009), would either negatively or positively mediate ingroup- or outgroup-oriented collective action (Mallett, Huntsinger, Sinclair, & Swim, 2008). Alternatively, research has also suggested inter-group contact play a significant role in inducing people to engage in collective action among members of the same group (Çakal, Eller, Sirlopú, & Perez, in press). It makes intuitive sense therefore to expect contact to contribute toward the formation of solidarity between the advantaged and disadvantaged groups via per-ceived shared threats (Subašić, Reynolds, & Turner, 2008). Conversely, if positive contact has a sedative effect, as our findings also attest, it would be interesting to investigate whether negative contact (Barlow et al., 2012) would motivate individuals to engage in collective action.
Limitations of the Present Research
Given our cross-sectional data, we acknowledge that the causal flow of our model—from inter-group contact and social identity to perceived threats, and then to collective action—should be interpreted with caution. It is possible for instance that both perceived threats and expres-sions of prejudice toward the ingroup, for example, discrimination or biased treatment, can positively influence the strength of identifi-cation with the threatened or discriminated group (Jetten, Branscombe, Schmitt, & Spears, 2001; Leach et al., 2008; Tajfel & Turner, 1979). Specifically, research suggests that perceiving the outgroup as a source of threat makes the group boundaries more salient (Branscombe et al., 1999; Doosje & Ellemers, 1997; Doosje et al., 2002) which, in turn, motivates the mem-bers of the threatened group to identify more with the threatened ingroup and protect or improve the image of the ingroup (Branscombe et al., 1999; Branscombe & Wann, 1994; Schmid & Muldoon, 2015). In a similar vein, being dis-criminated against on the basis of one’s group membership might also be associated with a heightened sense of identification with the ingroup in an attempt to reject the very same standards imposed upon the ingroup by the dominant outgroup (Jetten et al., 2001; Tajfel & Turner, 1979). Therefore, it is possible to con-ceptualize the relation between threats and social identity as a reciprocal one. However, as outlined earlier, we believe identification as Turkish and Kurdish precedes perceptions of threats due to the historical conditions which led to the displacement of Kurds which, in turn, triggered a range of perceived threats among our participants. Nevertheless, despite the fact that our findings tally with those of previous research on threats, results are prelimi-nary and should be replicated experimentally and longitudinally. Future research should seek to experimentally manipulate both threats and identification with the ingroup for a better understanding of the relationship between these two variables.
A second limitation of the present research relates to our choice of measures. Firstly, we did not control for other predictors of collective action. We could have measured the level of negative affect regarding the structural positions of the groups or perceptions of group efficacy. Similarly, including measures of collective disadvantage and empirically differ-entiating perceived threats from collective disadvantage would have provided a stronger test of the role of perceived threat. We need, therefore, more research to test simultaneously the combined effects of perceived threats and other predictors of collective action (i.e., per-ceived efficacy, anger, and relative deprivation). In addition, we did not measure other inter-group emotions, for example, fear. Perceiving the outgroup as strong and threatening could easily induce an inferiority complex and experi-ences of fear which, in turn, could trigger avoidance behaviour as well as negative evalua-tions of the outgroup (Spears et al., 2011). It would be particularly interesting to see whether threat appraisals would predict anger, if the ingroup is perceived as strong and efficacious, or fear, when the outgroup is perceived threat-ening, which in turn would be associated with either action-oriented or avoidance-oriented tendencies on the one hand, and negative evalu-ations of the outgroup, on the other hand, especially in high conflict situations. Finally, In Study 1 we used three items to measure collec-tive action tendencies but in Study 2 only two of these items performed reliably. This implies that there might be subtler differences in terms of willingness to engage in collective action between Turks and Kurds. A wider pool of items could have detected such differences.
A third limitation of our studies relates to our samples. We have used data from two nonrepresentative samples recruited through house visits. We are therefore unable to draw broad generalizations from our findings. On the other hand, our effort to go beyond the usual student samples or samples from WEIRD societies (Henrich et al., 2010) increases the gen-eralizability of our findings, and we emphasize
the difficulty of collecting any data in such a setting.
Last but not least, although we have reported on the impact of contact on action tendencies, and the similarity of relevant paths between the two studies, we should be cautious before assert-ing that such effects would be shown if we had assessed actual behaviours. Talk is said to be cheap, and matching one’s words to actual behav-iour in this setting could yield different results. With salient social identities, individuals in both groups may have been motivated to endorse items that sound like behaviours that would address threat, and to respond to such items in the manner that they believe a “good” in-group member should. However, the potential social costs of real-world behaviour are much higher for Kurds, who still today expose themselves to palpable risks with such statements and, espe-cially, such behaviours, especially when the level of perceived threat increases.
Conclusion
To conclude, we have refined and extended vari-ous lines of research on contact, perceived threats, and collective action and have done so in the context of a severe, ongoing conflict, between Turks and Kurds in Turkey. We have also shown that threats predict collective action tendencies among both the advantaged and the disadvantaged. Additionally, we have success-fully demonstrated that both contact and social identity can be related to collective action in alternative ways (i.e., directly; or indirectly, via threats). We still believe, though, that much work remains to be done to achieve a better understanding of the factors which mobilize and demobilize individuals, especially in con-flictual intergroup contexts.
Acknowledgements
We are most grateful to Martijn van Zomeren, University of Groningen; the editor Winnifred Louis, University of Queensland; and two anonymous review-ers for their invaluable comments on an earlier vreview-ersion of the manuscript.
Funding
This research received no specific grant from any fund-ing agency in the public, commercial, or not-for-profit sectors.
Notes
1. Earlier, the latent model in which we entered realistic and symbolic threats as individual constructs fit the data comparatively worse. However, we still wanted to test whether real-istic and symbolic threats would have different effects on our outcomes as previously discussed by ITT. Therefore we investigated a latent multiple mediator model in which we entered realistic and symbolic threats as unique con-structs. The results showed that both constructs have rather similar effects on our outcome variables (intergroup contact–realistic threat– collective action: point estimate (β = −.040, 95% CI [−0.053, −0.021]); intergroup contact– symbolic threat–collective action: point esti-mate (β = −.052, 95% CI [−0.065, −0.042]). 2. As in Study 1, we also tested a multiple
media-tion model in which we entered symbolic and realistic threats separately to the model. In line with Study 1, the results show that both variables have effect sizes similar to the effect of unitary perceived threat variable has on our outcome variables: intergroup contact–realistic threat– collective action: point estimate (β = −.046, 95% CI [−0.058, −0.022]); intergroup contact– symbolic threat–collective action (β = −.059, CI 95% [−0.070, −0.034]).
References
Aberson, C. L., & Gaffney, A. M. (2008). An integrated threat model of explicit and implicit attitudes.
European Journal of Social Psychology, 39, 808–830.
doi:10.1002/ejsp
Abrams, D., & Grant, P. R. (2012). Testing the social identity relative deprivation (SIRD) model of social change: The political rise of Scottish nation-alism. The British Journal of Social Psychology, 51,
674–689. doi:10.1111/j.2044-8309.2011.02032.x Akkaya, A. H., & Jongerden, J. (2011). The PKK in the
2000s: Continuity through breaks. In M. Casier & J. Jongerden (Eds.), Nationalisms and politics in Turkey: Political Islam, Kemalism and the Kurdish issue
Ashmore, R. D., Deaux, K., & McLaughlin-Volpe, T. (2004). An organizing framework for collective identity: Articulation and significance of multi-dimensionality. Psychological Bulletin, 130, 80–114.
doi:10.1037/0033-2909.130.1.80
Ayata, B., & Yükseker, D. (2005). A belated awaken-ing: National and international responses to the internal displacement of Kurds in Turkey. New Perspectives on Turkey, 32, 5–42. doi:10.1017/
S089663460000409X
Barlow, F. K., Paolini, S., Pedersen, A., Hornsey, M. J., Radke, H. R. M., Harwood, J., . . . Sibley, C. G. (2012). The contact caveat: Negative contact predicts increased prejudice more than posi-tive contact predicts reduced prejudice. Personal-ity and Social Psychology Bulletin, 38, 1629–1643.
doi:10.1177/0146167212457953
Barrett, P. (2007). Structural equation modeling: Adjudging model fit. Personality and Individual Differ-ences, 42, 815–824. doi:10.1016/j.paid.2006.09.018
Bentler, P. (2007). On tests and indices for evaluating structural models. Personality and Individual Differ-ences, 42, 825–829. doi:10.1016/j.paid.2006.09.024
Blackwood, L. M., & Louis, W. R. (2012). If it mat-ters for the group then it matmat-ters to me: Collective action outcomes for seasoned activists. The British Journal of Social Psychology, 51, 72–92. doi:10.1111/
j.2044-8309.2010.02001.x
Branscombe, N. R., Ellemers, N., Spears, R., & Doosje, B. (1999). The context and content of social iden-tity threat. In N. Ellemers, R. Spears, & B. Doosje (Eds.), Social identity: Context, commitment, content
(pp. 35–58). Oxford, UK: Blackwell.
Branscombe, N. R., & Wann, D. (1994). Collective self-esteem consequences of outgroup deroga-tion when a valued social identity is on trial.
European Journal of Social Psychology, 24, 641–657.
doi:10.1002/ejsp.2420240603
Brown, R., & Hewstone, M. (2005). An integrative the-ory of intergroup contact. In M. P. Zanna (Ed.),
Advances in experimental social psychology (Vol. 37, pp.
255–342). San Diego, CA: Academic Press. Çakal, H., Eller, A., Sirlopú, D., & Perez, A. (in press).
Intergroup relations in Latin America: Intergroup contact, common ingroup identity and activism among indigenous groups in Mexico and Chile.
Journal of Social Issues.
Çakal, H., Hewstone, M., Schwar, G., & Heath, A. (2011). An investigation of the social identity model of collective action and the “sedative” effect of intergroup contact among Black and
White students in South Africa. British Journal of Social Psychology, 50, 606–627.
doi:10.1111/j.2044-8309.2011.02075.x
Chen, F. F., Sousa, K. H., & West, S. G. (2005). Teach-er’s corner: Testing measurement invariance of second-order factor models. Structural Equation Modeling: A Multidisciplinary Journal, 12, 471–492.
doi:10.1207/s15328007sem1203_7
Corenblum, B., & Stephan, W. G. (2001). White fears and native apprehensions: An integrated threat theory approach to intergroup attitudes.
Canadian Journal of Behavioural Science/Revue Cana-dienne Des Sciences Du Comportement, 33, 251–268.
doi:10.1037/h0087147
Cottrell, C., & Neuberg, S. L. (2005). Different emo-tional reactions to different groups: A sociofunc-tional threat-based approach to “prejudice.”
Journal of Personality and Social Psychology, 88, 770–
89. doi:10.1037/0022-3514.88.5.770
Craig, M. A., & Richeson, J. A. (2014). On the precipice of a “majority–minority” America: Perceived status threat from the racial demo-graphic shift affects White Americans’ politi-cal ideology. Psychological Science, 25, 1189–1197.
doi:10.1177/0956797614527113
Crisp, R. J., Stone, C. H., & Hall, N. R. (2006). Recate-gorization and subgroup identification: Predicting and preventing threats from common ingroups.
Personality & Social Psychology Bulletin, 32, 230–243.
doi:10.1177/0146167205280908
Curşeu, P. L., Stoop, R., & Schalk, R. (2007). Preju-dice toward immigrant workers among Dutch employees: Integrated threat theory revisited.
European Journal of Social Psychology, 37, 125–140.
doi:10.1002/ejsp.331
DeSteno, D., Dasgupta, N., Bartlett, M. Y., & Cajdric, A. (2004). Prejudice from thin air: The effect of emotion on automatic intergroup attitudes. Psy-chological Science, 15, 319–324.
doi:10.1111/j.0956-7976.2004.00676.x
Dixon, J., Levine, M., Reicher, S., & Durrheim, K. (2012). Beyond prejudice: Are negative evaluations the problem and is getting us to like one another more the solution? Behavioral and Brain Sciences, 35,
411–425. doi:10.1017/S0140525X11002214 Dixon, J. C., & Ergin, M. (2010). Explaining
anti-Kurd-ish beliefs in Turkey: Group competition, identity, and globalization. Social Science Quarterly, 91, 1329–
1348. doi:10.1111/j.1540-6237.2010.00734.x Doosje, B., & Ellemers, N. (1997). Stereotyping under
Spears & P. J. Oakes (Eds.), The social psychology of stereotyping and group life (pp. 257–272). Oxford,
UK: Wiley-Blackwell.
Doosje, B., Ellemers, N., & Spears, R. (1995). Perceived intragroup variability as a function of group status and identification. Journal of Experimental Social Psy-chology, 31, 410–436. doi:10.1006/jesp.1995.1018
Doosje, B., Spears, R., & Ellemers, N. (2002). Social identity as both cause and effect: The develop-ment of group identification in response to antici-pated and actual changes in the intergroup status hierarchy. British Journal of Social Psychology, 41,
57–76. doi:10.1348/014466602165054
Drury, J., & Reicher, S. (2009). Collective psycho-logical empowerment as a model of social change: Researching crowds and power. Journal of Social Issues, 65, 707–725.
doi:10.1111/j.1540-4560.2009.01622.x
Duckitt, J., & Mphuthing, T. (1998). Group identification and intergroup attitudes: A longitudinal analysis in South Africa. Journal of Personality and Social Psychol-ogy, 74, 80–85. doi:10.1037/0022-3514.74.1.80
Effron, D., & Knowles, E. (2015). Entitativity and intergroup bias: How belonging to a cohesive group allows people to express their prejudices.
Journal of Personality and Social Psychology, 108, 234–
253. doi:10.1037/pspa0000020
Ethier, K., & Deaux, K. (1994). Negotiating social identity when contexts change: Maintaining identification and responding to threat. Journal of Personality and Social Psychology, 67, 243–251.
doi:10.1037//0022-3514.67.2.243
González, K. V., Verkuyten, M., Weesie, J., & Poppe, E. (2008). Prejudice towards Muslims in the Netherlands: Testing integrated threat theory. The British Journal of Social Psychology, 47, 667–685. doi:
10.1348/014466608X284443
Grant, P. R., & Brown, R. (1995). From ethnocentrism to collective protest: Responses to relative depri-vation and threats to social identity. Social Psychol-ogy Quarterly, 58, 195–212. doi:10.2307/2787042
Greenglass, E. R. (2002). Proactive coping. In E. Frydenberg (Ed.), Beyond coping: Meeting goals, vision, and challenges (pp. 37–62). London, UK: Oxford
University Press.
Hair, J. F., Black, W., Babin, B., & Anderson, R. E. (2010). Multivariate data analysis (7th ed.). Upper
Saddle River, NJ: Prentice-Hall.
Hayes, A. F. (2009). Beyond Baron and Kenny: Sta-tistical mediation analysis in the new millen-nium. Communication Monographs, 76, 408–420.
doi:10.1080/03637750903310360
Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world? Behavio-ral and Brain Sciences, 33, 61–83. doi:10.1017/
S0140525X0999152X
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indices in covariance structure analysis: Con-ventional criteria versus new alternatives. Struc-tural Equation Modeling, 6, 1–55. doi:10.1080/
10705519909540118
Islam, M. R., & Hewstone, M. (1993). Dimen-sions of contact as predictors of intergroup anxiety, perceived out-group variability, and out-group attitude: An integrative model. Per-sonality and Social Psychology Bulletin, 19, 700–710.
doi:10.1177/0146167293196005
Jetten, J., Branscombe, N. R., Schmitt, M. T., & Spears, R. (2001). Rebels with a cause: Group identifica-tion as a response to perceived discriminaidentifica-tion.
Personality and Social Psychology Bulletin, 27, 1204–
1213. doi:10.1177/0146167202289004
Jetten, J., Spears, R., & Manstead, A. S. R. (1997). Strength of identification and intergroup differ-entiation: The influence of group norms. European Journal of Social Psychology, 27, 603–609. doi:10.1002/
(SICI)1099-0992(199711/12)27:63.0.CO;2-# Kauff, M., & Wagner, U. (2012). Valuable therefore
not threatening: The influence of diversity beliefs on discrimination against immigrants. Social Psychological and Personality Science, 3, 714–721.
doi:10.1177/1948550611435942
Kline, R. B. (2011). Principles and practice of structural equa-tion modelling (3rd ed.). New York, NY: The
Guild-ford Press.
Lazarus, R. S. (1991). Emotion and adaptation. New York,
NY: Oxford University Press.
Leach, C. W., van Zomeren, M., Zebel, S., Vliek, M. L. W., Pennekamp, S. F., Doosje, B., & Spears, R. (2008). Group-level self-definition and self-invest-ment: A hierarchical (multicomponent) model of in-group identification. Journal of Personality and Social Psychology, 95, 144–165.
doi:10.1037/0022-3514.95.1.144
Louis, W. R., Taylor, D. M., & Douglas, R. M. (2005). Normative influence and rational con-flict decisions: Group norms and cost–ben-efit analyses for intergroup behavior. Group Processes & Intergroup Relations, 8, 355–374.
doi:10.1177/1368430205056465
Luhtanen, R., & Crocker, J. (1992). A collective self-esteem scale: Self-evaluation of one’s social iden-tity. Personality and Social Psychology Bulletin, 18,
