Special Thematic Section " Collective Memories and Present-Day Intergroup Relations : A Social-Psychological Perspective " Collective Nostalgia Is Associated With Stronger Outgroup-Directed Anger and Participation in Ingroup-Favoring Collective Action

Collective nostalgia refers to longing for the way society used to be. We tested whether collective nostalgia is associated with ingroup-favoring collective action and whether this association is mediated by outgroup-directed anger and outgroup-directed contempt. We conducted an online study of Hong Kong residents (N = 111) during a large-scale democratic social movement, the Umbrella Movement, that took place in Hong Kong in 2014 in response to proposed electoral reforms by the Chinese government in Mainland China. Reported collective nostalgia for Hong Kong’s past was high in our sample and collective nostalgia predicted stronger involvement in ingroup-favoring collective action, and it did so indirectly via higher intensity of outgroup-directed anger (but not through outgroup-directed contempt). We argue that collective nostalgia has implications for strengthening ingroup-serving collective action, and we highlight the importance of arousal of group-based emotions in this process.

We are also interested in understanding the potential mediating role of outgroup-directed anger and outgroup-directed contempt in the association between collective nostalgia and ingroup-favoring collective action.Collective nostalgia encompasses sentimental longing for the group's past.Such longing is likely to be accompanied by awareness that the group's present circumstances are not as favorable as its past ones (i.e., temporal relative deprivation; Smith, Pettigrew, Pippin, & Bialosiewicz, 2012).Temporal relative deprivation is often manifested in the form of outgroup-directed anger and outgroup-directed contempt (Becker & Tausch, 2015;Yzerbyt, Dumont, Wigboldus, & Gordijn, 2003).Hence, we hypothesize that collective nostalgia precedes outgroup-directed anger and outgroup-directed contempt, which in turn predict participation in ingroup-favoring collective action.Taken together, we examine whether: (1) collective nostalgia is heightened when the ingroup perceives a deterioration of present circumstances compared to past ones, (2) collective nostalgia predicts outgroup-directed anger and outgroup-directed contempt, (3) collective nostalgia has direct behavioral implications for ingroup-favoring collective action, and (4) outgroup-directed anger and outgroup-directed contempt mediate the putative relation between collective nostalgia and ingroup-favoring collective action.We address these issues in the context of a large-scale social movement, the Umbrella Movement, that unfolded in Hong Kong in 2014.

Collective Nostalgia
Nostalgia is "a sentimental longing or wistful affection for the past" (Pearsall, 1998(Pearsall, , p. 1266)).Nostalgia pertains not only to personal, but also to collective memories.Collective nostalgia, then, entails a sentimental longing for the collective past and the way society used to be.In bringing to mind positive collective memories, collective nostalgia may be more likely to emerge when people perceive the group as better off in the past than in the present.Specifically, collective nostalgia is defined as "nostalgic reverie…that is contingent upon thinking of oneself in terms of a particular social identity or as a member of a particular group (i.e., self-categorization at the collective level; Iyer & Leach, 2009) and concerns events or objects related to it" (Wildschut et al., 2014, p. 845).
Indeed, nostalgia can be studied on a group-level, because collective nostalgia meets the conceptual criteria of group-level emotions (Smith, Seger, & Mackie, 2007) to the extent that it motivates and regulates attitudes and behavior in relation to social groups (Wildschut et al., 2014;Wildschut, Sedikides, Van Tilburg, & Leunissen, in press).
Whereas identification with a group in the present promotes engagement in ingroup-favoring action (Van Zomeren, Postmes, & Spears, 2008), positive collective memories about a group should also be an important driver for such action.Given that collective nostalgia entails both a sense of belonging to a group that is valued positively at present (Wildschut et al., 2014) and fondness for the way things were for this group in the past (Prusik & Lewicka, 2016), collective nostalgia should have the capacity to drive ingroup-favoring action.There is support for the notion that collective nostalgia heightens favoritism toward the ingroup or one's nation.For example, British undergraduates who nostalgized about a collective event (as opposed to a personal event) expressed stronger behavioral intentions to volunteer time for the ingroup's publicity campaign (Wildschut et al., 2014, Study 2).Likewise, Irish undergraduates who nostalgized about a collective event (as opposed to engaging in autobiographical recollection of a collective event) spent more of their money to punish a transgression inflicted upon an ingroup member (Wildschut et al., 2014, Study 3).On a national level, nostalgia is positively associated with ethnic national identi-fication, protection of the national ingroup identity, and endorsement of autochthony (i.e., entitlement based on native privilege; Smeekes, 2015;Smeekes, Verkuyten, & Martinovic, 2015).

Collective Nostalgia as an Antecedent of Outgroup-Directed Anger and Outgroup-Directed Contempt
We propose that collective nostalgia shapes group-based emotions directed at the outgroup who poses a threat to the ingroup.We focus on two such emotions, outgroup-directed anger and outgroup-directed contempt.Anger and contempt are negative emotions that sometimes co-occur and involve an overlapping appraisal that others bear hostile intentions (Frijda, Kuipers, & ter Schure, 1989;Izard, 1971).These emotions also have unique features.
In an intergroup context, outgroup-directed anger arises when the ingroup receives unfair treatment, and it elicits action tendencies to act against the offender -typically through normative actions (e.g., peaceful protest; Tausch et al., 2011).People express contempt by rejecting and socially excluding the other person, with little intention to form any long-term or short-term reconciliation (Fischer & Roseman, 2007).In an intergroup context, outgroupdirected contempt entails the maintenance of psychological distance, and typically precipitates non-normative action (e.g., violence; Tausch et al., 2011).
Before providing support for the proposition that collective nostalgia predicts outgroup-directed anger and outgroupdirected contempt, we note that collective nostalgia is conceptually different from outgroup-directed anger and outgroup-directed contempt.First, nostalgia is a bittersweet emotion that is predominantly positive, albeit with tinges of sadness or longing (Hepper, Ritchie, Sedikides, & Wildschut, 2012;Sedikides & Wildschut, 2016;Stephan, Sedikides, & Wildschut 2012;Wildschut, Sedikides, Arndt, & Routledge, 2006), whereas anger and contempt are negative emotions that entail an antagonistic or dismissive interpretation of others' intention (Fischer & Roseman, 2007;Kuppens, Van Mechelen, Smits, & De Boeck, 2003).Second, collective nostalgia can emerge independently of outgroup-directed anger and outgroup-directed contempt.For instance, employees experience collective nostalgia after having worked for an organization for a considerable period of time (Gabriel, 1993), or when they undergo workplace relocation (Milligan, 2003).The experience of collective nostalgia increases social cohesion among the employees (Gabriel, 1993) and enhances identity continuity during the transitional period (Milligan, 2003).Such experiences do not necessarily concern any antagonism against outgroups and therefore do not need to involve outgroup-directed anger and outgroup-directed contempt.
Collective nostalgia may precede outgroup-directed anger and outgroup-directed contempt, because collective nostalgia involves longing for the collective past.Such longing may be accompanied by dissatisfaction due to the perception that present circumstances have deteriorated compared to past ones (i.e., temporal relative deprivation; Pettigrew, 2002;Smith et al., 2012).Previous research has shown that collective nostalgia is associated with dissatisfaction with the current political situation (Cernat, 2010;Sedikides, Wildschut, Gaertner, Routledge, & Arndt, 2008;Todorova & Gille, 2010).For example, in some postcommunist countries, people report experiencing nostalgia for the Commmunist regime decades after it has ended (Boyer, 2006;Cernat, 2010;Nikolayenko, 2008;Pobłocki, 2007;Todorova & Gille, 2010).In Poland, those who did not benefit financially from postcommunist transformation view the past (i.e., social welfare system, employment security, noncompetitive market) more fondly and yearn more for it (Prusik & Lewicka, 2016).Levels of collective nostalgia are linked to current cultural values diverging from past ones.For example, in Belarus, Russia, and Ukraine, people report nostalgia for the values of the Soviet Union (Gherghina & Klymenko, 2012).The link between collective nostalgia and perceived discrepancy of past and present values is also observed in Western culture.For example, previous research has shown that people in the U.S. report high levels of nostalgia for the 1960s, a decade known for its openness and freedom (Wilson, 2005).
Taken together, it appears that collective nostalgia is linked to perceived changes from a desirable past to a less desirable present (temporal relative deprivation; Pettigrew, 2002;Smith et al., 2012).Awareness of such changes is likely to prompt outgroup-directed anger and outgroup-directed contempt, especially when the ingroup perceives that there is an outgroup who is responsible for these changes.

Collective Nostalgia and Ingroup-Favoring Collective Action
Besides focusing on collective nostalgia as an antecedent of outgroup-directed anger and outgroup-directed contempt, we also examine the implications of collective nostalgia for ingroup-favoring collective action.In particular, we are concerned with actual behavior (i.e., action) rather than attitudes or behavioral intentions.The latter, after all, are not the most robust predictors of future behavior (Ajzen, 2001;Kraus, 1995).
We posit that collective nostalgia is associated with ingroup-favoring collective action aiming to restore a desirable past.This proposition is consistent with intergroup emotions theory (Mackie et al., 2000;Mackie & Smith, 1998).
According to this theory, when individuals view themselves as members of socially defined groups, their behaviors will be influenced by the way they feel about their own group and other groups.Further, group-based emotions entail specific action tendencies that regulate group-based behaviors (Maitner, Mackie, & Smith, 2006).Given that collective nostalgia is a group-level emotion (Wildschut et al., 2014, in press) that leads to group-level benefits (e.g., favoritism, strenghening of social identity; Kim, 2010;Smeekes, 2015;Smeekes et al., 2015;Volkan, 1999), it should render group membership salient, meaning the self would be defined primarily in terms of group membership.Collective nostalgia, then, should motivate collective action to counter threats to the ingroup-threats that appear to widen the gap between past and present.
The proposition that collective nostalgia is associated with ingroup-favoring collective action is also consistent with the broader regulatory function of nostalgia (Sedikides et al., 2015;Sedikides, Wildschut, Routledge, Arndt, & Zhou, 2009;Wildschut, Sedikides, & Cordaro, 2011).When an aversive psychological state threatens aspects of the psychological equilibrium, nostalgia is triggered to mitigate the negative influence and restore the equilibrium.
For example, Stephan et al. (2014, Study 1) investigated the link between personal nostalgia on the one hand, and approach and avoidance motivations on the other.They assessed both proneness and frequency of personal nostalgia (Barrett et al., 2010;Routledge, Arndt, Sedikides, & Wildschut, 2008) and personal nostalgia for specific objects of one's past (e.g., family, toys, TV shows; Batcho, 1995).Avoidance motivation (i.e., an aversive psychological state) was negatively associated with approach motivation, after controlling for nostalgia.Avoidance motivation was positively linked with nostalgia.Crucially, in a mediation model in which avoidance motivation predicts approach motivation indirectly via nostalgia, nostalgia suppressed the negative link between avoidance motivation and approach motivation.That is, the direct effect of avoidance motivation on approach motivation was negative, but the indirect effect of avoidance motivation on approach motivation via nostalgia was positive.In other words, in the mediation model, avoidance motivation was positively linked with nostalgia, and nostalgia, in turn, was positively associated with approach motivation.These findings align with the notion that nostalgia serves a regulatory function on a personal level.Similarly, on a group level, national nostalgia plays a role in managing group identity in the face of group continuity threat by strengthening ingroup identity and increasing support for outgroup exclusion (Abakoumkin, Wildschut, Sedikides, & Bakarou, in press;Smeekes & Verkuyten, 2015).

Outgroup-Directed Anger and Outgroup-Directed Contempt as Potential Mediators
Thus far, we proposed that collective nostalgia precedes outgroup-directed anger and outgroup-directed contempt, and that collective nostalgia has direct implications for ingroup-favoring collective action.A good deal of research has indicated that outgroup-directed anger and outgroup-directed contempt are strong predictors of collective action (Becker & Tausch, 2015;Folger, 1987;Smith & Ortiz, 2002;Tausch et al., 2011;Yzerbyt et al., 2003).For example, White Australians who feel relatively deprived experience outgroup-directed anger, which in turn mediates their higher willingness for political action against government policy that favors the outgroup (i.e., Aborigines; Leach, Iyer, & Pedersen, 2007).Also, when individuals view an outgroup as threatening to the ingroup (e.g., as challenging its values or rights), they may experience group-based emotions, which will likely motivate collective action that promises to offset the threat.For instance, in Hessen, a federal state in Germany, students engaged in different types of collective action when their right to cost-free university education was removed (Tausch et al., 2011).Specifically, those who experienced outgroup-directed anger were more likely to engage in normative collective action (e.g., demonstrations, writing flyers), but those who experienced outgroup-directed contempt were more likely to engage in non-normative collective action (e.g., throw stones or bottles, arson attacks on university buildings).In all, we hypothesize that collective nostalgia is directly associated with ingroup-favoring collective action, and that collective nostalgia is also associated with collective action indirectly via outgroup-directed anger and outgroup-directed contempt.

Umbrella Movement in Hong Kong
We test the abovementioned hypotheses in the context of the current political situation in Hong Kong, a former British colony.The transfer of Hong Kong's sovereignty from the United Kingdom back to Mainland China occurred in 1997 (Higgins, 1997).During the one and a half centuries of British rule, Hong Kong had a democratic and pluralistic political system unlike that of Mainland China, a socialist republic run by a single political party (Tsang, 2004;Zhao, 2010).The transfer entailed an arrangement, according to which Hong Kong would be under a "one country, two systems policy" during the first 50 post-handover years (McKirdy, 2014).Hong Kong would retain a different political system and have a high degree of autonomy in all areas except for foreign relations and military defence (So, Lin, & Poston, 2001).
The historical background and political differences between Hong Kong and Mainland China have implications for intragroup and intergroup dynamics.In a poll conducted in Hong Kong in December 2016 (N = 1001, with sample weighted according to gender and age distribution of the Hong Kong population), approximately 64% of respondents identified as "Hongkongers" or "Hongkongers in China," but only 34% as "Chinese" or "Chinese in Hong Kong" (The University of Hong Kong, 2016).This poll testifies to the distinctiveness of Hongkongers' identity or their disinclination to identify as Mainland Chinese.Such distinctiveness stems, at least in part, from differences in values between the two groups.Hongkongers seem to endorse more strongly values of freedom, justice, and democracy (Hong, Chiu, Yeung, & Tong, 1999), whereas these values are less prevalent in Mainland China (Malesky, Abrami, & Zheng, 2011;Zhao, 2010).Indeed, following the 1997 handover, there have been growing concerns in Hong Kong over the erosion of political autonomy, the clampdown on the freedom of assembly, and the suppression of free speech (Agence France-Presse, 2016a; Lee, 1998;Lee, 2007;Radio Free Asia, 2015).
The level of concern soared when, in 2014, the Chinese Government retracted its promise for Hong Kong to host direct elections for chief executive by 2017.Subsequently, in September 2014, a large-scale democratic protest, known as the Umbrella Movement, broke out and lasted for 79 days.Thousands of protesters demanded reforms to the electoral system and rejected the Chinese Government's practice of vetting candidates to run in the election of Hong Kong's chief executive.Although the protest was manifested in terms of peaceful demonstrations and occupation (Dissanayake, 2014), the Chinese government described the protests as illegal and supported the Hong Kong government's approach to disperse harshly the crowd with tear gas and pepper spray (Dearden, 2014) as well as brandishing of guns (Flanagan, 2014).The protestors used umbrellas as their most prominent form of defense-hence the name Umbrella Movement (Henley, 2014).Toward the end of the protest in 2014, 955 individuals, including opposition lawmakers, the three co-founders of the "Occupy Central" civil disobedience campaign, and leaders of two student groups, were arrested by the Hong Kong police (Amnesty International, 2016).A small proportion of them was subsequently convicted by the end of 2015.
After the Umbrella Movement, continued support for the general movement is evinced in the first election of the legislative council members in 2016.Any Hong Kong citizen can stand as a candidate in this election and, from a total of 70 seats, democrats received 30 seats, six of which went to social activists.Among these social activists, some were highly visible during the Umbrella Movement and supported the principle of democratic self-determination among Hong Kong people, whereas some supported permanent separation of Hong Kong and China (Ng, 2016;Phillips, 2016).These results reflects the positive change the Umbrella Movement brought to the political system in Hong Kong.

Overview of Hypotheses
Dissatisfaction with the present (e.g., the current political system) is associated with collective nostalgia (Cernat, 2010;Sedikides et al., 2008;Todorova & Gille, 2010).Collective nostalgia, then, will likely be high among Hongkongers, as the outgroup (Mainland China) appears to contribute to a dissatisfying present by imposing restrictions on political freedom and challenging core ingroup values.More direct involvement by the Chinese Government in the governance of Hong Kong should be experienced as particularly threatening by those Hongkongers who are nostalgic for a time when Hong Kong had more political autonomy.Indeed, with the Chinese government's growing pressure on Hong Kong's political system, collective nostalgia is featured more prominently in public and political discourse, including films, TV commercials, and protests (Chan, 2015;Cheung, 2012;Ho, 2016).Hongkongers who are nostalgic for the way their society used to be will likely view the increasingly direct involvement of the Chinese Government as a change for the worse (i.e., perceived temporal relative deprivation).
In comparison to their autonomous past, the present political situation looks bleak and unjust.This engenders outgroup-directed anger and outgroup-directed contempt toward the perceived agent of change (de la Sablonnière, Tougas, & Lortie-Lussier, 2009): the Chinese Government.In turn, outgroup-directed anger and outgroup-directed contempt motivate political action and protest (Smith et al., 2012).Putting it in the context of the Umbrella Movement, we hypothesize that: (1) collective nostalgia would be heightened when the ingroup perceives a deterioration of present circumstances compared to past ones, (2) collective nostalgia would predict outgroup-directed anger and outgroup-directed contempt, (3) collective nostalgia would have direct behavioral implications for ingroup-favoring collective action, and (4) the association between collective nostalgia and ingroup-favoring collective action would be accounted for (i.e., mediated) by outgroup-directed anger and outgroup-directed contempt.

Method Participants
One hundred and fifteen Hong Kong residents took part in our online study that was advertised on Facebook and Twitter from November 12-25, 2014-during the time when the Umbrella Movement took place.We excluded four participants who opted to abstain from four questions asking about involvement in the Umbrella Movement.From the remaining 111 participants, 58 were men, 51 were women, and one was a transgender person (one participant did not report their gender).Participant age ranged from 16-65 years (M = 29.50,SD = 9.53; five participants did not report their age).Before completing the study, all participants read an information page, confirmed that they are a Hong Kong citizen, and gave their consent to participate.In our sample, one participant was 16 years old, and another participant was 17 years old.The British Psychological Society guideline states that researchers need to seek additional parental consent for children under the age of 16.According to the Hong Kong Psychological Society, for adolescents aged 16 to 17, additional consent from their parents/guardians is optional in regard to studies involving minimal risk.Since our study does not involve much risk, we included all participants in our study.

Procedure and Measures
We administered the online survey in both Chinese and English.The Chinese version of the survey was translated by a bilingual speaker and was checked by another bilingual speaker.
We assessed collective nostalgia with a measure that we partly constructed for the purposes of this study and partly adapted from Smeekes et al. (2015).The measure comprised five items that are relevant to Hong Kong's historical context (1 = never, 5 = to a very large extent).We prefaced the items with the stem "When thinking about Hong Kong now, to what extent do you feel…".The items were: "…nostalgic about the way Hong Kong people were in the past i ," "…nostalgic about the values that Hong Kong people had in the past," "…nostalgic about the way Hong Kong society was in the past," "…nostalgic about the Hong Kong of old," and "…nostalgic about the sort of place Hong Kong was before 1997" (α = .92,M = 3.40, SD = 1.16).
Next, we assessed the extent to which participants experienced outgroup-directed anger and outgroup-directed contempt (1 = never, 5 = to a great extent).We prefaced the items with the stem "When thinking about the current democracy development in Hong Kong, to what extent do you feel…".The items were: "anger" (M = 4.10, SD = 1.27) and "contempt towards the Chinese government" (M = 4.27, SD = 1.29).
Finally, we assessed ingroup-favoring collective action by examining participation in the Umbrella Movement with four items.First, participants read the following instructions: "Using the scale below, please indicate how you would define your level of involvement in the following since 22 nd September, 2014."Then, they responded to the items (1 = never, 5 = frequently): "Occupation," "Hanging slogans," "Any other civil disobedience," and "Being active on social media to support the movement" (α = .70,M = 2.91, SD = 1.10).

Collective Nostalgia
A one sample t-test revealed that participants experienced collective nostalgia about the way Hong Kong society and Hong Kong people used to be (M = 3.40, SD = 1.16), with the mean being significantly different from the scale mid-point (3), t(110) = 3.61, p < .001.The level of collective nostalgia among Hongkongers can thus be considered high.

Correlations
Collective nostalgia was positively related to outgroup-directed anger, outgroup-directed contempt, and ingroupfavoring collective action (see Table 1).Outgroup-directed anger and outgroup-directed contempt were also positively related to ingroup-favoring collective action.Given that the results were not qualified by gender and age ii , we omitted these variables from subsequent analyses.

Mediational Analyses
The above correlational analyses revealed that collective nostalgia was associated with stronger outgroup-directed anger and outgroup-directed contempt.These analyses also revealed that outgroup-directed anger and outgroupdirected contempt, each in turn, had implications for ingroup-favoring collective action (i.e., participation in the Umbrella Movement).Next, we assessed whether collective nostalgia indirectly predicted ingroup-favoring collective action through outgroup-directed anger and outgroup-directed contempt.
We first tested the association between collective nostalgia and ingroup-favoring collective action, and the indirect path (denoted as ab) via outgroup-directed anger (10,000 bootstrap samples) by using the PROCESS macro (Hayes, 2013;Model 4).The direct path from collective nostalgia to ingroup-favoring collective action was null, B = 0.08, SE = 0.09, 95% CI [-0.097, 0.258].The indirect path was significant, ab = 0.12, SE = 0.05, 95% CI [0.040, 0.240].Hence, collective nostalgia was associated with higher levels of ingroup-favoring collective action, indirectly via outgroup-directed anger.In other words, the association between collective nostalgia and ingroupfavoring collective action was fully mediated by outgroup-directed anger.
We then tested the association between collective nostalgia and ingroup-favoring collective action, and the indirect path via outgroup-directed contempt.The direct path from collective nostalgia to ingroup-favoring collective action was null, B = 0.11, SE = 0.08, 95% CI [-0.047, 0.271].The indirect path was not significant, ab = 0.09, SE = 0.05, 95% CI [-0.006, 0.203].Taken together, these results mean that collective nostalgia was indirectly associated with ingroup-favoring collective action via outgroup-directed anger, but not via outgroup-directed contempt.In the following analyses, we focused on the indirect effect involving outgroup-directed anger.

Model Fit and Alternative Models
We assessed model fit (with AMOS within SPSS for Windows) using observed variables.We trimmed the nonsignificant direct path from collective nostalgia to ingroup-favoring collective action and then calculated fit indices for the resultant nonsaturated model.This model provided a good fit (Table 2).We also tested alternative mediational models.Within a set of models for the same data, the Akaike Information Criterion (AIC; Akaike, 1974) and Expected Cross Validation Index (ECVI; Browne & Cudeck, 1993) can be used to compare competing models that need not be nested (a smaller value indicates better fit).However, any two models that have the same paths between the same variables will have the same fit, even if some paths are in a different direction (see Cheung, Sedikides, & Wildschut, 2016).For example, consider an alternative model in which ingroup-favoring collective action precedes outgroup-directed anger.To test this model, one cannot simply reverse the order of outgroup-directed anger and ingroup-favoring collective action.Doing so would create an alternative model that differs only in the direction of the link between outgroup-directed anger and ingroup-favoring collective action, and would therefore have the same fit as the original model.Accordingly, we tested a series of models in which each variable predicted only the variable that immediately followed it in the postulated causal chain.This enabled us to evaluate which ordering of variables produced the lowest AIC and ECVI values.
With three variables in the model, there are six possible alternative sequences of variables.Among these alternative models, one model (ingroup-favoring collective action → outgroup-directed anger → collective nostalgia) differed from the original model (collective nostalgia → outgroup-directed anger → ingroup-favoring collective action) only in direction of the link between each variable.In other words, it was a mirror image of the original model and produced the same fit indices.Between these two models, we opted to retain the original one, in which collective nostalgia predicts ingroup-favoring collective action via outgroup-directed anger.Our decision was based on theoretical ground that emotions often precede behaviors (Forgas, 2001;George, 1996;Wildschut et al., 2014;Zhou, Wildschut, Sedikides, Shi, & Feng, 2012) and, more specifically, that outgroup-directed anger predicts col-Journal of Social and Political Psychology 2017, Vol.5(2), 301-319 doi:10.5964/jspp.v5i2.697 Collective Nostalgia, Anger, and Collective Action lective action (Van Zomeren, Spears, Fischer, & Leach, 2004;Yzerbyt et al., 2003).Of the remaining four alternative models, there were two pairs with the same fit indices.We assessed one from each pair of alternative models.
Both alternative models produced higher (i.e., indicating worse fit) AIC and ECVI values (see Table 2).

Discussion
Our study replicated previous findings that outgroup-directed anger and outgroup-directed contempt are associated with participation in collective action (Becker & Tausch, 2015;Yzerbyt et al., 2003).More important, our study revealed that collective nostalgia was a predictor of outgroup-directed anger and outgroup-directed contempt as well as collective action.This suggests that collective nostalgia has direct implications for both intergroup relations (e.g., anger at the outgroup) and intragroup relations (i.e., increasing support for ingroup-favoring collective action).
These findings were substantiated with measures of actual participation in a social movement.Specifically, during the Umbrella Movement, Hongkongers (the ingroup) felt that Mainland China (the outgroup) challenged Hong Kong's present way of life and core values, and Hongkongers felt nostalgic for the way society used to be.In this capacity, collective nostalgia (1) was prognostic of heightened outgroup-directed anger and outgroup-directed contempt, (2) was directly linked to mobilization toward collective action that aimed at countering the threat imposed on the ingroup, and this link was fully mediated by outgroup-directed anger (but not outgroup-directed contempt).
We perused the collective action literature in an effort to explain why the indirect path from collective nostalgia to collective action via outgroup-directed contempt was not significant.According to the recent literature on the role of emotions in collective action, outgroup-directed anger uniquely predicts normative collective action (e.g., peaceful protest), whereas outgroup-directed contempt uniquely predicts non-normative collective action (e.g., violence; Tausch et al., 2011).Our measure of collective action focuses on normative and peaceful actions (e.g., hanging slogans, civil disobedience) rather than non-normative and violent actions.This may be the reason why the indirect path from collective nostalgia to collective action via outgroup-directed anger was significant, but the indirect path via outgroup-directed contempt was not significant.Other phenomenological differences between anger and contempt may also account for our results.Anger is more typically expressed in situations where one has increased (rather than reduced) control over the situation.Contempt, on the other hand, is more typically expressed in situations where one has reduced control over the situation and is accompanied by an intention to distance oneself from the situation (Fischer & Roseman, 2007).Taken together, collective nostalgia may be especially relevant to participation in normative collective action and intentions to engage in positive social change.
Regardless, overall, our findings indicate that collective nostalgia has implications for present-day intragroup and intergroup relations.Below, we elaborate on these findings and contextualize them.
As an emotion that bolsters intragroup relations, collective nostalgia also features in social movements.In the 1980s, interest in the history of the civil rights movement rose in the United States.Lawson (1991) posited that such interest could be rooted in peoples' nostalgia for the historical era of their youth in the U.S., when the civil rights movement thrived and had instilled community cohesion and prosocial values.Further, such nostalgia "made many Americans feel better about themselves and what they might accomplish once again in the future" (Lawson, 1991, p. 456).Collective nostalgia is also relevant at times when continuity between the past and present breaks down.In this case, collective nostalgia "…may be used in an effort to maintain a collective sense of sociohistoric continuity, as a form of resistance to the hegemony of elites, and as a defense against anxiety" (Brown & Humphreys, 2002, p. 143).
Consistent with the above-stated role of collective nostalgia in the political arena, we showed that collective nostalgia is linked to involvement in collective action in the context of a large-scale social movement in Hong Kong.
The level of collective nostalgia among Hongkongers was high following the Chinese Government's announcement of a plan to restrict electoral rights in Hong Kong.In turn, collective nostalgia was related to stronger involvement in the Umbrella Movement.Put otherwise, collective nostalgia can invigorate collective action aimed at thwarting threats to the ingroup.
Nostalgia's capacity to stymie threat or aversiveness may explain why the general Hong Kong social movement did not cease following termination of the 79-day Umbrella Movement.Instead, the general movement sparked further support for more independence from Mainland China (Agence France-Presse, 2016b; Phillips & Cheung, 2016).Collective nostalgia may help maintain socio-historic continuity through anchoring present behavior driven by values that people endorsed in the past.The outcome will be the strengthening of such values.

The Indirect Effect of Outgroup-Directed Anger
Other than having an influence on intragroup dynamics, collective nostalgia is also relevant for intergroup dynamics.
Collective nostalgia precedes outgroup-directed anger, a negative emotion that entails hostility against the outgroup that poses a threat to the ingroup.Such hostility has downstream implications.Collective action often arises when the ingroup is perceived to be at a collective disadvantage (Walker & Smith, 2002).According to relative deprivation theory (Folger, 1987), appraisals of unfair and illegimate disadvantage lead to anger, which is crucial in fueling collective action against the outgroup deemed responsible for the disadvantage (Mackie et al., 2000;Miller, 2001).
Our results suggest that collective nostalgia may be directly related to relative deprivation.Relative deprivation concerns the judgment of being worse off and the accompanying emotions of anger or resentment (Smith et al., 2012).Such judgment and emotions originate in various types of comparisons, including comparisons to other people, other groups, and oneself or one's group in the past.Given that collective nostalgia entails longing for the group's past, it is particularly relevant to the type of relative deprivation that refers to the comparison between the group's present and the group's past (i.e., temporal relative deprivation).The notion that collective nostalgia is associated with temporal relative deprivation needs to be directly tested in future research.

Limitations and Broader Considerations
We recruited participants during the Umbrella Movement through social media (i.e., Facebook and Twitter), instead of at the demonstrations and occupation movement (i.e., road blockage of streets in the Central district) or from targeted populations that were not involved with the movement.Volunteers who took part in the study may possess traits or attitudes that are different from those who participated in the demonstrations and occupation for an extensive time period or those who did not participate in the Umbrella Movement.It is possible that individuals with different levels of involvement in the Umbrella Movement all experienced collective nostalgia, but engaged in ingroup-favoring collective action to different extents.We acknowledge this sampling limitation and treat the generalizability of the findings with caution.Nevertheless, we deem the current results informative for understanding the relevance of collective nostalgia for ingroup-favoring collective action.Future investigations would do well to build on these findings-by either sampling participants from a broader background (e.g., activists and non-activists) or using experimental and longitudinal designs to establish the direction of causality.
Further research can examine alternative mediators that may also account for the association between collective nostalgia and ingroup-favoring collective action.It is likely that collective nostalgia is not only linked to negative emotions, such as outgroup-directed anger and outgroup-directed contempt, but also to positive emotions, such as ingroup-directed pride.We would expect ingroup-directed pride to positively mediate the association between collective nostalgia and ingroup-favoring collective action.
Collective nostalgia boosts ingroup favoritism and increases the positivity of attitudes toward ingroup members (Wildschut et al., 2014, in press).Collective nostalgia, though, may also be associated with outgroup derogation.
One such circumstance is collective nostalgia for the nation (i.e., national nostalgia) experienced by the native majority.Prior research has shown that, in The Netherlands, national nostalgia prompts opposition to the rights of the Muslim minority to build mosques or Islamic schools, due to autochthony claims (Smeekes et al., 2015).
National nostalgia is also associated with more negative attitudes toward non-Western immigrants in The Netherlands (Smeekes, 2015).In a culturally diverse society, national nostalgia concerning native majority members could induce longing for a homogenous past.Yet, these findings do not necessarily imply that national or collective nostalgia (let alone personal nostalgia; see Cheung, Sedikides, & Wildschut, 2017) leads to outgroup derogation or prejudice.The key determinant may be the referent of nostalgia.Nostalgia concerning "the way the country was" is more likely to be associated with conservative attitudes toward immigrants or minorities.On the contrary, nostalgia concerning social values of openness and tolerance is more likely to be associated with liberal attitudes toward immigrants or minorities.Future research could manipulate the referent of collective nostalgia and assess the corresponding attitudinal outcomes toward distinct outgroups (e.g., immigrants or minorities).

Concluding Remarks
Collective nostalgia can shape present day intragroup and intergroup relations.Collective nostalgia, through its association with outgroup-directed anger, predicted higher involvement in ingroup-favoring collective action.In all, collective nostalgia has implications for the forging of collective action.

Notes
i) In Chinese, nostalgia refers to missing the past sentimentally and missing the way life was in the past.

Table 2
Comparison of Alternative Mediational Models