The Politicized Motivations of Volunteers in the Refugee Crisis : Intergroup Helping as the Means to Achieve Social Change

The refugee crisis in the summer of 2015 mobilized thousands of volunteers in Hungary to help refugees on their journey through Europe despite the government’s hostile stance. We conducted a survey (N = 1459) among people who were active in supporting refugees and providing services to them to test the hypothesis of whether volunteers in the context of this humanitarian crisis had social change motivations similar to those engaged in direct political activism. Hierarchical regression analysis and mediation analysis revealed the importance of opinion-based identity and moral convictions as predictors of volunteerism, while efficacy beliefs and anger only predicted political activism. Our findings suggest that volunteers engaged in helping refugees based on motivations previously described as drivers of mobilization for political activism, but chose volunteerism to alleviate the problems embedded in the intergroup situation. Although the context of the refugee crisis in Hungary may have been somewhat unique, these findings have implications for other asymmetrical politicized intergroup relations in which advantaged group members can choose to offer humanitarian aid, engage in political actions to change the situation, or do both.

Kulcsszavak: önkéntesség, társadalmi változás, politikai aktivizmus, csoportközi segítségnyújtás, többségi aktivizmus, véleményalapú identitás, morális meggyőződés involved in helping refugees (http://www.europeact.eu).Given the rapid rise in the number of refugees entering Europe, relying on volunteers to tackle the humanitarian crisis was inevitable.However, the statement quoted above makes it clear that volunteers involved in helping refugees also became actors for social change.The statement is notable because striving for social change is not necessarily a part of volunteerism, as volunteerism has been characterized as working toward social cohesion (Wright & Lubensky, 2009), and providing services (Penner, 2004) without disrupting the social structure.
There were crucial differences in the treatment of and in the official rhetoric about refugees across the countries of Europe.In countries where governments did not offer humanitarian aid, and treated refugees as enemies, authorities also perceived the work of volunteers as an expression of political dissent.Nevertheless, it is not clear what motivated people to volunteer during the crisis, and whether achieving political and social change was among the goals of their involvement in offering humanitarian aid.

Intergroup Helping and Volunteerism
People can show civic engagement in many different ways, such as engagement in traditional political participation (e.g., voting), in some form of volunteerism, or in political activism.Putnam (1993) argues that participating in civic organizations is an important first step toward engaging in political activities.However, this simple and direct connection between volunteerism and political activism is not supported empirically (van der Meer & van Ingen, 2009;van Stekelenburg, Klandermans, & Akkerman, 2016).This may be because of the charitable and philanthropic framing of volunteerism as opposed to the disruptive and confrontational nature of political activism (Bickford & Reynolds, 2002), related to setting different goals in terms of social change (Marullo & Edwards, 2000;Penner, 2004).
Volunteerism is the intentional engagement in long-term prosocial behavior that benefits others, mostly within an organizational framework (Penner, 2004), but it can also flare up in response to crises, such as a natural disaster or a terrorist attack (Penner, Brannick, Webb, & Connell, 2005).Volunteers are generally motivated to learn new skills, find meaning and personal growth (Omoto & Snyder, 1990;Snyder & Omoto, 2008), or help people with whom they have personal relationships (Russell, 2011).While volunteerism usually benefits ingroup members, there are many cross-group volunteer activities by which advantaged group members offer help to disadvantaged outgroups (Wilson, 2000), like in the current refugee crisis.
The motivations to engage in interpersonal and ingroup helping mostly also apply to the cross-group context.Intergroup helping can offer rewards to the helper that are similar to those of ingroup helping, and people help others out of sympathy and empathy regardless of their group membership (Batson, Turk, Shaw, & Klein, 1995).Positive behavioral intentions toward ingroup and outgroup members are both affected by perceived morality of the ingroup (Brambilla, Sacchi, Pagliaro, & Ellemers, 2013).Volunteers are motivated to engage in cross-group helping in the presence of a salient common ingroup identity, emphasizing similarities rather than differences between the groups (Dovidio et al., 1997;Levine, Prosser, Evans, & Reicher, 2005).
However, the perception of similarities and the disregard for relative deprivation based on power differences (Powers & Ellison, 1995) can reduce willingness to engage in social competition and actions challenging the status quo by both advantaged and disadvantaged groups (Wright & Lubensky, 2009).Additionally, if volunteers offer services and thereby relieve the pressure created by the situation, authorities may be even less inclined to pay sufficient attention to a social issue.This was evident in the case of the refugee crisis, when the lack of services provided by the state was less apparent because of the humanitarian aid offered by volunteers.Furthermore, intergroup helping and volunteerism can underscore the hierarchical differences between the helper and the help recipient: Helping relations are necessarily asymmetrical and convey the message that the helper is more competent and has the resources that the help recipient lacks.According to van Leeuwen and Täuber (2010), members of advantaged groups may help for strategic reasons, to affirm their power and independence, and create a positive group impression and group identity relative to the outgroup.Intergroup helping can also disprove negative stereotypes about the ingroup and restore group image (Hopkins et al., 2007).Furthermore, Nadler's (2002) model of intergroup helping as status relations suggests that intergroup help -and specifically dependency-oriented help -is offered in order to defend the group's privilege when existing status differences are threatened.The strategy of offering, rather than withholding help to the outgroup not only confirms the higher status, but it also helps to maintain the moral advantage of the helper.Thomas and McGarty (in press) distinguish between benevolent and activist forms of generosity to grasp the differences between service efforts to reduce the outgroup's disadvantage and efforts that are directed toward challenging the system that maintains disadvantage.From this perspective, volunteers can be perceived as service providers who strive for social cohesion and offer services within the domains of the existing intergroup structure, whereas social change activists attempt to disrupt the social order (Snyder & Omoto, 2008;Wright & Lubensky, 2009) and challenge intergroup relations (i.e., group boundaries, social distance, and intergroup hierarchies).
Therefore, intergroup volunteerism can only offer help that successfully reduces social distance if it also addresses the structural sources of inequalities (Case, Hensley, & Anderson, 2014;Case, Iuzzini, & Hopkins, 2012;Montgomery & Stewart, 2012).In contrast, paternalistic help and charity services (i.e., dependency-oriented forms of help) maintain the status differences between the groups.In sum, ingroup and intergroup helping and volunteerism share the motivations that are connected to the general rewards of helping, but stemming from the hierarchical differences between helper and help recipient, intergroup volunteerism can paradoxically be the means to maintain intergroup asymmetries and reinforce existing inequalities between the groups (Nadler, Harpaz-Gorodeisky, & Ben-David, 2009).
Nevertheless, volunteerism has social change potentials.Thomas, Rathmann, and McGarty (2017) argue that active volunteerism in contrast to nominal volunteerism can be understood as a form of collective action, based on similar motivations of engagement.Charity organizations and volunteers can offer services that draw attention to a problem and advance change, especially if their work addresses the structural causes of inequalities or empowers the outgroup (Penner, 2002).An example is AIDS activism, which started off as offering services during the AIDS crisis, but became the source of various types of politicized actions (Gould, 2001), and managed to achieve significant change in the social perception and public treatment of HIV/AIDS.Furthermore, directly changing intergroup dynamics is not the only way to achieve social change, and volunteers can work toward social change through for example the redistribution of financial resources and technologies (Hansen & Postmes, 2013).

Social Change Activism
The social psychological literature on collective action is primarily interested in why non-activists engage in actions that benefit their group and challenge an unjust intergroup situation (for a meta-analysis, see van Zomeren, Postmes, & Spears, 2008).Collective action is related to one's identification with a group that is disadvantaged, inferior, or unfairly treated, and it refers to engagement in social competition as an escape from a "negative social identity" (Tajfel & Turner, 1979, 1986) to improve the situation of the group (Wright, Taylor, & Moghaddam, 1990).
Non-activists involved in collective action can become activists following the development of a politicized collective identity.Politicized identities emerge through the recognition of injustices and the decision to engage in bringing about (or preventing) change (Simon & Klandermans, 2001).Activists can be defined as people who actively work toward social and political causes (Curtin & McGarty, 2016).We use the terms political activists and volunteers in the current paper, as our main goal is not to understand how non-active citizens became active in some form of collective action, but rather why active citizens choose a particular form of action (i.e., engagement in direct political action vs. volunteer helping).
i Specifically, we are interested in activities addressing intergroup injustices toward an outgroup, based on the assumption that identities are politicized by the recognition of injustices in an intergroup situation in highly similar ways-regardless of whether injustices are perceived in connection with the ingroup or an outgroup (Curtin, Kende, & Kende, 2016).
However, activists derive their moral motivations from their identification with their activist group, suggesting the existence of group-based moral motivations among them (Stürmer & Simon, 2004;van Zomeren, 2015).Moral convictions are therefore a primary force in the engagement for collective action to achieve change and remove the threat to one's moral principles; they can be both the source of identification with a politicized group (for nonactivists) and the consequence of it (for activists).
Identification with the outgroup or the particular movement is not necessary in the case of a strong opinion-based identity (Bliuc, McGarty, Reynolds, & Muntele, 2007).Identification with an opinion-based group mobilizes people for collective action when aligned with other positively evaluated social categories, or to put it differently: social categories can be successful sources of mobilization if they are also connected to moral stances that represent important social identities (McGarty, Bliuc, Thomas, & Bongiorno, 2009).Opinion-based identities may therefore be more accurate predictors for engagement in social change activism than other forms of group identification, especially when activism is not related to injustices experienced by the ingroup or preexisting membership in an activist organization.
Moral convictions and opinion-based identities are not independent concepts.Ally activism and engagement in political action for ideological goals are driven by both moral convictions and identification with people who share these convictions (for a summary see Curtin & McGarty, 2016).On the one hand, people develop identities based on opinions that reflect moral values, and on the other hand, moral values express identities.And although moral motivations can reflect both personal and group identities, shared moral convictions increase identification with the group (van Zomeren, 2013).
Despite the important differences in motivations in terms of social change, volunteerism and activism do not necessarily refer to essentially distinct activities, as they both fit into the larger category of engaging in social action for the benefit of other people (Kende, 2016;Snyder, 2009).Therefore, we ask whether volunteers of the refugee crisis were mobilized on the basis of social change motivations similar to those that previous research identified in connection with collective action for social change (see Bliuc et al., 2007;Stürmer & Simon, 2004;van Zomeren et al., 2008, van Zomeren et al., 2012), and similar to those who were engaged in direct political action related to the refugee crisis.Our question is justified by a context in which a humanitarian disaster required volunteer help, but at the same time the treatment of refugees by the authorities triggered political protests (Zalán, 2015).Moreover, volunteerism was mostly grassroots and opposed the government, suggesting a connection with political goals.
The refugee crisis was unique because of the scale of both the political tension and the humanitarian crisis.
However, the situation was not unique in the sense that intergroup conflicts involving groups with asymmetrical social status are often politicized, and require that members of the advantaged group help the disadvantaged.
We therefore use the case of the refugee crisis to draw attention to the importance of understanding the similarities and differences in the paths toward volunteerism and political activism, in terms of people's motivation to achieve change, as social movements are just as dependent on mobilizing allies for political actions (see Thomas, McGarty, & Mavor, 2010;Thomas et al., 2017) as they are on mobilizing volunteers.And specifically, these social changeoriented movements are dependent on mobilizing volunteers who not only provide services, but also critically reflect on the structural aspects of disadvantage (see Case et al., 2012).

The Context and Overview of the Study
Between June and August 2015, a growing number of migrants and refugees entered Hungary at the Serbian-Hungarian border, as Hungary became the primary entry point into the European Union.According to official police statistics, 430,607 migrants crossed the border in 2015 (National Headquarters of the Hungarian Police, 2015).
Hungary was considered a transit rather than a destination state, as asylum seekers intended to continue their journey toward Western Europe, where economic conditions and integration opportunities are more favorable.
Hungary had practically no experience with an immigrant population, having one of the lowest percentages of immigrants in Europe; however, anti-immigrant sentiments were running high (Sik & Szeitl, 2016).The Hungarian government was the first to react to the terrorist attack on Charlie Hebdo in Paris with a clearly anti-immigrant rhetoric, and the first to take an openly hostile stance during the refugee crisis in 2015.Even before refugees arrived in Hungary in greater numbers, authorities framed their actions as defense against an external threat, launched an anti-immigrant campaign, and built a barbed wire fence at the southern border.Despite the government's reluctance to accept refugees, free passage out of the country was also cut.As a result, the main railway stations of Budapest quickly filled up and became unofficial open air refugee camps in the summer of 2015, where refugees waited for the opportunity to leave the country.In these so-called transit zones, only civilian volunteers provided services.
Because the number of asylum seekers had been very low prior to this influx of refugees, there were hardly any organizations experienced in dealing with them.Therefore, aid was provided mostly by ad hoc, grassroots organizations recruiting their volunteers on Facebook, supported by the civilian population and some for-profit companies.
Migration aid, the largest of the pro-refugee organizations recruited new volunteers by starting a new (Hungarian) Facebook group in June 2015, which was joined by thousands of people within only a couple of days.Most of these groups did not become formal organizations, they did not have any formal leadership; the more active participants coordinated the actions.Most volunteer groups worked in a specific location (e.g., at a specific railway station), and their activities consisted of meeting the immediate needs of refugees in terms of food, shelter, medical care, and information.Besides these general aid groups, there were some specialized groups offering medical or legal help for refugees.Additionally, some religious organizations -Christian, Jewish, and Islamic groupswere also active in the pro-refugee movement.
Information exchange, coordination of activities, as well as (political) discussions took place online, and mobilization was therefore highly dependent on people's existing online social networks (see Dessewffy & Nagy, 2016).Social media as a tool for mobilization and online discussions can serve rapid identity politicization and the development of opinion-based identities (see McGarty, Thomas, Lala, Smith, & Bliuc, 2014) that we expected in the current context too.We set out to test the influence of such opinion-based identities on volunteering and activism in the context of the refugee crisis in Hungary.
The upsurge of volunteerism with thousands of people offering physical help or donations was surprising and unprecedented in Hungary, which generally has a low level of volunteer engagement, even in a Central European comparison (Voicu & Voicu, 2009).Volunteers received a lot of positive attention in the online media, but the government made explicitly negative remarks about refugee helpers, and portrayed them as political dissidents receiving support from external political forces on public and government-controlled media (Frayer, 2015).Volunteers were subject to frequent verbal and occasional physical attacks by the civilian Hungarian population, but less typically by the police or authorities.

Hypotheses
We hypothesized that engagement in volunteerism -offering direct help to refugees in the form of participating in aid work and logistics, or financially helping them -would be predicted by motivations similar to those for engagement in direct political actions and protests related to the refugee crisis, because helping refugees fulfilled the purpose of both offering humanitarian aid to an outgroup and attempting to change the authorities' treatment of refugees.Specifically, we hypothesized that people who were engaged in some pro-refugee activities (either volunteer work or online and offline political actions) were mobilized on the basis of moral convictions and their identification with the pro-refugee opinion group.Based on the collective action literature (see van Zomeren et al., 2008), we predicted that efficacy beliefs and group based anger would also predict both types of actions, however considering that we relied on a sample of already active people, we expected that these two variables would have less influence than other variables related to social change motivations (Stürmer & Simon, 2004;van Zomeren, 2015).
In order to test whether volunteerism was predicted by the same social change motivations as political activism, when predicting non-activists' collective action intentions.However, as we conducted our study among people who were already active, it was less relevant to understand whether identification with the pro-refugee opinion group was based on moral convictions than to understand if people's level of engagement depended on whether or not their opinion-based identity reflected their moral convictions (for a distinction between the role of moral motivations among non-activists and activists see Stürmer & Simon, 2004;van Zomeren, 2015), as well as efficacy beliefs and anger.

Method
We conducted a survey among people engaged in pro-refugee activities to test these hypotheses.Data was collected in October, 2015, a few days after the borders were closed, and refugees could no longer enter the country.
It was just after the largest wave of the crisis passed, when there was little to do in terms of actual help, but volunteer groups were still active on social media, and the topic continued to dominate public discourse.We recruited our participants using social media, advertising the call for participation in the widest possible range of groups including official NGOs, newly emerged local and national volunteer groups, and various religious organizations.
The study was approved by the IRB of Eötvös Loránd University, Budapest.We report all measures and data exclusions in the current paper.The questionnaire was administered in Hungarian, with adopted measures translated from and back-translated into English.

Measures Outcome Variables
Ten self-developed items (see Table 1) measured the frequency of engagement in different types of helping and political actions, both in online and offline forms, on a 5-point scale from 1 (never) to 5 (very often).As this measure was used to reflect actual participation, based on our knowledge of the context we indicated the types of activities that were prevalent forms of aid work, expression of politicized opinions, and protest during the refugee crisis of 2015.These self-developed items paralleled previous research measures only in so far as they represented actual types of involvement, for example similarly to the activities listed in connection with AIDS volunteering in the study by Simon, Stürmer, and Steffens (2000).We selected activities in which many people participated, and that did not require previous training, such as the work of doctors, lawyers, or translators.Nevertheless, to ensure variance in the responses the items reflected actions with different degrees of involvement and personal cost.Social media was widely used for both instrumental purposes (organizing aid work and protests) and expressing political opinion during the refugee crisis (see Dessewffy & Nagy, 2016), therefore many of our items measured online activities besides the more traditional forms of political actions, such as street protests or contacting politicians.Furthermore, the inclusion of online forms of involvement allowed us to represent the participation of those who were physically away from the events, either because they lived abroad or because they lived in areas distant from the route of refugees.Factor analysis using Maximum likelihood analysis with Promax rotation revealed two correlating factors (r = .44,p < .001,KMO = .82,p < .001)explaining 54.58% of the total variance of Volunteerism (4 items, α = .77)and Political activism (5 items, α = .75).One item had to be removed because it cross-loaded.The political activism scale consisted of three items referring to one type of online activity: posting and sharing political content online (hence the high correlation between the items), and two items assessing more traditional forms of political activism.
We chose to keep the two items on the political activism factor with loadings below .4 for conceptual reasons and because of the acceptable reliability of the scale.Removing these two items would have narrowed down the analysis to posting political content online, rather than a wider spectrum of political actions.
ii The decision to keep the items referring to online and offline political activism as one factor is also supported by research suggesting the connection between online and offline political activism (see McGarty et al., 2014;Valenzuela, Arriagada, & Scherman, 2012).
We asked participants with one item to also indicate the degree to which they self-identified as a helping volunteer and with another item to indicate whether they self-identified as a political activist which was used to validate the behavioral measures in this context (correlations between activities and self-identification as an activist or a volunteer are presented in the section below reporting descriptive statistics).

Predictors
All items -unless otherwise indicated -were measured on a 7-point scale from 1 (completely disagree) to 7 (completely agree).We relied on the 5-item scale used by Bliuc et al. (2007) for assessing opinion-based identity (α = .90),and referred to "refugee helpers" as the relevant opinion-based group.For example, one item read: "I am content about my choice of supporting refugee helpers."Refugee helpers was the term generally used in public discussions, media, and social media to refer to people with a pro-refugee opinion who approved of helping refugees as opposed to those who disapproved of it.
Moral convictions were measured by four items (α = .72)adopted from the scale of van Zomeren et al. (2011van Zomeren et al. ( , 2012) ) to correspond with the refugee crisis: "My opinion about refugees is an important part of my moral norms and values.";"There is only one true stance on this issue, and that is my stance.";"My opinion about refugees is a universal moral value that should apply to everywhere in the world."and "My opinion about refugees reflects an important part of who I am." Group efficacy beliefs were measured using two items -"I believe that we, refugee helpers as a group, can change the situation of refugees in Hungary." and "I believe that we, refugee helpers can achieve the common goal of changing the situation of the refugees in Hungary."-taken from van Zomeren, Spears, Fischer, and Leach ( 2004), r = .85,p < .001.
Anger was measured by one item asking respondents to rate how angry they felt about the current situation of refugees in Hungary.

Control Variables
We tested self-ascribed political orientation using two semantic differential items to measure left-right, liberal-conservative orientations; because of their acceptable correlations (r = .57,p < .001),left-right and liberal-conservative were combined for the analysis.We also asked about age and gender of participants, their financial status and current residence.
For exploratory purposes, we measured other variables connected to political activism and volunteerism to get a more accurate picture of the pro-refugee movement.However, these measures did not serve the purpose of testing our hypothesis, and therefore, they are not presented in the paper.More information on these measures is available in the Supplementary materials.

Sample
The online questionnaire was completed by 1479 respondents recruited on Facebook on the public pages and closed groups of NGOs, volunteer groups, and religious groups who were involved in helping refugees.We expressly invited those who engaged in volunteerism and/or political activism connected to the refugee crisis regardless of whether it was regular or occasional, within or outside organizations.We closed the link after five days because of the rapidly changing political context, the influence of which we could not account for in the cross-sectional design of our study.
After removing 20 respondents from the sample because they indicated no involvement in any of the listed activities, we ran the analysis on the remaining sample of 1459 people (M age = 43.6,SD age = 13.6,18-85).The sample consisted of volunteers and activists on the basis of their involvement as volunteers or activists in the refugee crisis, but it does not mean that they were otherwise engaged volunteers.Based on the measurement of previous engagement in volunteerism, 30.5% reported that they never engaged in a volunteer activity prior to the refugee crisis, and 6.6% never offered a donation or financial aid previously.
We did not use forced response, and analyzed only the available data.We used listwise deletions for the missing data.The majority of respondents were women (80%) which corresponded with the observation that most helpers were also women.59.2% of the sample were residents of Budapest, 34.4% lived in other towns and villages of Hungary, and 6.6% lived abroad.
iii We asked respondents to rate their financial status relative to the Hungarian average, and found that respondents scored somewhat above the midpoint on a 5-point scale from 1 (far below average) to 5 (far above average, M = 3.36, SD = .94).In terms of political orientation, respondents identified rather as left wing-liberal (the semantic differential of left-right, liberal-conservative used a 1 to 7 scale: M = 2.65, SD = 1.35).We did not collect information about formal membership in or informal association with any specific pro-refugee organization, and therefore relied on individual differences in the level of volunteer or political activities as our outcome measure.

Descriptive Statistics
Volunteerism, that is participation in activities that in some way directly help refugees, such as spending time at the railway stations, offering food, money, or information (as presented in Table 1), was highly correlated with respondents' self-identification as volunteers (r = .64,p < .001).The factor of political activism that is the expression of political opinion either in social media or by participating in street protests correlated with respondents' selfidentification as political activists (r = .34,p < .001).To check whether the self-identification also correlated with engagement in activities opposite to self-identification, we ran partial correlations (necessitated by the correlations between the two types of behaviors).We found that self-identification as a volunteer showed a negative and very weak correlation with engagement in political actions when engagement in volunteerism was controlled for (r = -.06,p = .020).Similarly, the partial correlations between identification as a political activist and engagement in volunteerism was negative and very weak when engagement in political actions was controlled for (r = -.07,p = .007).These findings provided external validation of our measures of political activism and volunteerism.Note.All steps of the models were significant at p < .001.We removed demographic variables from the tables that were nonsignificant in both steps.

Mediation Model
Our hypothesis that the connection between on the one hand opinion-based identity, and on the other hand, either volunteerism or political activism, is mediated by moral convictions, efficacy beliefs, and anger was tested using the Process macro (Hayes, 2013; testing Model 4 using a regression-based path model that allows the inclusion of multiple parallel mediators in the analysis).The significance of the indirect effects was assessed with 95% bias corrected bootstrap confidence intervals based on 1,000 re-samples (Hayes, 2013).
We entered opinion-based identity as the independent variable and either volunteerism or political activism as the dependent variable.Moral convictions, efficacy beliefs, and anger functioned as mediators.Total and indirect effects are reported in Table 4.In both models the results show a significant indirect effect of opinion-based identity via moral convictions, but not efficacy beliefs or anger.Thus, moral convictions mediates the effects of opinion-based identity and volunteerism, while neither efficacy beliefs nor anger functioned as significant mediators.For political activism both moral convictions and anger were significant mediators, while efficacy beliefs were not.

Discussion
We interpreted the refugee crisis as a politicized context in which immediate humanitarian help was required, and hypothesized that in this context, people would engage in volunteerism largely by the same social change motivations that the literature identified as motivations to engage in collective action to achieve social change, that is by identification with the relevant politicized group, moral convictions, efficacy beliefs, and anger (e.g., van Zomeren et al., 2011).The findings only partially supported the hypothesis about volunteerism, as anger and efficacy beliefs did not predict volunteer engagement.Nevertheless, the results, rather than showing similarities, suggested that mobilization for volunteerism was based on social change motivations more than mobilization for political activism.
A possible interpretation of this finding may be that in the context of the refugee crisis, volunteerism was the normatively appropriate response among those who held a pro-refugee opinions, based on both their attitudes and the norms of their ingroup (see theory of planned behavior: Fishbein & Ajzen, 1975).
Opinion-based identity and moral convictions were present as motivators of both types of actions, while efficacy beliefs and anger were less important predictors of both types of actions: not significant predictors of volunteerism, and although they were significant predictors of political activism, they added little to the explained variance.The results also supported our hypothesis about the mediating role of moral convictions, but refuted the hypothesis about a similar mediating role of anger and efficacy beliefs: Efficacy was not a mediator in either models, while anger was a weak but significant mediator for political activism only.These findings imply that in the context of the refugee crisis, the level of involvement could not be explained by anger and efficacy as it was in previous research on collective action in other contexts (van Zomeren et al., 2008).
Previous research suggests that anger is an important motivation for non-activists' initial mobilization and collective action intentions, but less important for already engaged activists (Stürmer & Simon, 2004;van Zomeren, 2015).
Efficacy beliefs were relatively low overall, suggesting that belief in change was not prevalent, and people did not offer volunteer help on the basis of efficacy beliefs related to change, while efficacy had some connection with engagement in direct political actions.The low level of efficacy beliefs and its weak predictive value may be more surprising for political activism than for volunteerism, as according to classical theories of social change and collective action, efficacy beliefs and belief in change are important conditions for engagement in collective action (Tajfel & Turner, 1986;van Zomeren et al., 2008), but they are less prevalent in volunteerism which does not necessarily focus on change (Penner, 2004).
The marginal role of efficacy beliefs in this context can be explained by prevalent political attitudes in Hungary.
The political system is predominantly perceived as stable, but illegitimate, and satisfaction with the democratic system is low (Szabó, 2013).There is a general mistrust in politics and institutions, and both just-world and system justifying beliefs are low.System justification beliefs are negatively connected to meritocratic ideology contrary to Western European or North American research findings (Berkics, Kóbor, & Karácsonyi, 2006;van der Toorn, Berkics, & Jost, 2010).These general attitudes explain a sense of loss of control and consequently a lack of interest in politics and low political participation (Lantos & Kende, 2015;Szabó, 2013).Accordingly, people may engage in protests in order to express their opinion even without a sense of control, influence, or hope for change, hence unrelated to efficacy motivations.
Our findings suggest not only that volunteers were motivated to act on their opinion-based identity and moral convictions, but also that these motivations were more important for engagement in volunteerism than in political activism.Our respondents -many of whom had not participated in any kind of volunteer activity previouslychose volunteerism to express their opinion-based identities and act on their moral convictions.Volunteerism may have therefore been perceived as a more appropriate response to a humanitarian crisis than the direct expression of political dissent, especially among women.
Men and women tended to choose different degrees of involvement in volunteer helping and activism.Political participation of women is low compared to men in most liberal democracies (Bari, 2005), while volunteerism, especially care-type services or those benefiting e.g., the poor or elderly, are more prevalent among women (Dittrich & Mey, 2015;Taniguchi, 2006).This too may have contributed to our findings about a closer connection between on the one hand, opinion-based identity and moral convictions, and volunteerism on the other.

Limitations and Future Research
Our study was conducted in a highly polarized societal context with the majority society -fueled by the actions of the government -strongly rejecting refugees (Sik & Szeitl, 2016).This context and the reliance on a convenience sample may bring some limitations to our results.We suspect some self-selection among respondents, with a higher representation of those who were motivated to express their political opinion.Therefore, our results may partially stem from an overrepresentation of volunteers with social change motivations.We therefore suggest that future research try to eliminate this bias by employing a more balanced sample of volunteers when examining social change motivations in intergroup helping within a politicized context.Nevertheless, the volume of responses reassured us that our questionnaire touched upon issues that respondents found relevant and in line with their experiences as volunteers or activists.Furthermore, reliance on a sample of already engaged volunteers and activists allowed us to reflect on mobilization differently from studies conducted among non-active populations that analyze mobilization intentions.
Opinion-based identity -identifying with the pro-refugee opinion group -was a strong predictor only for volunteerism.This may have been a consequence of the wording of the scale, in which we used the term "refugee helper" to designate the opinion group, rather than referring to the group merely as holding a pro-refugee opinion.Therefore, people who physically helped may have found it easier to state identification with other helpers.
The nonexperimental design of the current study could only highlight the connections between social change motivations and engagement in the two forms of action, but their causal connections should be further investigated using a longitudinal design.Our study pointed out that engagement in volunteerism can reflect the presence of social change motivations, but future research should establish the conditions under which people choose volunteerism over political activism and vice versa.
Our research focused on motivational variables of engagement in volunteerism and collective action.However, we acknowledge that individual motivations constitute only one aspect of the conditions of mobilization.This focus on individual motivations, and a lack of attention to other social psychological phenomena (e.g., intergroup attitudes) as well as societal level factors such as the political climate and the mobilizing potential of organizations, can be the reasons for the relatively low explained variance in our statistical analyses.
Finally, one potential way to distinguish between volunteers with or without social change motivations is to measure system justification (Jost & Hunyady, 2003) that was not included in the questionnaire.Individual differences in the endorsement of system justification beliefs, as well situational and cultural aspects of system justification or system critique could help explain some of our findings about the motivational aspects of volunteerism.

Conclusions
Based on our findings, we argue that a large group of people engaged in volunteer helping with opinion-based motivations and moral convictions which have previously been identified as important drivers for social change oriented collective action (Bliuc et al., 2007;McGarty et al., 2009;van Zomeren et al., 2012).Volunteerism was therefore not just cross-group helping out of sympathy or for the rewards of helping (e.g., Hopkins et al., 2007;Penner, 2004), but to some extent a means to express moral convictions on the basis of opinion-based identity.
Thus, we supplement the existing literature of volunteerism and social change activism with the claim that volunteers and political activists are not necessarily different by virtue of their motivations, but that they choose different actions to alleviate the problems embedded in the intergroup situation.The decisive difference is therefore not in the intention to achieve change, but rather in the interpretation of the situation and the perception of required action.
Our results support the hypothesis that under strong political pressure, in a situation in which intergroup help is an appropriate form of engagement, volunteers may be motivated to help disadvantaged outgroups due to underlying social change motivations.In other words, people motivated for collective action may choose volunteerism to restore the moral principles shared by their opinion-based group if volunteerism seems the more adequate answer to the problem, or fits better with their personal inclinations.The refugee crisis may seem unique in some ways because of its severity and because it happened as an unforeseen humanitarian crisis, but all intergroup conflicts between groups in hierarchical relationships share some of the important characteristics of the current situation.Members of advantaged groups are not only essential for participating in the intergroup struggle as ally activists (see Thomas et al., 2010), but they are also needed to alleviate the everyday financial, educational, or health difficulties of members of disadvantaged groups.While the activities of pro-refugee volunteers consisted mostly of aid work -offering food, information, shelter, and financial support -these acts became the means of expressing a desire for social change, and should therefore be recognized as such.

Notes
i) Admittedly, the term political activism does not differentiate between those who work toward social change or toward conserving the existing status quo.Therefore, conservative or reactionary forms of activism should also be recognized as political activism (Kende, 2016).However, this discussion falls outside the scope of this paper, as in the context of the refugee crisis, political activism was clearly directed toward expressing political discontent with the way authorities dealt with the refugee crisis and had the explicit goal of achieving change.
ii) We checked whether the results remain similar if we use the 3-item online activism subscale, and found that efficacy was not a significant predictor of online activism, but all other variables remained significant predictors in the regression model after controlling for demographic variables (R 2 = .13,F(7, 1373) = 29.84,p < .001,opinion-based identity: β = .14,p < .001moral convictions: β = .17,p < .001,efficacy beliefs: β = .01,p = .936),and the mediation model had the same significant paths as with the 5-item scale presented in the text.
iii) Although living abroad may be a physical obstacle to some forms of volunteerism or political activism, we opted for the inclusion of participants living abroad despite the fact that their pattern of involvement was somewhat different than the engagement of those who lived in Hungary.Firstly, because emigrant Hungarians living mostly in Western Europe played a visible role in the refugee crisis.They participated not only on an individual basis, but also as groups using their existing channels of communication, and newly formed ones.They carried out a lot of activities that did not require personal participation, such as translation of information, financial donations, food deliveries, buying train tickets online, background coordination of activities, etc. Removing them from the sample would have distorted the representation of the movement.Secondly, since the refugee crisis peaked in the summer months, some of the people living abroad were also physically present, and participated in person during their summer vacations.Specifically, in response to the item: "I helped in the field…" 42.7%, and in connection with the item "Did you participate in some street demonstration…?",38.6% answered that they participated once or more.Nevertheless, t-test shows that participants living abroad were somewhat less active in Volunteerism: M abroad = 2.14, SD abroad = 0.87, M local = 2.47, SD local = 1.00, t(1445) = 3.13, p = .002,while the level of Political activism was not different: M abroad = 2.13, SD abroad = 0.80, M local = 2.03, SD local = 0.91, t(1445) = -1.10,p = .271).Removal of this subsample does not produce any changes in the patterns of the regression models, the same predictors are significant with or without respondents living abroad.
we tested a mediation model including the variables outlined byvan Zomeren et al. (2012): politicized identity, moral convictions, group based anger, and group efficacy beliefs.The mediation model, as shown in Figure1, depicts our prediction that identification with the pro-refugee opinion group plays an important role in motivating people to engage in both volunteerism and political activism; however, the mobilizing potential of opinion-based identity influences perceived violations of moral principles (i.e., by moral convictions), anger, and efficacy beliefs, which in turn influence the level of engagement in pro-refugee actions.Van Zomeren et al. (2012) found support for a different mediation model between moral convictions and collective action mediated by politicized identity

Figure 1 .
Figure 1.Predicted mediation of moral convictions, group based anger and efficacy beliefs in the effect of opinion-based identity on both volunteerism and political activism.

Table 1
Items Measuring Volunteerism and Political Activism With Factor Loadings Based on the Pattern Matrix of the Factor AnalysisI helped refugees in the field (e.g., train stations, refugee camps, headquarters of civil organizations, I had refugees stay in my house, I transported refugees in my car).

Table 2
Means, Standard Deviations and Correlations for All Variables

Table 3
Hierarchical Regression Analysis With Volunteerism and Political Activism as Dependent Variables

Table 4
Analyses of Moral Convictions, Efficacy Beliefs, and Anger as Mediators in the Relationship Between Opinion-Based Identity and Either Volunteerism or Political Activism