One World in Diversity – A Social-Psychological Intervention to Foster International Collective Action Intention

Although effective interventions to increase international collective action for human rights are highly desirable, the validation of theory-based interventions and their transfer to this practical field is still scarce. We investigated whether collective action intention can be improved by using a diversity intervention technique rooted in social psychology. The intervention builds on the ingroup projection model, postulating that negative intergroup relations are based on the perception of ingroups as more typical of a common superordinate group than outgroups (i.e., relative prototypicality). Thereby, the (quasi-)experimental study tested the ingroup projection model’s theoretical assumptions in the context of Model United Nations (MUN) conferences. We hypothesized that the diversity intervention leads to a higher perceived diversity within the superordinate group (the United Nations, UN) as well as identification with the superordinate group (UN). Furthermore, we hypothesized an indirect effect of the intervention on collective action intention mediated by perception of diversity of, and identification with, the superordinate group. In comparison to the control group (n = 45), those participants who received the diversity workshop intervention (n = 55) perceived less relative prototypicality and more diversity of the UN. In addition, we provide evidence of a serial mediation: Compared to the control group, the diversity workshop group perceived the UN as relatively more diverse, facilitating identification with the UN. In turn, this was associated with a stronger intention to act collectively. This study shows the importance of including psychological theories in the field of international relations.


Zusammenfassung
The world is currently facing the highest number of refugees since the aftermath of World War II.Filippo Grandi (2017), United Nations High Commisioner, stated via twitter: "In 2015 the EU agreed to relocate 66,400 refugees from Greece.Less than 8000 have been relocated".This indicates a lack of action, or action intention.Also, one year ago 193 member states of the UN worked on a Declaration for Refugees and Migrants.But experts say the obstacles to achieving the objectives are not tackled enough by the member states.Regarding the inadequate response to climate change one can also state that there is no intention to act sufficiently.When UN member states' representatives gathered in Paris in 2015, they agreed on a new climate contract.Although this was seen as an important turn in reducing climate change, a press release in November 2016 by the United Nations Environment Programme stated that the goals were still not sufficient to stop the destructive consequences of climate change.Additionally there are country leaders who openely say they want to opt out of the Paris agreement.This begs the questions: 1) Why is there a lack of collective action intention in the face of important crises such as climate change and refugee relocation? 2) How can one increase collective action intention concerning human rights issues among people in charge?Many political and economic factors could help explaining these questions.However, social-psychological factors could also inhibit or promote collective action intention at the global level."Sometimes you just don't get to a solution because you don't like each other" (anonymous diplomat in a UN discussion, personal communication).
In the last few decades, much research about fostering intergroup relations originated in psychological science (Messick & Mackie, 1989;Tajfel, 1982;Tajfel, Billig, Bundy, & Flament, 1971).The aim of the present study is to construct and evaluate a diversity workshop to endorse and facilitate collective action intention among (future) leading negotiators.Based on practical considerations, we will work with a set of simulated international negotiations, called the Model United Nations (MUN; see Methods section).Our goal is on the one hand to go beyond investigating interventions for improving attitudes, but extend research to collective action, and on the other hand to look for practical applications of social-psychological knowledge.

The Ingroup Projection Model (IPM) as a Basis for Diversity Interventions
Intergroup relations are influenced by group members' perception of a superordinate group that includes the ingroup as well as outgroup(s).Therefore, it has been argued that intergroup relations can be improved by changing that representation (Wenzel, Mummendey, & Waldzus, 2007).The ingroup projection model (Wenzel, Waldzus, & Steffens, 2016) states that it is not sufficient to simply share a superordinate category as suggested by the common ingroup identity model (Gaertner, Dovidio, Anastasio, Bachman, & Rust, 1993).Rather, the specific cognitive representation of that common superordinate group is also important.Specifically, if someone holds a more diverse representation of the superordinate category, intergroup relations will be improved (Mummendey & Wenzel, 1999).This is because the way the superordinate category is represented affects perceived relative prototypicality.Relative prototypicality is the phenomenon that group members tend to perceive their ingroup as more prototypical for a common superordinate group than an outgroup.A prototype is defined as "the ideal-type member of a category that best represents its identity in a given context and frame of reference" (Wenzel et al., 2007, p. 335).That is, members 'project' characteristics of their own group to the superordinate category, and therefore feel more prototypical than members of the outgroup.Thus ingroup projection is used as "a label for the perception, or claim, of the ingroup's greater relative prototypicality for the superordinate group" (Wenzel et al., 2007, p. 337).As Wenzel et al. (2007) reviewed, there is a correlation between intergroup attitudes and relative prototypicality.Specifically, as perceived relative prototypicality increases, intergroup attitudes worsen.For example, Reese, Berthold, and Steffens (2012) showed that people from developed countries perceived their group as more proto-typical of the world community, and this was in turn related to beliefs that global inequality is legitimate.As developed countries are often in more powerful positions in a wide range of negotiation settings, perceived prototypicality could result in unequal living, trade, and environmental conditions or higher perceived entitlement in general.Interventions that reduce perceived relative prototypicality could therefore be useful to arrive at more equal negotiation processes and results.As Waldzus, Mummendey, Wenzel, and Weber (2003) showed, increasing perceived diversity of the superordinate category reduces the perception of relative prototypicality among (sub)groups.

Diversity Training, Identification, and Collective Action
By inducing a diverse representation of the superordinate category, Waldzus et al. (2003) successfully reduced ingroup projection (also see Waldzus, Mummendey, & Wenzel, 2005).The authors randomly assigned their German participants to one of two conditions.In the 'complex' condition participants had to think about the diversity of the superordinate category (in this case Europe), whereas participants in the control condition had to think about its unity.Those participants who were in the 'complex' condition rated the relative prototypicality of Germans lower than those in the control condition.In line with these findings, it has been reasoned that when the superordinate group is represented as complex and consisting of many different prototypes, a "single group cannot reasonably claim to be the part that represents the whole" (Wenzel et al., 2007, p. 366).Thus, diversity training based on the IPM seemed to be useful for our study.We adopt Pendry, Driscoll, and Field's (2007) definition of diversity training "as any discrete programme, or set of programmes, which aims to influence participants to increase their positive or decrease their negative intergroup behaviors, such that less prejudice or discrimination is displayed towards others perceived as different in their group affiliation(s)" (p.29).A previous study already corroborated that the ingroup projection model can be used as a theoretical basis for diversity training to improve intergroup relations (see Ehrke, Berthold, & Steffens, 2014).Ehrke and colleagues implemented a diversity-training programme as a get-to-know-you exercise for first-year students (Experiment 1), and as a one-day training programme (Experiment 2).The training increased perceived diversity of the superordinate group, which mediated its positive effect on attitudes.As findings were limited to effects on attitudes, the main aims of the present research were to extend the effects to collective action intention and to test them in an applied setting.
In line with Becker (2012), we define collective action "as any action that promotes the interests of one's ingroup or is conducted in political solidarity" (p.19).The global fight against climate change can be seen as a collective action task.The countries' leaders need to work out an agreement (and a course of action) to prevent further escalation of anthropogenic climate change.Past research suggests that social identification impacts on the likelihood of collective action being organized and executed by the given group.Group identification has been shown to predict not only members' motivation to ensure group success, but also the likelihood of translating this motivation into action (Stürmer & Simon, 2004).Kawakami and Dion (1993) showed that the salience of the individual's social identity, rather than personal identity, increased his or her collective action intentions.In particular, social identification was a key predictor of collective action as it directly and indirectly (by injustice and efficacy perceptions) affects collective action (van Zomeren, Postmes, & Spears, 2008).Thus, on the one hand social identification directly fosters collective action and on the other hand it increases efficacy and injustice perceptions, which, in turn, enhance collective action.Further, a multinational study focusing on intergroup behaviour in a public goods game demonstrated the importance of subscribing to a global social identity over one based on nationality in terms of finding a fair solution to a common problem (Buchan et al., 2011).Because of the fact that subgroup identity is very salient during UN discussions, interventions increasing identification with a common ingroup could be a promising avenue for collective action.
While past research has shown that diversity training can mitigate ingroup bias (Ehrke et al., 2014), it is unclear whether such training can facilitate identification with the superordinate category (and thus collective action).For example, Peker, Crisp, and Hogg (2010) found that a complex representation of the superordinate category reduced ingroup projection as well as superordinate category identification.This would be detrimental for fostering collective action as high social identification was found to be the key predictor of collective action (van Zomeren et al., 2008). However, van Knippenberg, Haslam, andPlatow (2007) presented a moderator of the relationship between diversity and identification with the superordinate category.They found that in an organizational context where diversity was valued, identification with the organization strengthened.It seems that in order to induce increased identification with a superordinate group through diversity training, one must foster an appreciation and recognition of the value of diversity first.Based on the reasoning that diversity perceptions and social identification should be influenced by our diversity workshop and be positively related to collective action intention, we assume that this results in a serial mediation.We hypothesize that diversity perception of and social identification with the UN should mediate the effect of the diversity workshop on collective action intention.

The Current Study
In the current study, we conducted and evaluated the effectiveness of a diversity training, based on the ingroup projection model, on action intentions concerning human rights issues.We held several workshops before simulated United Nations negotiations took place to test if the diversity training workshops increase collective action intention.
Four hypotheses were tested.
Hypothesis 1: Participants in the diversity training condition will perceive the UN as more diverse compared to those in the control condition.
Hypothesis 2: Participants in the diversity training condition will identify more strongly with the UN (superordinate identification) compared to the control condition.
Hypothesis 3: Participants in the diversity training condition will have a higher intention to act collectively compared to the control condition.
Hypothesis 4: The impact of the diversity training on participants' intention to act collectively will be serially mediated by their perception of the UN as a diverse organization as well as by their identification with the UN.

Method Pretest and Measurement of Relative Prototypicality
Ingroup projection was operationalized based on a pretest (n = 10).Previous operationalizations of ingroup projection have varied among studies (e.g.Wenzel et al., 2007).In the present multigroup context, it can be assumed that widely recognized attributes of the superordinate category (UN) are projected onto the ingroup (Wenzel et al., 2007, p. 337).Thus, attributes describing the work of the United Nations were gathered from the UN's homepage (http://www.un.org/en/aboutun/index.shtml).Then, participants were asked: Please indicate how much you link the following attributes with the United Nations (1 = I connect the UN very little with; 7 = I connect the UN very much with).Three particular attributes were selected, which had the highest mean rating and indicated a positive evaluation of the UN.The three attributes were: providing humanitarian assistance, being diplomatic, promoting human rights.These attributes were used to construct a measure of perceived relative protoypicality (see below).

Setting: Model United Nations
The diversity training was undertaken in a Model United Nations (MUN) setting.MUN conferences are educational programmes organized by many universities in which students are assigned to function as representatives of member states of the UN (for more information, see http://www.nmun.org).Students discuss world politics, bearing in mind their designated nation's recent history and policies.Important negotiations between different nations are simulated so students learn to take the perspective of another nation as well as become privy to current international issues.During the last decades, MUN have been developed extensively.Today almost every big city around the world hosts MUN, with one of its goals being to function as a diplomatic training for future international leaders.
The workshops conducted in this study were held before the participants of the MUN started their negotiations.

Participants and Design
Participants were invited to take part in workshops (in English language) on the United Nations with an accompanying evaluation.They were requested to answer a survey prior and after the workshop.Due to low data return at the first measurement time, we can only report these data for part of the sample (see below).Additionally because of insufficient sample size we conducted three workshops.Data were aggregated in the end (see Table 1).The three diversity training workshops did not differ in content, but they differed in three aspects (also see Figure 1).First, participants in diversity Workshop 1 and 3 were assigned different nationalities (e.g.Germany, Great Britain, India, United States, Russia, South Africa, which is common during MUN negotiations), whereas participants in the diversity Workshop 2 all participated as German citizens, which was their actual citizenship.Second, only participants in Workshop 2 and 3 completed a MUN simulation after the workshop.Participants in Workshop 1 did not participate in a MUN simulation afterwards (but received the same preparation materials; see below).
There were also three control groups.Control groups 1 and 3 did not participate in a workshop, and thus only answered the survey assessing the dependent variables.Control group 2 consisted of randomly assigned participants who participated in a control workshop (see below for information on content).Additionally, the themes of the subsequent UN simulations differed.Part of the participants signed up to simulate the United Nations Environment Programme (Control group 1 and 3, Diversity group 3, see Table 1), others the Security Council (Control group 2 and Diversity group 1 and 2).The research design was (quasi-)experimental.We randomly assigned participants to Workshop 2 and Control group 2 as well as to Workshop 1 and Control group 1. Diversity Workshop 3 and Control group 3 could not be randomly assigned as they were based on two different UN simulations taking place at different locations.All groups received the same surveys.The independent variable was the condition (diversity training workshop/control), the measurement of relative prototypicality was used as a manipulation check, diversity perception and UN identification were mediators, and collective action intention was the main dependent variable.

Procedure
Several weeks before the main study, participants received preparation materials from the MUN conference organizers or from us, informing them of their political position for their simulated committee.For example, the participants simulating UNEP (United Nations Environment Programme) received a paper with general information about climate change and the distribution of CO 2 emissions among the participating countries.Consequently, each participant received a paper and had to prepare the negotiation position of their country in advance of the simulation.
We sent participants an online survey for Measurement point 1 at least one week prior to the workshop.This contained demographics, identification measures, and pretest prototypicality measures.People participating in a workshop (diversity or control) filled out a paper and pencil survey at Measurement point 2 after the workshops, whereas control participants without a workshop did it online.At the end, all participants were thoroughly debriefed via e-mail about the goals of the study.

Measures
We administered two surveys (Measurement point 1 and 2) -Measurement 1 did not contain diversity perception and collective action intention measures because that could have affected responding.Apart from that the surveys were identical.Answers ranged from 1 (strongly disagree) to 7 (strongly agree).

Negotiation Position
The survey began with a task to list key elements of the negotiation position of their (assigned) member state in order to ensure that participants were prepared to take on the negotiation position.

Prototypicality
Ingroup projection was measured using the three attributes selected after the pretest (humanitarian assistance, being diplomatic, promoting human rights).Participants were instructed to evaluate how much they link these three attributes with each participating country (scale 1 "very little connected" to 7 "very much connected").Hence, participants indicated indirectly how much they associated each country (their own and the 12 other countries that took part in the simulation) with the UN.The attributes were consequently used as an indirect measure of group prototypicality.The difference in mean rating scores between the ingroup and the other groups was used as a prototypicality index.A positive index indicated that the person evaluated "his/her" country as more prototypical for the UN compared to the average of all other countries.

Diversity Perception
To measure perceived diversity of the superordinate category (UN), we asked: I think the United Nations are a very diverse forum.

Identification Measures
Participants indicated their identification with the national (i.e., assigned country; α = .88)and superordinate category (α = .91)with two items each that were averaged: I identify with my nation; I have a sense of belonging to my nation and correspondingly: I feel that I am part of the United Nations; I identify with the United Nations (Postmes, Haslam, & Jans, 2013).

Collective Action Intention
The survey included a nine-item scale to measure collective action intention (α = .88;e.g., "I am willing to engage in joint action even if this implies we suffer economic losses in the short-run; all nations need to work together to be able to combat the consequences of climate change").This scale also included items that measured collective action intention for other international problems such as handling migration movements and reducing poverty.
Due to the specific context of this study the items were self-constructed (see full list Table A.1 in the Appendix).
Means of participants in Workshop 3 only consisted of items measuring collective action intention concerning climate change (see Table 1).

Socio-Demographic Data
Finally, we asked for demographic information, including age, gender, study major, and career aspiration.

Workshops Diversity Training
In order to avoid triggering threat by reflecting on the diversity of the United Nations (Steffens, Reese, Ehrke, & Jonas, 2017), the workshop started with a get-to-know-you game.Self-affirmation theory (Cohen & Sherman, 2014) postulates that reminding people of their individual abilities functions as a buffer against group threat.
Therefore we asked participants to think of one of their main abilities or characteristics that best suits the goal of describing themselves to a yet unknown partner.After the self-affirmative intervention participants received input about diversity-related international developments.In particular, it was outlined that increased migration and technical development will lead to more diverse societies.Then, the diversity induction began.Similar to the laboratory studies presented by Wenzel et al. (2007), participants were invited to think about the diversity of the  United Nations as well as diversity in general.After collecting participants' ideas about diversity we posed the question of which ideologies about diversity can have the best results for intergroup relations.The two strategies people and governments may use in dealing with diversity, colorblindness and multiculturalism, were contrasted and empirical research was presented that supported a multiculturalism strategy (Correll, Park, & Smith, 2008;Wolsko, Park, Judd, & Wittenbrink, 2000).In addition, the UNESCO declaration of cultural diversity (UNESCO, 2001) was presented in order to illustrate the pro-diversity climate within the UN.Finally, participants were asked to work in groups of four or five representatives.They received the following instruction: "Reflect on the diversity of the United Nations, and think of policies that can be developed within the UN to acknowledge diversity".When they were finished they were asked to present their results to the plenum.The workshop took 90 minutes.

Control Workshop
The control workshop also was described as a workshop on the UN.With regard to content, the control workshop did not discuss diversity of the UN.Instead participants received input about social psychological research on theories and concepts such as social identity theory (Tajfel & Turner, 1979), the self-fulfilling prophecy (Rosenthal & Jacobson, 1966), and the ultimate attribution error (Pettigrew, 1979).Divided into groups, they received the task to think about situations in a UN context in which these theories and concepts may apply, and reflect on possible consequences of psychological biases in negotiation processes.This workshop also took 90 minutes.

Preliminary Analyses Differences Prior to Intervention
We originally intended to analyze a pre-/post-intervention design, but due to the low quota of data return we can only investigate with a subset of the participants whether the two groups already differed on the dependent variables prior to the workshops.For those participants for whom data were available, neither identification with the UN (diversity M = 4.33, SD = 1.68; control M = 4.02, SD = 1.63; t(56) = -.70,p = .48)nor relative prototypicality (diversity M = 1.47,SD = 1.72; control M = 1.68,SD = 1.03, t(48) = 0.53, p = .60)differed significantly between the two conditions prior to the workshop.

Correlational Analyses
Table 2 shows the correlational pattern in both groups seperately.The relationships appear to be in the hypothesized direction for both conditions.Identification, diversity, and collective action all correlated positively (albeit not always significantly), suggesting the importance of diversity perception and collective action intentions.However, there were no relationships between perceived prototypicality and the other three variables.

Manipulation Check
Table 3 provides the means and standard deviations for prototypicality separately for the experimental conditions.Consistent with our assumption, a t-test revealed that the difference index score was higher in the control group (M = 1.77,SD = 1.00) than in the workshop group (M = 1.19,SD = 1.26), t(70) = 2.11, p = .04,d = 0.51.Consequently, participants in the diversity workshop condition perceived their country as relatively less prototypical for the UN than those in the control group did.

Effects on Main Outcome Variables Perceived Diversity of the UN
To determine whether the diversity workshop group perceived the UN as more diverse than the control group (Hypothesis 1), we computed a t-test which yielded a significant difference, t(90) = -2.82,p < .01,d = 0.60.In line with our prediction, the diversity workshop group perceived the UN as more diverse after the intervention compared to the control group (see Table 3).

Identification With UN
A t-test was conducted to compare participant identification with the UN post intervention between the two conditions.
The difference between the two conditions was not significant, t(87) = -0.98,p = .33.Controlling for pre-identification, the difference was in the same direction, larger (workshop M = 4.65, SD = .24;control M = 3.99, SD = .27),but not statistically significant, F(46) = 3.40, p = .07.Thus, while the relevant associations in our data appear to be in the hypothesized direction, we found no statistically significant effect of the diversity intervention on identification with the superordinate category.

Collective Action Intention
We hypothesized that the diversity workshop would lead to higher collective action intention compared to the control group (Hypothesis 3).However, there was no significant difference between the diversity group and the control group, t(90) = -0.27,p = .79(see Table 3).

Analysis of Indirect effects
In order to test whether there is an indirect effect of the diversity training on collective action intention by participants' perception of diversity within the UN and their identification with the UN (Hypothesis 4), multiple regression analyses were conducted with the Process macro (Hayes, 2012).The results are displayed in Figure 2. First, they indicated that the diversity training (0 = control group, 1 = diversity group) led to a more diverse perception of the UN, b = 0.80, SE = 0.25, p < .01,compared to the control group.Diversity perception was positively associated with identification with the UN, b = 0.38, SE = 0.13, p < .01.Moreover, higher identification with the UN was associated with stronger intention to act collectively on international issues, b = 0.22, SE = 0.07, p < .01.We used a bootstrapping method with bias-corrected confidence interval estimates to assess whether the diversity training impacted on collective action intention through participants' diversity perception and identification with the superordinate category (Shrout & Bolger, 2002).Only including diversity perception as a mediator in the model results in b = 0.17, SE = 0.10.The 95% confidence interval of the indirect effect was estimated with 1,000 bootstrap resamples, and indicated a significant association (CI [.02,.43]).Adding identification with the UN to the mediation model resulted in a serial mediation on collective action intention with an effect of b = .07,SE = 0.04, CI [.01, .20].
Consequently there is an indirect effect of the diversity training on collective action intention through diversity perception and identification with the UN.

Discussion
Extensive research in psychology has investigated mechanisms and interventions to foster intergroup relations.
The present study took place in a (simulated) multinational context, the United Nations.The main aims were to extend knowledge of the effects of diversity trainings on collective action intention and to test this in an applied setting.We conducted a diversity training workshop based on the ingroup projection model with the intention to decrease relative prototypicality.We hypothesized that the workshop would increase the perception of diversity within the UN, identification with the superordinate category (UN), as well as the intention to engage in collective action regarding human rights issues.Results indicated that the workshop reduced perceived relative prototypicality, as intended.Furthermore, we provide evidence that the diversity training workshop increased perceived diversity of the UN, which in turn was positively related to collective action intentions.Our data are also in line with the idea that there was an indirect effect of the diversity training workshop on participants' intention to engage in collective action, mediated by the perceived diversity of the UN and the extent of superordinate category (UN) identification.To our knowledge, this is the first study that extends research from effects on attitudes to collective action intentions in evaluations of a diversity training workshop.
It is promising that the diversity workshop group showed less ingroup projection after the workshop than the control group.The diversity workshop therefore seems effective in reducing ingroup projection.

Theoretical Implications, Limitations, and Future Research Directions
Although we cannot provide conclusive evidence that relative prototypicality did not differ between subgroups prior to the intervention, subanalyses with partial data suggest that this was not the case.Nonetheless, future efforts should be invested in replicating and extending our results with a full pre-post randomized design.Further, the finding of reduced ingroup prototypicality should be tested in a real setting (real UN negotiations), as it would corroborate the effects of ingroup projection in negotiation settings.The missing relationship between relative prototypicality and collective action intention and perceived diversity needs to be noted as a limitation.If there was a negative relationship between relative prototypicality and collective action we could corroborate the importance of reducing ingroup projection to facilitate collective action intention more clearly.In contrast to other research (e.g.Waldzus et al., 2003), we measured relative prototypicality deductively instead of inductively.Specifically, we used attributes of the superordinate category that were projected to the subgroup instead of using subgroup attributes projected to the superordinate group.As the direction of projection has not been fully investigated yet, the direction from superordinate to subgroup seemed plausible to us in this specific context.However, measuring relative prototypicality in a rather unconventional way could be one reason for the missing relationship in our data.Also, error variance was introduced because of the different ingroups for different participants, so effects were harder to detect.Nevertheless, we consistently found that perceived diversity within the UN was higher in the diversity group compared to the control group.Further, perceived diversity, identification with the UN, and collective action intention were positively related to each other.Consequently there are indications for the importance of diversity perceptions induced by the diversity workshop for identification with a superordinate group and collective action intention.
Another limitation lies in the fact that the concept of diversity was mentioned only in the diversity groups.Future research should investigate whether the observed effect is more than mere conceptual priming.Although we did not find a statistically significant effect of the diversity training on identification with the UN, tendencies are consistent with our hypotheses.Additionally, correlational analyses yielded a positive relation between diversity perception and identification with the UN.The workshop was constructed in such a way that a positive organizational climate regarding diversity was induced in order to not reduce identification.Nevertheless, it cannot be tested empirically whether the positive organizational climate towards diversity led to a higher level of identification.To further strengthen this assumption, diversity beliefs (van Knippenberg et al., 2007) or the successful implementation of diversity norms should be included as potential moderator variables.
The present study also replicated the finding that the higher the identification the higher the collective action intention.Regression analyses yielded a significant effect of identification with the UN on collective action intentions in response to global issues.This underlines the importance of identity management in the international sphere.
However, mediator and dependent variables were assessed concurrently, which does not permit causal inferences.
Therefore, further investigation on the mediators is needed.Moreover, we only measured collective action intention as a proximal predictor for actual action (e.g.Ajzen & Fishbein, 1977).Future research should include actual indicators of behavior.In the Model United Nations context this could be the final resolution participants compile at the end of the simulation as delegates that serves as a draft law for the countries.It could also be beneficial to add attitude measures.
Another drawback of our pattern of findings is that the diversity training workshop did not have a direct effect on collective action intention.Thus, while no direct evidence of Hypothesis 3 was obtained, the indirect effect, mediated by diversity perception and identification, was established.A methodological explanation for this pattern could be that there are additional variables accounting for the relation between mediator and outcome that are not due to the experimental manipulation (see MacKinnon & Pirlott, 2015).Alternatively, the diversity training may have had unknown negative side effects (see Hayes, 2013, for methodological discussion).Future research is needed to determine which negative effects these could be.One assumption could be that threat plays a role, a process identified in previous research on the effects of diversity interventions on outgroup attitudes (e.g.Ehrke & Steffens, 2016).Possibly, our intention to buffer identity threat with a self-affirmation induction (Cohen & Sherman, 2014) was not successful.Self-affirmation has been shown to be an effective strategy against psychological threat.It is based on the assumption that people include several roles in their self-concept.Sherman and Cohen (2006) state that "people can respond to threats using the indirect psychological adaptation of affirming alternative selfresources unrelated to the provoking threat" (p.190).If social identity (here: national identity) is threatened by the diversity induction, reassuring positive personal characteristics independent of concepts being related to the negotiation or UN situation can serve as a buffer.That we did not find a direct effect in the mediation analysis could imply that the buffer did not work as intended.Future research should therefore investigate whether the self-affirmation technique used here empirically serves as an effective tool to buffer threat in diversity training or if there are other options to achieve that (e.g.see limitation for identity threat buffer for advantaged group members, Shnabel, Purdie-Vaughns, Cook, Garcia, & Cohen, 2013).Consequently, the workshop modules should be tested individually in order to detect their specific effects.Although every module was designed on the basis of empirical research, their effects on the indirect ingroup projection measurement, identification, and collective action intention cannot be assessed separately.Until now, laboratory research induced diversity simply by instructing participants to think about the diversity of a superordinate category (e.g., Waldzus et al., 2003).Thus, it has to be tested separately whether the positive induction of diversity, as done in the present study, has the same effects, as we did not measure whether the diversity training induced a positive evaluation of diversity.In addition, the focus on the superordinate category could be further strengthened.Besides the options used in our workshop (like verbal emphasis and showing the UN flag on several occasions), the UN flag could be added next to the name of the country on participants' table plaques.
Another extension for the future development for a diversity training to foster collective action intention could be to investigate its effects on injustice perceptions and efficacy beliefs, which are theorized as the other two predictors of collective action in addition to social identification (van Zomeren et al., 2008).For example, regarding injustice beliefs, based on literature showing that relative prototypicality is associated with a higher perceived legitimacy of social inequality (Reese et al., 2012), one could assume that a method reducing perceived relative prototypicality of one's own country decreases perceived legitimacy of social inequality.This could lead to an increase in perceptions of injustice.
Due to the very difficult data collection process and as fewer participants showed up than had previously registered, we had to collect data on several occasions.The control group consisted of participants from randomized control workshops and non-intervention samples.Consequently, the effect of the diversity training is contrasted against several different operationalizations of the control group.

Practical Implications
In addition to this discussion of theory and methodological limitations, we want to highlight that this study focused on (future) leaders' responsibility for collective action concerning human rights issues.But even political leaders' having established a new ambitious treaty in November 2015 does not necessarily affect the behavioral intentions of the general public.Consequently workshops like the one presented could be adapted to be included in broader education and training settings, so that potential change for supporting human rights issues is directed at a broader audience.Alternatively, based on psychological research, van der Linden, Maibach and Leiserowitz (2015) point out five policy recommendations that should increase public engagement: 1) emphasizing climate change as a present, local and personal risk; 2) delivering less abstract or numerical information but more experiential engagement; 3) promoting pro-environmental behavior within a community to establish group norms; 4) shifting policy conversation from potentially negative consequences in the future of not acting to positive benefits of immediate action in the present; and 5) focusing on intrinsic motivations.Our research could add with cautious first evidence a recommendation of increasing diversity beliefs for the superordinate category to affect collective action intention.These recommendations could be included in campaigns, aimed at fostering public awareness of a given issue.We hope this study encourages other researchers to adapt their knowledge to real-world applications regarding human rights.
Finally, we should mention the potential of the chosen research setting.A Google Scholar search yielded hardly any scientific peer-reviewed articles investigating Model United Nations as a research setting.Only a few articles investigated single MUN in spheres of educational relevance (McIntosh, 2001).The potential of this setting has therefore not yet been fully acknowledged.MUN could serve as an economic tool for investigating peace processes within negotiation settings.Furthermore, there is no systematic analysis and evaluation of the results of the MUN conferences -for example in terms of MUN conferences possibly informing real UN negotiators.Nevertheless it has to be stated that MUN conferences are only simulations by students or young academics aspiring to hold a position in international relations in the future.For this reason behavioural intentions, as shown in the simulations, may not transfer to real interactions in international negotiation settings.

Conclusion
This quasi-experimental study attempted to apply psychological research on intergroup relations and collective action intentions to the sphere of the United Nations.It provides evidence that a diversity workshop decreased ingroup projection compared to a control condition.Furthermore, the diversity workshop had an indirect effect on collective action intention mediated by perceived diversity and identification with the UN.This is the first study to investigate the effect of a diversity training workshop on perceived diversity, identification with the superordinate category, and collective action intention in an applied setting.Additionally, it is the first study combining Model United Nations with psychological theories of intergroup relations.As negotiation settings within the political areas have major implications for people around the world, methods for reducing biases in perception seem especially useful.With reference to the research questions posed in the introduction, based on our first evidence, one could cautiously assume that diversity perceptions and identification with a superordinate group influence and foster collective action intention.Policies simultaneously aiming at diversity and identity management could pave a fruitful avenue improving intergroup relations and collective action in general.Although the United Nations are already advocating a pro-diversity culture, this study supports the assumption that it should be integrated not only in public policies but also in leaders' training and negotiation settings.

Figure 1 .
Figure 1.Illustration of research design.

b
Control 3 and Workshop 3 scores of collective action are based on the mean of climate action items (items 1-5 in the Appendix), as the conference simulated the United Nations Environment Programme (UNEP).

Figure 2 .
Figure 2. Mediation model with unstandardized values.Analysis of the hypothesized indirect effect of the training on collective action via diversity perception and identification with the United Nations.Coefficients of dotted paths are non-significant.

Table 1
Summary of Main Results and Demographic Composition of Subgroups in the Two Conditions, all at Measurement Point 2The total number of participants does not always match the indicated number of each gender, due to missing data regarding gender. a

Table 2
Correlations of Main Variables in the Two Conditions (Measurement Point 2)Note.Below the diagonal are values for the control condition (n = 42), above the diagonal are values for the diversity condition (n = 50).

Table 3
Effect of Diversity Training on Outcomes at Measurement Point 2All variables were measured on scales ranging from 1-7, with high numbers indicating higher levels of the construct.Means are shown with standard deviation in parentheses.