Group Norms as Moderator in the Effect of Cross Group Friendship on Outgroup Attitude: a Study on Interreligious Group in Indonesia

Past studies indicate that the effect of intergroup contact on outgroup attitude is not isolated to contextual factors. One of the contextual factors that has begun to be studied is group norm. However, group norm in these studies is still merely conceptualized as the perception of how ingroup members evaluate outgroup members. In fact, according to norm focus theory, in a given context, individuals are influenced, at least, by two types of group norms, namely injunctive norms (i.e., what most people morally accept to do) and descriptive norms (i.e., what most people do). To fill the gap, present studies attempt to answer the question of how two types of group norms might have different effects on the relationship of intergroup contact and outgroup attitude. Built on past studies, it was hypothesized that both quality (H1) and quantity (H2) of cross-group friendship would positively affect outgroup attitude. Further, built on the fact that the nature of attitude in present studies is more utilitarian than hedonic, it was predicted that injunctive norms would be more likely to function as moderator in the effect of cross-group friendship on outgroup attitude, either in dimensions of quality (H3) or quantity (H4). 110 Muslim students were recruited as participants and asked to fill in a self-report questionnaire regarding their interactions with Christians. The findings partly support the hypotheses.


Introduction
One of the most challenging issues to face in current diverse society is how to build a harmonious intergroup relationship. This is also the case for Indonesia, a Moslem majority country with the most populous Moslem population in the world, but at the same time has "Pancasila" ideology (i.e., the official ideological foundation of Indonesia as a country, with its very first principle as "Believe in the one and only God"). This ideology underlines that Indonesia is not an Islamic country and consequently requires all religions to respect each other in practicing their beliefs. However, a national survey recorded, in 2013, that there were 245 cases of intolerance in Indonesia (The Wahid Institute, 2014). This included sealing worship houses, coercing beliefs to other believers, discrimination, etc. The tensions were largely found between Moslems and Christians. The question is: How can such conflict be resolved? In current studies, this question was addressed by focusing on the role of cross-group friendship and contextual factors.
Over 60 years, considerable research in social psychology have been conducted to find possible ways to reduce, resolve, and prevent intergroup conflicts. A notable contribution initially came from Gordon W. Allport (1954) with his contact hypothesis. He maintained that a more harmonious intergroup relation could be established by contact between members of different groups under the right conditions. In this sense, Allport defined direct intergroup contact as cross group friendship which involved intimacy within the relationship. Some studies have supported the important of cross group friendship on outgtroup attitude in different settings, such as in Catholic-Protestant in Northern Ireland (Paolini, Hewstone, Cairns, & Voci, 2004;Hewstone, Cairns, Voci, Hamberger, & Niens 2006) and in school (Barlow, Louis, & Hewstone, 2009;Beelmann & Heinemann, 2014;Tropp, Hawi, van Laar, & Levin, 2012).
Despite evolving, the development of the intergroup contact theory and attitude is still criticized. One of the main criticisms is the view that this theory originally used simplistic explanations of the effect of intergroup contact on prejudice reduction (Ata, Bastian, & Lusher, 2009;Christ et al., 2014). This criticism points out that the effects of contact on intergroup attitudes are not only influenced by factors of with whom one interacts, but also contextual factors.
To understand the role of contextual factors, Christ et al. (2014) conducted a series of studies comparing intergroup contact experiences of people who live in diverse and less diverse areas (high vs. low diversity). The study shows that those who live in diverse areas, where there are greater chances to perceive its people positively interact with out-group members, will be less likely to have prejudice. This study suggests that despite lack of direct intergroup contact, positive out-group attitude could be enhanced when individuals perceive that their neighborhoods have positive intergroup contact. Christ et al. attribute their findings to the fact that diverse social environment would provide knowledge to individuals that in-groups members make contact with outgroup members. Consequently, individuals perceive that the intergroup contact is normative and supported by the ingroup members (i.e., form positive social norms). When intergroup harmony is normative within a society, in other words there are a supportive ingroup norms, it would provide more space for individuals to engage in various type of intergroup contact (see Pettigrew, 1998 for a review).
The studies of group norms were originally developed within social identity studies. In this perspective, group norms are defined as unwritten rules that guide ingroup member behavior. The basic idea is that when individuals identify to a group and feel that being a member of that group is important to their self-conceptualization, they will bring their behavior in accordance with the perceived norms and standards of the groups (Smith & Louis, 2009). Research has supported the idea of referent group norms could influence how people behave toward an object (e.g., Norman, Clark, & Walker, 2005;Terry & Hogg, 1996;White, Terry, & Hoog, 1994). The referent groups could vary, depending on the context. For example, behavior in office is shaped by the perception of what coworkers expect to do and with peers by the perception of what peers expect to do (Louis, Davies, Smith, & Terry, 2007).
Some studies have attempted to understand the effect of societal norms on the relationship of intergroup contact and outgroup attitude. They generally test two hypotheses with different directions: (1) group norms moderate the effect of intergroup contact on outgroup attitude (cf., Cook, 1984;Pettigrew, 1998;Tropp & Prenovost, 2008), and (2) intergroup contact moderate the effects of group norms on outgroup norms (Pettigrew & Tropp, 2006). In the former hypothesis, group norms serve information about whether intergroup contacts are acceptable or not. When people perceive it is acceptable or even encouragable, people then would be more likely to take benefits from the intergroup relationship. When people perceive that their ingroup members disapprove intergroup relationship, the intergroup contact's experience would lead to intergroup anxiety as the fear of being criticized by fellow ingroup members for doing something unacceptable in the ingroup (Stephan & Stephan, 1992). In the latter hypothesis, it is presumed that group norms would affect positive outgroup attitude only when people are involved in positive intergroup contact (Christ et al., 2014;Hewstone et al., 2008;see Schmid, Tausch, Hewstone, Hughes, Jenkins, & Cairn, 2008 for review).
In the present study's context, most Indonesians have interacted with religious outgroups since they have lived in a diverse environment. Nonetheless, in some groups, such religious intergroup contact is disapproved. In Indonesian cases, this is largely found in the relationship between Muslims and Christians. One of the main reasons might be because of the fear of Christianization that has been long rooted since 1967, when many Muslim leaders explicitly defense to social and political expansions of Christian and foreigners (Mujiburrahman, 2006). Based on this context, the present study seeks to answer the question of whether the effect of intergroup contact on outgroup attitude among Indonesian would be moderated by religious group norms. Although the role of group norms has been quite largely investigated in the study of intergroup contact and outgroup attitude, none of them have been studied in Indonesian contexts. Further, the measurement of group norms remains focused on only one type of group norms. Christ et al. (2014), for example, measures ingroup norms as ingroup members' support or opposition towards prejudice or tolerance. Similarly, in other studies, ingroup norms were measured by asking individuals about their perception on how happy their ingroup friends interact with outgroup members (e.g., Turner, Hewstone, Voci, & Vonokafau, 2008) or how positive their attitude is in general (Mahonen, Jasinskaja-Lahti, & Liebkind, 2011). Taken together, those studies examined ingroup norms as individuals perception about how ingroup members treat outgroup members.
In fact, according to a norm focus theory, Cialdini, Reno, and Kallgren (1990) suggested that in a given situation, people are influenced, at least, by two types of group norms: injunctive (i.e., what people should do) and descriptive norms (what people actually do). This separation would allow a better understanding on the influence of group norms on human's behavior. Injunctive norms motivate behavior with social rewards and punishments associated with the behavior. Descriptive norms motivate behavior by providing information about effective or adaptive behavior in a given context.
One theory that could explain the divergent findings is the norm focus theory by Cialdini, Reno, and Kallgren (1990) which suggests that individuals' response to group norms is goal directed. When individuals want to behave effectively, they might refer to the descriptive norms. Meanwhile, when individuals want to build and maintain social relationships, they would be more likely to focus on injunctive norms. In the case of corruption, for example, individuals would be more likely to cultivate information that corrupt behavior is feasible. When they observe that many people perform that behavior, it must be then feasible to do. Alternatively, as suggested by Manning (2009) in a meta-analysis study, the relative power of injunctive and descriptive norms depend on the utility of behavior. He found that the relationship of descriptive norms and behavior was stronger than injunctive norms and behavior, but only when the behavior is hedonic rather than utilitarian. While the hedonic behavior means the the behavioral engagement is intended to fulfil a short term or pleasure need, utilitarian behavior requires more thought and has useful functions.
Built on the theories of intergroup contact and group norms, the main question of the present study is to understand how the two types of group norms (i.e., injunctive and descriptive norms) can moderate the effect of cross-group friendship and outgroup attitude. It is predicted that both quality (H1) and quantity (H2) of cross-group friendship positively affect outgroup attitude. Further, built on the categorizations of behavior in terms of their utility, it could be known that intergroup related behaviors are more utilitarian rather than hedonic. Therefore, it is also predicted that injunctive norms would be more likely to function as moderators in the effect of cross-group friendship on outgroup attitude, either in the dimensions of quality (H3) or quantity (H4).
By conducting the present study, at least two contributions could be given. First, it highlights the need to extend the study of intergroup contact in non-Western countries. There are only few studies outside of Western contexts, for example interreligious relationship among Muslim and Hindu in Bangladesh (Islam & Hewstone, 1993), in India (Tausch, Hewstone, & Roy 2009), and interethnic relationship in Malaysia (Al Ramiah & Hewstone, 2011;Al Ramiah, Hewstone, Little, & Lang, 2013). In Indonesia, in particular, there are only two studies, namely in the context of Muslim majority and Christian minority's relationship in Ambon and Jogjakarta (Kanas, Scheepers, dan Sterkens 2015), and impact evaluation of intergroup contact's intervention during post-conflict reconciliation in Aceh (Schiller, 2012). Second, it would be the first study that attempts to understand the moderating role of group norms on the relationship of cross-group friendship and outgroup attitude. Finally, it would extend the understanding on the role of two types of group norms, in particular within the intergroup contact's study.

Methods
Participant and Procedure. 110 university Muslim students were recruited as participants (Male = 40, Female =70, nineteen years old = 50, twenty years old = 60). Seventeen of the total participants were involved in a religious organization. Participants were asked to fill in a set of questionnaires in a class with the guidance from a research assistant.
Measures Predictor variables. (Cross-group friendship was measured by asking participants to report quantity and quality of contact with their Christian friends. The measures were adapted from past research (e.g., Turner et al., 2008;Tausch, Hewstone, Kenworthy, Cairns, & Christ, 2007) by measuring quantitative and qualitative dimensions of contact. Quantity of contact was measured by four items, such as: "How often do you spend time with your Christian friends?" (response scale: from 1 = Never to 5 = Always) and "How many Christian friends do you have? (response scale from 1 = None to 5 = more than 10). The items formed a realiable scale (α = 0.713). Higher values indicate more contact. Quality of contact was measured by two items asking participants how pleasant (1 = very unpleasant to 6 = very pleasant) and positive (1 = very negative to 6 = very positive) they perceive contact's experiences with Christian. The items were highly correlated (r = 0.742, p < 0.001).
Higher values indicate more qualified contact.
Moderators. Group Norms. Injunctive norms were examined by two questions, adapted from Smith and Louis (2008): "Do you think your Moslem friends support every Moslem to make a friendship with Christians?", "Do you think your Moslem friends support you to have a good relationship with Christians?" (response scale: 1 = very unsupporting to 6 = very supporting). The items were significantly correlated (r = 0.37, p < 0.01). Descriptive norms items were also developed in accordance with the  guideline on group norms measurements. There were two questions employed to examine descriptive norms: How many your Moslem friends that you think have a good contact with their Christian friends? (1 = None to 6 = Most). The items were significantly correlated (r = 0.35, p < 0.01).
Criterion Variable. Outgroup Attitude. On six semantic differential scales, based on the General Evaluation Scale (Wright, Aron, McLaughlin-Volpe, & Ropp, 1997), participants reported their general evaluation toward Christians using bipolar adjectives, such as: warm-cold, and friendly-unfriendly. The items formed a reliable scale (α = 0.834). Higher values reflected more favorable attitude towards Christians.
Covariate Variable. Gender was entered as control variable (1 = Female, 2 = Male). In addition, given that the participants in this study were first year students which were at late adolescence developmental stage, parents approval was also included in control variable.
Past studies indicate adolescence still rely on parent's attitude in developing their own outgroup attitude moral judgment via the degree of intimacy individuals feel in the relationship (Edmonds & Killen, 2009;Smetana & Asquith, 1994).  Table 1 shows descriptive statistics of these variables and intercorrelation between them. As presented by the table, preliminary results from the correlational analysis provide initial evidence for the idea that quantity and quality of contact is related to injunctive norms and descriptive norms. These preliminary results also suggest that intergroup contacts and group norms were related to outgroup attitude.

Predicting Outgroup Attitude.
To conduct multiple regression analysis, Hayes PROCESS macro was employed (Model 2 to test the effect of two moderators simultaneously, see Hayes, 2013). By using this program, it also allowed estimation of the bias-corrected coefficients from a series of 5000 bootstrap samples (see Preacher, Rucker, & Hayes, 2007). Before conducting the analysis, all variables values were first mean centered. To test interaction effects, two interaction terms were formed by multiplying the values with the mean centered variables.
Based on past studies that suggest that quality and quantity of intergroup contact have different effects on outgroup attitude (Islam & Hewstone, 1993;  (1) mean centered of quality of contact, injunctive norms, and the interaction terms (quality of contact x injunctive norms) were entered; (2) mean centered of quality of contact, decriptive norms, and the two interaction terms (quality of contact x descriptive norms) were entered; model 2: (1) mean centered of quantity of contact, injunctive norms, and the interaction terms were entered (quantity of contact x injunctive norms); (2) mean centered of quantity of contact descriptive norms, and the interaction terms were entered (quantity of contact x descriptive norms).
In all model, gender and parents approval were entered as covariate variables.
As shown in the  . This model explained 36% of variance in outgroup attitude. These analysis supports Hypotheses 1 and 2 that suggest that either quality or quantity of cross-group friendship have significant positive effects on outgroup attitude. Given that the beta weights of quality cross-group friendship were greater than the quantity, this indicates that the effects of quality dimension were stronger than the quantity dimension.
Moderated regression analysis indicated that hypothesis 3 was rejected. That is, neither injunctive norms nor descriptive norm functioned as a moderator for the effect of quality of cross-group friendship on outgroup attitude ( As presented in Table 3, the conditional effect analysis revealed that the effect of a good quantity of contact on positive outgroup attitude was consistently positive and statiscally significant when the level of supportive injunctive norm were moderate and high among the low, moderately, and highly supportive descriptive norm. The effect were larger when the injunctive norm was highly supportive (b = 0.89, p = 0.001; b = 0.70, p < 0.001; b = 0.51, p = 0.002) than moderate (b = 0.54, p = 0.003; b = 0.36, p = 0.005; b = 0.18, p = 0.366) across the three levels of support in descriptive norm. The interaction plot can be seen in figTure 1.
General Discussion. Previous studies that attempted to understand the relationship between intergroup contact, group norms, and outgroup attitude mainly tested two possible hypotheses: (1) group norms function as moderators in the effect of intergroup contact on outgroup attitude, (2) intergroup contact moderate the effect of group norms on outgroup attitude. The present study re-tested the first hypothesis, mainly due to the fact that intergroup contact has been supported and even promoted by the government in the current study context. However, it does not necessarily lead to a more (2) injunctive norms, but not descriptive norms, would function as moderators in the effect of quality and quantity of contact on outgroup attitude The analysis revealed that some of hypotheses were supported.
Supporting hypothesis 1 and 2, this study revealed that quality and quantity of cross-group friendship had a positive significant effect on outgroup attitude. It indicated that the more positive and the more frequent cross-group friendship, the more positive individuals' evaluation toward outgroup would be. However, the effect weight was found larger in quality dimension of cross-group friendship than in the quantity dimension. This finding was in line with past studies (e.g., Islam & Hewstone, 1993;Stangor et al., 1996;Turoy-Smith, Kane, Pedersen, 2013) that suggest that the outgroup attitude is more strongly associated with quality of intergroup contact rather than the amount of actual intergroup contact. It might be attributable to the fact that frequency does not necessarily lead to a positive intergroup contact. As criticized by Hewstone (2009)  The analysis of moderator effect revealed the prediction that injunctive norms would moderate the effect of quality dimension of contact on outgroup attitude (hypothesis 3) was partially rejected. The analysis showed that positive contact would lead to a more positive outgroup attitude regardless of the content of the group norms (i.e., injunctive and descriptive norms). Meanwhile, in line with hypothesis 4, the effect of quantity of contact on outgroup attitude would depend on the injunctive norms.

Figure 1. Simple Slope Analysis
When the injunctive norms supported the intergroup relationship, the frequency of contact would positively increase the outgroup attitude. These findings indicate two things. First, as discussed earlier, the quality dimension of contact has relative power to the quantity dimension on outgroup attitude. Therefore, the positive evaluation of intergroup contact's experiences are more important than the frequency of the contact. Second, this study showed the importance of injunctive norms over descriptive norms in facilitating the effect of quantity of contact on outgroup attitude.
The importance of injunctive norms over descriptive norms have been found in another context, such as in the context of environmental behavior (Cialdini et al., 2006;Reno et al., 1993;Smith & Louis, 2009;Schultz et al., 2007), pro-social behavior (Raihani & McAuliffe, 2014), compliance to authority expectation (Savani et al., 2012). One main theory that could explain the relative power of injunctive norm is a norm focus theory by Cialdini et al., (1990Cialdini et al., ( , p 1015) that suggested that "although they are said to characterize and guide behavior within a society, they should not be seen as uniformly in force at all times and in all situation". In five series of experimental studies, they found that the norms that would be more influential were the salience norms in a particular context. In the case of intergroup relationship, where there is a tone of moral judgment in it, individuals' attention would be more likely to focus on what other ingroup members expect. This is because individuals will be more likely to be driven by the goal of building social relationship, to get social approval, rather than to behave effectively. This is in line with previous studies on prejudice and social norms: evaluation toward outgrouop is mainly developed based on information about appropriate attitude and behavior that are shared within a group (Crandall & Schaller, 2005;Stangor, Scherist, & Jost, 2001).

Future Directions, Limitations and Implications.
Although the present study has provided strong support for the effect of quality and quantity of intergroup contact on outgroup attitude and the moderating role of injunctive norms, in particular regarding the effect of quantity dimension, some limitations must be acknowledged. First, the present study was cross-sectional. Therefore, a selection bias, namely the effects of group norms on outgroup attitude that are moderated by intergroup contact, cannot be ruled out. As suggested by past studies, further analysis should be conducted on alternative relationships. Second, all the participants were university students, thus the findings might be not strong enough to generalize a wider population. However, previous studies indicated that the effect of intergroup contact among university students and general population are very similar (see Paolini et al., 2004). Third, it should be noted that the context of the present study allows individuals to frequently interact with outgroups. Consequently, findings in the present study may not be fully generalizable to other contexts where formal or informal support for intergroup contact do not exist. For future studies, it would be important to examine whether the moderator effects of group norms are found in such contexts.
Despite its limitations, the present study may have some policy implications to lessen intergroup conflicts in Indonesia. Although the government strongly supports intergroup interactions and norms of tolerance, as set in Pancasila ideology, the government still needs to look at how this ideology has been transmitted within society. It means that although there might be positive institutional injunctive norms, this does not necessarily create positive subgroup injunctive norms. The present study suggests that injunctive norms are important to facilitate the effect of quantity of intergroup contact, which was found to have a less strong effect than quality of contact. Therefore, it is important for society and institutions to establish positive injunctive norms, for example, by clarifying student's perception about what ingroup members expect them to do in terms of intergroup relationships, from less supportive to being more positively supportive. Further, built on the finding that quantity of contact had less strong effects on outgroup attitude than the quality dimension of intergroup contact, institution needs to consider how a more positive intergroup contact experiences could be achieved. Past studies show that it can be done, for example, through school activities that invite students from different groups to work together and rely each other (Tropp & Prenovost, 2008)

Conclusions
This study has shown evidence that contact with outgroup religious members would increase positive outgroup attitude. In an attempt to understand the relationship, present study has extended past studies by indicating that the effects of quantity dimension of intergroup contact on outgroup attitude would vary as a function of group norms (i.e., injunctive norms). However, the effect of quality of contact would be less likely to rely on group norms. These findings highlight the important group norms in the context where people have more opportunities to engage in intergroup contact but not necessarily a positive intergroup contact.