Moral Values That Thwart Intergroup Interactions: an Investigation on the Interaction Between Indonesian Moslems and Chinese-Indonesian Christians

As one of the most established theories, the contact hypothesis has been well-researched throughout decades of investigations. However, there have been few attempts to investigate individual factors that may influence interaction processes that may lower prejudice. The present study attempts to find the individual factors that can moderate the contact – prejudice effect, that is, individual moral values. Previous researches have noted that individuals with high moral loyalty, authority, and sanctity may resist interacting with outgroups. Consequently, these individuals may possess higher prejudice. Thus, we hypothesize that individuals with higher levels of those three moral values may experience the contact effect more profoundly, in which there is stronger contact – prejudice effect. 594 Moslem participants participated in the online survey we administered. We found that moral authority and purity can moderate the contact – prejudice effect, consistent with our hypotheses. These patterns were found only for the contact – subtle prejudice effect. However, moral loyalty cannot moderate this effect. We discuss the implications by examining the Indonesian current sociopolitical conditions and how the three moral values influence the dynamics of intergroup contact.


Introduction
Ask: "Can we befriend non-Moslems? Is there any decree from Al-Qur'an and Hadits?" Answer: "To befriend non-Moslems is Haram (forbidden)" -This conversation was taken from the questions and answers section of http://al-atsariyyah.com/ and was accessed at December 23, 2016 These days, such narratives are exceptionally uncommon. Google Searches using the keywords "do not get along with kafirs (Non-Moslems)" resulted in many similar statements (see almanhaj.or.id and muslim.or.id website attached in the references, for example). In fact, while this article was being written, Indonesia witnessed the emergence of a scholar (psychologist) who refused to accept money containing the image of Non-Moslem figures (Batubara,21 Desember 2016 on news.detik.com). As a part of Indonesian historical context, a never ending tension between Moslem and Christian individuals has persisted. Under these conditions, prejudice tends to flourish (Allport, 1954).
If there are certain opportunities to interact with the outgroups, prejudice can be reduced. Several decades of research and numerous studies in the umbrella of contact hypothesis have confirmed that contact between groups can indeed improve intergroup attitudes and lower prejudice (Pettigrew & Tropp, 2006). In fact, contact itself not only reduces the blatant prejudice of a person, but also reduces implicit prejudice (Hamberger & Hewstone, 1997). The effectiveness of contact in reducing prejudice has been demonstrated through meta-analysis of 515 studies (Pettigrew & Tropp, 2006). However, this does not mean that the theory is free from criticism. One critic highlighted the lack of representation of individual differences that might influence group dynamics when contact occurred (Vorauer, 2006: Hodson, Costello, & MacInnis, 2013. So far, most research has focused on objective factors that improve contact effectiveness, without using subjective responses and analyzing individual differences across the occurrence of contact (Pettigrew & Tropp, 2011). In its development, various individual differences factors have been known to increase or even inhibit the effectiveness of contact in reducing prejudice. Factors such as right wing authoritarianism (Dhont & Van Hiel, 2009;Hodson, Harry, & Mitchell, 2008), social dominance orientation (Hodson, 2008), and biological identity (Hodson, Harry, & Mitchell, 2008) have gained empirical support.
In this study, we argue that there are other individual factors that may improve or hinder the effectiveness of contact. Research by Van Leeuwen and Park (2009) finds that people with moral values that emphasize loyalty to ingroup members, obedience to authority within ingroups, as well as the perseverance of the sanctity of ingroups tend to perceive outgroups as a source of danger. Through the terminology of moral foundation theory, such individuals tend to hold the binding moral values (characterized by the moral domain of loyalty, obedience, and sanctity) (Haidt, 2012). Thus, individuals with higher levels of these moral domains may be anxious about interacting with outgroup members since outgroup members may pose certain threats to their ingroup. Consequently, they avoid contact with outgroups (Plant & Devine, 2003;Crisp & Turner, 2009). When contact indeed happens, these individuals may be profoundly affected because of lowered anxiety, which is a common mediator for the contact -prejudice effect (Pettigrew & Tropp, 2008), especially when they are ready to experience direct contact (Dovidio, Eller, & Hewstone, 2011).
Ergo, we attempt to answer the main question: "Can the contact -prejudice effect be moderated by the three moral values of moral loyalty, moral sanctity, and moral authority?". More specifically, "Will contact lower prejudice more significantly when individuals have higher levels of moral loyalty, moral sanctity, and moral authority?" Prejudice. What exactly is prejudice? In his monumental work 'The Nature of Prejudice', Gordon Allport offers a definition that is widely used in social psychology research. He defined prejudice as "... antipathy based on faulty and inflexible generalization. It may be directed toward a group as a whole or toward an individual because he is a group member" (Allport, 1954 p. 9). In other words, prejudice is a negative attitude that is felt or expressed towards a person of a group (or the group itself).
Prejudice can be often distinguished into two forms, namely blatant/explicit prejudice and subtle/implicit prejudice (Wittenbrik, Judd, & Park, 1997). Explicit prejudice is defined as the perceived threat from interacting with an outgroup while implicit prejudice can be considered as perseverance of traditional ingroup values regarded as very different from outgroup values (Gaertner & Dovidio, 1986). While the latter tends to be more socially acceptable, the former is often scrutinized by social norms. Although both are considered distinct, it may be that both types of prejudice are actually a single entity (Hamberger & Hewstone, 1997). Nevertheless, it is still important to differentiate the two in research because while explicit prejudice is often absent, implicit bias remains present (Dovidio, Kawakami & Gaertner, 2002).

Contact and Prejudice.
There is a condition that is often utilized as a panacea for hatred against outgroups, which is the contact hypothesis. This hypothesis was basically developed by Gordon Allport in 1954. According to the contact hypothesis, negative attitudes toward outgroups can be reduced when individuals in groups interact with individuals between groups. This is considered a powerful way to reduce intergroup conflict (Pettigrew & Tropp, 2011).
It is believed by Allport (1954) that contact is able to achieve its optimum effect when there are four conditions. First, there needs to be equality of status between the ingroup and outgroup. Second, ingroups and outgroups must have a common goal. Third, there must be institutional support such as legality, tradition, norms, and others. Finally, there is also a need for cooperation between ingroups and outgroups. Research proves that having these four structured conditions does increase the effect of contact in lowering prejudice when compared to unstructured conditions. However, these conditions often act as facilitators rather than necessary prerequisites (Pettigrew & Tropp, 2006). Since Allport, the contact theory has stimulated many empirical studies. In 2006, Pettigrew and Tropp (2006) conducted a meta-analysis of 515 studies that examined contact and prejudice relationships. It was found that 94 percent of the studies supported the assumption of contact theory. Moreover, the contact theory can be generalized to a wide variety of situations (neighbors, schools, workplaces, etc.), different research designs, as well as varying age groups. There is a reason why contact is considered a powerful method to reduce prejudice across different contexts.
Hypothesis 1: Individuals who are in contact with other group members tend to have lower prejudices. In other words, higher implicit or explicit prejudice is predicted by less contact.
Through another meta-analysis, Pettigrew and Tropp (2008) added that the crucial mechanisms or mediators in the effect of contact on prejudice are: 1. Knowledge about outgroup, 2. Anxiety towards outgroup, and 3. Empathy towards outgroup. For the first mediator, prejudice decreases as the individual has knowledge of other groups (Allan & Johnson, 2008;Stephan & Stephan, 1984). However, the mediation effect of this knowledge tends to be weak compared to the other two mediators (Pettigrew & Tropp, 2008) and unreliable (Gries, Crowson, & Cai, 2011). Meanwhile, anxiety (Page-Gould, Mendoza-Denton, Turner, Hewstone, & Voci, 2007;Blair, Park, & Bachelor, 2003) and empathy towards outside members of the group (Pettigrew & Tropp, 2011;Batson et al., 1997) can be considered consistent mediators. Thus, prejudice may be reduced when contact occurs, because contact can reduce anxiety and increase empathy to those outgroups whom they interact with. As an important mechanism, reducing anxiety is actually very important to explain how contact reduces prejudice. Those who experience contact tend to have their anxiety lowered which consequently reduces prejudice.
However, the contact theory is not free from criticism. One of the potential flaws to take note of is whether or not individuals enjoy the contact they experience. A recent study by Mallett et al. (2016) concluded that cross-group interaction tends to be less favored compared to the same-group interactions. This means that individuals tend to avoid contact with outgroups whenever possible and this might lead them to experience more prejudice. In one of their experiments, participants were given PDAs to record cross-group and fellow interactions with at least 10 minutes of interaction each. Participants reported positive and negative emotions experienced as soon as the interactions occurred. It was found that cross-ethnic interactions caused less positive emotions when compared to the interactions between same-ethnic groups. Individuals tend to enjoy more interaction with their own group than outgroups. In other words, people tend to seek less interaction with outgroups than with ingroups.
In addition, according to the theory of integrated threats (Stephan & Stephan, 2000), outgroups often give signals of danger for ingroups and therefore ingroup members avoid them. When anxiety is felt by individuals who interact, they are prone to fear and nervousness (Whitley & Kite, 2010). Contact full of anxiety and prejudice does not seem to be an effective form of contact. In this study, we argue that there are individual factors that may facilitate this anxiety. The factors that we emphasize here are the three moral values of loyalty to ingroup, respect for ingroup's authority, and purity of ingroup.
The Role of Moral Loyalty, Moral Sanctity, and Moral Authority as Moderator. We suggest that moral values which emphasize social order and cohesion (Graham et al., 2009) tend to benefit ingroups alone rather than intergroup relationships. This is because these moral values may strengthen tendencies of selfsacrifice for one's ingroup, obedience to authority within a group, and the cleanliness of contamination or social sins (Lewis & Bates, 2011). According to the moral foundation theory, these are the values held by conservative political groups, which often emphasize the need for stability and order as opposed to progress and change (Haidt, 2012;Lewis & Bates, 2011;Graham, Haidt, & Nosek, 2009). Because they hold such moral values, people with conservative world views tend to resist the things that threaten social stability and security (Haidt, 2012), including threats from outgroups (Van Leeuwen & Park, 2009). Here, the moral values consist of loyalty, authority, and sanctity (Lewis & Bates, 2011).
Basically, the loyalty foundation determines whether a person likes those who are loyal to his group and reject those who betray him. Evolutionarily, this foundation is the value that is responsible for stability in society. After all, groups with individuals that are loyal may survive better compared to those who are disloyal. Despite that function, this foundation is also responsible for wars and assassinations between groups (Haidt, 2012). If we recall the theory of integrative threats described, loyalty goes hand in hand with ingroup favoritism (assuming ingroup superiority) where this condition triggers prejudice and hatred towards outgroups (Whitley & Kite, 2010). However, a research by Marylin Brewer (2007) indicated that the fondness or loyalty towards outgroups may predict ingroup favoritism more compared to the outgroup derogation. Thus, while loyalty has the potential to create prejudiced attitude toward outgroups, it is also possible that loyalty may not influence the intergroup relations.
The authority foundation arises from the need to maintain order in society. The foundation determines whether a person likes those who are obedient and respectful to the authorities or hates those who rebel and disrupt. Authority is an important element in society. Without authority, anyone would be able to teach advanced statistics or quantum physics and everyone would believe them (Haidt, 2012). But obedience to authority can sometimes exert a socially undesirable effect, as demonstrated by the Stanley Milgram experiment (Milgram, 1963). In his famous laboratory experiment, participants were instructed to press a button that triggered an electric shock to a victim. Participants were notified beforehand that electric shocks could endanger the lives of victims. But about 60 percent of participants still obeyed orders to push the button. This shows that there are times when obedience to authority precisely guides individuals to harm others.
The sanctity foundation is actually more complicated. This moral foundation determines whether we like people who are physically and socially clean or reject those who are contaminated and sinful (Haidt, 2012). Physical and social contamination may vary from bodily stains, HIV / AIDS, penetrated vaginas (or non-virgins), unhealthy food, cannibalism, homosexuality, and many more. Evolutionarily, this foundation was manifested from the disgust aroused by poisonous food (Haidt, 2012).
However, social contaminations like homosexuality or individuals that are regarded as impure may also ignite the same emotional response. At first glance it appears that the foundation of sanctity is not related to intergroup relations. But imagine the relationship between religious fundamentalists and homosexual groups. The values of religious purity in fundamentalist groups may trigger disgust towards homosexuals (Koleva et al, 2012, Haidt, 2012.
Hypothesis 2: Individuals in contact with other group members tend to have lower prejudice. This effect tends to be stronger for those with higher moral Loyalty.
Hypothesis 3: Individuals in contact with other group members tend to have lower prejudice. This effect tends to be stronger for those with higher moral Authority.
Hypothesis 4: Individuals in contact with other group members tend to have lower prejudice. This effect tends to be stronger for those with higher moral Sanctity.
What is the rationale behind these effects? There are two mechanisms that we propose. First, individuals with such values tend to feel that outgroups are a source of danger that threatens ingroup survivability. In this mechanism, the anxiety felt towards outgroups can obstruct intergroup relationships (Crisp & Turner, 2013). Those with such moral values tend to believe that the world is not a safe place. They believe that there are dangers around, one of which comes from foreign groups or outgroups (Van Leeuwen & Park, 2009). For the second mechanism, threats toward outgroups may activate a process called morality shifting. The threat will divert the morality of caring and fairness into moralities that emphasize loyalty and adherence to ingroups. A study by Leidner and Castano (2012) proves that when ingroups are threatened, their morality also tends to shift from behaving with care and fairness toward an outgroup into only behaving with loyalty and obedience toward one's ingroup.

Methods
To answer the research questions, we adhere to the NHST (Null Hypothesis Statistical Testing) paradigm. This paradigm attempt to test whether the hypothesis is acceptable when tested over and over again (usually 95% of the time). Therefore, it tries to reject the null hypothesis or hypothesis where there is no relationship and effect between variables (usually in the 5% probability level). All measuring instruments were translated back-to-back and have been tested for psychometric indexes to fulfill the cross-cultural requirements (Beaton, 2000).

Research Participants.
Generally, the effect of contact on prejudice should be experienced by various groups with the assumption of occurring conflicts or history of conflicts between ingroups and outgroups (Allport, 1954). Therefore, the tests should be possible in a variety of intergroup relational contexts. Nevertheless, the researchers focused this research on prejudice against Chinese Indonesian Christians because the issue is quite salient to the current Indonesian context. The prejudices against ethnic Chinese Indonesians were the political consequences of the Soeharto era that proclaimed the differences between indigenous and Chinese communities (Freedman, 2003;Turner & Allen, 2007). In addition, the prejudice against ethnic Chinese Indonesian Christians is quite salient through the issue of blasphemy accusations towards the Chinese Christian governor of Jakarta in 2017. Meanwhile, prejudice against Christian individuals in Indonesia itself is actually considered to be strong as annual data has reported (see KBB's annual report The Wahid Institute).
Thus, prejudice against Chinese Indonesian Christians became the focus of this research. The researchers managed to successfully collect respondents with the main criteria of "Moslem Indonesian citizens at least 17 years of age". The age was chosen because prejudice tends to be more influenced by social factors (rather than biological and maturational factors) in late adolescence or early adulthood (Raabe & Beelman, 2011). 594 Moslems participated in this study (N Women = 385, 64.8%) with the age range of 17 to 39 years (Mean Age = 20.34, Standard Deviation = 2.38). Of all these participants, the majority were senior high school graduates (N = 466 or 78.5%). Only 17 people were junior high school graduates (N = 17 or 2.9%). Since the majority of data was obtained from the LINE account of official student pages, it can be assumed that the majority of participants were college students, and the rest had graduated with bachelor degrees. In accordance with this demography, the majority of participants had an income range of Rp. 3,000,000 or less (N = 539 or 90.7%). Only 9 participants claimed to have more than Rp. 12,000,000 (N = 9 or 1.5%). The rest had an income of Rp. 3,000,000 to Rp. 9,000,000 (N = 46 or 7.8%).
Measures. Explicit and implicit prejudice against Chinese Indonesian Christians was measured using blatant and subtle prejudice scales from Hamberger and Hewstone (1997). This instrument consists of 16 items where 8 items measure blatant/explicit prejudice while 8 others measure subtle/implicit prejudice. The examples of items that measure explicit prejudice are: "I do not mind if a well-qualified Chinese Indonesian Christian is appointed as my leader" and "Most Chinese Indonesian Christians should be able to live without government assistance." Examples for items that measure implicit prejudice include "Many other groups in Indonesia can live in harmony to be accepted in Indonesia. Chinese Indonesian Christians should be able to do the same without the help of certain parties" and "Chinese Indonesian Christians living in Indonesia teach their children values or skills which are different to the needs of Indonesia ". Participants responded using a Likert scale of 1 to 6 (1 = very unfavorable while 6 = very favorable). The higher the score, the higher the explicit and implicit prejudices. The reliability index for the explicit prejudice measure is Cronbach-Alpha = 0.79 while the reliability index for the implicit prejudice tool is Cronbach-Alpha = 0.68.
Contact with Chinese Indonesian Christians was measured by a three-item contact measure adapted from Miller, Smith, and Mackie (2004). This instrument consists of the following items: 1. "How many Chinese Indonesian Christians have spoken to you more than twice during the last 6 months?" (Likert scale response with 1 = '0 to 3 people' and 6 = 'More than 20 people'), 2. "On average, how close are you to the Chinese Indonesian Christians you know?" (Likert scale response with 1 = 'not close at all' and 6 = 'Extremely close'), And 3. "Which one best describes your relationship with the Chinese Christians?" (Participants answered based on the choice of answers shown in figure 1). The reliability of this instrument is Cronbach-Alpha = 0.70.

Figure 1. Indonesian version of Item 3 for Contact Measure ("Saya" = "Me", "Mereka" = "Them")
Individual Moral Values were measured using the Moral Foundation Sacredness Scale (MFSS) originally created by Graham and Haidt (2012). This measuring instrument was adapted and now consists of 9 items that measure the three moral domains of loyalty, sanctity, and authority. In the scenario where participants would be paid to behave in a certain way (without negative consequences and guaranteed confidentiality), participants were asked to write down how much money they were willing to accept at minimum amount (within the range of IDR 0 to IDR 10,000,000). Participants were also given the choice of not wanting to behave in a certain way even though they would be paid to do so. An example of a scenario for the authority moral value is: "Show impolite hand movements against your boss, your professor, or your teacher". An example for the loyalty moral value is: "Saying bad things about Indonesia in a conversation on foreign radio". Finally, an example for the sanctity moral value is: "Having plastic surgery to change your genitals (you can change again 1 week later)". Before participants fill the form, participants were primed with the memories of the desired items that they can only buy with money. This was aimed at making participants aware of the value of money.
Social desirability was measured through the Social Desirability Scale that Widiarso adapted in Bahasa Indonesia from the Marlowe-Crowne Social Desirability Scale (Widhiarso, 2012). This measuring instrument is considered feasible for use in Indonesian culture. Examples of the items are: "I never really hate someone". Participants responded on a Likert scale (1 = 'very unsuitable' and 6 = 'very appropriate'). The higher the total score on this scale, the higher the tendencies of people to respond according to social propriety. Lastly, the demographic data taken was gender, age, religion, prior education, and income per month.

Mahalanobis Distance Analysis and Descriptive
Statistics. The Mahalanobis Distance Analysis was performed to detect outliers (a highly unlikely response or odd response) in the overall distribution of the data. This was done to clean the deviant responses. With Chi-Square (df = 4) > 18.47, p <.001, we eliminated a total of 32 samples because they were considered to have multivariate response patterns (4 variables) which were highly unlikely to occur. Thus, the total number of participants included was 562. For the independent variable, we obtained the mean score (M) = 9.56 (SD = 3.42). Meanwhile, we obtained the mean for the blatant prejudice with the score of M = 22.84 (SD = 6.23) and subtle prejudice with the score of M = 28.12 (SD = 3.89). Finally, we obtained mean scores as follows: Loyalty M = 23.48 (SD = 1.19), Authority M = 21.46 (SD = 3.34), and Sanctity M = 20.72 (SD = 3.74).
Main Effect. The Pearson correlation results indicated that the higher the contact with the Chinese Indonesian Christians, the lower the blatant prejudice against the them (N = 562, r = -0.331, p <.001). The linear regression test was performed with blatant prejudice as outcome and contact as predictor. In accordance with our first hypothesis, it appears that the decrease in blatant prejudice against Chinese Indonesian Christians is predicted by an increase in contact (b = -0.33, t (561) = -8.29, p <.001). The proportion of variance in blatant prejudice against Chinese Indonesian Christian is significantly explained by the contact score.
Meanwhile, when tested on subtle prejudice as the dependent variable, we also found results consistent with our first hypothesis. The decrease in scores of subtle prejudice was predicted by an increase in the contact scores (b = -0.15, t (561) = -3.57, p <.001) where the predictor variance explained variance in outcome significantly. The correlation index between the two variables is r(562) = -0.15, p <.001. Although both are significant at LOS 99.99%, the effect of contact on subtle prejudice is weaker compared to blatant prejudice.

Moral Loyalty as Moderator.
To test the second hypothesis, we conducted the Hayes Process SPSS (Hayes, 2013) that tests the interaction between IV and moderators in explaining the DV. Moderation testing on the loyalty domain showed no interaction between contact with that value in predicting blatant prejudice (F (3.558) = 1.76, p = .18, = 0.03, b = -0.11, t (561) = -1.33, p =. 18). The Johnson-Neyman test showed that the effect is only observed when the sample size is 6.4% below the reference value of deviation and 93.6% above the deviation reference. When tested on subtle prejudices, similar results were found (F (3.558) = 1.15, p = .28, = 0.002, b = 0.05, t (561) = 1.07, p = .28). The Johnson-Neyman test showed that the effect occurs when the sample size is 2.1% below the deviation reference and 97.9% above the deviation reference. Further details can be seen in table 1.
Moral Authority as Moderator. The test results on the moral authority domain as a moderator are significant towards subtle prejudices but not towards blatant prejudice. Testing the effect of contact interaction and authority on blatant prejudice obtained the result of F (3.558) = 1.01, p = .32, = 0.002, b = -0.02, t (561) = -1.01, p = .32. Johnson-Newman analysis showed that significance is obtained when the sample is in deviation value <3.4% and> 96.6%. Meanwhile, testing of contact interaction effect and authority on subtle prejudice obtained F (3.558) = 9.22, p <.01, = 0.013, b = -0.04, t (561) = -3.04, p <.01. This suggests that the effect of contact on subtle prejudices is weakened when individuals have high authority values (b = -0.27, p <.0001) in comparison to when individuals have low authority values (b = -0.03, p = .60). This supports the claim on Hypothesis 3. The whole regression equation can be seen in Table 1.
However, for DV subtle prejudices, statistical significance was obtained when the sample was in deviation values <35.1% and> 64.9%. This suggests that moderation effects occur when the upper limit of low Sanctity moral values is below or above the 35.1% sample size (and not just +1 SD / -1 SD). Thus, our fourth hypothesis is confirmed.

Discussion
The results show that contact predicts prejudice, meaning that the more Moslem individuals interact with Chinese Indonesian Christians, the lower their prejudice towards them (Hypothesis 1 is confirmed). The effect persists when we take into account the different types of prejudice (subtle and blatant). Moreover, the effect of contact to subtle prejudice is moderated by the moral values of authority (Hypothesis 3 is confirmed) and sanctity (Hypothesis 4 is confirmed). However, we do not find the same pattern for loyalty as a moderator (Hypothesis 2 is not confirmed). In this section, we will discuss the implications of these findings.
In accordance with Pettigrew and Tropp's meta-analysis (2006), contact has indeed been shown to reduce prejudice. Additionally, in accordance with Hamberger and Hewstone's (1997) research, contact not only affects explicit prejudice (blatant) but also implicit (subtle) prejudice. Thus, explicit prejudice which involves feeling threatened by Chinese Indonesian Christian groups tends to be lower when people of different groups (Moslem groups) have adequately interacted with that outgroup. Other than that, implicit prejudice such as attempts to discriminate between Chinese Indonesian Christian groups and indigenous groups of Indonesian society (Moslems Pribumi or Indigeneous) is also affected.
However, it should be noted that the effect of contact towards implicit or subtle prejudices tends to be weaker than to more blatant prejudice. This suggests that there are distinct characteristics between implicit prejudice and explicit prejudice. Since explicit prejudice may have similar dimensions with contact, the results might be attributed to the overlapping dimensions between explicit prejudice and contact. The results of this research seemed to be consistent with previous research from Hamberger and Hewstone (1997) in which they found that explicit prejudice contains components like: "avoid intimate contact with other groups". Moreover, it should be noted that the reliability index for the implicit prejudice tool is lower than that of explicit prejudice.
It could be that this is not just the result of inadequacy of our measuring instruments alone. It may be that contact is not capable of completely separating ingroups (IG) from outgroups (OG). Contact only helps to empathize, better understand, or not feel anxious about OG (Pettigrew & Tropp, 2008) which is exactly what explicit prejudice means. However, contact does not really separate IG with OG in which IG may feel that there are fundamental differences between IG and OG and thus can never be united. This is the core of implicit prejudice. This seems to be associated with the finding that strong identification of IG is able to moderate the effects of prejudice because they are more likely to feel intense anxiety when contact occurs (Tausch, Tam, Hewstone, Kenworthy, & Cairns, 2007). This intense anxiety will inhibit the effect of contact on prejudice, let alone prejudice regarding fundamental differences between IG and OG. In the context of the Indigenous Islamic and Chinese Christian relations in Indonesia, the rise of the religious blasphemy case by the governor of Jakarta in 2016 shows that there are Muslim ideological groups that emphasize differences with Christians, leading to not selecting non-religious leaders (see Sinaga, 28 February 2017). This makes people who strongly identify with Islamic groups adopt rigid views on OG, no matter how often they interact with them. With this in mind, we suggest that future studies use a scale of explicit prejudice measurement with higher reliability values. In addition, future studies should address the Moslem ideological group members characteristics to better understand the context in which this phenomenon applies.
For the Authority moral value, it was found that the value moderated the effect of contact on implicit prejudice but not on explicit prejudice. People who possess high moral authority values tend to feel stronger contact -implicit prejudice effect. Conversely, people who have lower moral authority score tend not to feel the effects of contact -implicit prejudice effect (consistent with Hypothesis 3). From this result it can be interpreted that individuals with higher moral authority tend to be affected more by interaction with other group members because they do not interact early with OG due to obedience to authority (Haidt, 2012). In accordance with previous assumptions, people with authority moral values tend to be more obedient to their group's authorities and norms, including the norms that are detrimental to IG and OG relationships. Milgram has proved in his experiments that blind obedience to authority can lead individuals to harm others. It is not impossible, in the context of prejudice against the Chinese Indonesian Christians, that certain authorities within the IG voiced negative prejudices against Non-Moslem groups and that this was obeyed by the people within the IG. This is more likely to happen when they have strong moral authority (Haidt, 2012). Future studies should address how the authority might impose their prejudice on group members that possess high obedience value.
But why is this effect strong only on implicit prejudice and not on explicit prejudice? Explicit or blatant prejudice may not really be socially normative while implicit prejudice is much more normative and socially acceptable (Hamberger & Hewstone, 1997). In other words, blatant prejudice may not be perpetrated by individuals simply because there are other stronger norms, namely, norms of politeness and compassion. However, subtle prejudice may be freely expressed. It is possible that individuals have their blatant prejudice lowered by Indonesian norms of politeness. But it does not necessarily lower subtle prejudice. When contact happens, prejudice is more profoundly affected. Thus, the moderating effect is stronger for implicit prejudice, in which individuals in IG feel that OG will never be one with IG.
As for moral sanctity, it appears that this moral value can moderate the contact -prejudice effect but only for implicit prejudice. The result might be attributed to the nature of intergroup relationships between Islam and Christianity. One might claim their religion as holier while the opposing religion is regarded as impure. Thus, the idea of 'oneness' between Moslem and Christian individuals can never sound very convincing. However, this moral value cannot moderate the contact -blatant prejudice effect.
It appears that loyalty does not moderate the effect of contact on both types of prejudice. The result might be attributed to how loyalty might not affect OG derogation because loyalty to the group is less related to attitudes to OG. Research by Marylinn Brewer shows that IG's favoritism, which is the basis of individual loyalty to his group, is independent of OG derogation. Apart from that, the research also concluded that the loyalty and sense of ownership of the group were better able to predict IG's favoritism but were unable to predict attitudes toward OG (Brewer, 2007).

Does contact affect prejudice or vice versa?
This research cannot answer with certainty because the research was done cross-sectionally and did not try to discover the cause-effect relationships. Therefore, further studies are expected to provide answers to assumptions about causal relationships through experiments on the same issue. In addition, the reliability of the measuring instrument needs to be properly addressed to ensure a more accountable outcome.
The other limitation is about data retrieval online, where researchers do not know whether participants really do not respond more than once. Also the researchers can't really enforce this, we hope that Mahalanobis distance analysis is sufficient to eliminate outliers. There is also the possibility of common method bias in which the results appear not due to actual representation of reality but because of the method of measurement. Consensus in psychology lately agrees that the use of measuring instruments of self-report should be accompanied also by other measurement methods. This is absent in this study, and further studies should consider this.

Conclusion
The present study suggests that those who possess authority and sanctity value tend to exert more negative attitudes towards outgroups especially when they have never experienced any contact with such outgroups. In other words, the contact -prejudice effect is moderated by the moral values of authority and sanctity. Future studies should address why this moderating effect happens. Also, future studies should address the deeper dynamics which happen in various groups, because Islam consists of many different subgroups.