Prosocial Intentions towards Religious Ingroup and Outgroup Members among Adolescents from Public and Religious Schools

. Prosocial behaviour is a very important component in harmonious social relationships, especially within a diverse community such as Indonesia. Public and religion-based schools provide different opportunities for students in terms of interaction with peers from different religions, and in turn may affect attitudes toward ingroup and outgroup religious members. This study compared prosocial intentions towards religion-based ingroup and outgroup peer members in adolescents from heterogeneous (public) schools and homogenous (Islamic) schools. Participants were 220 Muslim middle school students aged 11-15 years, who attended public schools (N= 130; M=48, F=82) and religious schools (N= 90; M=46, F=44). They were recruited through snowball sampling and convenience sampling. This research is a quasi-experimental study, by using the Scenario-based Prosocial Intention Questionnaire (SBPIQ) to measure prosocial intentions, which was modified to include manipulations on the targets of prosocial intention. Two-way mixed ANOVA results showed that the level of prosocial intention was highest toward the ingroup target (M=3,411; SD=0,392), followed by those toward the outgroup (M=3,357; SD=0,403) and neutral targets (M=3,234; SD=0,411). Furthermore, prosocial intentions towards ingroup, outgroup, and neutral targets were all higher in students from public schools. These results implied that there were biases in students’ prosocial behaviors toward ingroup/outgroup religious members. It is crucial for both public and Islamic schools to facilitate positive interactions with members of other religious group in order to overcome these biases.


INTRODUCTION
Statistics Indonesia calculated that Indonesia is the place for a population of 268,074,600 people (BPS, 2020) and consists of 1,340 ethnic groups, 2,500 languages, and acknowledges six major religions (BPS, 2010). The existing diversity will lead to the development of diverse traditions, cultural values, and customs (Afandi, 2018), which in turn becomes a distinctive feature and advantage for the nation. However, this diversity may trigger various potential conflicts that endanger the unity (Lestari, 2016).
One of the diversities that frequently arises as a motive behind conflicts in Indonesia is religious diversity. The results of a survey conducted by the Indonesian Survey Circle (LSI) and Denny JA Foundation found that in the course of 1998-2012, out of 2398 violence cases, 65% were due to religious conflicts (Asril, 2012). The results of the LSI survey in 2019 found that more than 50% of the community rejected non-Muslims in government as well as in everyday life (LSI, 2019), indicating a large potential for horizontal conflict among religious believers.
Religious conflicts and intolerance do not only occur in the public context but also in the school setting. Research results from the Center for the Study of Islam and Society (PPIM) UIN Syarif Hidayatullah with 1859 teenage participants showed that 43.88% of them supported discrimination against adherents of other religions (Suyanto, 2021). Casuistically, it was found a bullying case where Public Elementary School (SDN) student became a victim of bullying by his friends due to different beliefs. These friends told him that he had to convert if he wanted to play together (Shelavie, 2017).
In Indonesia, a strong religion-related social identity often leads to hostility and social division. This is very unfortunate considering that history has written that the birth and growth of Indonesia have happened thanks to its unity and integrity. During the independence era, for instance, interfaith communities worked together to win independence. After independence, Indonesia's progress was achieved, among other things, because of the strong cooperation and mutual assistance in the society, including among religious believers. This concept of helping each other in society, within psychology can be approached through the concept of prosocial behavior.
Prosocial behavior can be defined as any activity that provides benefits and merits to others or is aimed to establish harmonious relationships with others (Jackson & Tisak, 2001). Prosocial behaviors develop since childhood. In the adolescent phase, prosocial behavior becomes one of the developmental tasks that must be fulfilled (Eisenberg et al., 2015). It is because, during this phase, a person must internalize values and norms, including those related to social relations. Furthermore, social skills in the adolescent phase are a determinant of personal quality at later developmental stages (Henera et al., 2017).
This study looks upon the prosocial behavior of adolescents in relation to religious diversity in educational settings, by examining the differences between prosocial behavior in public and religious schools. Public school in this study is represented by state schools, while religious school is represented by Islamic school for the reason that Islamic schools constitute the largest religious-based schools in Indonesia. According to statistical data from the Ministry of Education and Culture, 17% of schools in Indonesia are Islamic schools (Kemendikbud, 2018). Between both types of school, it is assumed that there are different opportunities for interaction with interfaith students (Jannati, 2018) which are perceived to influence prosocial behavior. Research from Monash University, UIN Walisongo, and Gadjah Mada University in 2017 found that Islamic schools that appeared to be exclusive would increase student discriminatory behavior towards members of different religions (Mutohar, 2018).
Given the different interaction opportunities with religious outgroups (members of different religious groups) in public and religious schools as aforementioned, it becomes important to examine whether there is a distinction in prosocial behavior towards members of the same religion (ingroup) and members of different religions (outgroup) among students in both school settings. The dynamics of inter-religious cooperation among adolescents based on school background is essential to study as an effort to maintain unity and prevent social conflicts in Indonesia.

Prosocial behavior
Prosocial behavior is defined as intrinsically motivated voluntary behavior to benefit others or establish harmonious relationships with others (Eisenberg & Mussen, 1989;Jackson & Tisak, e-ISSN: 2541-450X p-ISSN: 0854-2880 Indigenous: Jurnal Ilmiah Psikologi 2021, 6(3), 73-91 Azzahra & Ampuni | 75 2001). Jackson and Tisak (2001) categorize prosocial behavior into four types of behavior, which are helping, sharing, cooperating, and comforting. Helping is defined as attempts to assist in overcoming the difficulties or shortcomings of others. Helping may occur in emergencies such as when rescuing accident victims; and in nonemergency situations such as when helping mother washing dishes. Sharing is defined as providing personal resources to share with others, such as giving some of own snacks to classmates. Cooperating is an effort to perform a job with others to achieve a shared goal, such as doing class cleaning collectively. Meanwhile, comforting is actions taken to boost others' moods, such as comforting a sad friend (Jackson & Tisak, 2001).
In this study, prosocial behavior was measured through its intentions. Intention is a desire to continuously perform certain behaviors (Ramdhani, 2011). Intention is one of the closest predictors in foreseeing one's behavior (Balau & Madalina, 2018). The stronger one's intention, the greater the tendency to perform a behavior is (Ajzen, 1991). Giandi et al. (2020) in their research found that there was a significant relationship between intention and behavior. Thus, prosocial intentions are expected to represent prosocial behaviors that were not possible to be measured directly, such as through observation, in this study.

Adolescent prosocial behavior
Prosocial behavior begins to emerge in childhood and continues to develop until the older adult stage (Eisenberg et al., 2015;Rosi et al., 2019). During adolescence, development is marked by transitions in various aspects, one of which is related to prosocial behavior ( Van der Graaff et al., 2018;van Hoorn et al., 2016). In the adolescent phase, individuals begin to cultivate concern for others ( Van de Groep et al., 2020). The target of prosocial behavior expands beyond the family environment, into school environment, peers, and the wider society (Bruner et al., 2014).

The concept of social identity and the emergence of ingroup-outgroup perception
In discussing the definition of ingroup and outgroup, we need to first discuss about social identity. Naturally, an individual tends to identify himself with a group, which then results in social identity (Everett et al., 2015). According to Social Identity Theory, social identity is one of the aspects of "self-image" coming from certain social categories that are considered part of them (Tajfel & Turner, 2019). For example, if one considers himself being a part of certain religious group, then he has the social identity of being the member of that religious group. Social identity motivates individuals to differentiate their group from other groups (Everett et al., 2015), leading to the perception of "we" or ingroup and "they" or outgroup (Ho & Yeung, 2019).
Different attitudes towards ingroup and outgroup members will be more obvious when social identity comes to the surface or becomes salient (Ho & Yeung, 2019). According to the concept of depersonalization, when social identity is activated, individuals tend to behave based on group values or social identity rather than individual values, which leads to benefiting own group (Stets & Burke, 2000).
The social identity of an individual can strengthen or weaken. Social identity will be strengthened when an interaction among members of ingroup is happening (Nezlek & Smith, 2005), when there is a competition between groups, and when there is an assumption that outgroup is a threat (Pamungkas, 2015). Meanwhile, social identity will weaken when other social identities within an individual are more superior or dominant (Hogg & Dominic, 1998a) and when the person does not perceive other groups as a threat (Pamungkas, 2015).

The difference between public and religious schools
In Indonesia, there are several types of educational institutions or schools. This study focuses on public and religious schools. There are fundamental differences between the two school backgrounds. In public schools, religious education and religious practices are not the main focus of learning (Yuniarrahmah & Rachmah, 2016). In religious schools, for instance Islamic schools, religious learning and practices are given a larger portion. Religion is the value adhered to and becomes the main focus of learning in schools (Gustiyana, 2019).
With regard to interactions, there are more opportunities for interaction between members of different religious groups in public schools, whether in or outside learning activities (such as extracurricular activities, campsites, and field trips) (Jannati, 2018). Whereas in religious schools, the opportunities for interaction between member of different religions are very limited or even absent (Jannati, 2018). This is because the majority of religious schools indicate that the religion of students must be in accordance to the school's identity (Baidhawy et al., 2014).
Religious education and practices taught in schools, as well as close interactions among students with a shared religion within the school, will strengthen the social identity of students, which in this context is the religion-based social (Hamburg & Hamburg, n.d.). This may cause adolescents from religious schools to have a strong religion-based social identity compared to students from public schools. (Gustiyana, 2018). On the other side, interactions among students of different religions in public schools are assumed to create a shared social identity as a larger group such as "school identity" or even "national identity". According to Hogg and Abrams (1998a) social identity at a higher/general level may suppress the other social identity at a lower/narrower level, such as that as a member of a particular religion.

Relationship between school background and prosocial intention toward ingroup-outgroup members
Prosocial behavior is affected by many factors, including the school environment (Eisenberg & Mussen, 1989). The school environment plays a role in the interaction process of various groups (Jannati, 2018). A school environment that facilitates intergroup cooperation may increase positive attitudes towards other group members, resulting in acceptance of outgroup (Hogg & Abrams, 1998b). This corresponds to the research conducted by (Scacco & Warren, 2018) where members of heterogeneous classes that facilitated interactions between groups showed more helpful and cooperative attitudes and behaviors towards outgroups than members of homogeneous classes.
Public schools and religious schools (Islamic schools) provide different opportunities for interactions with outgroup members, in this case members of other religions. in public schools, students come from diverse religions, and religious identity is not really prominent. Meanwhile, in religious schools, students come from homogeneous religion, therefore interaction with interfaith students is limited, and religious identity is highlighted. Such differences in the characteristics of these two school settings then provoke the question of whether there are differences in prosocial behavior towards members of the same religion (ingroup) and members of different religions (outgroup) among students from public and religious schools.
Several previous studies have examined prosocial intentions towards ingroup and outgroup in children and adolescents, in which the intentions were found to be higher towards ingroup than outgroup targets (Everett et al., 2015;Hackel et al., 2017;Preston & Ritter, 2013). However, the studies only involved children and adolescent participants in general, and have not considered their school background, especially religion-related backgrounds. Furthermore, research on the social interactions between students of different religions, which focuses on adolescents in religious e-ISSN: 2541-450X p-ISSN: 0854-2880 Indigenous: Jurnal Ilmiah Psikologi 2021, 6(3), 73-91 schools, is still rarely carried out in Indonesia, This study compares prosocial intentions towards ingroup and outgroup targets among adolescents who are students of public schools and religious schools. The results of this study are expected to contribute to preserving the spirit of cooperation and tolerance, especially in interfaith interactions among adolescents. Moreover, this research is expected to contribute knowledge that can be applied as a basis for educational interventions related to religious tolerance in public and religious schools.
Two research questions were proposed. The first question was whether there are differences in prosocial intentions towards ingroup and outgroup targets in general. It was hypothesized that regardless of school background, prosocial intention towards ingroup members will be higher than prosocial intention towards outgroup members. The level of prosocial intention towards neutral targets will stand between the levels of prosocial intentions towards ingroup and outgroup members (Hypothesis 1). The second question was whether school background influences prosocial intention towards ingroups and outgroups. The second hypothesis was as follows: Prosocial intention towards ingroups will be higher in adolescents from religious school (Hypothesis 2a); prosocial intention towards outgroup will be higher in adolescents from public school (Hypothesis 2b), while prosocial intention towards neutral targets will not significantly differ between the two groups (Hypothesis 2c).  Table 1 summarizes the participant descriptions. Participants in this study included male and female adolescents aged 11-15 years from the provinces of East Java, Central Java, and the Special Region of Yogyakarta (DIY), coming from Public Junior High Schools (SMPN) and Integrated Islamic Junior High Schools (SMPIT). In order to control participants' Islamic identity and interactions with outgroups, they were required to have the corresponding elementary school background, meaning that subjects from public junior high school must be graduated from public elementary school, and vice versa.

Participants
The recruitment of participants from public schools was performed using snowball sampling method, in which the researchers contacted key-persons -which were students from various junior high schools researchers are acquainted with. The key persons were asked for their consent to become participants and to help recruit the next participants by passing research information through online conversation media. The next participants also passed the information to other students or groups, and so on. Meanwhile, the sampling technique used to recruit participants from religious schools was convenience sampling. We asked a teacher from an SMPIT to pass the research information to students. Even though the information was distributed by the teacher, we ensured to inform that student participation in the research was voluntary, thus students had the right to choose to participate or otherwise. Data collection was performed online via the SurveyMonkey platform.
Informed consent containing research information was displayed at the beginning of the online survey. Since the research involved adolescents (minors), a checklist of parental consent was added. This checklist was to be fill out by parents to state their agreement for their children to participate in the study. Participants received rewards in the form of a credit of ten thousand rupiah electronic money or snacks for those who filled out the questionnaire completely. Electronic money was given to participants from Public Junior High Schools since participants came from schools throughout three provinces. Meanwhile, Islamic Junior High School participants were rewarded with snacks because they came from the same school, thus facilitating the distribution. All procedures described above have been approved by the Ethics Committee of the Faculty of Psychology, Universitas Gadjah Mada (UGM).

Research design and procedures
The independent variables in this study was the school backgrounds represented by public schools (heterogeneous schools) and Islamic schools (homogeneous schools). Meanwhile, the dependent variable was prosocial intention including (1) prosocial intention towards ingroup target, namely members of the same religion, (2) prosocial intention towards outgroup target, namely members of different religions, and (3) prosocial intentions towards neutral target, namely targets without specified religious identity.
This study employed a quasi-experimental quantitative method with a within-subject and one-group post-test-only design. Within-subject method (repeated measures) is a condition in which research subjects receive research treatment more than once (Charness et al., 2012). Each subject will have their prosocial intention measured on three targets, which are ingroup, outgroup, and neutral, after receiving treatment in the form of names of targets of prosocial behaviors in the story/scenario (explained further in the Treatment and Instrument section). The one-group posttest-only design method is a measurement carried out after treatment and without a control group (Knapp, 2016) (Figure 1).

Treatment and instrument
To measure prosocial intentions towards ingroup, outgroup, and neutral targets, we used a modified Scenario-based Prosocial Intention Questionnaire (SBPIQ). This measuring tool was first developed by Buwono (2018), then developed further by Ampuni and Buwono (in press). The difference between the modified SBPIQ and the original version is that the scenarios in the modified SBPIQ was adapted to the research objectives by assigning names of characters representing ingroup, outgroup, and neutral religious identities.
The treatment or manipulation in this experiment was attached to the research instrument by manipulating the names of the target of prosocial behaviors in the SBPIQ items. The prosocial targets from ingroup members were marked by names of figures or characters which indicate that they are Muslim, for instance, Azizah and Ahmad. The prosocial outgroup targets were signified by names of characters that imply a religion other than Islam (Christian or Catholic), for example, Maria and Antonius. We only used Christian and Catholic names for the outgroup figures because these two religions are more familiar given that they are the second and third major religions embraced after Islam. The names used in this research instrument has been selected through a preliminary study (Ampuni & Buwono, in press). For the neutral prosocial targets, the scenarios were not assigned with name of a certain religious character or identity. For an example of the scenarios and items, see Figure 2.
The scenario in this example ( Figure 2) can be translated as follows: Today, you and your group would do a group work on History after school in the hall room. Your group was assigned to create short drama about Indonesian Independence history. The group distribution was decided by teacher and your group consisted of you, Budi, Nicholas, and Andreas. You are heading to hall room which was quite far from your class in the second floor. While heading to the hall room, Nicholas who just recovered from illness looked in trouble carrying his big and heavy bag. Would you help Nicholas to carry his bag? (This scenario represents Helping aspect for outgroup targets, because Nicholas is a Christian name) . intentions towards ingroup (4 scenarios), outgroup (4 scenarios), and neutral targets (4 scenarios). We added 2 additional dummy scenarios containing stories that are not related to prosocial behavior (e.g., there was a power outage situation and the subject is asked if he will play outside for not being able to play electronic games). The dummy scenarios were included in the measuring instrument to conceal the research objectives, and were excluded from the analysis. To avoid gender bias in filling out the questionnaire, the names of the characters in the scenarios were adjusted to the gender of the participants; male participants received a questionnaire with male character names, and female participants received a questionnaire with female character names. To facilitate the distribution of the questionnaire based on gender, the cover of the questionnaire for male participants was illustrated with a picture of a boy, and vice versa. Each scenario was followed by four sub-scenarios that tapped the aspects of helping, sharing, cooperating, and comforting. Following each sub-scenario, a question item to measure intention was attached, in which participants were asked to choose among four alternatives that indicate their prosocial intentions, namely "no" (score = 1), "less likely" (score = 2), "more likely" (score = 3), and "very likely" (score = 4). A preliminary study involving 69 adolescents (N = 69; Female = 41; Boy = 28) from public and religious schools has been conducted previously to ensure that the items measuring prosocial behaviors for the three different targets have balanced weights of sacrifice needed to perform.
This scale in the original and modified versions has been used several times and tested for reliability. The reliability test on the original version conducted by Buwono (2018)  Before measuring prosocial intentions, priming was carried out to activate the subject's religious identity. A priming can activate the memory representation and individual self-concept, and may lead the person being primed to perform the behaviors in accordance to the representations being primed (Batara et al., 2016).
The priming used in this study consists of four short questions related to religious identity and activities, accompanied by choices of answer in the form of pictures of religious identities from six religious groups, namely Islam, Christianity/Catholicism, Hinduism, Buddhism, and Confucianism. It is expected that by responding to these questions, the participant's religious identity will become salient. Thus, they will respond to the SBPIQ while their social identity as a Muslim is awakened. The priming questions presented include: (1) My place of worship is...; (2) I worship by..; (3) When I celebrate my religious holidays, I usually see people wearing ..; (4) When I celebrate my religious holidays, I usually see… .An example of priming question is presented in Figure 3.

Data analysis
The main analysis in this study was to investigate the differences in prosocial intentions towards ingroup, outgroup, and neutral targets among adolescents from public and religious school backgrounds. The analysis was conducted through a two-way (3 (prosocial intention target: ingroup, outgroup, and neutral) x 2 (school background: public and religious)) mixed ANOVA using Statistical Package for Social Science (SPSS) version 23. Prosocial intention was assigned as within-subject component, while school background was assigned as between-subject component.

Data Description
Prosocial intention scores were obtained from the prosocial intention scale, with each prosocial target (ingroup, outgroup, neutral) encompasses 16 items. Each item has a score that ranges from 1-4, with a maximum hypothetical score of 64 and a minimum hypothetical score of 16. The data is presented in Table 2.

Assumption test
Before performing the main analysis using a two-way mixed ANOVA, assumptions including normality (Table 3) and homogeneity (Table 4) were tested. Normality test was administered using Shapiro-Wilk (Ghasemi & Zahediasl, 2012). As seen in Table 2, some of the data are normally distributed and some are not normally distributed. However, data that are not normally distributed can be ignored due to the large sample size (more than 100 samples) (Mishra et al., 2019). Homogeneity test was carried out using Levene's test (Sharma & Kibria, 2013). Based on the results of Levene's test (See Table 3), the variance of prosocial intention scores for ingroup and neutral targets deviated significantly from homogeneity (not homogenous), while the score for prosocial intentions towards outgroup was homogeneous. Isnawan (2020) states that in a quasiexperimental design, the analysis can still be conducted regardless of the homogeneity of the data.

Hypothesis testing
Two-way mixed ANOVA indicates that there was no interaction effect between prosocial intentions x school background, F(2.1932) = 0.056, p = 0.94. The main effect of prosocial intentions (within-subject) was found to be significant, F(2.1932) = 83.90, p < 0.001. These findings imply that there were significant differences in the levels of prosocial intention towards ingroup, outgroup, and neutral targets within each individuals (Table 5). Subsequently, Bonferroni's post hoc analysis was carried out (see Table 5.) to find general prosocial intentions (regardless of the school background) towards ingroup, outgroup, and neutral  The finding that prosocial intentions towards ingroup targets were higher than prosocial intentions towards outgroup targets corresponds to the hypothesis, but the finding that prosocial intentions towards neutral targets was the lowest is not in line with the hypothesis (Figure 4). Between-subject analysis (Hypothesis 2a) showed that based on type of school, prosocial intention towards ingroups were significantly higher in adolescents from public schools than those from Islamic schools (ΔM = 0.139; SD = -0.094). This finding was not consistent with the hypothesis where prosocial intention towards ingroups was assumed to be higher among adolescents from the Islamic junior high school.
With regard to prosocial intentions towards the outgroup target (Hypothesis 2b), prosocial intention of adolescents from public schools were significantly higher than that from Islamic school (ΔM = 0.136; SD = -0.075). This finding is in line with the hypothesis. In terms of prosocial intentions towards neutral targets, there was a significant difference between prosocial intentions in public school students and Islamic school students, where prosocial intentions towards neutral targets were higher in public schools than Islamic schools (ΔM = 0.145; SD = -0.075). This finding is inconsistent with Hypothesis 2c, where prosocial intention toward neutral targets was predicted to be similar across the two school backgrounds. Table 1 and Graph 2 summarize the differences in the means of prosocial intentions towards ingroup, outgroup, and neutral targets among students in public and Islamic schools.
In summary, this study yielded several findings. First, regardless of school background, adolescents' prosocial intentions were found to be highest toward ingroup targets, followed by those toward outgroup targets, and the lowest toward neutral targets (Hypothesis 1 was partially accepted). Second, based on school background, prosocial intentions toward ingroups were higher in adolescents from public school (inconsistent with Hypothesis 2a); prosocial intentions towards outgroups were higher in adolescents from public school (consistent with Hypothesis 2b), while prosocial intentions towards neutral targets were found to be higher in adolescents from public school (inconsistent with Hypothesis 2c). The discussions below elaborate further these results.

Comparing prosocial intentions in ingroup, outgroup, and neutral targets
The results of within-subject analysis showed that prosocial intentions towards ingroups were found higher than prosocial intentions towards outgroups, which are in line with the previous findings (Fiedler et al., 2018;Hackel et al., 2017;Preston & Ritter, 2013). The difference in prosocial intentions towards ingroup and outgroup targets might be attributed to the social identity which leads to ingroup favoritism (Everett et al., 2015). When individuals consider themselves a member of certain group (Everett et al., 2015), positive interactions between individuals and the group members will foster ingroup favoritism, by evaluating and attaching positive values to the group (Rothgangel & Dreyer, 2016;Verkuyten, 2016).
Strong ingroup favoritism is usually followed by outgroup prejudice, which is a mechanism to make ingroup looks more positive (Burch-Brown & Baker, 2016;Moulin, 2016). Prejudice then may lead to discrimination against outgroup members (Prati et al., 2016). Among the forms of discrimination is not providing benefits and assistance to outgroup members (Hamley et al., 2020) and avoiding cooperation with outgroup members (Sosis & Ruffle, 2004). This would result in prosocial intentions towards the ingroup target being higher than towards the outgroup target.
Meanwhile, prosocial intentions towards neutral targets were found the lowest, lower than prosocial intentions towards outgroup targets. We assume that this relates to the importance of reciprocality for adolescents in their social interactions (Groep et al., 2020). Adolescents prefer to distribute resources to ingroup members and outgroup members rather than neutral targets in order to have more opportunities to obtain reciprocal benefits from the prosocial behavior that they perform. When dealing with a neutral target, there is no clear social identity, this might make adolescents perceive that their chances for obtaining an exchange from their prosocial behavior are smaller.

Comparing prosocial intentions based on school backgrounds
Between-subject analysis was carried out to see the prosocial intentions toward ingroup, outgroup, and neutral targets based on school background. Prosocial intentions towards ingroup targets were found to be higher in adolescents from public schools than in adolescents from Islamic school. This result was not consistent with Hypothesis 2a, in which prosocial intentions towards the ingroup target were expected to be higher in adolescents from religious school. Meanwhile, prosocial intentions towards outgroup targets were found higher in public school, which was consistent with Hypothesis 2b.
These findings also showed that prosocial intentions towards ingroup, outgroup, and neutral targets were overall higher in adolescents from public school . This result can be explained from at least two perspectives: the perspective of social identity and the perspective of interaction. From a social identity perspective, public school students may develop a more varied and equally strong social identity than religious school students, making them more flexible in responding to various groups (Hamburg & Hamburg, 2004;Verkuyten, 2007). Furthermore, adolescents from public schools are more likely to interact with various groups than adolescents in religious schools. Adolescents who are accustomed to interacting with various groups are found to be more likely to exhibit positive behavior when interacting with inter-groups (Rothgangel, 2016).
Prosocial intentions towards ingroup targets were found to be higher in public school students. This may be in line with a research by Johnson, Rowatt, and LaBouff (2012) which found that individuals will increasingly activate their religious identities and increase support for ingroup members when they feel the presence of outgroup members as a threat, both physically and emotionally. In public schools, students feel the presence of the outgroup (i.e. students of different religions); this situation unconsciously raises the perception of a threat from outgroup members and, in turn, makes their religious identity more salient.
On the other hand, prosocial intentions towards outgroup targets were also found to be higher in adolescents from public school . This is in accordance to the previous research in which class homogeneity is associated with an increase in discriminatory behaviors such as a decrease in cooperation and helping behavior towards outgroup members (Scacco & Warren, 2018;Waillet & Roskam, 2012). In Islamic schools where students are homogeneous in religion, students have limited opportunities to interact with students of different religions (Azizah, 2006;Rahayu & Nadir, 2018). This condition can increase prejudice against outgroup members (Aboud & Amato, 2008;Baidhawy et al., 2014)and lead to reduced keenness to help outgroup members (Hamley et al., 2020). Exclusive interaction with own religious group members will strengthen religious identity, and in turn may lead to higher tendency to reject giving support to other religious group members (Verkuyten, 2007).
Overall, the results of this study complement the previous studies conducted by Everett et al. (2015), Hackel et al. (2017), and Preston and Ritter (2013), confirming that prosocial behaviors towards ingroup target tend to be higher than those towards outgroup target, and provided data in the Indonesian context. However, the results of this study need to be interpreted carefully due to the limited number and background of participants and problems with the normality and homogeneity of the data.
Further research should expand the sample from various religious backgrounds, regions, and school backgrounds in order to achieve a better power for generalization. Mediating and moderating variables such as level of social identity (religious identity and national identity), prejudice, and intergroup contact need to be explored to understand the dynamics of prosocial intentions towards ingroup, outgroup, and neutral targets. The level of social identity before and after priming should be measured to ensure the effect of priming in awakening the individual's religious social identity.
Apart from the above drawbacks, the results of the study showed that there was bias in intergroup relations among adolescents, indicated by the differences in prosocial intentions towards ingroups (member of the same religion) and outgroups (members of different religion). This fact should be taken seriously into account, as a consideration for the authorities in developing school curricula and programs, as well as creating school environments that are more anticipatory towards the potential conflicts between groups and promote harmony and tolerance.
Both public and religious schools should accommodate positive interactions between interfaith students. The positive interaction in public schools will ensure that religion-based social identity is not evoked in negative fashion. Religious schools should provide opportunities for students to interact with interfaith groups, which in turn will ensure students develop positive attitudes towards other religiius group members.

CONCLUSION
It can be concluded that the level of prosocial intentions was generally highest toward the ingroup target, followed by those toward outgroup and neutral targets. Based on the school background, prosocial intentions towards ingroup, outgroup, and neutral targets were found to be higher in public school students than religious school students. The implication of this research is both public and religious schools should accommodate positive interactions between interfaith students. This positive interaction is expected to reduce bias between groups and increase prosocial spirit within the community and ensure harmonious relations within the society.