Teachers’ ratings of social exclusion among students: The role of situational information and the ethnic origin of the excluded student based on the example of Syrian refugees

: In the current study, we examined teachers’ ratings of social exclusion among students. 120 teachers ( M age =24.00, SD =3.71, 88% female) evaluated a hypothetical exclusion scenario in which the excluded student’s origin (German vs. Syrian refugee) was varied as well as whether participants received additional situational information about prior norm-violating behavior of the excluded student or not. Additionally, participants rated how likely they would intervene in the situation. For the evaluation of exclusion and the likelihood of intervention, there was a main effect of additional situational information, revealing that participants evaluated exclusion as less acceptable and were more likely to intervene if there was no additional situational information. Regarding the evaluation of exclusion, there was an interaction of additional situational information and the origin of the excluded student as the effect of additional situational information was bigger if the excluded student was German. For the likelihood of intervention, this interaction was not significant; but descriptively a similar pattern emerged. Results indicate that information about prior norm-violating behavior is more relevant for teachers’ reactions to social exclusion than the origin of the excluded student. However, in situations with an understandable reason for exclusion, teachers do include the origin of the excluded student in their considerations.


Introduction
Since the influx of refugees in Europe and especially in Germany starting in 2015 (Eurostat, 2021), refugees and their integration have become a much-discussed topic in Germany within the academic community (Funk, 2016) and in society as a whole (Borsutzki & Schenk, 2020). Today, more than 100,000 refugee children live in Germany (DESTATIS, 2019) and face specific challenges as a consequence of being refugees. They experienced various psychological stressors or physical threats in their home countries or during their escape, and many of them suffer from trauma as a consequence (Kien et al., 2019;Nesterko, Jäckle, Friedrich, Holzapfel, & Glaesmer, 2019). Given these circumstances, it is important to understand that the group of refugees is markedly different from other groups of immigrants and that children from refugee families can be considered to be an at-risk population (Gavranidou, Niemiec, Magg, & Rosner, 2008).
The various psychological conditions combined with a typical lack of German language skills lead to a range of challenges and disadvantages in educational contexts (Henschel et al., 2019;Schipolowski, Edele, Mahler, & Stanat, 2021). However, besides these specific challenges, children from refugee families have to face social challenges, similar to children from other migrant groups. One of these social challenges is social exclusion. Riva and Eck (2016, p. ix) define social exclusion as "the experience of being kept apart from others physically or emotionally." Social exclusion can manifest itself in a variety of ways such as hurtful laughter or outright rejection in the form of exclusion from social activities. Given the strong human need to belong, social rejection causes emotional distress (Baumeister & Leary, 1995) and is associated with a variety of negative psychological outcomes (Wesselmann et al., 2016). Research has also shown that children from ethnic minorities experience more social exclusion than others (Plenty & Jonsson, 2017;Verkuyten & Thijs, 2002). Given the mental health problems refugee children already suffer from and their increased vulnerability, exclusion might be particularly challenging for them, and schools should be a place where children from refugee families feel safe and included.
One approach for addressing the issue of social exclusion among students is to focus on teachers. As teachers are role models for their students (Evans, 1992), teachers' attitudes and behavior can have a variety of effects on student attitudes and behaviors, going beyond merely improving student performance in school subjects (Blazar & Kraft, 2017;McAuliffe, Hubbard, & Romano, 2009). Teachers shape their students' attitudes and behaviors with their expectations regarding different social groups in the classroom and their treatment of these groups (Muntoni & Retelsdorf, 2020). Research by Verkuyten and Thijs (2002) on children in the Netherlands showed that children report fewer personal experiences of discrimination and exclusion based on their ethnicity when they think that they can tell their teachers about ethnic victimization and that their teachers will listen and take action. This shows that teachers' behavior and reactions to interethnic social exclusion have the potential to impact the behavior of their students. Thus, research is needed that systematically investigates teachers' reactions to interethnic social exclusion, especially regarding the particularly vulnerable group of refugees. In the current study, we specifically focus on Syrian refugees as they represent the largest group of refugees in Germany (DESTATIS, 2019).

Teachers' reactions to interethnic social exclusion
So far, there has been little research on teachers' reactions to social exclusion. To our knowledge, no study has specifically focused on the exclusion of refugees. Three studies (Beißert & Bonefeld, 2020;Beißert, Staat, & Bonefeld, 2021;Kollerová & Killen, 2021) examined teachers' ratings of the social exclusion of children with migration backgrounds from countries that are typically associated with less recent immigration (such as typical labor migration countries). Both studies used experimental designs to assess teachers' ratings of hypothetical social exclusion scenarios in which the ethnic origin of the excluded student was manipulated. Beißert and Bonefeld (2020) analyzed teachers' ratings of the exclusion of a German vs. a Turkish student in Germany while Kollerová and Killen (2021) compared reactions to the exclusion of a Czech vs. an Arab student in the Czech Republic. Kollerová and Killen (2021) found that teachers showed a slight ethnic bias when reasoning about the wrongfulness of exclusion, raising more moral concerns when speaking about a native student being excluded. Beißert and Bonefeld (2020) demonstrated that teachers generally reject social exclusion among students, and that they evaluate the exclusion of a Turkish boy as even more reprehensible than the exclusion of a German boy. Both studies suggest that teachers take into account an excluded student's origin when rating social exclusion scenarios. In the following sections, we present some considerations as to why and how the ethnic origin of an excluded student might affect teachers' reactions to social exclusion -especially for the group of refugees as a particular group of immigrants.

Awareness of the vulnerability of ethnic minorities or refugees in particular
Social exclusion based on ethnicity is a serious problem faced by many immigrants and refugees in a new host country (Minority Rights Group International, 2010). Minority groups are especially likely to be confronted with stereotypical mindsets and behavior, which can result in exclusion (Killen, Mulvey, & Hitti, 2013). Given the numerous psychological stressors children from refugee families face and the fact they have had to leave most of their friends and family, the need to belong might be especially strong in these children, and social exclusion might be particularly stressful for them.
Further, it has been shown that refugee children are even more disadvantaged in the educational system than other groups of immigrants. They are overrepresented in lower school tracks (Henschel et al., 2019) and have lower competencies than non-refugee foreign-born children (Schipolowski et al., 2021). In addition to factors related to the children (e.g., low language skills, traumatization), teachers' attitudes, stereotypes, and (unconscious) biases might be relevant in this context, as prior research has indicated for children with migration backgrounds in general (Bonefeld & Dickhäuser, 2018;Bonefeld, Dickhäuser, & Karst, 2020;Kleen & Glock, 2018a).
While discrimination is still an issue in Germany (Bonefeld & Dickhäuser, 2018;Kristen & Granato, 2007;Steinmann, 2019), it is widely considered to be morally wrong. Germany adopted the General Act on Equal Treatment with the goal of stopping discrimination, e.g., based on ethnic origin (German Federal Anti-Discrimination Agency, 2012a) and established a Federal Anti-Discrimination Agency (German Federal Anti-Discrimination Agency, 2021b). With regard to teachers more specifically, one of the two largest and most important associations representing the interests of teachers and educators in Germany stresses the value of diversity and the importance of integrating students with a migration background and refugees (Verband Bildung und Erziehung, n.d.). Additionally, the challenges of refugee children are widely discussed in the media (Zeit Online, n.d.). Given these considerations, it can be assumed that teachers are generally aware of the above-mentioned issues.
Awareness of the situation of students from refugee families might lead to enhanced feelings of empathy and sympathy for refugee children in general, especially in teachers who have refugee students in their classes. As empathy is associated with prosocial behavior towards a victim after witnessing social exclusion (Barchia & Bussey, 2011;Masten, Morelli, & Eisenberger, 2011), empathy could guide teachers' reactions regarding the social exclusion of refugee students. Teachers might consider the negative emotions that refugee children generally have to deal with and want to support them even more than other children when they are socially excluded. Acknowledging that they need special attention, teachers might tend to reject the exclusion of refugee students more strongly than the exclusion of German1 students. This would mean that they evaluate the exclusion of a refugee student as more negative and would be more likely to intervene in situations where a refugee student is excluded.
However, based on social categorization and social identity theory (Tajfel & Turner, 1979), the opposite assumption, i.e., the assumption that teachers tend to reject the exclusion of German students more strongly, is also conceivable.

Social categorization and ingroup bias
It is a well-studied phenomenon that humans use social categories to help them process information, and these social categories inform their behavior (Macrae, Stangor, & Milne, 1994;Trepte & Loy, 2017). Social categorization refers to the process of classifying different people into different social categories or social groups (e.g., based on ethnicity or gender) in order to structure the complex social world and free cognitive resources (Macrae, Milne, & Bodenhausen, 1994;Macrae, Stangor, & Milne, 1994). When we make these categorizations, we also differentiate between ingroups (groups with which we psychologically identify as being a member) and outgroups (groups with which we do not identify).
However, certain biases are associated with this classification into ingroups and outgroups. People tend to have a positive bias for members of their ingroup which is manifested in their evaluations and behavior (Mullen, Brown, & Smith, 1992;Tajfel & Turner, 1979). This tendency to favor or promote an ingroup, often at the expense of other groups, is called ingroup bias (Tajfel & Turner, 1979). To explain this phenomenon, Tajfel and Turner (1979) introduced the concept of social identity. According to social identity theory, people define their social identity based on group memberships (Tajfel, 1982Tajfel & Turner, 1979. In order to enhance their self-esteem, people want to identify with and belong to social groups that are seen as superior to others (Tajfel, 1982). Accordingly, social identity theory postulates that people engage in ingroup bias in order to enhance the positive distinctiveness of their own social group and consequently strengthen their social identity and self-esteem.
Ingroup bias is manifested in peoples' evaluations and behavior (Mullen, Brown, & Smith, 1992;Tajfel & Turner, 1979). In the educational context, ingroup bias has been found to influence teachers' judgments of students' social behavior (McGrady & Reynolds, 2013) and their sanctioning behavior (Schuchart, Glock, & Dunkake, 2021). Further, it has been shown that ingroup bias can generally affect the evaluation of exclusion (Rudert, Sutter, & Greifeneder, 2018). Taken together, this leads to the assumption that ingroup bias might also affect teachers' ratings of interethnic social exclusion. In line with that, teachers might react differently to the social exclusion of a German student compared to the exclusion of a Syrian refugee student. According to ingroup bias, German teachers' attitudes towards German students should be more positive and, consequently, teachers should reject the exclusion of ingroup members (German students) more strongly than the exclusion of outgroup members (Syrian refugee students). As a result, they would be expected to evaluate the exclusion of a German student more negatively and be more likely to intervene in situations where a German student is excluded.

Attitudes regarding social groups
Besides ingroup bias, social categorization can also lead to specific attitudes. Attitudes are defined as the positive or negative evaluation of an object (Gawronski & Bodenhausen, 2006;Greenwald & Banaji, 1995). Attitudes can guide judgments and behavior (Fazio, 1986).
Given that social categories guide human behavior (Tajfel & Turner, 1979), the specific attitudes teachers have towards Syrian refugees could help explain and predict their reactions to the social exclusion of Syrian refugees. Although no studies have specifically focused on teachers' attitudes towards Syrian refugees to date, there are several studies that have investigated stereotypes about refugees in Germany and studies on teachers' attitudes towards students with migration backgrounds in general. Stereotypes are broadly shared assumptions in a society about the characteristics of members of certain groups (Hilton & Hippel, 1996) and can be understood as the cognitive part of attitudes (Eagly & Mladinic, 1989). In line with the Stereotype Content Model (SCM; Fiske, Cuddy, Glick, & Xu, 2002), stereotypes regarding immigrants can be analyzed along the dimensions competence and warmth. It has been shown that Syrian immigrants (along with other migrant groups that recently moved to Germany) are considered to have low competence and moderate warmth (Froehlich & Schulte, 2019;Kotzur, Friehs, Asbrock, & Zalk, 2019).
These stereotypes of middle to lower competence are in line with research on teachers' perceptions of students with a migration background Glock & Krolak-Schwerdt, 2013). For instance, it has been shown that a mere indication that students have a migration background leads teachers to grade their work worse than students who do not seem to have a migration background who show the same performance (Bonefeld & Dickhäuser, 2018). Moreover, teachers' attitudes can have an impact on social interactions in class (Glock & Kleen, 2017;Kleen & Glock, 2018b).
Overall, these findings suggest that teachers' negative attitudes towards students from ethnic minorities impact their decisions and behavior related to their students. In the current paper, we want to investigate whether this also holds for teachers' ratings of social exclusion among students.

The role of situational information
Teachers' reactions to social exclusion are not only affected by who is the target of exclusion. Situational aspects also play an important role in teachers' perception of exclusion scenarios. In line with this, in a study by Beißert and Bonefeld (2020), many teachers stated that it was difficult to evaluate an exclusion scenario because they had so little information about the situation.
While, in general, humans typically reject social exclusion (Beißert & Bonefeld, 2020;Rudert, Ruf, & Greifeneder, 2020), there are circumstances under which social exclusion can be more accepted, e.g., if individuals engage in exclusion as a punitive measure for the violation of social norms (Rudert et al., 2020). Social norms are shared rules or standards which are considered acceptable in a certain group or society (e.g., lying, being rude, or not cooperating in social situations) (Cialdini & Trost, 1998;Lapinski & Rimal, 2005). Rudert et al. (2020) found that observers of social exclusion would punish the excluding group when the excluded target's behavior was consistent with social norms. On the other hand, they would punish the target of social exclusion, if his or her behavior was not in line with social norms. Beyond causing others to engage in exclusion as a punitive measure, such norm-violating behavior could also trigger considerations about group functioning; that is, the group might expect the norm-violating person to negatively affect the functionality of the group. Research shows that children and adolescents often refer to aspects related to smooth group functioning when justifying social exclusion (Horn, 2003;Park & Killen, 2010). A person violating social norms might be considered a poor fit for a cooperating group because he or she might hinder smooth group functioning, thus further justifying his or her social exclusion. Overall, this demonstrates that people do not blindly reject social exclusion by any means, but take into account additional information about the previous behavior of an excluded person when evaluating social exclusion.

The interplay of situational information and the ethnic origin of an excluded student
Furthermore, additional situational information might affect the impact that the origin of an excluded person has on the evaluation of social exclusion. As described in previous sections, teachers might show ingroup bias and/or use stereotypes about refugees when they rate situations in that refugee students are being excluded. This would especially be expected in situations that are ambiguous where they do not have information about the situation, because people tend to utilize stereotypes more if they do not have enough relevant information to judge a situation otherwise (Duncan, 1976). Previous research has shown that peoples' judgments are less guided by stereotypes when they have more information about the situation or person that they are evaluating (Locksley, Borgida, Brekke, & Hepburn, 1980;Locksley, Hepburn, & Ortiz, 1982). Also, discriminatory behavior (Singletary & Hebl, 2009) and ingroup bias (Dunham & Emory, 2014) have been shown to be stronger in ambiguous situations. These results indicate that the social group as a source of information should be less relevant in situations with little ambiguity.

Theoretical framework
This research is embedded in several theoretical considerations. Social identity theory (Tajfel & Turner, 1979), on the one hand, provides a model for examining how identification with the ingroup can lead to ingroup bias and affect intergroup judgments (see Chapters 1.3 and 1.4). Social domain theory, on the other hand, postulates that individuals use three distinct domains when reasoning about social situations: the moral domain (concerns about fairness or others' welfare), the social-conventional domain (references to group functioning, norms, traditions, or conventions), and the personal domain (references to personal choice, autonomy, and self-interest) (Turiel, 1983;Smetana, 2006). In line with this, depending on the specific situation, sometimes exclusion might be seen as a moral issue (acknowledging the unfair nature of exclusion or the harm it can cause), but in other situations exclusion might be understood as a social-conventional issue (e.g., exclusion aiming to protect certain social norms or enabling smooth group functioning) or as a personal issue (e.g., with references to personal choice or autonomy).
Integrating social identity theory (Tajfel & Turner, 1979) and social domain theory (Turiel, 1983;Smetana, 2006), the social reasoning developmental model (SRD, Rutland & Killen, 2015;Rutland, Killen, & Abrams, 2010) represents a suitable theoretical framework for research on social exclusion in intergroup contexts. According to the SRD model, when evaluating or reacting to the exclusion of others, individuals weigh up information about group memberships, norms, and group functioning against moral considerations about fairness or others' welfare (Rutland & Killen, 2015;Rutland et al., 2010). This framework can guide our research on social exclusion in interethnic contexts and help us to understand the specific findings.

Current study
To extend research on teachers' reactions to interethnic exclusion, hypothetical scenarios were used to examine preservice and in-service teachers in the role of observers of exclusion among students. More specifically, we assessed teachers' evaluations of hypothetical exclusion scenarios and their anticipated reactions and interventions, i.e., how likely they were to intervene in such a situation and what they would specifically do. We examined the role of an excluded student's ethnic origin and additional situational information (information about the excluded student's prior norm-violating behavior) for teachers' reactions to the exclusion scenarios. We wanted to extend prior research by focusing on refugees as a particularly vulnerable group of immigrants and focused on students with Syrian background because the most common origin of refugees in Germany is Syria (DESTATIS, 2019).

The role of additional information regarding the excluded student
Social domain theory highlights the important role of situational aspects for the evaluation of social exclusion. It has been shown that norm-violating behavior leads to higher acceptance of social exclusion. Thus, in our study, we expected additional situational information suggesting that a group has a justifiable reason for excluding another student (e.g., as a punitive measure for the transgression of social norms or to enhance smooth group functioning) to lead to a higher acceptance of that specific instance of social exclusion and a lower likelihood of intervention (H1).

The role of the ethnic origin of the excluded student
Regarding the role of the excluded student's ethnic origin, two opposing hypotheses can be assumed. Given the overall norm of inclusivity and avoidance of discrimination paired with empathy for the struggles that refugee children face, it can be assumed that teachers are particularly sensitive regarding the social exclusion of refugees (H2a). On the other hand, teachers' ratings might be biased by negative stereotypes and attitudes about refugees. Additionally, social identity theory suggests that they might prefer their ingroup. Based on these considerations, it can be assumed that teachers reject the exclusion of a German student more than the exclusion of a student who is a Syrian refugee (H2b). These tendencies should manifest themselves in teachers' evaluation of the exclusion scenario as well as in their likelihood to intervene in the scenario.

The interplay of ethnic origin and additional information
When presented with an ambiguous social exclusion scenario, teachers are more likely to rely more on stereotypes and/ or ingroup bias when judging a situation. The provision of additional situational information decreases the ambiguity of the situation. Thus, it can be assumed that teachers' ratings are less affected by the origin of the excluded student when additional situational information is provided (H3).

Sample and recruiting
An a priori power analysis in G*Power 3 (Faul, Erdfelder, Lang, & Büchner, 2007) revealed that a sample size of 128 would be required to achieve a power (1-β) of 0.8 at α = .05 (two-tailed) with an assumed medium effect size of f = .25 for an ANOVA with four factors (two independent variables: additional situational information, origin; two control variables: participant gender, professional status).
In total, 173 participants participated in the online survey. Due to the hypotheses related to social identity theory, we aimed at participants who were likely to identify themselves at least to some degree as being German. Thus, we aimed at participants from families who have been living in Germany for at least two generations. For this purpose, we excluded persons who were not born in Germany and/or who had at least one parent who was not born in Germany (n = 20). Furthermore, only participants who fully completed the questionnaire were considered for the analyses. So, the sample used in the statistical analyses included 120 participants (M age = 24.00, SD = 3.71, range 18-37 years, 88% female). Approximately one-fifth of the participants were in-service teachers (n = 10 primary school teachers, n = 13 secondary school teachers); the rest of them were pre-service teachers. Pre-service teachers are university students who will become teachers after graduation. Of the 97 pre-service teachers, 32 were majoring in primary school education, 50 in secondary school education, and 15 in special needs education. Most of them (n = 84) had practical teaching experience, having already completed a teaching internship in a school.

Design and procedures
The experiment was conducted as an online survey using the software Unipark. The participants were recruited via different mailing lists and online groups on social media platforms (e.g., Facebook groups).
The study was conducted in accordance with the ethical guidelines of DGPs (German Psychological Society). All subjects gave written informed consent in accordance with the Declaration of Helsinki and in line with the General Data Protection Regulation of the EU. At the beginning, the participants were informed about the expected length of the study (10-15 minutes) and told that it was a study on social situations in school. In order to obtain unbiased responses, we did not mention that the study was about interethnic social exclusion in the introduction. Before the assessments started, participants were informed of their data protection rights. They were also told that participation in the study was anonymous and voluntary, and that there would be no disadvantages if they decided not to participate or leave the study early without completing it. After the participants had confirmed that they understood the information and were willing to participate in the study, they completed a questionnaire collecting demographical data (gender, age, professional status, migration history in the family). Next, they were presented with a hypothetical scenario in which a student is excluded from a learning activity by other students in his class. In this scenario we varied the ethnic origin of the excluded student. Additionally, the participants were either provided with further information about prior norm-violating behavior of the excluded student or not. The experiment was realized as a 2 (student origin: German vs. Syrian) x 2 (situational information: insult observation vs. no additional information) between-subjects design. Participants were assigned randomly to one of the four experimental conditions (German student -insult observation: n = 27, Syrian student -insult observation: n = 27, German student -no additional information: n = 32, Syrian student -no additional information: n = 34).

Material
The approach of using hypothetical exclusion scenarios which can be experimentally varied has been used extensively in research with children and adolescents (e.g., Killen & Stangor, 2001;Malti, Killen, & Gasser, 2012;Cooley, Burkholder, & Killen, 2019) and was adapted for examining teachers' ratings of social exclusion by Beißert and Bonefeld (2020).
The hypothetical exclusion scenario consisted of a situation in which one student was excluded from a learning group by his classmates. The excluded student was called either Lukas or Rami. Lukas was described as a new boy in the class who had moved there from another German city. Rami, too, was described as a new boy in the class, but as having escaped from Syria with his family. Additionally, the participants were either provided with additional situational information about the prior norm-violating behavior of the excluded student or not. As norm-violating behavior, we selected an everyday situation in school: a student insulting another student. Therefore, the participants either learned that the excluded student had insulted another student in the class prior to the exclusion situation or they did not receive any additional situational information. See Table 1 for the English translation of the exact wording of the hypothetical exclusion scenarios.

No additional information Insult observation
German protagonist While packing your belongings after a class in 7 th grade, you observe some students making an appointment to study together. You observe that Lukas, who is new in the class after having moved here from Hamburg with his family, joins the conversation. Lukas wants to join the study group. The other students tell him that he can't join.
While packing your belongings after a class in 7 th grade, you observe some students making an appointment to study together. You observe that Lukas, who is new in the class after having moved here from Hamburg with his family, joins the conversation. Lukas, who insulted another student earlier on in the day, wants to join the study group. The other students tell him that he can't join.

Syrian refugee protagonist
While packing your belongings after a class in 7 th grade, you observe some students making an appointment to study together. You observe that Rami, who is new in the class after having come to Germany with his family as a Syrian refugee, joins the conversation. Rami wants to join the study group. The other students tell him that he can't join.
While packing your belongings after a class in 7 th grade, you observe some students making an appointment to study together. You observe that Rami, who is new in the class after having come to Germany with his family as a Syrian refugee, joins the conversation. Rami, who insulted another student earlier on in the day, wants to join the study group. The other students tell him that he can't join.
Prior research has shown that the gender of an excluded person can also affect observers' evaluations of social exclusion (Beißert, Bonefeld, Schulz, & Studenroth, 2019;Beißert, Staat, & Bonefeld, 2022). As the focus of the current study was the role of the ethnic origin of the excluded students above and beyond their gender, we kept the gender of the excluded student constant in the scenarios. Thus, the excluded student was a boy in all scenarios.

Measures
After being presented with the scenario, the teachers were asked for two kinds of ratings. Firstly, they indicated on a seven-point Likert-type scale how (1) not okay/okay, (2) unfair/fair, and (3) unjustifiable/justifiable they evaluated the scenario to be. A score indicating a participant's evaluation of the exclusion was created by calculating mean values across these three items (Cronbach's alpha = 0.92). High numbers indicate high acceptance of exclusion; low numbers indicate strong rejection of exclusion. Secondly, the participants were asked to rate how likely it was that they would intervene in the situation if they witnessed such a situation in their class. This was also assessed using a seven-point Likert-type scale (1 = very unlikely to 7 = very likely). In order to better understand their underlying considerations, we additionally asked them to justify their decision (open-ended question). Additionally, participants were asked about their anticipated reactions, i.e., they were asked what specifically they would have done when intervening in the situation (open-ended question).

Coding of open-ended questions
The coding systems for the open-ended questions (justification of the likelihood of intervention, anticipated actions) were based on prior research (e.g., Beißert & Bonefeld, 2020;Beißert et al., 2021) and were extended with additional categories inductively developed from the surveys themselves (see Tables 2 and 3 for an overview and examples). The coders were allowed to code up to three relevant justifications for each statement. Coding was completed by two independent coders. On the basis of 25% of the interviews, interrater reliability was high, with Cohen's kappa = .90 for the justifications of the likelihood of intervention, and kappa = 1.00 for the anticipated actions. We included the categories that were used by more than 10% of the participants in the analyses. Learning from own behavior "Children need to learn from their behavior." 9 Social norm of inclusion and participation/equity

47
Mediation/avoid escalation "I would talk to them, trying to mediate the conflict.", "I would talk to them in order to prevent the quarrel from escalating."

19
Class-oriented perspective "The sense of belonging within the class is very important." "Exclusion negatively affects the class climate."

Other
Meaningful, but single statements 4 Undifferentiated Meaningless statements 5

Results
Univariate ANOVAs were used to test for differences in the evaluation of exclusion and the likelihood of intervention between the different experimental conditions. Repeated measures ANOVAs on the use of categories were conducted to analyze reasoning data from the openended questions. ANOVA frameworks are appropriate for analyzing reasoning data because ANOVAs are robust to the problem of empty cells, whereas other data analytic procedures require cumbersome data manipulation to adjust for empty cells (see Posada & Wainryb, 2008, for a more detailed explanation and justification of this data analytic approach).
Participants' gender was used as a control variable in all analyses because in previous research women and men differed significantly in their ratings of social exclusion scenarios (e.g., Beißert & Bonefeld, 2020). Furthermore, participants' professional status (pre-service vs. in-service teacher) was included as a control variable because it has been shown that teachers' attitudes regarding diversity can change over the course of their training (Kumar & Hamer, 2013).

Evaluation of exclusion
In order to test for differences in the evaluation of exclusion based on the additional situational information as well as on the ethnic origin of the excluded student, we conducted a 2 (situational information: insult observation vs. no additional information) X 2 (origin: German, Syrian) X 2 (participant gender: male, female) X 2 (professional status: preservice teacher, in-service teacher) univariate ANOVA with the evaluation of exclusion as dependent variable.
As expected, there was a main effect of the additional situational information, F (1, 107) = 56.47, p < .001, ηp² = .35. In line with Hypothesis 1 and across both protagonists, the participants evaluated the exclusion situation as much less acceptable in the condition with no additional situational information (M = 1.89, SD = 0.96) compared to the conditions where they learned about the prior norm-violating behavior of the excluded student (M = 4.01, SD = 1.42).
While the participants who did not know about the prior norm-violating behavior clearly rejected the exclusion, the means in the condition with additional situational information were situated more in the middle and in part even on the acceptance side of the scale (see Figure 1). Furthermore, as expected, there was a main effect of the origin of the excluded student, F (1, 107) = 10.15, p = .002, ηp² = .09. In line with Hypothesis 2a, the participants rejected the exclusion of the Syrian student (M = 2.58, SD = 1.39) more strongly than the exclusion of the German student (M = 3.12, SD = 1.73). However, this effect was mainly driven by an interaction effect of the origin of the excluded student and situational information, F (1, 107) = 4.62, p = .034, ηp² = .04. Post-hoc tests revealed that the exclusion of the Syrian was more strongly rejected than that of the German protagonist only in the condition where additional situational information was provided. In other words, in the condition where no additional situational information was provided, the participants' evaluations of the exclusion of the Syrian protagonist did not differ (M = 1.84, SD = 0.93) from their evaluations of the exclusion of the German protagonist (M = 1.95, SD = 1.01), p > .05. However, in the condition with information about prior norm-violating behavior, the teachers rejected the exclusion of the Syrian protagonist (M = 3.51, SD = 1.34) more strongly than the exclusion of the German protagonist (M = 4.51, SD = 1.33), p < .001. These results are displayed in Figure 2. Thus, in contrast to our expectations (Hypothesis 3), the origin of the excluded student was less relevant in the condition where no additional situational information was given.
To sum up these main and interaction effects, exclusion was more accepted in the condition with information about the insult observation for both protagonists, but this effect of additional situational information was less strong for the Syrian protagonist.

Likelihood of intervention
To test for differences in teachers' ratings of their likelihood to intervene in the situation based on the additional situational information and on the ethnic origin of the excluded student, we conducted a 2 (situational information: insult observation vs. no additional information) X 2 (origin: German, Syrian) X 2 (participant gender: male, female) X 2 (professional status: pre-service teacher, in-service teacher) univariate ANOVA with the likelihood of intervention as dependent variable.
As expected in Hypothesis 1, there was a main effect of the situational information, F (1, 107) = 13.27, p < .001, ηp² = .11, revealing that participants were more likely to intervene in scenarios where they did not receive any additional situational information (M = 5.1, SD = 1.78) compared to in the scenarios where they were informed that the excluded student had insulted another student prior to the exclusion situation (M = 4.50, SD = 1.70).
There was neither a main effect of the excluded student's ethnic origin, nor an interaction effect of the ethnic origin and additional situational information. However, although the interaction effect was not significant, the descriptive data revealed a similar pattern for the interplay of situational information and the excluded student's ethnic origin regarding the evaluation of exclusion. While there was no difference between the two protagonists in the condition without additional situational information (M German = 5.09, SD = 1.61, M Syrian = 5.12, SD = 1.95), in the condition with  additional situational information about the excluded student's prior norm-violating behavior, participants were more likely to intervene when a Syrian student was excluded (M = 4.81, SD =1.84) than when a German student was excluded (M = 4.19, SD =1.52). It could be due to the lack of power that the interaction effect was not significant.

Justification of likelihood of intervention
The justifications of participants' likelihood to intervene in the situation were analyzed based on proportional category use. We included all categories that were used by at least 10% of the participants. These categories were: "need for information/lack of context information," "children's autonomy," "empathy for the victim," "social norm of inclusion and participation/equity," "mediation/avoid escalation," and "class-oriented perspective." To test for differences in participants' justifications based on the excluded student's ethnic origin and on the additional situational information, we ran a 2 (situational information: insult observation vs. no additional information) X 2 (origin: German, Syrian) X 6 (justification: need for information/lack of context information, children's autonomy, empathy for the victim, social norm of inclusion and participation/equity, mediation/avoid escalation, and classoriented perspective) ANOVA with repeated measures on the factor "justification."² There was a main effect of justification, F (4.79, 527.08) = 5.81, p < .001, ηp² = .05, revealing that references to the norm of inclusion and participation/equity were more frequent than all other codes, except for children's autonomy.
Furthermore, there was an interaction effect of situational information and justification, F (4.79, 527.08) = 5.85, p < .001, ηp² = .05. More specifically, teachers made many more references to the social norm of inclusion and participation/ equity in the condition where no additional situational information was given (M = 0.55, SD = 0.50) compared to in the condition where they received additional situational information about the norm-violating behavior (M = 0.24, SD = 0.43), p = .001. In contrast, the participants stated that they would intervene in order to mediate or avoid escalation almost exclusively in the condition where information about prior norm-violating behavior was given (M = 0.30, SD = 0.46) and would hardly intervene in the condition where no additional situational information was provided (M = 0.06, SD = 0.24), p = .001. See Table 4 for all means and standard deviations.  Note. M = mean, SD = standard deviation.
2 As the analyses did not reveal any effects based on participants' gender or professional status, we omitted these two variables and calculated a reduced model for the sake of simplicity.

Anticipated actions
Again, analyses were conducted on proportional category use, and all categories that were used by at least 10% of the participants were included. These categories were "ask for reasons," "help to find inclusion-oriented solution," "explain norm of inclusion," "find alternative solution for the excluded student," "mediation between students," and "talk to students." To test for differences in participants' anticipated actions based on the excluded student's ethnic origin and on additional situational information, we conducted a 2 (situational information: insult observation vs. no additional information) X 2 (origin: German, Syrian) X 6 (anticipated action: ask for reasons, help to find inclusion-oriented solution, explain norm of inclusion, find alternative solution for the excluded student, mediation between students, and talk to students) ANOVA with repeated measures on the factor "action."3 There was a main effect of the anticipated action, F (4.82, 535.22) = 6.69, p < .001, ηp² = .06, revealing that teachers would ask for reasons (M = 0.40, SD = 0.49) and help to find an inclusion-oriented solution (M = 0.32, SD = 0.47) more frequently than find an alternative solution for the excluded student (M = 0.10, SD = 0.31), mediate between students (M = 0.17, SD = 0.37), or talk to students (M = 0.20, SD = 0.40).
Further, there was an interaction effect of situational information and anticipated action, F (4.82, 535.22) = 3.74, p = .0031, ηp² = .03. More specifically, teachers stated that they would ask for reasons or help to find an inclusionoriented solution more frequently in the condition where no additional situational information was provided. See Table  5 for all means and standard deviations.

Discussion
Teachers have been shown to impact their students' attitudes and behaviors with their expectations regarding different social groups in the classroom. Thus, it is important to study teachers' reactions to interethnic social exclusion. Especially with regard to the integration of refugees as a specific group of immigrants, teachers play an important role, as refugee youth and children typically make their first social contacts at school -very early after their arrival in Germany, and teachers need to deal particularly prudent with exclusion in this context. Therefore, the current study examined whether teachers' ratings of hypothetical exclusion scenarios differ based on an excluded student's ethnic origin (Syrian refugee vs. German) and/or information about the prior norm-violating behavior of the excluded student.
The evaluation of the exclusion situation, the likelihood to intervene in such a situation, and teachers' anticipated actions were examined.

The role of situational information
In line with the SRD framework, the current study confirms our assumption that situational information about the exclusion situation plays an important role for pre-and in-teachers' ratings of social exclusion. Without any further information about the situation, teachers clearly rejected social exclusion and had a strong tendency to intervene in the situation. This is in line with prior findings that observers generally reject exclusion if they do not see any obvious reason for it (Rudert et al., 2020). However, if teachers were provided with additional information about prior normviolating behavior (i.e., the excluded student had insulted another student earlier in class), they rejected the exclusion behavior less strongly and were less likely to intervene. We interpret this as an indication that teachers weigh up moral and social-conventional concerns and understand norm-violating behavior as a valid (or at least an understandable) reason for excluding a student who has engaged in such behavior. They thus tend to consider leaving the situation to the students more often compared to situations where they do not have any further information and strongly reject exclusion due to moral considerations.
However, we cannot know for sure if the teachers really take the prior norm-violating behavior as justification for their ratings in the sense of an understandable reason for exclusion as elaborated by Rudert and colleagues (2020). Another interpretation could be that in the condition with the insult observation getting this additional piece of information just made the teachers aware of how little they know about the situation and that they cannot make their decisions without further knowledge (which might not have happened to those in the condition without additional situational information). However, this would be in contrast to prior studies with scenarios without the additional situational information in that teachers strongly expressed their lack of information, too (Beißert & Bonefeld, 2020;Beißert et al., 2021). This shows that teachers -also in the condition without an additional piece of information -are very aware of the little information they have.
But no matter which specific interpretation is correct, what we can clearly see is that the provision of additional situational information -in our context the prior insult of the student as a typical norm-violating behavior -makes the situation less clear and more difficult to rate for our participants. Thus, the current results demonstrate that additional situational information impacts teachers' ratings of social exclusion situation.
In line with this, situational information also afffected teachers' reasoning about their decisions and anticipated actions. As long as they did not know about much about the exclusion situation, they had a strong focus on social inclusion as a social norm and tried to find inclusion-oriented solutions (moral domain). However, this focus faded a little when teachers were informed about the norm-violating behavior of the excluded student; then mediation and de-escalation became the main motives for them to intervene (social-conventional domain).
Interestingly, the provision of additional situational information had a much bigger impact on the evaluation of exclusion than on the likelihood to intervene. This underlines the fact that persons' perception of a situation does not necessarily predict their behavior (Ajzen & Fishbein, 2000;Glock & Krolak-Schwerdt, 2014). This means that, although the teachers evaluated the exclusion as more acceptable, they were still quite likely to intervene. This might be due to a sense of responsibility based on their educational mandate in school. However, it must be kept in mind that we did not assess real behavior, but only ratings of behavioral intentions and anticipated actions. And although there is evidence that the behavior specified in response to vignettes corresponds with behavior in real-life situations (Hainmueller, Hangartner, & Yamamoto, 2015), it is not clear to what extent this is the case in the current context.
Taken together, these results demonstrate how important it is to consider and examine situational information in the context of social exclusion. As proposed by the SRD model, teachers integrate situational information into their considerations, and this can even lead them to reconsider aspects of the excluded person such as their origin. This will be discussed in the next section.

The interplay of situational information and ethnic origin of the excluded student
As expected, participants' ratings of exclusion scenarios differed based on the ethnic origin of the excluded student. Namely, pre-and in-service teachers evaluated the exclusion of a Syrian refugee student as more negative than the exclusion of a German student. However, this difference was mainly driven by the interplay of the excluded student's origin and additional situational information, as the difference between the Syrian and the German student was only significant if teachers were provided with additional information about prior norm-violating behavior of the excluded student. That is, when teachers do not have any situational information about the reasons for exclusion, they generally reject exclusion and are likely to intervene. In this case, they justify their intervention with the general norm of inclusion, equity, and fairness. This holds for both protagonists and is in line with prior research (Beißert & Bonefeld, 2020;Kollerová & Killen, 2021). However, the situation changes if teachers know that the excluded student has engaged in prior norm-violating behavior. Then, they accept exclusion much more and are less likely to intervene in the situation. In addition, in this case, they also make a difference between the two protagonists. However, even though they recognize that there is a reason to evaluate the exclusion of both protagonists as more acceptable, the effect of additional situational information is stronger for the German protagonist. Further, in the condition where teachers learned about the protagonist's prior norm-violating behavior, they still considered the exclusion of a Syrian refugee to be more reprehensible than the exclusion of a German student. We understand this to mean that teachers take into account the vulnerability of refugee children in these situations. Even though they might consider exclusion to be justified to some extent, they may still feel empathy for a Syrian refugee, acknowledging that exclusion might affect a refugee child much more strongly than his or her native peers. This again reflects how teachers balance socialconventional with moral considerations when evaluating social exclusion.

Limitations and methodological discussion
The current study provided first evidence on the role and interplay of situational information and an excluded student's ethnic origin for teachers' ratings of social exclusion scenarios. However, the study has some limitations which should be kept in mind when interpreting the results.
First, the method should be discussed. The experimental approach using hypothetical scenarios clearly has the advantage that the influence of the variables of interest can be systematically investigated, allowing causal interpretations of the results and enabling us to focus on the sole effect of situational information and refugee status in the participants' ratings which is difficult to realize in field studies. However, this approach, comes at the cost of limited ecological validity and bares the risk of social desirability. The scenarios explicitly state that the protagonist is a Syrian refugee, which might lead to self-deception and impression management among respondents (Paulhus, 2002). As being tolerant and inclusive are generally favorable traits (Mann & Kawakami, 2012), the participants might have answered the questionnaires in a socially acceptable manner in order to shed a positive light on themselves. Although the use of written scenarios has been associated with low external validity (Eifler & Petzold, 2020) and the risk of social desirability, there is evidence that, at least for children and adolescents, evaluations of hypothetical social situations and reasoning related to such situations correspond to those in real-life situations and thus are comparable (Mulvey et al., 2018;Turiel, 2008). Yet, it is unclear if this also holds for adults, or more specifically for teachers. Thus, future studies should assess measures of social desirability in order to include them as control variables in the analyses. Additionally, observational studies in naturalistic settings or at least retrospective studies about real exclusion situations in class would be very helpful to replicate the current findings in real-life situations and to examine the extent to which selfreports correspond with actual behavior.
A further restriction is that we only used one specific scenario, so the generalizability of the results is limited. For the current research questions, we did not want to apply a within-subjects design in order to avoid contrast effects. When participants have to rate several scenarios, they might adapt their answers in relation to the scenario they rated first. However, future research should try to replicate this study using several scenarios in order to replicate the current findings across different situations and contexts.
Another issue is that the Syrian nationality and the status of being a refugee are confounded in the scenarios. Theoretically, these two groups could be associated with different stereotypes (Froehlich & Schulte, 2019;Kotzur et al., 2019). However, as the majority of Syrians in Germany are refugees, and the group of Syrians without a flight background is very small (DESTATIS, 2019), it is not very helpful to distinguish between Syrian refugees and other Syrian immigrants in our scenarios. Additionally, we assume that the refugee status might overshadow the nationality. Nevertheless, as it has been shown that specific stereotypes are linked to Syrian refugees in particular (Kotzur et al., 2019), further research should replicate this study and include refugees from other nationalities in the scenarios.
Also, it would be interesting to explore the role of teachers' prior personal and professional contacts to refugees. For instance, some teachers may have already had students from refugee families in their classes, or others may have refugees among their friends or neighbors, raising their awareness of their situation. Thus, future research should also explore the role of the quantity and quality of teachers' contacts to refugees.
Another interesting aspect for future studies would be a stronger focus on the identity (origin, gender) of the excluding students. In the current study, we did not further describe this group of students. However, prior research on social exclusion suggests that the ethnic composition of the excluding group might affect the observers' evaluation of the situation (Rudert et al., 2018). Thus, future studies should systematically vary the origin and composition of the excluding students in order to explore whether their identity impacts teachers' reactions to social exclusion among students.
Finally, some limitations related to the sample should be mentioned. Because we conducted an online experiment with minor knowledge about our participants, it is difficult to determine whether our effects only apply to a certain subgroup of pre-and in-service teachers that participated in our study, thus making our results harder to generalize. The sample itself included mostly pre-service teachers and comprised only very few male participants. There is evidence that pre-service teachers' attitudes about diversity change over the course of their training (Kumar & Hamer, 2013) and that there are gender differences in teachers' reactions to social exclusion (Beißert & Bonefeld, 2020). Thus, a replication with a larger sample of in-service teachers and a more balanced gender distribution is needed in order to generalize our results. In line with this, it would be helpful to recruit participants directly in schools (not online) as to realize a more controlled sample with more accurate information about their backgrounds and their school contexts. Besides this, it would be very interesting to include teachers with a migration background in the analyses as they might play an important role in the inclusion of children from ethnic minorities (Strasser & Steber, 2010). Especially in the context of social exclusion, teachers with a migration background are of great interest. As persons who themselves might have experienced (group-based) exclusion, it would be interesting to see if their reactions differ from teachers without an immigration history in the family.

Implications and conclusion
Overall, this study provides promising results regarding the integration and inclusion of refugees in German schools. Encouragingly, the current results did not indicate the presence of ingroup bias, stereotype-based or discriminatory decisions in pre-and in-service teachers' reactions to social exclusion. Quite the opposite, teachers seem to be particularly sensitive when it comes to the exclusion of refugee children. Even though situational information was generally more important for teachers' reactions to social exclusion, they did not completely disregard the origin of the student who was being excluded, as the effect of additional situational information was less strong for the Syrian protagonist. This shows that, although exclusion is possibly justified in some situations, teachers keep the specific situation of refugees in mind. This is very positive as teachers can establish social norms in class and thus contribute to the feeling of inclusion in schools (Verkuyten & Thijs, 2002). Thus, one approach to promote the successful integration of refugees in schools is to train teachers accordingly (German Federal Anti-Discrimination Agency, 2019).
Altogether, additional situational information was much more important for pre-and in-service teachers' reactions to social exclusion than the origin of the excluded student. This is important since, in real-life exclusion situations, teachers often have at least some background information. However, their knowledge might be incomplete, selective, or even biased by student characteristics. Also, what teachers are witnessing may be only part of the whole story. For instance, that one student insults another could already be a reaction and not the beginning of the matter. Thus, teachers should be encouraged to always analyze the reasons for exclusion in order to acquire a deeper understanding of the situation. Yet, given the severe consequences exclusion can have teachers should be encouraged to generally intervene in exclusion situations -no matter if exclusion is possibly justified in some situations.

Conflict of interest: Authors state no conflict of interest.
Financial Support: This research received no specific grant from any funding agency, commercial or nonprofit sectors.
Ethics Statement: Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. The study was conducted in accordance with the ethical guidelines of DGPs (German Psychological Society). All subjects gave written informed consent in accordance with the Declaration of Helsinki and in line with the General Data Protection Regulation of the European Union.