Graduate School of Psychology
RESEARCH MASTER’S PSYCHOLOGY THESıS REPORT
Teachers’ differing classroom behaviors: The role of emotional sensitivity and cultural tolerance
Ceren Su Abacıoğlu
Supervisor: prof. dr. Agneta Fischer Second supervisor: dr. Disa Sauter External Supervisor: prof. dr. Monique Volman
Research Master’s, Social Psychology Ethics Committee Reference Code: 2016-SP-7084
Abstract
Ethnicity is a key element of school segregation in the Netherlands, hindering society’s goals of social integration and equality in educational achievement. In order to investigate determinants of integration in schools, we turned to the classroom, where fulfillment of socio-emotional needs of pupils may affect their sense of integration, belongingness, and consequently their school engagement (e.g., Ryan & Patrick, 2001). The current study explored how teachers’ emotional sensitivity and cultural tolerance relate to teachers’ differing behavior towards pupils’ misbehaviors. It was expected that cultural tolerance and emotional sensitivity would predict differing behaviors of teachers towards pupils with different ethnic backgrounds, especially for severe reactions. Based on our results, we have rejected our hypotheses. However, our analysis yielded a significant relationship between emotional sensitivity and cultural tolerance of teachers, and our exploratory findings signal that emotional factors might be even more important than cultural tolerance for ethno-cultural intergroup relations.
Teachers’ differing classroom behaviors: The role of emotional sensitivity and cultural tolerance
Societal integration relies on positive relationships among different social groups (Blau, 1974). Improving social integration means reducing distance between individuals from different origins, by promoting interethnic friendships, diversifying social networks, and facilitating coexistence (Van Houtte & Stevens, 2009). Schools are ideal places for this integration to take place. However, with the influx of migrants during the last half-century, the Dutch school system reinforced segregation, which also hindered society’s goals of equality in educational achievement. One frequently used policy measure attempting to surmount these problems has been motivating schools and parents towards dispersing minority pupils in order to increase the number of schools with children of mixed ethnic backgrounds (Peters & Walraven, 2011). The goal was enhancing minorities’ educational achievement and occupational success by fostering integration and mutual understanding (Van Houtte & Stevens, 2009). Yet, empirical research shows that mixing pupils does not automatically lead to integration. Consequently, there seem to be few to no academic advantages to mixing pupils, especially if these pupils do not feel “at home” in school contexts (Agirdag, Demanet, Van Houtte, & Van Avermaet, 2011). The goal of this research was to investigate why the interethnic schools might have failed to have the desired effects, and to this end, we have focused on the teachers.
Social integration and Educational Achievement
Since World War II, there has been an influx of immigrants coming to the Netherlands mainly for economic reasons (Driessen, 2012). While this immigration was initially seen as a temporary situation, with the reuniting of families (Van Houtte & Stevens, 2009), the Dutch society has quickly evolved into a multi-ethnic society
(Rijkschroeff, ten Dam, Willem Duyvendak, de Gruijter, & Pels, 2005). This change in population was particularly noticeable in an educational context. Large numbers of immigrant children entered the Dutch schools starting from the 1980s (Driessen, 2012). Currently, over 50% of the pupil population in primary and secondary school in the large cities in the Netherlands is comprised of first and second generations of migrant pupils from Surinam, Turkey, Morocco, the Antilles and African countries. However, with this new development, schools became segregated along the lines of ethnicity. For instance, half of all primary schools in the four biggest cities of the Netherlands (Amsterdam, Rotterdam, The Hague, and Utrecht) are now considered “black schools”. In other words, more than 50% of the pupils have a non-western ethnic background, compared to the 14% of the total Dutch population (OECD, 2010). These numbers are noteworthy since early integration to society starts in the classroom and integration into social networks and institutions is likely to affect pupils’ engagement in these institutions during a school career, which is a strong predictor of educational achievement (Wu, Schimmele, & Hou, 2012).
Moreover, learning the language and the culture of the host society from an early age, through interethnic friendships, may ease access to networks that are predominately native Dutch (e.g., “higher” academic tracks and higher education) at later stages in life. Indeed, educational achievement of these pupils often lag significantly behind their ethnic majority peers already when starting the primary education, and it remains lower than of the ethnic majority peers over the years, limiting minority pupils’ chances in “higher” academic tracks and education. As a consequence, researchers, policy makers, and politicians had concerns about the effects this ethnic segregation had on educational achievement (e.g. for a discussion Gijsberts & Dagevos, 2005). If ethnic segregation continues such that minority and
majority members only have relationships within their own group, the education gap between the ethnic groups will grow in accordance with the Matthew effect (Merton, 1968; Van Tubergen, 2010). In other words, those who started off well will continue to do well, while those who did not will be unlikely to catch up, which deepens the gap between ethnic groups as their school careers continue.
Due to the findings aforementioned, policy makers strived (and are still striving) towards the dispersal of immigrant pupils, creating schools with children from mixed ethnic backgrounds, with the goals of fostering integration and enhancing minority pupils’ educational achievement (Driessen, Doesborgh, Ledoux, Van der Veen, & Vergeer, 2003). Although mixing increases the opportunity to form majority-minority group relationships, empirical research finds that mixing often fails to increase actual interethnic integration, however (Feld & Carter, 1998). Moreover, scientific evidence for positive effects of mixing on minority pupils’ educational performance is inadequate (Stark, 2011). Thus, an important question that comes to mind is why interethnic schools have failed to benefit integration and minority pupils’ educational achievement? In order to gain more insights into this societal issue, our study draws from two influential theories, namely the Contact Theory (Allport, 1954) and the Ecological Systems Theory (Bronfenbrenner, 1979).
Contact Theory
Schools, especially classrooms, are often considered places that can best promote interethnic integration of children since classmates spend a considerable proportion of their days together and can have a crucial effect on each others’ social development (Kassenberg, 2002). However, the opportunity of contact itself, in other words simple exposure, does not naturally lead to integration. Allport’s contact theory (1954) posits that contact may enhance interethnic relations by, for instance, reducing
prejudice. However, a substantial enhancement is only possible if there are adequate opportunities for people to get to know each other, and if this is supported by social and institutional authorities (e.g., teachers). Conversely, there can be drawbacks of interethnic contact, such as strengthening of group boundaries and confirming of previously held stereotypes and hostilities (Brown, 1995; Paolini, Harwood, & Rubin, 2010). Therefore, characteristics of the contact setting contribute to enhancing or inhibiting the positive effects of contact (Patchen, 1999; Pettigrew, 1998; Riordan, 1978; Stephan, 1987). As previously mentioned, the most frequent contact setting for young children of different ethnic backgrounds is the classroom. Classroom climate - the socio-emotional atmosphere of the learning environment (Moos, 1974) – being a prominent characteristic of the contact environment can thus play a major role in enhancing interethnic relations (Yoneyama & Rigby, 2006), which would in turn, facilitate integration.
Ecological Systems Theory
Classrooms are by definition social environments, and the social interactions with classmates and teachers shape the learning process (Urdan & Schoenfelder, 2006); thus, the quality of relationships within a classroom may be essential for pupils’ motivational and learning outcomes (Ryan & Patrick, 2001). This view is supported by the influential Ecological Systems Theory (or Bioecological Systems Theory; Bronfenbrenner, 1979), which posits that a person's development is affected by the system of relationships in their surrounding environment. Schools and in particular classrooms are part of the microsystem in Bronfenbrenner's theory.
Microsystem is the first and the most influential of the five levels of the ecological
systems, referring to the groups and institutions that are most immediate to the child and that can directly and greatly impact his/her development. Within the classroom
microsystem, children interact with peers and teachers, and the quality of these interactions may substantially influence children’s developmental outcomes (e.g., educational achievement).
Classroom cultural climate and pupil outcomes
In line with the two aforementioned theories, there is increasing evidence suggesting that the climate of classrooms, specifically the extent to which it fulfills children’s need for being socially connected, influences pupils’ school engagement and subsequent educational achievements (e.g., Eccles, Wigfield, & Schiefele, 1998; Ryan & Patrick, 2001; Connell, 1990). Pupils who have better relationships with peers and teachers (Klem & Connell, 2004), who feel socially supported and accepted, are also more likely to be engaged in school and learning activities (Deci & Ryan, 1985; Wentzel, 1997), whereas exclusion and conflictual relationships in the classroom undercut engagement (Merton, 1953; Newmann, 1991). School engagement, referring to the quality of pupils’ connection or involvement with school-related people, activities, goals, values, and the places that contain these, has a strong influence on academic retention, achievement, and resilience (Skinner, Kindermann, & Furrer, 2009).
The classroom climate, however, might not benefit all pupils equally well. The quality of relationships, thus the socio-emotional atmosphere of the learning environment, may suffer from stereotypical expectations and/or cultural differences. Minority pupils, for instance, are more likely to be victims of discrimination, social exclusion, and name-calling than Dutch pupils as reported from a nation-wide study (Verkuyten & Thijs, 2002). Victims of classroom discrimination find few occasions for peer-communication, feel rejected, isolated, and not integrated within the class (Cerezo & Ato, 2010). As mentioned in the previous paragraphs, research in a variety
of areas has documented that social connectedness has an important impact on school engagement and subsequent educational achievement (e.g., Eccles, Wigfield, & Schiefele, 1998; Ryan & Patrick, 2001; Connell, 1990). In light of these findings, we believe that the social integration and educational achievement issues of minority pupils are inter-correlated. In order to gain more insights into why the interethnic schools have failed to have the desired outcomes, there is a need for investigating the cultural aspects of social connectedness, which we consider as interethnic integration, in relation to school engagement. We therefore specifically focus on and conceptualize aspects of the classroom climate that are important for integration as classroom cultural climate. This is an abstract term that we introduce in order to highlight the importance of cultural aspects of a classroom for minority pupils’ developmental outcomes.
The role of teachers
Teachers are not only specialists who deliver the curriculum. They also help creating the classroom cultural climate in which peer relationships develop wherein pupils pursue social goals next to academic goals (Juvonen & Murdock, 1995; Urdan & Maehr, 1995; Wentzel, 1993). Ideally, teachers should help pupils reach these social goals, by setting rules and norms for pupils’ social behavior, giving explicit messages about pupils’ interaction with their peers (Ryan & Patrick, 2001), and acting as a role model in how to engage in convenient and respectful communication, and prosocial behavior (Jennings & Greenberg, 2009). Indeed, research found that the undesired outcomes of mixed education (e.g., ethnic bullying) can be counteracted when teachers react constructively to negative incidents (Verkuyten & Thijs, 2002). Moreover, ethnic composition, which is thought to be one of the reasons why minority pupils in mixed education are subjected to more bullying, does not affect
victimization when the number of interethnic conflicts are taken into account (Agirdag et al., 2011), indicating that the problem is not in the numbers but in the quality of relationships, which is highly dependent on the teachers’ reactions (Verkuyten & Thijs, 2002). Thus, teachers set the tone of the classroom by implementing behavioral guidelines in ways that should promote intrinsic motivation, and therefore facilitate reaching both social and academic goals.
However, when teachers lack the necessary resources to adequately manage social and emotional challenges, it is directly manifested in children’s motivation and educational achievement (Marzano, Marzano, & Pickering, 2003). Teachers, thus, need to have awareness, knowledge, and skills that will allow them to better comprehend and teach their pupils while promoting an open and tolerant classroom climate towards being different. This is especially important for minority pupils given that most teachers in the Netherlands are native Dutch (Thijs, Westhof, & Koomen, 2012) and their cultural misperceptions may negatively affect conceptions about pupils’ reactions, skills, and abilities (Banks, 1989). Research shows that teacher expectations affect pupils’ achievement (e.g., via stereotype threat or differing treatment) and thereby worsen the achievement gap (Brophy & Good, 1970; Dusek & Josef, 1983; Good & Weinstein, 1986; McKown & Weinstein, 2002). Moreover, teachers may affect their pupils’ outgroup attitudes by the norms they express (e.g., cultural diversity, multiculturalism; Banks & Banks, 2004; Vogt, 1997), and they can influence classroom climate with their attitudes and expectations (Graybill, 1997). In order to look into the mechanisms by which teachers may communicate certain attitudes and beliefs, we have focused on emotional and cognitive predictors of teacher behaviors.
The current study
The current study focuses on answering the research question “what is the teachers’ role in classrooms’ cultural climate?” in order to investigate certain teacher factors that can influence pupils’ sense of integration and their school engagement. Our target group was primary school teachers. Teachers’ behaviors in pupils’ early years of schooling can have implications for children’s entire life course considering the high developmental importance of this period (Rocque & Paternoster, 2011). The most contact between children from mixed ethnic backgrounds takes place in primary schools and the classrooms are crucial for interethnic integration (Kassenberg, 2002). Thus, understanding the teacher behaviors that can affect minority pupils’ integration during this period is very important, not least because ethnic minority pupils already lag behind their peers starting at primary school (Driessen, 2012) and integration might play a big role in this.
We investigated both cognitive and emotional factors as predictors of teachers’ classroom behavior: emotional sensitivity and cultural tolerance. The first set of key variables related to teachers’ cultural tolerance, which is their positive or negative beliefs and attitudes about other cultures (Gasser & Tan, 1999). Teachers’ own beliefs about cultural diversity and their awareness of cultural differences can play important roles in creating an open and tolerant classroom climate (Garcia, 1994) since they value cultural diversity and are more likely to address multicultural issues in their classrooms (Ponterotto & Pedersen, 1993). We operationalized this by using a multiculturalism questionnaire.
The second set of key variables related to emotional sensitivity. We defined emotional sensitivity as the ability to recognize and understand one’s own and others’ emotions. Emotion-related cues have been proposed to regulate interpersonal
interactions (e.g., Fischer & Manstead, 2008; Fridlund, 1994; Keltner & Haidt, 1999; Scherer, 1980, 1988, 1994; Van Kleef, De Dreu, & Manstead, 2004). The interpretation of these cues, however, might differ depending on whether the displayer is an ingroup or outgroup member (Tajfel & Turner, 1986; Turner Hogg, Oakes, Reicher, & Wetherell, 1987). For example, teachers may interpret looking away either as a sign of shame or indifference, depending on the ethnic background of the pupil (Kommattam, Jonas, & Fischer, 2016). Responses to others’ emotions can consequently serve a distancing or a bonding function (Fischer & Manstead, 2008; Keltner & Haidt, 1999). Attending to and correctly interpreting emotional signals is part of a more general emotional intelligence, also including the ability to recognize, understand, and manage one’s own emotions (Salovey & Mayer, 1990). We operationalized this by using a well-validated self-reported emotional intelligence questionnaire, as well as an emotion recognition test that measures the ability to recognize emotions in others.
We expected that cultural tolerance and emotionally sensitivity would predict teachers’ differencing behaviors towards pupils with different ethnic backgrounds. This expectation was based on the findings that pupil behavior may be misinterpreted and misunderstood when certain cultural differences are poorly acknowledged. There is evidence that teachers often classify disruptive behaviors differently for majority and minority group pupils (Ferguson, 2000). Thus, teachers’ sensitivity to cultural aspects would benefit their behaviors when dealing with problematic situations that are prone to misinterpretation (Ladson-Billings, 1995; Radstake, 2009), and teachers’ ability to recognize and deal with pupils’ emotions would help dealing with these situations. In addition, teachers’ ability to manage their own emotions would strongly determine the reactions they would give (Brackett & Katulak, 2006). Thus, in this
exploratory study, our aim was to examine the relationship between the two sets of variables (cultural tolerance and emotional sensitivity) and whether these predict teachers’ differing classroom behaviors towards pupils with different ethnic backgrounds. Different classroom behaviors consisted of self-reported behavioral reactions, either rewarding or punishing, to problematic situations.
In order to back up these assumptions, we first conducted a pilot study, which is described below in more detail (see Measures section). In this study, we have asked primary school teachers the kind of behaviors they find problematic in the classroom and the kind of reactions they give to these behaviors. Our analysis revealed that the common teacher reactions to pupils’ classroom misbehaviors only included punishing behavior and not rewarding behavior, which gave us more insights into the direction of the difference in teacher behaviors. Based on previous studies that consistently found that ethnic minority pupils are more frequently punished in schools (Skiba & Rausch, 2004; Fenning & Rose, 2007), and on the evidence that teachers respond more harshly towards misbehavior of ethnic minority pupils than towards identical behavior of ethic majority pupils (Ferguson, 2000); we have adjusted our hypotheses as follows.
(H1) Cultural tolerance and emotional sensitivity predict differences in frequency of punishment between majority and minority pupils, such that the difference in frequency would decrease when cultural tolerance and emotional sensitivity would increase.
(H2) The predictive power of cultural tolerance and emotional sensitivity should be higher for more severe punishments.
Lastly, we wanted to explore whether emotional sensitivity would have an additional value for predicting teachers’ differing classroom behaviors after
controlling for cultural tolerance. Notwithstanding the success of the previous studies that focused on the cognitive factors (e.g., attitudes, beliefs, stereotypes) that indicate cultural tolerance (e.g., Kawakami, Young, & Dovidio, 2002; Stangor & Lange, 1994), there is important empirical evidence in the literature pointing out the significance of emotions in interpersonal interactions and the studies focusing on only cognitive factors fail to account for emotions when considering teacher-pupil interactions in a cultural context. To the best of our knowledge, the current study is the first to look at emotional factors as possible predictors of teachers’ differing behaviors towards pupils with an ethnic minority versus majority background.
Method Participants
Our target group was primary school teachers. The participants were recruited from cities in all regions of the Netherlands through a Facebook advertisement targeting our specific sample in addition to approaching schools using the Dutch national listing of primary schools. In total 591 primary school teachers participated in the study. Six participants were excluded from our analyses due to missing data, and 1 participant was excluded for being a secondary school teacher, reducing our sample to 52 participants. Fifty-one participants were native Dutch, whereas 1 participant indicated being Hungarian (49 females, mean age = 43.44, SD = 10.81).
Procedure
The two sets of key variables explicated above were investigated in relation to each other and teachers’ differing classroom behaviors. To this end, we first conducted a pilot study in which we asked teachers about the kind of behaviors they
1 Our aimed sample size was between 56 to 87 primary school teachers based on the power analysis
using Gpower (Bonferronni corrected α = .01, power = .80; Gpower 3.1; Faul, Erdfelder, Lang, & Buchner, 2007). The power analysis was conducted for an effect size of f2 = .21 - .35, based on similar research by Christ and colleagues (2014) investigating contextual effects of positive intergroup contact on prejudice.
find problematic in a classroom and how they react to those behaviors. The answers in this pilot study were used to form vignettes that depict possible problematic situations and reactions in classrooms, which were used to assess teacher behaviors (see Measures).
For the main study, participants filled in the online survey composed of the three measures2 and that lasted 20 minutes. In order to recruit participants, we have offered 50 euros to one sixth of the participants that were randomly chosen with a lottery. The study was performed according to the ethical committee guidelines of the University of Amsterdam and the participation to the study was voluntary and anonymous. As such, participants indicated their consent before starting the survey. Participants were told that the objective of the study was to investigate emotional predictors (e.g., emotion recognition) of teacher behaviors that might affect pupils’ educational outcomes. At the end of the survey, participants were thanked and debriefed about the true nature of the study.
Measures
Pilot study pupil behaviors-teacher reactions. We have conducted an online pilot study with 22 primary school teachers in which we have asked them the kind of behaviors they find problematic in a classroom and how they react to those behaviors. The answers were analyzed in order to identify the most common responses both for the problematic situations and teacher reactions. In line with the previous research on common classroom misbehaviors (Sun & Shek, 2012), the themes included, 1) not cooperating with others (e.g., during a class exercise), 2) verbal/physical aggression, 3) hindering others (e.g., having disruptive conversations), 4) disrespecting teachers
2 Since all the measures were presented in Dutch, The Schutte Self-Report Emotional Intelligence Test
(SSEIT; Schutte et al., 1998) and Teacher Multicultural Attitude Survey (TMAS; Ponterotto, Baluch, Greig, & Rivera, 1998) have been translated from English to Dutch, and back-translated for accuracy. All the three measures were brought together under one link and administered using the Qualtrics online survey system.
(e.g., rudeness/Talking back, arguing with teacher), 5) non-attentiveness/daydreaming/idleness/sleeping. Building on previous research, we have also included another common classroom misbehavior, namely 6) being out of seat (e.g., running away from the classroom). The common teacher reactions to these behaviors included, 1) warn/express disapproval, 2) give detention/ time-out/ send pupil out of class, 3) discuss the misbehavior and the class rules with the pupil in private (immediately or later on), 4) contact the parents if the behavior persists (after 3 times). These reactions fell on a continuum such that the severity of the reactions increased depending on the severity of the pupil misbehavior. To illustrate, teacher first would warn the pupil but would not have the need to send the pupil out of class if the misbehavior is trivial. To have a better understanding on where the teacher reactions fall on this continuum, we have also added the reaction “do nothing” as a starting point (Figure 1). The answers in this pilot study have been used to form vignettes that were used to assess self-report teacher behaviors in the classroom.
Teacher’s Differing Behavioral Reactions. Teacher behaviors were measured by providing participants with twelve vignettes in total, describing common classroom classroom misbehaviors (see the pilot study results), 6 of these involved native and 6 involved non-native pupils, signaled by the pupils’ names (e.g., Pieter, Hassan respectively). There were two slightly different descriptions of each category of misbehaviors. For each vignette (see Appendix A), the same reaction options (five in total) were provided. The participants were asked to indicate the extent to which
Do nothing Warn Send out of
class
Discuss the misbehavior and class rules
Contact the parents
they would engage in each of the reactions (e.g., “do nothing”, “send out of class”) on a scale from 0 to 100 (0: never, 100: always). The presentation of the vignettes was counterbalanced. Every participant received both versions of each vignette randomly with either a native or a non-native name (i.e., either version 1 for vignette 1 as native or version 2 for vignette 1 as native) such that half of the versions 1 for each vignette were presented with a native name and the other half with a non-native name. Independent from this randomization, half of the matching situations were randomly assigned a male name while the other half was assigned a female name (either native or non-native name depending on the version). The non-native names were chosen from Moroccan names for they are one of the largest ethnic minority groups in the Netherlands, and they face the highest level of prejudice and discrimination (Gijsberts & Dagevos, 2010). The names were as follows. Native-female: Marloes, Claudia, Anouk; native-male: Pieter, Jeroen, Dennis; non-native-female: Fatima, Naima, Meryem; non-native-male: Hassan, Ahmed, Farouk. As a last step, the presentation orders of the 12 situations were randomized per participant. For each participant, the average difference scores between native and non-native versions for all the 6 matching situations were calculated per reaction. Moreover, per reaction, average difference scores were calculated separately for male and female pupils, as well as native and non-native pupils for each teacher reaction.
Teacher’s Cultural Tolerance. We have used Teacher Multicultural Attitude Survey (TMAS; Ponterotto, Baluch, Greig, & Rivera, 1998) to assess the cultural tolerance of teachers. TMAS (see Appendix B) comprises of 20 statements, seven of which are reverse-scored (3, 6, 12, 15, 16, 19, and 20) (α = .86). Participants replied to the statements on a 5-point Likert-type scale (1: strongly disagree, 5: strongly
agree). For our analyses, mean ratings for all 20 items were calculated per participant and the data have been treated as continuous.
Teacher’s Emotional Sensitivity. We have used (1) The Schutte Self-Report Emotional Intelligence Test (SSEIT; Schutte et al., 1998), and (2) The Amsterdam Emotion Recognition test (AERt; Van der Schalk, Hawk, Fischer, & Doosje, 2011) to assess teachers’ emotional sensitivity.
(1) SSEIT is a 33- item questionnaire (see Appendix C) developed based on the Emotional Intelligence (EI) model by Salovey and Mayer (1990). It measures general EI on a 5-point Likert-type scale (1: strongly disagree, 5: strongly agree) using four sub-scales, namely, utilizing emotions, emotion perception, managing others’ emotions, and managing self-relevant emotions (α = .90). Amongst the 33 items, three of them are reverse-scored (5, 28, and 33). For our analyses, mean ratings for all 33 items from all the subtests were calculated per participant.
(2) AERt assesses the ability to correctly infer basic emotional expressions via prototypical communicative facial signals (see Appendix D). Stimuli are derived from the Amsterdam Dynamic Facial Expression Set, including both North-European and Mediterranean faces displaying anger, contempt, fear, joy, pride, shame, disgust, surprise, and sadness (α = .90, .93). For each of the nine emotions, one male and one female North-European and Mediterranean face was randomly presented to the participants (36 in total). The emotion displays were similar for all faces reflecting low intensity, which were interpreted according to the facial action units (for unit codes Ekman, Friesen, & Hager, 2002; Figure 3). The provided answers for the emotion displays were as follows: anger, contempt, fear, pride, shame, disgust, or something else. Participants’ percentages of correct responses were calculated (correct responses to total responses). Moreover, because we included faces from
different ethnic groups, we were also able to compute an Index of Outgroup Recognition based on previous research that has shown that people are less accurate in recognizing outgroup emotions (Elfenbein & Ambadi, 2002). This was done by separately calculating the percentages of correct responses for native and non-native faces and dividing one with the other (correct responses for native relative to correct responses for non-native faces).
Figure 2. Disgust at different levels of expression intensity (Wingenbach, Ashwin, &
Brosnan, 2016).
Demographics. At the end of the survey, participants were asked to report on certain variables that might affect teachers’ behaviors. The variables included their age, sex, gender, ethnicity, years of teaching experience, minority percentage in their school, minority percentage in their classroom, and the grade they teach. Some of the variables have been excluded from the analyses (explained below), and the variables that were included have been treated as continuous.
Results Testing for covariates
Prior to proceeding with the main data analysis, we investigated the relationships between the demographics (as they are possible covariates) within themselves and with the dependent variables (Figure 4). Teachers reported on their age, sex, gender, ethnicity, years of teaching experience, minority percentage in their
school, minority percentage in their classroom, and the grade they teach. Because there were only three male participants, and only one non-Dutch participant, gender and ethnicity (respectively) were excluded from the analyses due to low power. Moreover, since most of the teachers taught pupils from various age groups, “the grade teachers teach” was also excluded as a possible covariate.
To begin with, ethnic composition of the classroom and ethnic composition of the school both significantly correlated with the dependent variable differing teacher reactions in discussing the misbehavior with the child, and thus should be included in the analyses as covariates, r(50) = -.31, p = .03 and r(50) = -.30, p = .03 respectively. The ethnic composition of the classroom and ethnic composition of the school,
Figure 4. Correlations between demographic variables and DVs plotted.
Note that the circles symbolize significant correlations (blue for positive and red for negative), with smaller and lighter-colored circles indicating weaker relationships.
however, correlated strongly with each other, r(50) = .95, p = .00. Because it is very likely that they have measured the same construct, it was logical to continue with only one of the two variables as a possible covariate. In our case, continuing with the classroom ethnic composition is more appropriate since our research focuses on teachers’ classroom behaviors. The rest of the demographics did not significantly influence the dependent variables, and therefore were ruled out as covariates from further analyses in order not to lose power (Becker, 2005).
Confirmatory Analyses
Our planned analyses included investigating relationships between the independent variables, and examining how the independent variables relate to the dependent variables (Figure 5).
Figure 5. Model of planned analyses. The arrows represent the investigated
relationships.
Relationships between IVs
Looking at the correlation matrix of independent variables (Figure 6), we could see that the mean scores on cultural tolerance measure TMAS and the emotional intelligence measure SSEIT correlated significantly, r(50) = .43, p = .00. Moreover, mean TMAS scores significantly correlated with ethnic composition of the classroom, r(50) = ..37, p = .01. After correcting for multiple comparisons with the Bonferroni method, however, the correlation between TMAS and ethnic composition of the classroom, r(50) = .37, p = .07 was no longer significant, while the means scores on TMAS and SSEIT remained significant, r(50) = .43, p = .01.
Figure 6. Correlations between the IVs plotted. Note that the circles symbolize
significant correlations (blue for positive and red for negative), with smaller and lighter-colored circles indicating weaker relationships.
Multivariate Multiple Regression Analyses
In order to predict values of our dependent variables from predictor variables, we conducted three separate multivariate multiple regression analyses3. All the analyses were conducted with the same dependent variables while the independent variables have changed. The dependent variables were the averaged difference scores between native and non-native versions of all the 6 matching situations, calculated per reaction (thus 5 in total namely, do nothing, warn the pupil, send out of class/expel, discuss the behavior with the pupil, and contact parents). At this point, the reader should focus on the Wilk’s Lambda –the most commonly used multivariate test statistic (Stevens, 2012).
(1) First, the four independent variables class ethnic composition, cultural tolerance, overall emotion recognition, and emotional intelligence were simultaneously regressed on the five averaged different scores from each of the five reactions.
The overall model fit was non-significant as indicated by the multivariate test of significance (Table 1). Although it is conventional to stop with further investigations if the multivariate test is non-significant (Braver, MacKinnon, & Page, 2003), we further explored the univariate test of significance, together with the raw regression coefficients4 because of the exploratory nature of this study. However, the univariate follow up analyses were non-significant after correcting for multiple
3 Before starting with our analyses, the assumptions of multivariate multiple regression were tested for
each of the models described below. The assumption of multivariate normality of residuals did not hold for the models. However, only three data points were not in line with normality, we have thus decided to continue with our analyses (see Appendix E for values and visualization). Yet, the reader should be aware of this problem, and as a solution, prospective research can consider a better-tailored model for non-normality.
4 Significance of the regression for each DV individually, and raw regression coefficients for
predicting each DV from the 4 IVs.
comparisons with the Bonferonni method (Please see Table 4.1 and 4.2 for the non-corrected test results in the tables section after References).
(2) In the second analysis, the variable overall emotion recognition was exchanged by the emotion recognition index (based on the different perception of ingroup and outgroup faces), while keeping the other three independent variables and the five dependent variables as described for the first analysis.
The overall model fit was non-significant as indicated by the multivariate test of significance (Table 1). Exploring the univariate analyses, there were no significant influences on the DV’s after correcting for multiple comparisons with the Bonferonni method (Please see Table 5.1 and 5.2 for the non-corrected test results in the tables section after References).
(3) In the third analysis, class ethnic composition, cultural tolerance, and emotion recognition index (relative percentage of correct AERt scores from native vs. non-native faces) have been regressed on the five dependent variables. This analysis left out emotional intelligence variable due to its significant correlation with cultural tolerance reported in previous paragraphs, given that the likelihood of the variance accounted for in dependent variables is not being apportioned correctly increases with every independent variable added to the regression.
Although the overall model fit improved (Table 1), the improvement was not significant. Exploring the univariate tests of significance and the raw regression coefficients no significant influences were found after correcting for multiple comparisons with the Bonferonni method. Since the values were comparable to the ones of the second analysis, they are not further reported in a table.
Multivariate Test of Significance
Note. * p < .05, ** p < .01.
Exploratory Analyses
Ingroup and Outgroup Emotion Recognition
Previous research shows that people are less accurate in recognizing outgroup emotions (Elfenbein & Ambadi, 2002). In order to replicate this finding, we have compared teachers’ emotion recognition abilities for native and non-native Dutch faces. However, testing this hypothesis showed that emotion recognition from native Dutch faces (M = .64, SD = .12) was not significantly different from emotion recognition from non-native Dutch faces (M = .65, SD = .13); t(101.29)= -.13659, p > .05. Conversely, emotion recognition for the non-native Dutch faces was slightly better than emotion recognition for the native Dutch faces, which is a finding that is not in line with the previous research.
Frequency of Punishment for Native versus Non-native Pupils
In order to investigate whether our results are in line with the previous literature indicating minority pupils are more likely to be punished and more harshly punished (Skiba & Rausch, 2004; Fenning & Rose, 2007; Ferguson, 2000), a Hotteling’s T analysis was employed, which compared the frequency of punishment for the five possible reactions (do nothing, warn, send out of class/expel, discuss the misbehavior, contact parents) between native and non-native groups. The two groups did not significantly differ from each other, F(5,98) = .173, p > .05. For further exploring the difference between the two groups for each dependent variable Analysis Wilks
Lambda
Approx. F Hypoth. Df Error df Significance of F
1 .62335 1.0994 20 143.56 .3564
2 .54876 1.4237 20 143.56 .1202
separately, individual t-tests were performed; however, since none of the dependent variables significantly differed from each other for the two groups, they are not further reported.
Frequency of Punishment for Male versus Female Pupils
Previous research shows that male pupils’ misbehavior are more frequently punished than of female pupils (Reynolds, Miller, & Weiner, 2003). Testing this argument in our sample, we have conducted paired-samples t-tests to compare the frequency of giving each of the five teacher reactions for male and female pupils. For none of the teacher reactions except for “contacting the parents” did males and females differ from each other within the native and non-native categories (e.g., native-Dutch females vs. native-Dutch males). Moreover, females in native and non-native categories, as well as males in non-native and non-non-native categories did not significantly differ from each other (e.g., native-Dutch females vs. non-native females).
For native Dutch pupils, however, there was a significant difference in the average frequencies for contacting the parents for males (M = 60.45, SD = 28.74) and females (M = 50.27, SD = 30.94); t(101.45) = 1.7384, p = .043. Nevertheless, for non-native pupils, the difference in the average frequencies for contacting parents for males (M= 61.03, SD = 26.66) and females (M = 58.10, SD = 30.21) was not significant; t(100.45) = .52321, p = > .05. As a follow-up, we looked whether the difference in frequencies between contacting the parents of males and females was significantly greater for native-Dutch (M = 10.18, SD = 21.38), pupils than of non-native pupils (M = 2.93, SD = 24.91). The difference was not significant; t(99.707) = 1.5939, p > .05.
Additive Predictive Power of Emotional Sensitivity
We aimed to explore whether emotional sensitivity has additive predictive value of teachers’ differing classroom behaviors after controlling for cultural tolerance. However, since both cultural tolerance and emotional sensitivity failed to predict teachers’ differing behaviors, we did not further explore the additive predictive value of emotional sensitivity after controlling for cultural tolerance. Instead, we have conducted a forward stepwise regression, which implies starting off with no variables in the model and choosing the variables5 that best improve the model until the model cannot be further improved6. This analysis was performed for each of the dependent variables. The forward stepwise regression based on Akaike Information Criteria (AIC) values (Table 2) revealed that for most of the teacher reactions, emotional sensitivity variables improved the regression model. Since AIC (information criterion) is not a “test”, we cannot make further inferences about the significance of the change in the AIC values.
Table 2
Change in AIC values
Variables AIC
DV1 Difference score “Do nothing” Start
+ emotion recognition index + emotional intelligence
DV2 – Difference score “Warn the pupil” Start
None of the variables improved the model DV3- Difference score “Send out of class/expel”
Start
+ emotion recognition index
227.69 225.60 223.84 276.59 276.59 259.31 257.72
5 The set of candidate predictor variables included: classroom ethnic composition, cultural tolerance,
overall emotion recognition, emotion recognition index, and emotional intelligence.
6 This is a procedure that is automatically undertaken by the statistical program. The sequence of
inclusion of a variable is determed by examining which variable improves model the most, followed by the next “best” variable that improves the model the most, until inclusion of additional variables does not improve the model further.
DV4- Difference score “Discuss the misbehavior” Start:
+ class ethnic composition
+ emotion recognition index DV5- Difference score “Contact the parents”
Start
+ class ethnic composition
277.11 273.93 275.38 267.18 265.48
Note. Addition of another variable to the model is symbolized by a plus sign before
the name of the added variable.
With this information in mind, we have conducted two more explorative multivariate multiple regression analyses. The first analysis included the ethnic composition of the classroom, emotional intelligence, and emotion recognition index as independent variables, which were simultaneously regressed on the five averaged different scores from each of the five teacher reactions (as in the previous multivariate multiple regression analyses). The multivariate test statistics indicated a marginally significant overall model fit (Table 3). The second analysis included only emotional sensitivity variables namely, emotional intelligence and emotion recognition index as independent variables, which were regressed on the same dependent variables. The multivariate test statistics indicated a significant overall model fit (Table 3). However, further univariate test of significance failed to yield any significant results, and thus is not reported. Non-significant univariate tests and a significant multivariate test can indicate that there is accumulated evidence from the individual variables in the overall test (Manly, 1994).
Table 3
Multivariate Test of Significance, Exploratory Analyses
Note. * p < .05, ** p < .01.
Test Wilks
Value
Approx. F Hypoth. Df Error df Significance of F
1 .58767 1.7253 15 121.87 .05434 .
Discussion Aim and Key Findings
The aim of this study was to investigate how teachers’ self-reported behaviors towards majority versus minority pupils differed, depending on their cultural tolerance and emotional sensitivity. Contrary to our expectations, we did not find support for our predictions. To explicate our findings, we will compare our results to that of previous studies, discuss our findings, and detail what could have restricted our interpretation of our results.
Firstly, our prediction that cultural tolerance and emotional sensitivity would predict difference in frequency of punishment between majority and minority pupils, therefore hypothesis 1, was not supported. Diverging from our results, previous research found that teachers often classify disruptive behaviors differently for majority and minority group pupils (Ferguson, 2000), and their sensitivity to cultural matters and emotion related cues benefit their behaviors when dealing with these disruptive behaviors (Ladson-Billings, 1995; Radstake, 2009), while their abilities to manage their own emotions would strongly determine the reactions they give (Brackett & Katulak, 2006). Secondly, we did not find evidence for our prediction that predictive power of cultural tolerance and emotional sensitivity would be higher for more severe punishments. Therefore, hypothesis 2 was also not supported. Thus, previous findings that teachers respond more harshly towards misbehavior of ethnic minority pupils than towards identical behavior of ethic majority pupils (Ferguson, 2000) was not replicated here.
Notwithstanding the lack of evidence for our predictions, we observed a significant correlation between the teachers’ cultural tolerance and emotional intelligence. This finding may not be exceptionally surprising since emotions are
inseparable from the way one thinks about an issue or a situation, and emotional intelligence has been found to correlate with various traits as openness to feelings, empathy, and verbal intelligence (e.g., Ciarrochi, Chan, & Caputi, 2000; Mayer, Caruso, & Salovey, 1999; Schutte et al., 1998). Thus, emotionally intelligent teachers might exhibit certain interpersonal strengths that may allow them to better grasp and/or adjust to the experiences and issues of others from different ethnical and cultural backgrounds (Constantine & Gainor, 2001). Nevertheless, this finding suggests that it could be noteworthy to investigate certain traits such as openness to experience, openness to feelings, and empathy as underlying factors for culture and emotion related teacher traits.
Exploratory Findings
In scope of our analyses, we have explored whether our results were in line with certain findings in the literature. More specifically, we have investigated whether emotion recognition was better for ingroup than outgroup members (Elfenbein & Ambadi, 2002), whether pupils from minority groups were more likely to be punished than pupils from majority groups (Skiba & Rausch, 2004; Fenning & Rose, 2007), and whether male pupils were more often punished than female pupils for the same behaviors (Reynolds et al., 2003). Our results, with only one exception, were not in line with these previous findings in the literature, possibly due to the limitations that are detailed below. As the exception, we did find a significant difference in frequency of contacting parents for males and females for the native-Dutch pupils, while this difference was not observed for non-native Dutch pupils. Looking at the mean punishment rates, it appears that female native-Dutch pupils have an advantage over both male-Dutch pupils and native pupils in general. Given that female non-native pupils are punished at the same rate as both non-native and non-non-native male pupils
only for the most severe form of punishment in our study (contacting the parents), our finding might signal a very subtle negative attitude towards non-native female pupils. For instance, there is evidence that boys are usually believed to be naturally more aggressive than girls (Maccoby & Jacklin, 1974). Not differentiating between boys and non-native girls, in this case, might signal that teachers in our sample might have perceived non-native girls being as aggressive in their behaviors as boys in general.
In addition, we have explored which of the variables included in the study would form the “best” subset of predictors of our data. Explaining our data in the simplest way, and removing the noise added with unnecessary predictors, our variable selection showed that for most of the teacher reactions, (one or both of) the emotional sensitivity variables constituted the best subset. The lack of cultural tolerance variable in the best subset could be due to its moderate correlation with emotional intelligence. While one should approach this finding with caution due to the limitations of our study (described below), our results once again signal the importance of emotion related factors in interethnic group interactions in educational context, even more than cultural tolerance.
Strengths, Limitations, and Directions for Future Research
Before all else, within a very limited time available, we have conducted two studies in which we have successfully managed to reach the sample sizes we have aimed for. This was the case even though we were interested in a very specific, “research-tired”, and busy group of participants. Our pilot study results have provided us with good insights into what kind of misbehaviors take place in primary school classrooms and what kind of reactions teachers give to these behaviors. Our study was the first to investigate pupil misbehaviors and teacher reactions in the Netherlands context. Moreover, the vignettes that were used to measure teacher behaviors can be
further improved and validated since if vignettes are constructed effectively, they can reduce social desirability and defensiveness of responses (Tierney, 2010), and allow approaching sensitive topics (e.g., discrimination) in a subtle manner (Rose, 2001). This can, in return, render a useful tool to predict behavior (Jenkins, Bloor, Fischer, Berney, & Neale, 2010). Lastly, our results have revealed a relationship between cultural tolerance and emotional intelligence of teachers. While teacher effects literature mostly focuses on cultural variables when accounting for teachers’ behavioral differences, our results suggest that cultural factors are not isolated from emotional factors; and, in fact, our analyses showed that emotion related variables were the best to predict teachers’ differing behaviors, albeit this prediction was not significant. It could be, thus, fruitful to further investigate emotion related variables in relation to a broader range of culture related factors.
Several limitations constrain the interpretation of our study’s findings. To start with, we recognize the limitations of relying on self-reported data, which might have led to socially desirable answers especially for the teacher behaviors measure. Within this measure, merely using different pupil names from different ethnical backgrounds might not have been enough for teachers to change their behavioral reactions, perhaps because it was too revealing of our intentions. One solution to this issue could be to use videos (together with or instead of vignettes) of the situations depicting different misbehaviors. Next, the possible teacher reactions to different pupil misbehaviors have only included punishing behaviors and not rewarding behaviors. This is because we have relied on our pilot study results in which teachers only reported punishing behaviors to the common problematic situations they faced in their classrooms. In order to obtain the best results from vignette studies, it was crucial to provide good approximations to realistic social situations (Simon & Tierney, 2011); thus, we have
tried to keep the possible reactions for the pupil misbehaviors as close as possible to the ones reported in our pilot data. Providing a broader range of possible reactions that also include positive ones might not only better conceal the aim of the study and lead to less social desirability, but also reduce defensiveness in participants and increase their motivation. One way to achieve this could be to have a pilot study with more participation in order to gather information about real-life positive behavioral reactions.
Another limitation of our study is related to the cultural tolerance test. Since the measure we have used was an explicit measure assessing attitudes and beliefs about multiculturalism in teachers, the social desirability concerns apply. While explicit attitude measures are important in predicting behaviors that are carried out with some degree of deliberation (e.g., Dovidio, Kawakami, Johnson, Johnson, & Howard, 1997), using implicit measures in addition to explicit measures at this point could be important since they circumvent social desirability concerns and disclose different evaluations than explicit self-reports (e.g., spontaneous information processing, judgment, and behavior; for a review see Petty, Fazio, & Brinol, 2009).
Last limitations of our study are related to the emotion recognition test. This test measured teachers’ emotion recognition abilities from static photographs depicting adult faces. However, dynamic expressions enhance emotion recognition compared to static photographs since action is part of natural expressions (Ekman, 1994; van der Schalk et al., 2011). Therefore, using dynamic stimuli could be ecologically more valid than using motionless photographs. Moreover, since we were interested in teachers’ behaviors towards children, it would be more optimal to use an emotion recognition test measuring abilities to recognize emotions from children’s faces. Development and validation of such a measure could be useful for future
research that investigates emotions in educational context. Lastly, we have received feedback from the participants that the options for choosing which emotions are displayed in the faces were too limited. We could obtain more information about this ability by providing a broader range of options.
Practical Contributions
Besides its limitations, the current study highlights the importance of emotion related factors in the educational context and in relation to cultural tolerance. Next to these theoretical contributions, our study makes several practical contributions. Investigating the factors that might affect children’s (feelings of) integration has great societal relevance for several reasons. Firstly, while the existing minority groups are still having problems with their integration and educational achievement, there are new minority groups forming due to the influx of refugees, which might go through the same problems in the future. Secondly, the recent religion inspired terrorism has been increasing hostility between ethnic groups. Thirdly and on a more general note, empirical research suggests that the perceived relationship with a majority member teacher is more important for attitudes towards the majority group, compared to what is taught about ethnic diversity and multiculturalism within the curriculum (Thijs & Verkuyten, 2012). Lastly, perceived teacher bias is associated with student dropout rates (Wayman, 2002), and lack of school engagement and educational achievement are strongly correlated with involvement in the criminal justice system –a problem possibly caused by experienced bias in school discipline during the primary school years (Rocque & Paternoster, 2011). Prospective research can build on this study and might benefit from exploring the same study design after overcoming the above-mentioned limitations.
Conclusion
All things considered, while our expectations were not confirmed, the central point that emerges from our research is that being mindful of emotion related factors in ethno-cultural intergroup relations is essential. Prospective research should build on our findings, improve our measures, further investigate how a broader range of emotion related factors link to culture related factors, and whether if there are other factors underlying them. It is of paramount importance to identify the factors that might affect teacher behaviors in order to be able to benefit minority pupils’ integration and educational achievement within the Dutch education system.
References
Agirdag, O. (2010). Exploring bilingualism in a monolingual school system: Insights from Turkish and native students from Belgian schools. British Journal of Sociology of
Education, 31, 307–321.
Agirdag, O., Demanet, J., Van Houtte, M., & Van Avermaet, P. (2011). Ethnic school composition and peer victimization: A focus on the interethnic school climate.
International Journal of Intercultural Relations, 35(4), 465-473.
Allport, G. W. (1979). The nature of prejudice. Reading, MA: Addison-Wesley
Banks, J. A., & Banks, C. A. M. (1989). (Eds.). Multicultural education: Issues and
perspectives. Boston: Allyn and Bacon.
Banks, J. A., & Banks, C. A. M. (Eds.). (2004). Handbook of research on multicultural education. San Francisco, CA: Wiley.
Becker, T. E. (2005). Potential problems in the statistical control of variables in organizational research: A qualitative analysis with recommendations. Organizational
Research Methods, 8(3), 274-289.
Blau, P. M. (1974). Presidential address: Parameters of social structure. American
Sociological Review, 615-635.
Bonferroni, C. E. (1935). Il calcolo delle assicurazioni su gruppi di teste. In Studi in Onore
del Professore Salvatore Ortu Carboni (pp. 13-60). Rome: Italy.
Bonferroni, C.E. (1936). Teoria statistica delle classi e calcolo delle probabilità.
Pubblicazioni del R Istituto Superiore di Scienze Economiche e Commerciali di Firenze, 8, 3-62.
Brackett, M. A., & Katulak, N. A. (2006). Emotional intelligence in the classroom: Skill- based training for teachers and students. Applying emotional intelligence: A
Brandt, M. J., Chambers, J. R., Crawford, J. T., Wetherell, G., & Reyna, C. (2015). Bounded openness: The effect of openness to experience on intolerance is moderated by target group conventionality. Journal of personality and social psychology, 109(3), 549.
Braver, S. L., MacKinnon, D. P., & Page, M. (2003). Levine's guide to SPSS for analysis of
variance. Psychology Press.
Bronfenbrenner, U. (1979). The ecology of human development: Experiments by design and
nature. Cambridge, Ma: Harvard University Press.
Brophy, J., & Good, T. (1970). Teachers communication differential expectations for children’s classroom performance: Some behavioral data. Journal of Educational
Psychology, 61, 365-374.
Brown, R. (1995). Prejudice: Its Social Psychology. Oxford: Blackwell Cerezo, F., & Ato, M. (2010). Social status, gender, classroom climate and bullying among
adolescents pupils. Anales de psicología, 26(1), 137-144.
Christ, O., Schmid, K., Lolliot, S., Swart, H., Stolle, D., Tausch, N., ... & Hewstone, M. (2014). Contextual effect of positive intergroup contact on outgroup prejudice.
Proceedings of the National Academy of Sciences, 111(11), 3996-4000.
Ciarrochi, J. V., Chan, A. Y. C., & Caputi, P. (2000). A critical evaluation of the emotional intelligence construct. Personality and Individual Differences, 28, 539–561.
Connell, J. P. (1990). Context, self, and action: A motivational analysis of self-system processes across the life span. In D. Cicchetti & M. Beeghly (Eds.), The self in
transition: Infancy to childhood (pp. 61–97). Chicago: University of Chicago Press.
Constantine, M. G., & Gainor, K. A. (2001). Emotional intelligence and empathy: Their relation to multicultural counseling knowledge and awareness. Professional School
Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self- determination in human
behavior. New York: Plenum.
Dovidio, J. F., Kawakami, K., Johnson, C., Johnson, B., & Howard, A. (1997). On the nature of prejudice: Automatic and controlled processes. Journal of experimental
social psychology, 33(5), 510-540.
Driessen, G. (2012). 4 Combating Ethnic Educational Disadvantage in the Netherlands. The
Politics of Education: Challenging Multiculturalism, 65, 31.
Driessen, G., Doesborgh, J., Ledoux, G., Van der Veen, I., & Vergeer, M. (2003). Sociale integratie in het primair onderwijs. Nijmegen/Amsterdam: ITS/SCO-Kohnstamm
Instituut.
Dunham, Y. (2011). An angry= outgroup effect. Journal of Experimental Social
Psychology, 47(3), 668-671.
Dusek, J., & Josef, G. (1983). The bases of teacher expectancies: A meta-analysis. Journal
of Educational Psychology, 75, 327-346.
Eccles, J. S., Wigfield, A., & Schiefele, U. (1998). Motivation to succeed. In W. Damon & N. Eisenberg (Eds.), Handbook of child psychology: Vol. 3: Social, Emotional, and
personality development (5th ed., pp. 1017–1095). Hoboken, NJ: Wiley.
Ekman, P. (1994). Strong evidence for universals in facial expressions: a reply to Russell's mistaken critique. Psychological Bulletin, 115(2), 268-287.
Ekman, P., Friesen, W. V., & Hager, J. C. (2002). Facial action coding system: The manual
on CD-ROM Instructor’s Guide. Salt Lake City: Network Information Research Co.
Elfenbein, H. A., & Ambady, N. (2002a). On the universality and cultural specificity of emotion recognition: A meta-analysis. Psychological Bulletin, 128, 203–235.
Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences.
Behavior Research Methods, 39, 175-191.
Feld, S. L., & Carter, W. C. (1998). When Desegregation Reduces Interracial Contact: A Class Size Paradox for Weak Ties 1. American Journal of Sociology, 103(5), 1165-1186.
Fenning, P., & Rose, J. (2007). Overrepresentation of African American students in exclusionary discipline the role of school policy. Urban Education, 42(6), 536-559. Ferguson, A. A. (2001). Bad boys: Public schools in the making of black masculinity.
University of Michigan Press.
Fischer, A. H., & Manstead, A. S. R. (2008). Social functions of emotions. In M. Lewis, J. M. Haviland-Jones, & L. Feldman Barret (Eds.), Handbook of emotions (3rd ed. pp. 456 – 468). New York: Guilford Press.
Fridlund, A. J. (1994). Human facial expression: An evolutionary view. New York: Academic Press.
Gasser, M. B., & Tan, R. N. (1999). Cultural tolerance: Measurement and latent structure of attitudes toward the cultural practices of others. Educational and psychological
measurement, 59(1), 111-126.
Garcia, E. (1994). Understanding and meeting the challenge of student cultural diversity. Boston: Houghton Mifflin.
Gijsberts M., en Dagevos J., (2005) Uit elkaars buurt. De invloed van etnische concentratie
op integratie en beeldvorming. Den Haag: Sociaal Cultureel Planbureau, SCP). [In
Dutch, Report on the effect of ethnic concentration on integration and interethnic perceptions, Dutch Bureau for Social and Cultural Planning].
Gijsberts, M., & Dagevos, J. (2010). At home in the Netherlands? Trends in integration of non-Western Migrants, Annual Report on Integration 2009. Social Cultural Planning
Office, The Hague.
Graybill, S. W. (1997). Questions of race and culture: How they relate to the classroom for African American students. Clearing House, 70, 311-320.
Good, T., & Weinstein, R. (1986). Teacher expectations: A framework for exploring classrooms. In K. Zumwalt (Ed.), Improving Teaching (pp. 63- 85). Alexandria, VA: Association for Supervision and Curriculum Development.
Heyneman, S. P. (2000). From the party/state to multiethnic democracy: Education and social cohesion in Europe and Central Asia. Educational evaluation and policy
analysis, 22(2), 173-191.
Hugenberg, K., & Bodenhausen, G. V. (2004). Ambiguity in social categorization: The role of prejudice and facial affect in racial categorization. Psychological Science, 15(5), 342−345.
Hugenberg, K., & Bodenhausen, G. V. (2003). Facing prejudice: Implicit prejudice and the perception of facial threat. Psychological Science, 14, 640−643.
Jenkins, N. A., Bloor, M., Fischer, J., Berney, L., Neale, J. (2010). Putting it in context: the use of vignettes in qualitative interviewing. Qualitative Research, 10(2), 175-198. Jennings, P. A., & Greenberg, M. T. (2009). The prosocial classroom: Teacher social and
emotional competence in relation to student and classroom outcomes. Review of
Educational Research, 79, 491–525.
Juvonen, J., & Murdock, T. B. (1995). Grade-level differences in the social value of effort: Implications for self-presentation tactics of early adolescents. Child Development, 66, 1694-1705.
