Emotionally supportive classroom interactions and students' perceptions of their teachers as caring and just

Learning and Instruction xxx (2017) 1e11

Contents lists available at ScienceDirect

Learning and Instruction journal homepage: www.elsevier.com/locate/learninstruc

Emotionally supportive classroom interactions and students' perceptions of their teachers as caring and just* Luciano Gasser a, *, Jeanine Grütter a, Alois Buholzer a, Alexander Wettstein b a b

University of Teacher Education Lucerne, Switzerland University of Teacher Education Berne, Switzerland

a r t i c l e i n f o

a b s t r a c t

Article history: Received 5 July 2016 Received in revised form 19 August 2017 Accepted 22 August 2017 Available online xxx

Emotionally supportive classrooms represent an important interactional context for the development of positive perceptions of studenteteacher relationships. The present study investigated how students' perceptions of teacher care and justice develop over the upper elementary grades. Moreover, we studied to what degree teachers' observed emotional support predicts changes in perceptions of teacher care and justice in students with and without risks for academic failure. The study consisted of 1209 upper elementary grade students who were followed from grade five (T1) to grade six (T2, 1009 students). Multilevel analyses revealed that, over the year, student perceptions of their teacher as caring decreased in classrooms with low-quality teacher-student interactions in the emotional domain. Moreover, high emotional support from teachers protected students with high academic disengagement from developing negative perceptions of teacher justice. These results suggest a protective function of teacher's emotional support. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Teacher justice Teacher care Studenteteacher relationship CLASS (Classroom Assessment Scoring System) Academic risks

1. Introduction Students' perceptions of teacher care and teacher justice represent important dimensions of studenteteacher relationships (Wentzel, 2002). Even though teachers feel strongly committed to the relational aspects of the teaching profession (Gasser & Althof, 2017), children and adolescents often experience unjust treatment and low emotional care in schools (Pianta, Hamre, & Mintz, 2012; Ruck & Wortley, 2002). The perception of low teacher care and low teacher justice can have serious implications for students' future academic and social-emotional adjustment, such as depression, conduct problems or school failure (e.g., Greene, Way, & Pahl, 2006; Patrick, Ryan, & Kaplan, 2007; Rueger, Malecki, & Demaray, 2010; Wentzel, 1997, 2002; Zee, Koomen, & van der Vee, 2013). Importantly, the developmental significance of perceived studenteteacher relationships increases as students progress through the elementary grades (Roorda, Koomen, Spilt, & Oort, 2011).

* The study was funded by the Swiss National Science Foundation. The authors thank teachers and students for their study participation. We also thank our student teachers for their help during data collection. * Corresponding author. Institute for School and Diversity, University of Teacher Education Lucerne, 6004, Lucerne, Switzerland. E-mail address: [email protected] (L. Gasser).

In particular, the upper elementary grades represent a challenging period with regard to the formation of positive perceptions of studenteteacher relationships because classroom interactions tend to become less personal and more focused on academic performance (Jerome, Hamre, & Pianta, 2009; Maldonado-Carreo & Votruba-Drzal, 2011; Spilt, Hughes, Wu, & Kwok, 2012). Moreover, in several European nations such as Germany or Switzerland, secondary school is organized according to a tracking system. Consequently, upper elementary grade teachers decide which ability level students will transition to in secondary school. This high pressure to academically perform during the upper elementary grades can negatively affect students' wellbeing and classroom social relationships (Gasser, Grütter, Torchetti, & Buholzer, 2017; Grütter, Gasser, Zuffiano, & Meyer, 2017). However, children and adolescents differ in their risk for negative perceptions of the relationship with their teachers (Nurmi, 2012; Sabol & Pianta, 2012). For example, students with high academic risks (e.g., low academic achievement and engagement) are more likely to develop negative perceptions of the relationships with their teachers than students with low academic risks (Nurmi & Kiuru, 2015). Therefore, classroom conditions that prevent these children from developing negative relational outcomes over time are of particular importance (Hamre & Pianta, 2005). Recent research suggests that teachers' quality of emotional support, such

http://dx.doi.org/10.1016/j.learninstruc.2017.08.003 0959-4752/© 2017 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Gasser, L., et al., Emotionally supportive classroom interactions and students' perceptions of their teachers as caring and just, Learning and Instruction (2017), http://dx.doi.org/10.1016/j.learninstruc.2017.08.003

2

L. Gasser et al. / Learning and Instruction xxx (2017) 1e11

as their warmth or responsiveness to students' individual needs, may protect children from developing negative relational outcomes (O'Connor, 2010). Accordingly, the goal of the present study was to investigate whether teachers' emotional support moderated the risk for negative changes in students' perceptions of teacher care and justice during the upper elementary grades. In addition, we investigated if this moderation hypothesis particularly applied to students with high academic risks (e.g., whether students with high academic risks in classrooms with high emotional support would show less or no decline of perceived teacher justice compared to students with high academic risks in classrooms with low emotional support). 1.1. Students' perceptions of the teacher as caring and just This study focused on two key aspects of perceived studenteteacher relationships: teacher care and teacher justice. The perception of teacher care includes students' experience with whether the teacher provides individual support and shows genuine interest and empathy towards the individual student (e.g., “My teacher really cares about me”). In contrast, perceptions of teacher justice include the experience of equal treatment by the teacher (e.g., “My teachers generally treat me fairly”). While teacher care has been subject to intensive research (also referred to as “emotional care” or “personal support”) (e.g., Patrick et al., 2007; Wentzel, 1998), the perception of teacher justice has received less attention. However, care and justice represent two complementary principles that characterize the ethical dimension of human relationships (Frankena, 1973; Gibbs, 2003): equal treatment without affection and attention to a person's individual needs is perceived as being “cold”, whereas the uneven distribution of affection and attention is perceived as being unjust. Studenteteacher relationships have been assessed from both the teachers' and students' perspectives (Murray, Murray, & Waas, 2008). Whereas measures from the teachers' perspective have mainly included dimensions such as closeness and conflict (e.g., Jerome et al., 2009), student measures are more likely to focus on the experience of individual care and support (e.g., Wentzel, 1997). Although both information sources capture relevant aspects of studenteteacher relationships, students' perceptions represent a more proximal predictor of students' motivation and classroom behavior. Consequently, students' perceptions of their relationships with teachers consistently relate to a wide range of cognitive and social outcomes (e.g., Donat, Umlauft, Dalbert, & Kamble, 2012; Rueger et al., 2010). Pianta, Hamre, and Stuhlman (2003) conceptualized students' perceptions of the teacher as representations through which students experience their relationships with their teachers. Based on developmental system theory (Lerner, 1998), these researchers proposed a model of studenteteacher relationships where perceptions of studenteteacher relationships are influenced by multiple proximal and distal systems that interact with each other in complex ways. Features of studenteteacher interactions (e.g., affective classroom climate), individual characteristics of students and teachers (e.g., students' academic achievement), or external influences such as structural aspects of the educational system (e.g., smaller versus larger classes, one versus several teachers) reciprocally influence how students perceive their relationships with their teachers. In this study, we assumed that teachers' concrete and observable practices in the emotional domain constitute proximal mechanisms that contribute to students' formation of positive perceptions of studenteteacher relationships. Reciprocal interactions between students and teachers that are characterized by respect, warmth, sensitivity and responsiveness lay the foundation for the formation of positive perceptions of relationships

(Brock & Curby, 2014; O'Connor, 2010; Pianta et al., 2003; Sabol & Pianta, 2012). These primary mechanisms may function as protective factors against contextual and individual risk factors, and they may decrease the likelihood of dysfunctional relational outcomes (e.g., Hamre & Pianta, 2005; O'Connor, 2010). 1.2. The upper elementary grades in Switzerland In this study, we followed students from the fifth to the sixth grade, which represent the upper elementary grades in Switzerland. Specifically, we focused on the time before the transition to secondary grades. Longitudinal research has revealed that the quality of studenteteacher relationships decreases over the elementary school years (e.g., Jerome et al., 2009; Spilt et al., 2012), which is a trend that continues into middle school (Eccles et al., 1993) and in part explains why older children or adolescents are less academically engaged and develop higher levels of problem behavior (Maldonado-Carreo & Votruba-Drzal, 2011). Researchers have suggested that transformations in the teacher's role (i.e., a stronger focus on instruction and less on emotional support) and the educational context (e.g., increased teacher-student ratio, several teachers per class) are possible reasons that studenteteacher relationships become more distant (Jerome et al., 2009). Moreover, factors related to early adolescence, such as students' growing interests in peer relationships and autonomy development, might explain why the quality of studenteteacher relationships declines as students enter the upper elementary grades (Zee et al., 2013). Finally, the upper elementary grades in Switzerland represent a transitional phase between elementary (first to sixth grade) and secondary school (seventh to ninth grade), which has crucial implications for students' future academic and occupational careers. Whereas elementary schools in Switzerland are inclusive (i.e., students with different academic abilities are educated in the same classroom), the secondary school is structured according to a tracking system (i.e., students with different academic abilities are educated in different schools). During the upper elementary grades, teachers decide on the basis of students' academic achievement which secondary level students will attend. Consequently, the upper elementary grades in Switzerland are characterized by increasing academic expectations, which might result in more negative classroom relationships (Gasser et al., 2017) and more internalizing and externalizing problems (Ball, Lohaus, & Miebach, 2006). Consequently, some upper elementary grade students might be more likely to perceive the teacher as a source of increasing academic demands rather than as a source of emotional care. 1.3. Academic risk factors and students' perceptions of teacher's care and justice In this study, we included two academic risk factors as individual predictors for upper elementary grade students' perceptions of teacher care and justice: low academic achievement (teacher ratings) and academic disengagement (student ratings). Teacherrated academic achievement is an important indicator for students' academic risk because teachers' evaluations of students' academic performance determines which ability level students' will be assigned to in secondary school. In addition, self-ratings of students' academic disengagement reflect elementary grade students' difficulties in coping with the pressure to academically perform (Tuominen-Soini & Salmela-Aro, 2014). Various studies showed that academic achievement and (dis) engagement relate to students' perceptions of teacher care and justice during early adolescence (e.g., Ko
sir & Tement, 2014; Murray et al., 2008; Patrick et al., 2007; Quin, 2017; Wentzel, 1997). Nurmi

Please cite this article in press as: Gasser, L., et al., Emotionally supportive classroom interactions and students' perceptions of their teachers as caring and just, Learning and Instruction (2017), http://dx.doi.org/10.1016/j.learninstruc.2017.08.003

L. Gasser et al. / Learning and Instruction xxx (2017) 1e11

and Kiuru (2015) referred to “evocative effects” in order to explain the impact of students' academic characteristics on studenteteacher relationships. According to this transactional model, it is not only the teacher that initiates certain types of dyadic interactions but also students who contribute to the studenteteacher relationship quality. From this perspective, students' academic and motivational characteristics activate certain cognitive and affective reactions from the teachers, which results in the teachers' differential treatment of students. As upper elementary grade students are highly sensitive to teachers' differential treatment (Weinstein, Marshall, Sharp, & Botkin, 1987), low achieving or academically disengaged students might perceive more negative treatment by their teachers in terms of less support, more criticism or lower expectations. Consequently, these students might be less likely to perceive their teachers as caring and just compared to their high achieving and engaging peers. 1.4. Studenteteacher interactions in the emotional domain The Teaching through Interaction framework by Hamre and Pianta (2005) assigns the interactions between students and teachers a key role for developmental change in schools. This framework organizes studenteteacher interactions into three broad domains, namely emotional support, classroom organization and instructional support, which differentially influence various aspects of students' learning (Hamre et al., 2013; Rimm-Kaufman, Curby, Grimm, Nathanson, & Brock, 2009). Emotional support includes interactions that reflect the emotional climate in the classroom as well as the teachers' attempt to respond individually to students' needs, including their need for autonomy. Interactions in the domain of classroom organization focus on how well teachers manage students' attention, time and behavior by using proactive approaches to discipline, routines and clear directions. Finally, the instructional support domain captures teachers' attempts to promote higher-order thinking skills and interconnected knowledge through open-ended and complex tasks, high-quality feedback and classroom discussions. An increasing body of literature demonstrates that teachers' observed emotional support importantly contributes to teacherrated quality of studenteteacher relationships (Brock & Curby, 2014; Hamre & Pianta, 2005; Hamre, Pianta, Downer, & Mashburn, 2008; Lee & Bierman, 2015; O'Connor, 2010; Rudasill et al., 2016). Teachers who consistently engage in building strong connections and are responsive to students' developmental needs provide students with an interaction context that is highly stimulating for the development of positive perceptions of studenteteacher relationships (Buyse, Verschueren, Doumen, Van Damme, & Maes, 2008). In addition, studies have shown that emotionally supportive classrooms protect students at risk for developing conflicting relationships with their teachers (e.g., Hamre & Pianta, 2005; Rudasill et al., 2016). Thus, classrooms that are characterized by high-quality emotional support do not have the same significance for all children; children with higher academic, behavioral or emotional risks are less likely to develop negative relationships with their teachers in classrooms where teachers provide students with high-quality emotional support than in classrooms where teachers provide students with lowquality emotional support. However, most of these studies have focused on younger children (preschool, K-3), and rather few studies included older elementary school children (for exceptions, see Brackett, Reyes, Rivers, Elbertson, & Salovey, 2011; O'Connor, 2010; Madill, Gest, & Rodkin, 2014). One of the exceptions is the study by O'Connor (2010) who longitudinally followed the quality of teacherestudent relationships (from the teacher's perspective) from the first

3

through the fifth grade. He found that teachers' perceptions of studenteteacher relationships showed a less rapid decline from the first through the fifth grade in classrooms that were characterized by high-quality observed emotional support than in classrooms that were characterized by low-quality observed emotional support. Therefore, high-quality emotional support protects students from developing highly conflicting relationships with teachers, despite challenging transformations across the elementary grades (e.g., several teachers). 1.5. Hypotheses Based on the literature review above (e.g., Maldonado-Carreo & Votruba-Drzal, 2011), we expected that students' perceptions of teacher care and justice would decrease during the upper elementary grades (hypothesis 1). Moreover, students at risk for academic failure develop more negative perceptions of their relationships with teachers (e.g., Ko
sir & Tement, 2014). Accordingly, we expected that low academic achievement and high academic disengagement would be related to lower perceptions of teacher care and justice (hypothesis 2). Previous research has shown that teachers' emotional support might represent a resiliency factor for students at risk (e.g., Hamre & Pianta, 2005). We thus expected cross-level interactions between teachers' observed emotional support, students' academic risks (i.e., low academic achievement and high academic disengagement) and the upper elementary grade context (i.e., transitioning from the 5th to the 6th grade). More specifically, we expected that teachers' observed emotional support would function as a protector for decreases in perceived teacher care and justice during the upper elementary grades, especially for those students with higher academic risks (hypothesis 3). In all of the analyses, we controlled for the effect of observed teacher quality in the domain of classroom organization and instructional support to test whether teachers' emotional support would uniquely relate to changes in students' perceptions of teacher care and justice. Even though previous research has revealed that the three domains of studenteteacher interactions differentially predict specific relational outcomes, the findings are inconsistent. Whereas some studies have revealed that only emotional support significantly relates to perceptions of studenteteacher relationships (Brackett et al., 2011; Hamre & Pianta, 2005), other studies have found that classroom organization also positively predicts perceived studenteteacher relationships (Brock & Curby, 2014; O'Connor, 2010). Most studies found no associations between instructional support and perceived studenteteacher relationships (e.g., Hamre & Pianta, 2005). Finally, we controlled for students' sex and migration background. Previous research has shown that boys compared to girls perceive their teachers as less supportive (e.g., Wentzel, Battle, Russell, & Looney, 2010) and as less just (e.g., Donat et al., 2012). Moreover, students from minority groups have more difficult relationships with their teachers than students from majority groups (Sabol & Pianta, 2012). As Switzerland has one of the highest immigration rates in Europe (Malti, Killen, & Gasser, 2012), we also controlled for students' immigration background. 2. Method 2.1. Sample, attrition, and missing values In the first wave of the study (T1), the sample consisted of 1209 Swiss students from 61 school classes, and 1009 students from 54 classes remained in the sample during the second wave of the study (T2). At T1, students were on average 11 years old (M ¼ 11.54,

Please cite this article in press as: Gasser, L., et al., Emotionally supportive classroom interactions and students' perceptions of their teachers as caring and just, Learning and Instruction (2017), http://dx.doi.org/10.1016/j.learninstruc.2017.08.003

4

L. Gasser et al. / Learning and Instruction xxx (2017) 1e11

SD ¼ 0.56), and at T2, the mean age of students was 12 years (M ¼ 12.58, SD ¼ 0.56). We assessed students' migration background by asking students to indicate their parents' first language. The proportion of students with a migration background in each wave was 41%, and 38%, respectively, which is representative of Switzerland (36%) (Swiss Federal Statistical Office, 2017). We estimated the educational level of the families based on information from the Swiss Federal Statistical Office about the type of community in which the parents lived in (for a similar procedure see Gasser, Malti, & Buholzer, 2014). Of the parents at T1, 24% had completed obligatory school, 50% had obtained a post-secondary diploma, and 20% had achieved a bachelor's degree or higher (the Swiss average regarding parents' education status is 23%, 49%, and 22%) (Swiss Federal Statistical Office, 2017). The average class size was 19.6 (T1) and 20.1 (T2) (mean class size in Switzerland is 19.1) (Swiss Coordination Centre for Research in Education, 2014). The parents received an information letter explaining the study, which was translated into the official languages and the most spoken foreign languages of Switzerland. We obtained informed consent from the parents, and only 0.4% (T1) and 1% (T2) of the parents did not give their consent. The sample attrition was 17%. To test whether there was a systematic pattern in the sample attrition, we used the MCAR (Missing Completely at Random) Test from Little (1988). The result of this test, including all the control and study variables, was not significant. Therefore, we concluded that the missing values on the main study variables were not related to any other measured or unmeasured variables. Sixty-one teachers participated in T1, and 54 teachers participated in T2. Eighty percent of the teachers that participated in both study waves were female, and most of the teachers (44%) belonged to the age group of 26e30 years old. Overall, their ages ranged from 20 to 55 years. Additionally, the teachers reported, on average, 11 years of teaching experience (M ¼ 11.31, SD ¼ 10.53). 2.2. Procedure and measures While the perceptions of teacher care, teacher justice, and academic disengagement were assessed by the student questionnaire, teacher ratings were used to assess the students' academic achievement. The teachers' emotional support was assessed by classroom observations. Research assistants attended a one-day training where they received instructions about the procedures for data collection (e.g., how to instruct the students) and the materials. Moreover, the research assistants received a two-day training on how to conduct the CLASS (Classroom Assessment Scoring System) observations, and they were required to pass a reliability test (for details, see chapter 2.2.5). The research assistants guided the students and teachers through the data collection, for which the teachers and students simultaneously completed the questionnaires (30e40 min). The students received a student magazine after they completed the questionnaire. A separate date was arranged with the teachers to conduct the CLASS observations. The study design is illustrated in Fig. 1. 2.2.1. Students' perceived teacher care The scale to assess perceived teacher care focused on the students' experience of individualized support and affection and included four items. One item was adapted from Battistich, Solomon, Watson, and Schaps (1997) (“My teacher really cares about me”), and the other three items were adapted from Ryan and Patrick (2001) (e.g., “My teacher understands how I feel about things”). The response format ranged from 1 (strongly disagree) to 4 (strongly agree). The scale had a high reliability at both measurement times, a ¼ 0.78 (T1) and a ¼ 0.80 (T2) (test retest stability: rtt ¼ 0.44, p ¼ 0.001).

CLASS (observations) Academic achievement (teacher ratings) Academic disengagement (student ratings) Perceived teacher care and justice (student ratings)

Perceived teacher care and justice (student ratings)

T1: 5th grade

T2: 6th grade

(N = 1,209)

(N = 1,009) Fig. 1. Study design.

2.2.2. Students' perceived teacher justice The scale to assess the students' perceptions of teacher justice was adapted from Peter, Kloeckner, Dalbert, and Radant (2012) and consisted of six items (e.g., “My teacher often treats me unfairly” [reversed]). The original scale includes 10 items and was constructed for secondary school students. In a pilot study, we therefore examined which items were difficult to understand for upper elementary school students and excluded four items. The items were answered on a four-point scale from 1 (strongly disagree) to 4 (strongly agree). The shortened scale that was used in this study also had a high reliability (a ¼ 0.78 [T1], a ¼ 0.80 [T2]) and a high stability across the two study waves rtt ¼ 0.46 (p ¼ 0.001). Moreover, the results of an explanatory multilevel factor analysis suggested that teacher care and teacher justice represent two independent scales. For teacher justice, the geomin rotated factor loadings of the six items ranged from 0.31 to 0.74 at the within-level, and from 0.51 to 1.03 at the between level. For teacher care, the loadings ranged from 0.53 to 0.73 for the within-level and from 0.81 to 0.99 for the between-level. 2.2.3. Academic achievement Teachers rated students' academic achievement by three items (e.g., “Performing academically at grade level”, “Able to read grade level material and answer questions about what he/she has read”, “Able to solve grade level math problems”) (Hughes, Dyer, Luo, & Kwok, 2009). The items were rated on a five-point response scale from 1 (almost never) to 5 (almost always). The measure longitudinally relates to math and reading test scores (Hughes et al., 2009). The Cronbach's alpha was a ¼ 0.94 (T1). 2.2.4. Academic disengagement We assessed academic disengagement by school burnout in which disengagement represents a central aspect (Tuominen-Soini & Salmela-Aro, 2014). The items that were used to assess academic disengagement at T1 were adapted from the school burnout scale (Salmela-Aro, Kiuru, & Nurmi, 2008). As the original scale was developed for secondary students, we excluded one item referring to sleeping disruptions (“I often sleep badly because of matters related to my schoolwork”) because school-related sleeping disruptions are more typical for secondary school students than nez, Carbonell-Baeza, elementary school students (Segura-Jime ~ ero, 2015). We thus included a scale Keating, Ruiz, & Castro-Pin with eight items (e.g., “I feel overwhelmed by my schoolwork”, “I'm continually wondering whether my schoolwork has any meaning”). The items were answered on a four-point scale from 1 (strongly disagree) to 4 (strongly agree). The Cronbach's alpha was a ¼ 0.81

Please cite this article in press as: Gasser, L., et al., Emotionally supportive classroom interactions and students' perceptions of their teachers as caring and just, Learning and Instruction (2017), http://dx.doi.org/10.1016/j.learninstruc.2017.08.003

L. Gasser et al. / Learning and Instruction xxx (2017) 1e11

(T1). The scale positively relates to teacher-rated emotional problems (r ¼ 0.23***) and negatively relates to self-reported school identification (r ¼ 0.46***) and teacher-rated academic achievement (r ¼ 0.28***), supporting the concurrent validity of the scale. 2.2.5. Emotional quality of studenteteacher interactions Classroom interactions were assessed at T1 with the Upper Elementary CLASS instrument (Pianta, Hamre, & Mintz, 2012). The measure includes 12 dimensions: (1) positive climate, (2) teacher sensitivity, (3) regard for students' perspectives, (4) behavior management, (5) productivity, (6) negative climate, (7) instructional learning formats, (8) content understanding, (9) analysis and inquiry, (10) quality of feedback, (11), instructional dialogue, and (12) student engagement. These dimensions are organized within three domains (Pianta et al., 2012): emotional support (dimensions 1e3), classroom organization (dimensions 4e7) and instructional support (dimensions 8e11). Student engagement represents a separate dimension, and it is not assigned to one of the domains (dimension 12). The research assistants observed four 15-min cycles. Based on these observations, each dimension was rated on a seven-point scale (1 ¼ low quality, 7 ¼ high quality). To obtain stable assessments of classroom interactions, a minimum of four cycles is required (Pianta et al., 2012). The research assistants completed a two-day training course that involved the review and coding of video examples. Prior to the data collection, the research assistants proved their ability to use the CLASS instrument in an online reliability test. In this test, the examinees must rate five 15-min segments of different classrooms; they pass the test and become certified CLASS observers when 80% of their ratings on all the 12 CLASS dimensions are within 1 point of the master codes (Pianta et al., 2012). In this study, we averaged the mean scores of the four observation cycles for each dimension, and in the next step, we calculated the mean scores for each of the three domains (e.g., Mashburn et al., 2008). The internal consistency of the three domains was high (emotional support: a ¼ 0.84, instructional support: a ¼ 0.92, classroom organization: a ¼ 0.79). 2.3. Data analysis As the students who participated in this study belonged to different classrooms, and the study consisted of two waves, the measurements of the students were not independent of each other. First, it is likely that students from the same classroom are more similar in their perception of their teacher than students from different classrooms. Second, repeated measures of the students are more similar within children than between children. A recommended method of handling this complex data structure is the use of hierarchical linear models (Bliese & Ployhart, 2002). In this study, the data structure had three levels: first, the repeated measures of students across the two waves; second, students in school classes; and third, school classes. We first examined whether there were significant differences between school classes in the two main study variables of teacher care and justice. The results of these analyses revealed that the classrooms significantly differed from each other in perceived teacher care, F(60, 2132) ¼ 4.48, p < 0.001, and perceived teacher justice, F(60, 2133) ¼ 4.53, p < 0.001. These differences between the classrooms explained 3% of the total variance in perceived teacher care and in perceived teacher justice, which reflects an ICC(1) value of 0.03 for both dependent variables. The ICC(2) denotes the homogeneity of teacher perceptions within classrooms and reflects the reliability of the group means. This value was ICC(2) ¼ 0.78 for teacher care and ICC(2) ¼ 0.79 for perceived teacher justice. In addition to differences between classrooms, the measurements

5

across the two waves depended on the characteristics of the students (teacher care: F(1205, 987) ¼ 2.28, p < 0.001; teacher justice: F[1205, 988] ¼ 2.59, p < 0.001). The differences between these students in their school classes explained 44% of the total variance in perceived teacher justice (ICC[1] ¼ 0.44, ICC[2] ¼ 0.61) and 42% of the total variance in perceived teacher care (ICC[1] ¼ 0.42, ICC [2] ¼ 0.56). These results suggest that, although students within classrooms are more similar in their perception of the teacher, perceptions of teacher care and teacher justice are more likely to represent subjective and individual variables. The hierarchical linear models were calculated with the Rpackage multilevel (Bliese, 2002). The data-set contained two rows per child: one row for teacher care (resp. justice) at T1 and one row for teacher care (resp. justice) at T2. We followed the recommendations of Bliese and Ployhart (2002) for growth modeling using random coefficient models in R. According to this procedure, we first conducted simple models and added more complex terms to the models in a stepwise procedure. We first tested whether models with random intercepts fit the data significantly better than simple models without any random terms (by using likelihood tests). The random intercepts represent different initial values in teacher care and justice between students and classrooms in grade five (T1). In a second step, we examined whether there was significant slope variation (i.e., both slopes, see below) among the children in the classrooms. The random slopes represent different changes between individuals and classrooms in teacher care and teacher justice from the fifth to the sixth grade. Therefore, we tested whether a model with random intercepts and random slopes fit the data better than the previous model that contained only random intercepts. After choosing the best model for each dependent variable, we tested whether there were significant mean level changes in our dependent variables from the fifth to the sixth grade. Next, we examined whether these changes over time and the variation between students and classrooms in these changes could be explained by our predictor variables at the individual level (i.e., academic achievement, school disengagement), at the classroom level (i.e., emotional support) and by their interaction (i.e., crosslevel interactions). As 41% of the students had a migration background, we controlled for this variable in our analyses. We also controlled for students' sex. As the results revealed no significant interaction effects including students' sex and migration background, we only included the main effects of students' sex and migration background in the analyses (e.g., girls were more likely to perceive teachers as just and caring compared to boys). Lastly, we controlled for observed instructional support and classroom organization to investigate whether the relationships between teachers' emotional support and perceptions of teacher care and justice were unique. The predictor variable time was centered at the first measurement point (i.e., the intercept values of our analysis reflect the initial values of teacher care and justice in the fifth grade). The predictor variables at the individual and at the classroom level were centered at the grand mean of the sample because this procedure enables the comparison of results between the different classrooms (Enders & Tofighi, 2007). We chose this procedure because we were interested in the contextual effects of the learning environment (i.e., emotional support from teachers) on students' development of teacher perceptions from the fifth to the sixth grade, while taking into account individual differences in academic achievement and academic disengagement. To estimate such contextual effects, grand-mean centering is recommended in three-level models (Brincks et al., 2017). The descriptive information for the study and the control variables are provided in Table 1.

Please cite this article in press as: Gasser, L., et al., Emotionally supportive classroom interactions and students' perceptions of their teachers as caring and just, Learning and Instruction (2017), http://dx.doi.org/10.1016/j.learninstruc.2017.08.003

6

L. Gasser et al. / Learning and Instruction xxx (2017) 1e11

Table 1 Descriptive results of the study variables and correlations between descriptive and study variables at level 2 (across children, N ¼ 1210 [T1] & 978 [T2]) and at level 3 (across classrooms, N ¼ 61).

1. Sex (0 ¼ girls) 2. Migration background (0 ¼ no) 3. Academic achievement 4. Academic disengagement 5. Perceived teacher care (T1) 6. Perceived teacher care (T2) 7. Perceived teacher justice (T1) 8. Perceived teacher justice (T2) 9. Emotional support 10. Classroom organization 11. Instructional support

M

SD

e e 3.87 1.95 3.47 3.47 3.12 2.97 5.21 6.39 3.90

e e 1.13 0.63 0.53 0.54 0.63 0.66 0.87 0.56 0.92

(1)

(2)

(3)

(4)

(5)

(6)

(7)

0.03 0.05 0.16*** 0.09** 0.03 0.15*** 0.10**

0.26*** 0.06* 0.10*** 0.07* 0.07* 0.07*

0.31*** 0.11*** 0.09** 0.24*** 0.18***

0.15*** 0.14*** 0.43*** 0.31***

0.44*** 0.34*** 0.21***

0.22*** 0.45***

0.46***

(8)

(9)

(10)

0.52*** 0.72***

0.47***

Note. T1 ¼ first measurement time, T2 ¼ second measurement time. * ¼ p < 0.05, ** ¼ p < 0.01, *** ¼ p < 0.001, two-tailed.

3. Results

3.2. Perceived teacher justice

3.1. Perceived teacher care A model with random intercepts (differences in initial values between students and classrooms in the fifth grade) and random slopes (differences in changes from the fifth to the sixth grade between students and classrooms) fitted the data better than a model with solely random intercepts (Dc2(4) ¼ 33.21, p < 0.001). In line with hypothesis 1, the results of this basic model showed that perceived teacher care significantly decreased from the fifth to the sixth grade (see Step 1, Table 2). Moreover, in line with hypothesis 2, students with low academic engagement had lower perceptions of teacher care. However, contrary to hypothesis 2, students with low academic achievement perceived the teacher as more caring than students with higher academic achievement (see step 2, Table 2). Since the different variance components of the model showed that there were significant differences in initial values and changes from the fifth to the sixth grade between students and between school classes (see step 1, Table 2), we predicted changes from the fifth to the sixth grade by adding the interactions of each predictor variable with the time variable to the model (see step 3, Table 2). We hypothesized that high emotional support from the teacher in the fifth grade would have a protective function for decreases in students' perceptions of teacher care, particularly for students with higher academic risks. To test this hypothesis, we added two threeway interaction terms between time, academic achievement (or academic disengagement), and emotional support to the previous model. Contrary to our assumption, there were no significant threeway interactions between time, academic achievement respective academic disengagement, and observed emotional support.1 Still, there was a significant two-way interaction between time and teacher emotional support (see Table 2, step 3). Fig. 2 shows that students in classrooms with low emotional support experienced a significant decrease in the perception of their teacher as caring when they transitioned from the fifth to the sixth grade (slopes test: B ¼ 0.32, SE ¼ 0.06, p < 0.001), while perceived teacher care did not decrease in classrooms with high emotional support (slopes test: B ¼ 0.01, SE ¼ 0.05, p ¼ 0.935). In other words, emotional support had a protective function for the decrease in perceived teacher care. Emotional support explained 29% of the variance in changes in perceived teacher care between classrooms. No significant effects for classroom organization and instructional support were found (see Table 2, step 3).

A model with random intercepts (differences in initial values between students and classrooms in the fifth grade) and random slopes (differences in changes from the fifth to the sixth grade between students and classrooms) fitted the data significantly better than a model with only random intercepts (Dc2(4) ¼ 16.34, p ¼ 0.003). On average, there was no significant change in perceived teacher justice (see step 1, Table 3). Therefore, hypothesis 1, which assumed a decrease in teacher justice over the year, was rejected. In line with hypothesis 2, students with low academic achievement and low academic engagement expressed significantly lower perceptions of teacher justice than their classmates with higher academic achievement and higher academic engagement (see step 2, Table 3). Since the different variance components of the model showed that there were significant differences in initial values and changes from the fifth to the sixth grade between students and between school classes (see step 1, Table 3), we predicted changes from the fifth to the sixth grade by adding the interactions of each predictor variable with the time variable to the model (see step 3, Table 3). To test hypothesis 3 (high emotional support would prevent the decrease in perceived teacher justice over the year, particularly for students at risk for academic failure), we used the same procedure as for teacher care (see above). The results (see step 4, Table 3) revealed no significant associations with academic achievement2 but a significant three-way interaction between time, academic disengagement, and emotional support. In order to better understand this interaction, we plotted changes in perceived teacher justice from the fifth to the sixth grade, depending on students' academic disengagement and emotional support by teachers. As observed in Fig. 3, students with high levels of academic disengagement who were in classrooms with high observed emotional support (i.e., above one standard deviation from the grand-mean of the sample) experienced a significant increase in the perception of teacher justice over time (slopes test: B ¼ 0.20, SE ¼ 0.05, p < 0.001), while such students experienced a significant decrease in the level of perceived teacher justice if they were in classrooms with low perceived emotional support (i.e., below one standard deviation from the grand-mean of the sample; slopes test: B ¼ 0.10, SE ¼ 0.05, p ¼ 0.035). Moreover, students with low levels of academic disengagement perceived more teacher justice than their classmates with high academic disengagement. Still, if these students attended classrooms with low observed emotional teacher

1 For reasons of power, these two three-way interactions were not included in the final model.

2 To enhance the power of our statistical analyses, this three-way interaction was not included in the final model.

Please cite this article in press as: Gasser, L., et al., Emotionally supportive classroom interactions and students' perceptions of their teachers as caring and just, Learning and Instruction (2017), http://dx.doi.org/10.1016/j.learninstruc.2017.08.003

L. Gasser et al. / Learning and Instruction xxx (2017) 1e11

7

Table 2 Perceived teacher care as predicted by academic achievement, academic disengagement, emotional support, instructional support, and classroom organization (N ¼ 2194 observations of 1209 children in 61 classrooms). Step 1

g (SE) Fixed effects level 1 Constant 3.12 (0.03)*** Time 0.16 (0.03)*** Fixed effects level 2 (students in classrooms) Sex Migration background Academic achievement Academic disengagement Fixed effects level 3 (between classrooms) Emotional support Instructional support Classroom organization Cross-level interactions Aa * time Ad * time Es * time Is * time Co * time Variances of the random effects and ICC(1) values for basic model Initial values of students 0.32; ICC(1) ¼ 0.42 Changes of students 0.32 Initial values of classrooms 0.02; ICC(1) ¼ 0.03 Changes in classrooms 0.03 Residuals 0.06 AIC 4016.38 BIC 4067.61

Step 2 df 1145 986

Step 3

g (SE)

g (SE)

df

df

3.14 (0.03)*** 0.16 (0.03)***

1099 962

3.14 (0.03)*** 0.16 (0.03)***

1099 957

0.07 0.05 0.08 0.17

1099 1099 1099 1099

0.07 0.05 0.09 0.18

1099 1099 1099 1099

(0.03)* (0.03) (0.02)*** (0.03)***

0.06 (0.04) 0.04 (0.04) 0.02 (0.05)

0.30 0.32 0.02 0.03 0.05 3878.76 3969.29

57 57 57

(0.03)* (0.03) (0.02)*** (0.03)***

0.02 (0.04) 0.02 (0.04) 0.01 (0.05)

57 57 57

0.01 (0.02) 0.04 (0.04) 0.18 (0.05)*** 0.09 (0.05) 0.09 (0.06)

957 957 957 957 957

0.30 0.33 0.02 0.03 0.05 3899.54 4018.31

Note. Control variables are sex (0 ¼ female, 1 ¼ male) and migration background (0 ¼ no, 1 ¼ yes). * p < 0.05, ** p < 0.01, *** p < 0.001, two-tailed. Aa ¼ Academic achievement; Ad ¼ Academic disengagement; Es ¼ Emotional support; Co ¼ Classroom organization; Is ¼ Instructional support.

students in classrooms with high values for classroom organization (i.e., above one standard deviation from the grand-mean of the sample) decreased in their perceptions of teacher justice over the year (slopes test: B ¼ - 0.08, SE ¼ 0.03, p ¼ 0.027), while students in classrooms with low values for classroom organization (i.e., below one standard deviation from the grand-mean of the sample) did not change in their perception of their teacher over the year (slopes test: B ¼ 0.05, SE ¼ 0.03, p ¼ 0.163). Classroom organization explained 16% of the variance in changes over the year between classrooms. 4. Discussion

Fig. 2. Development of students' perceived teacher care from grade five to grade six predicted by teachers' emotional support (classroom observation) in grade five (bold line ¼ significant slope, dotted line ¼ non-significant slope).

support, they experienced a significant decrease in their perceptions of teacher justice (slopes test: B ¼ 0.13, SE ¼ 0.05, p ¼ 0.007; explained variance of the three-way interaction in changes of students in classrooms: 1%). Furthermore, no significant relationship between changes in teacher justice and instructional support were found (see step 3, Table 3). Surprisingly, classroom management significantly predicted changes in teacher justice. Simple slope tests showed that

We investigated whether students' perceptions of teacher care and justice decrease during the upper elementary grades and whether students' academic risks (i.e., low academic achievement and high disengagement) predict students' perceptions of teacher care and justice. Moreover, we assumed that the observed quality of studenteteacher interactions in the emotional domain functions as a protector against the decline of positive perceptions of teacher care and teacher justice during the upper elementary grades, specifically for children with academic risks (i.e., low academic achievement and high academic disengagement). The previous research has mainly focused on how children's and adolescents' perceptions of teacher care and justice relate to their socialemotional and academic functioning mostly using cross-sectional designs (e.g., Patrick et al., 2007; Peter et al., 2012; Rueger et al., 2010). This study extends the previous research by focusing on how objective aspects of studenteteacher interactions in the emotional domain longitudinally relate to changes in teacher perceptions in the upper elementary grades. The hypothesis that students' perceptions of their relationship with the teacher decrease over the upper elementary grades was partially supported, as we only found decreases regarding students'

Please cite this article in press as: Gasser, L., et al., Emotionally supportive classroom interactions and students' perceptions of their teachers as caring and just, Learning and Instruction (2017), http://dx.doi.org/10.1016/j.learninstruc.2017.08.003

8

L. Gasser et al. / Learning and Instruction xxx (2017) 1e11

Table 3 Perceived teacher justice as predicted by academic achievement, academic disengagement, emotional support, instructional support, and classroom organization (N ¼ 2194 observations of 1209 children in 61 classrooms). Step 1

g (SE)

Step 2 df

Fixed effects level 1 Constant 3.46 (0.02)*** 1145 Time 0.01 (0.02) 987 Fixed effects level 2 (students in classrooms) Sex Migration background Academic achievement Academic disengagement Fixed effects level 3 (between classrooms) Emotional support Instructional support Classroom organization Cross-level interactions Aa * time Ad * time Es * time Is * time Co * time Es * Ad Ad * Es * time Variances of the random effects and ICC(1) values for basic model Initial values of students 0.22; ICC(1) ¼ 0.44 Changes of students 0.22 Initial values of classrooms 0.02; ICC(1) ¼ 0.03 Changes in classrooms 0.01 Residuals 0.04 AIC 3181.68 BIC 3232.91

Step 3

g (SE)

Step 4

g (SE)

df

g (SE)

df

df

3.52 (0.03)*** 0.01 (0.02)

1099 963

3.52 (0.02)*** 0.01 (0.02)

1099 958

3.52 (0.02)*** 0.02 (0.02)

1098 957

0.08 0.03 0.05 0.28

1099 1099 1099 1099

0.08 0.03 0.05 0.31

1099 1099 1099 1099

0.08 0.02 0.05 0.32

1098 1098 1098 1098

(0.02)*** (0.03) (0.01)*** (0.02)***

0.02 (0.03) 0.01 (0.03) 0.01 (0.04)

57 57 57

0.01 (0.03) 0.02 (0.03) 0.04 (0.04) 0.01 0.10 0.11 0.03 0.11

0.18 0.22 0.01 0.01 0.03 2885.44 2975.98

(0.02)*** (0.03) (0.01)*** (0.02)***

0.18 0.22 0.01 0.01 0.03 2897.77 3016.55

(0.02) (0.03)*** (0.04)** (0.04) (0.05)*

57 57 57 958 958 958 958 958

(0.02)*** (0.03) (0.01)*** (0.02)***

0.01 (0.03) 0.02 (0.03) 0.03 (0.04) 0.01 0.10 0.12 0.03 0.11 0.03 0.09

(0.02) (0.03)*** (0.04)** (0.04) (0.05)* (0.02) (0.03)**

57 57 57 957 957 957 957 957 1098 957

0.18 0.22 0.01 0.01 0.03 2904.63 3034.70

Note. Control variables are sex (0 ¼ female, 1 ¼ male) and migration background (0 ¼ no, 1 ¼ yes). * p < 0.05, ** p < 0.01, *** p < 0.001, two-tailed. Aa ¼ Academic achievement; Ad ¼ Academic disengagement; Es ¼ Emotional support; Co ¼ Classroom organization; Is ¼ Instructional support.

Fig. 3. Development of students' perceived teacher justice from grade five to grade six predicted by students' disengagement (student ratings) and teachers' emotional support (classroom observation) in grade five (bold line ¼ significant slope, dotted line ¼ non-significant slope).

perception of teacher care. This finding is in line with longitudinal research that has revealed that the quality of studenteteacher relationships decreases over the elementary school years (Jerome et al., 2009). Over the elementary grades, teachers increasingly focus on teacher-directed activities and individual work, which results in less personalized interactions between students and

teachers. Moreover, upper elementary grade students in Switzerland are educated in several classrooms and thus have less contact with their main teacher. Finally, the upper elementary grades in Switzerland represent a transitional phase between elementary and secondary school during which teachers decide which secondary level students will attend in the seventh to tenth grades. Consequently, students are faced with heightened expectations about their academic performance in school, which might conflict with the perceptions of their teachers as a source of emotional care and support (Gasser et al., 2017). As the quality of studenteteacher relationships remains important for students' psychological well-being and academic growth in early adolescence (Maldonado-Carreo & Votruba-Drzal, 2011), researchers have argued that some classrooms and schools fail to adequately respond to students' developmental needs (Eccles et al., 1993). Early adolescence represents a phase of increasing developmental vulnerability, which might result in negative psychological adjustment such as depression or antisocial behavior (Wang, Brinkworth, & Eccles, 2013). Moreover, early adolescence is characterized by increased levels of parent-child conflict and a stronger need for autonomy (Eccles et al., 1993). During this period of enhanced emotional insecurity, teachers might represent a compensating emotional resource for early adolescents, which might protect them from developing emotional and academic problems (Pianta et al., 2003; Wentzel, 1997). Contrary to our prediction, we did not find that perceived teacher justice generally decreased from the fifth to the sixth grade. Possibly, a more extensive longitudinal study including a time span longer than one year would be necessary to identify changes in teacher justice. Moreover, the developmental literature on children's and adolescents' experiences of justice, equality and discrimination mostly focused on students from minority groups

Please cite this article in press as: Gasser, L., et al., Emotionally supportive classroom interactions and students' perceptions of their teachers as caring and just, Learning and Instruction (2017), http://dx.doi.org/10.1016/j.learninstruc.2017.08.003

L. Gasser et al. / Learning and Instruction xxx (2017) 1e11

(e.g., Ruck & Wortley, 2002; Greene et al., 2006). For example, Greene et al. (2006) found that changes in adolescents' perceptions of adult discrimination only occurred for black adolescents but not for other minority groups (e.g., Asian-American adolescents). It is therefore possible that intra-individual change in perceived teacher justice only occurs for specific groups of students. As predicted and in accordance with the previous research (e.g., Sabol & Pianta, 2012), academic disengagement was found to be negatively related to students' perceptions of teacher care and teacher justice. This finding might indicate that academically disengaged students evoke negative interactions from their teachers, which results in negative perceptions of studenteteacher relationships. In turn, these negative student perceptions might increase subsequent disengagement in school (Skinner & Belmont, 1993). Moreover, consistent with our hypotheses and the previous research (e.g., Wentzel, 2002), low-achieving students perceived less teacher justice than high-achieving students. Weinstein et al. (1987) have shown that upper elementary grade students compared to younger children are more sensitive to differential treatment of classmates with low or high academic abilities. Consequently, the lower justice ratings of the students with high academic risks compared to students with low academic risks in our sample might mirror teachers' differential treatment of students with different academic abilities. However, contrary to our hypothesis, academic achievement negatively predicted perceptions of teacher care. A possible explanation for this unexpected finding might be that low-achieving students are more in need of teachers' support compared to high-achieving students and thus are more likely to report high teacher care. Despite this, they might believe that classmates receive more support and care than they do and consequently rate the teacher as less just than their highachieving peers. Moreover, students' justice ratings might refer to different experiences than the experience of teachers' equal or unequal distribution of affection and care (e.g., experiences of unjust disciplinary or grading practices). Our findings further suggest that teachers' observed emotional support moderated the effect (a) of grade and (b) of academic disengagement on students' perceptions of their teachers. First, decreases in perceived teacher care only occurred in students from classrooms that were characterized by a low quality of emotional support from teachers. Second, the perceptions of teacher justice only increased in academically disengaged students who were in school classes with high emotional support. In contrast, if teachers provided low emotional support, the perceptions of teacher justice decreased in both students with high and low academic disengagement. Similarly, O'Connor (2010) has shown that high emotional support from teachers protects students from developing negative studenteteacher relationships from first through fifth grade. If students experience the teacher supporting their autonomy, treating them with warmth and being responsive to their academic and emotional needs, they might view the teachers as supportive and just, despite changes during the elementary grades that might challenge the formation of positive studenteteacher relationships (e.g., increasing academic demands). Moreover, the research that has included younger children has shown that teachers' observed emotional support did not have the same relevance for all children with regard to the quality of the relationships with their teachers (e.g., Hamre & Pianta, 2005; Rudasill et al., 2016). For example, teachers who provided children at risk for school failure with high emotional support reported less conflict with these children than teachers who provided low emotional support (Hamre & Pianta, 2005). Therefore, high-quality emotional support might buffer students with academic risks from developing negative studenteteacher relationships. It should be noted that our hypothesis that teachers' emotional

9

support protects students with academic risks from developing negative perceptions of their teachers was only partially supported. For example, we did not find that teachers' emotional support protects low-achieving students from forming negative perceptions of teacher justice. Additional classroom conditions, such as teachers' attempts to treat students justly (e.g., with regard to grading or disciplinary practices), should be considered in future studies. Moreover, the upper elementary grades are characterized by increasing levels of academic expectancies (e.g., Gasser et al., 2017). It is therefore possible that emotionally supportive classroom environments are equivalently important for all students during the upper elementary grades. This interpretation is consistent with findings from Madill et al. (2014) who studied first, third, and fifth graders based on a cross-sectional design. They found that teachers' emotional support did not moderate the effect of social risks (i.e., low social peer preference) on teacher-perceived closeness (Madill et al., 2014). Thus, more research is required to better understand these inconsistencies in moderation effects in upper elementary grade students. The effects of teachers' emotional supports were unique, i.e., they remained after controlling for classrooms in the domains of instructional support and classroom organization. While no effects were found for teachers' instructional support, classroom organization significantly predicted perceptions of teacher justice. Surprisingly, we found that teacher justice decreased when the classroom was observed to be highly organized, whereas teacher justice remained unchanged when classroom organization was low. An explanation for this unexpected finding might be that students perceived highly organized teachers to be controlling and less engaged in creating a just classroom community. However, contrary to this conclusion, the previous research that has included younger children has revealed that classroom organization was not related or even positively related to perceptions of studenteteacher relationships (e.g., Brock & Curby, 2014; O'Connor, 2010). As early adolescents develop a growing need for autonomy, a classroom environment that is highly controlling might be perceived more negatively by early adolescents than by younger students (Eccles et al., 1993). Therefore, classroom organization might be differentially related to students' perceptions of their relationships with the teacher in different age groups. To test this hypothesis, longitudinal research that includes a wider age range is warranted. The following limitations should be considered in the interpretation of the results. First, our study only included two measurement points. More extensive longitudinal studies are necessary to obtain a more comprehensive picture of how students' perceptions of teacher care and justice develop over the elementary grades. Second, this study did not consider how teachers' emotional support might affect students' social and academic adaption through students' perceptions of their teacher as caring and just. To investigate whether students' perceptions might function as an important social-cognitive mechanism through which teacher practices relate to students' future social and academic outcomes, research that focuses on mediational analyses is necessary. Third, we did not control for the stability of observed emotional support; thus, we were not able to predict changes in teacher perceptions while controlling for possible changes in observed emotional support. Future research could control for potential co-developments of changes in emotional support and changes in perceived teacher perceptions. Fourth, the few existing longitudinal studies suggest bidirectional relations between studenteteacher relationships and academic achievement or engagement (Ko
sir & Tement, 2014; Quin, 2017). Therefore, the perceived quality of studenteteacher relationships is not only the result but also the cause for students' academic characteristics (Maldonado-Carreo & Votruba-Drzal, 2011).

Please cite this article in press as: Gasser, L., et al., Emotionally supportive classroom interactions and students' perceptions of their teachers as caring and just, Learning and Instruction (2017), http://dx.doi.org/10.1016/j.learninstruc.2017.08.003

10

L. Gasser et al. / Learning and Instruction xxx (2017) 1e11

Extensive research confirms that students' academic engagement decreases during adolescence and that one explanation for this finding is that studenteteacher relationships become less personal during the upper elementary and secondary grades (e.g., Eccles et al., 1993; Wang et al., 2013). The present study suggests that supporting pre- and in-service teachers in improving their emotionally supportive behaviors could represent a key strategy for promoting positive changes in studenteteacher relationships. It is certainly important that pre- and in-service-teachers acquire detailed knowledge in their university courses about the role of teachers' emotional support, characteristics of positive studenteteacher relationships, how they develop and how they affect children's and adolescents' cognitive and social development in different student populations (e.g., students with and without academic risk factors). However, pre- and in-service teachers should also be supported to apply this knowledge in actual classroom situations (Gasser & Althof, 2017). For example, video reflections with regard to teachers' emotional support allow students to transfer their knowledge about effective emotional support to highly contextualized situations, without being exposed to the full complexity of real-life classroom contexts (Blomberg, Renkl, Gamoran Sherin, Borko, & Seidel, 2013). Moreover, teacher educators could support pre- and in-service teachers increasing their emotional supportive behavior by providing them specific feedback and suggestions for improving their teaching on the basis of reliable and valid observational or student survey data (Kane & Staiger, 2012; Pianta & Hamre, 2009). Altogether, teacher trainings that help pre- and in-service teachers to improve their emotionally supportive behaviors in contexts that are closely related to their actual or future teaching practice (e.g., video reflections), might represent a promising way to buffer the negative effects of individual or school risk factors on changes in perceived studenteteacher relationships in adolescence. References Ball, J., Lohaus, A., & Miebach, C. (2006). Psychische Anpassung und schulische Leistungen beim Wechsel von der Grundschule zur weiterführenden Schule [Psychological adjustment and school achievement during transition from elementary to secondary school]. Zeitschrift für Entwicklungspsychologie und €dagogische Psychologie, 38, 101e109. http://dx.doi.org/10.1026/0049Pa 8637.38.3.101. Battistich, V., Solomon, D., Watson, M., & Schaps, E. (1997). Caring school communities. Educational Psychologist, 32, 137e151. http://dx.doi.org/10.1207/ s15326985ep3203_1. Bliese, P. D. (2002). Multilevel modeling in R: A brief introduction to R, the multilevel package, and the NLME package. Washington, DC: Walter Reed Army Institute of Research. Bliese, P., & Ployhart, R. (2002). Growth modeling using random coefficient models: Model building, testing, and illustrations. Organizational Research Methods, 5, 362e387. http://dx.doi.org/10.1177/109442802237116. Blomberg, G., Renkl, A., Gamoran Sherin, M., Borko, H., & Seidel, T. (2013). Five research-based heuristics for using video in pre-service teacher education. Journal for Educational Research Online, 5(1), 90e114. Brackett, M. A., Reyes, M. R., Rivers, S. E., Elbertson, N. A., & Salovey, P. (2011). Classroom emotional climate, teacher affiliation, and student conduct. Journal of Classroom Interaction, 46, 27e36. Brincks, A. M., Enders, C. K., Llabre, M. M., Bulotsky-Shearer, R. J., Prado, G., & Feaster, D. J. (2017). Centering predictor variables in three-level contextual models. Multivariate Behavioral Research, 54, 149e163. http://dx.doi.org/ 10.1080/00273171.2016.1256753. Brock, L. L., & Curby, T. W. (2014). Emotional support consistency and teacherechild relationships forecast social competence and problem behaviors in prekindergarten and kindergarten. Early Education and Development, 25, 661e680. http:// dx.doi.org/10.1080/10409289.2014.866020. Buyse, E., Verschueren, K., Doumen, S., Van Damme, J., & Maes, F. (2008). Classroom problem behavior and teacher-child relationships in kindergarten: The moderating role of classroom climate. Journal of School Psychology, 43, 367e391. http://dx.doi.org/10.1016/j.jsp.2007.06.009. Donat, M., Umlauft, S., Dalbert, C., & Kamble, S. V. (2012). Belief in a just world, teacher justice, and bullying behavior. Aggressive Behavior, 38, 185e193. http:// dx.doi.org/10.1002/ab.21421. Eccles, J. S., Midgley, C., Wigfield, A., Buchanan, C. M., Reuman, D., Flanagan, C., et al. (1993). Development during adolescence: The impact of stage-environment fit

on young adolescents' experiences in schools and in families. American Psychologist, 48, 90e101. http://dx.doi.org/10.1037/0003-066X.48.2.90. Enders, C. K., & Tofighi, D. (2007). Centering predictor variables in cross-sectional multilevel models: A new look at an old issue. Psychological Methods, 12, 121e138. http://dx.doi.org/10.1037/1082-989X.12.2.121. Frankena, W. K. (1973). Ethics. Englewood Cliffs, New Jersey: Prentice-Hall. Gasser, L., & Althof, W. (2017). Developing teachers' cognitive strategies of promoting moral reasoning and behavior in teacher education. In D. J. Clandinin, & J. Husu (Eds.), International handbook of research on teacher education (pp. 387e402). London: Sage Publishers. Gasser, L., Grütter, J., Torchetti, L., & Buholzer, A. (2017). Competitive classroom norms and exclusion of children with academic and behavior difficulties. Journal of Applied Developmental Psychology, 49, 1e11. http://dx.doi.org/10.1016/ j.appdev.2016.12.002. Gasser, L., Malti, T., & Buholzer, A. (2014). Swiss children's moral and psychological judgments about exclusion of children with disabilities. Child Development, 85, 532e548. http://dx.doi.org/10.1111/cdev.12124. Gibbs, J. C. (2003). Moral development and reality. Beyond the theories of Kohlberg and Hoffman. Thousand Oaks, CA: Sage. Greene, M. L., Way, N., & Pahl, K. (2006). Trajectories of perceived adult and peer discrimination among black, Latino, and Asian American adolescents: Patterns and psychological correlates. Developmental Psychology, 42, 218e236. http:// dx.doi.org/10.1037/0012-1649.42.2.218. , A., & Meyer, B. (2017). Promoting inclusion via crossGrütter, J., Gasser, L., Zuffiano group friendship: The mediating role of change in trust and sympathy. Child Development. http://dx.doi.org/10.1111/cdev.12883. Hamre, B. K., & Pianta, R. C. (2005). Can instructional and emotional support in the first-grade classroom make a difference for children at risk of school failure? Child Development, 76, 949e967. http://dx.doi.org/10.1111/j.14678624.2005.00889.x. Hamre, B. K., Pianta, R. C., Downer, J. T., DeCoster, J., Mashburn, A. J.Jones, S. M., … (2013). Teaching through interactions: Testing a developmental framework of teacher effectiveness in over 4,000 classrooms. The Elementary School Journal, 113, 461e487. http://dx.doi.org/10.1086/669616. Hamre, B. K., Pianta, R. C., Downer, J. T., & Mashburn, A. J. (2008). Teachers' perceptions of conflict with young students: Looking beyond problem behaviors. Social Development, 17, 115e136. http://dx.doi.org/10.1111/j.14679507.2007.00418.x. Hughes, J. N., Dyer, N., Luo, W., & Kwok, O. M. (2009). Effects of peer academic reputation on achievement in academically at-risk elementary students. Journal of Applied Developmental Psychology, 30, 182e194. http://dx.doi.org/10.1016/ j.appdev.2008.12.008. Jerome, E., Hamre, B., & Pianta, R. C. (2009). Teacherechild relationships from kindergarten to sixth grade: Early childhood predictors of teacher-perceived conflict and closeness. Social Development, 18, 915e945. http://dx.doi.org/ 10.1111/j.1467-9507.2008.00508.x. Kane, T. J., & Staiger, D. O. (2012). Gathering feedback for teachers: Combining highquality observations with student surveys and achievement gains (Research paper prepared for the Bill and Melinda Gates Foundation). Retrieved from MET project website http://metproject.org/downloads/MET_Gathering_Feedback_ Research_Paper.pdf . Ko
sir, K., & Tement, S. (2014). Teacherestudent relationship and academic achievement: A cross-lagged longitudinal study on three different age groups. European Journal of Psychology of Education, 29, 409e428. http://dx.doi.org/ 10.1007/s10212-013-0205-2. Lee, P., & Bierman, K. L. (2015). Classroom and teacher support in kindergarten: Associations with the behavioral and academic adjustment of low-income students. Merrill-Palmer Quarterly, 61, 383e411. http://dx.doi.org/10.13110/ merrpalmquar1982.61.3.0383. Lerner, R. M. (1998). Theories of human development: Contemporary perspectives. In W. Damon, & R. M. Lerner (Eds.), Handbook of child psychology: Vol. XX (5th Ed.): Theoretical models of human development (pp. 1e24). New York: John Wiley & Sons, Inc. Little, R. J. A. (1988). A test of missing completely at random for multivariate data with missing values. Journal of the American Statistical Association, 83, 1198e1202. http://dx.doi.org/10.1080/01621459.1988.10478722. Madill, R. A., Gest, S. D., & Rodkin, P. C. (2014). Students' perceptions of relatedness in the classroom: The roles of emotionally supportive teacher-child interactions, children's aggressive-disruptive behaviors, and peer social preference. School Psychology Review, 43, 86e105. Maldonado-Carreo, C., & Votruba-Drzal, E. (2011). Teacherechild relationships and the development of academic and behavioral skills during elementary school: A within-and between-child analysis. Child Development, 82, 601e616. http:// dx.doi.org/10.1111/j.1467-8624.2010.01533.x. Malti, T., Killen, M., & Gasser, L. (2012). Social judgments and emotion attributions about exclusion in Switzerland. Child Development, 83, 697e711. http:// dx.doi.org/10.1111/j.1467-8624.2011.01705.x. Mashburn, A. J., Pianta, R. C., Hamre, B. K., Downer, J. T., Barbarin, O., Bryant, D., et al. (2008). Measures of classroom quality in pre-kindergarten and children's development of academic, language and social skills. Child Development, 79, 732e749. http://dx.doi.org/10.1111/j.1467-8624.2008.01154.x. Murray, C., Murray, K. M., & Waas, G. A. (2008). Child and teacher reports of teacherestudent relationships: Concordance of perspectives and associations with school adjustment in urban kindergarten classrooms. Journal of Applied Developmental Psychology, 29, 49e61. http://dx.doi.org/10.1016/

Please cite this article in press as: Gasser, L., et al., Emotionally supportive classroom interactions and students' perceptions of their teachers as caring and just, Learning and Instruction (2017), http://dx.doi.org/10.1016/j.learninstruc.2017.08.003

L. Gasser et al. / Learning and Instruction xxx (2017) 1e11 j.appdev.2007.10.006. Nurmi, J.-E. (2012). Students' characteristics and teacherechild relationships in instruction: A meta-analysis. Educational Research Review, 7, 177e197. http:// dx.doi.org/10.1016/j.edurev.2012.03.001. Nurmi, J. E., & Kiuru, N. (2015). Students' evocative impact on teacher instruction and teacherechild relationships: Theoretical background and an overview of previous research. International Journal of Behavioral Development, 39, 445e457. http://dx.doi.org/10.1177/0165025415592514. O'Connor, E. (2010). Teacherechild relationships as dynamic systems. Journal of School Psychology, 48, 187e218. http://dx.doi.org/10.1016/j.jsp.2010.01.001. Patrick, H., Ryan, A. M., & Kaplan, A. (2007). Early adolescents' perceptions of the classroom social environment, motivational beliefs, and engagement. Journal of Educational Psychology, 99, 83e98. http://dx.doi.org/10.1037/0022-0663.99.1.83. Peter, F., Kloeckner, N., Dalbert, C., & Radant, M. (2012). Belief in a just world, teacher justice, and student achievement: A multilevel study. Learning and Individual Differences, 22, 55e63. http://dx.doi.org/10.1016/j.lindif.2011.09.011. Pianta, R. C., & Hamre, B. K. (2009). Conceptualization, measurement, and improvement of classroom processes: Standardized observation can leverage capacity. Educational Researcher, 38, 109e119. http://dx.doi.org/10.3102/ 0013189X09332374. Pianta, R., Hamre, B., & Mintz, S. (2012). Classroom assessment scoring system [class] manual, upper elementary. Baltimore: Brookes. Pianta, R., Hamre, B., & Stuhlman, M. (2003). Relationships between teachers and children. In W. M. Reynolds, G. E. Miller, & I. B. Weiner (Eds.), Handbook of psychology: Vol. 7. Educational psychology (pp. 199e234). Hoboken, NJ: Wiley. http://dx.doi.org/10.1002/0471264385.wei0710. Quin, D. (2017). Longitudinal and contextual associations between teacher-student relationships and student engagement: A systematic review. Review of Educational Research, 87, 345e387. http://dx.doi.org/10.3102/0034654316669434. Rimm-Kaufman, S. E., Curby, T. W., Grimm, K. J., Nathanson, L., & Brock, L. L. (2009). The contribution of children's self-regulation and classroom quality to children's adaptive behaviors in the kindergarten classroom. Developmental Psychology, 45, 958e972. http://dx.doi.org/10.1037/a0015861. Roorda, D. L., Koomen, H. M., Spilt, J. L., & Oort, J. (2011). The influence of affective teacherestudent relationships on students' school engagement and achievement: A meta-analytic approach. Review of Educational Research, 81, 493e529. http://dx.doi.org/10.3102/00346543114. Ruck, M. D., & Wortley, S. (2002). Racial and ethnic minority high school students' perceptions of school disciplinary practices: A look at some Canadian findings. Journal of Youth and Adolescence, 31, 185e195. http://dx.doi.org/10.1023/A: 1015081102189. Rudasill, K. M., Hawley, L., Molfese, V. J., Tu, X., Prokasky, A., & Sirota, K. (2016). Temperament and teacherechild conflict in Preschool: The moderating roles of classroom instructional and emotional support. Early Education and Development, 27, 859e874. http://dx.doi.org/10.1080/10409289.2016.1156988. Rueger, S. Y., Malecki, C. K., & Demaray, M. K. (2010). Relationship between multiple sources of perceived social support and psychological and academic adjustment in early adolescence: Comparisons across gender. Journal of Youth and Adolescence, 39, 47e61. http://dx.doi.org/10.1007/s10964-008-9368-6. Ryan, A. M., & Patrick, H. (2001). The classroom social environment and changes in adolescents' motivation and engagement during middle school. American Educational Research Journal, 38, 437e460. http://dx.doi.org/10.3102/ 00028312038002437.

11

Sabol, T. J., & Pianta, R. C. (2012). Recent trends in research on teacherechild relationships. Attachment & Human Development, 14, 213e231. http://dx.doi.org/ 10.1080/14616734.2012.672262. Salmela-Aro, K., Kiuru, N., & Nurmi, J. E. (2008). The role of educational track in adolescents' school burnout: A longitudinal study. British Journal of Educational Psychology, 78, 663e689. http://dx.doi.org/10.1348/000709908X281628. nez, V., Carbonell-Baeza, A., Keating, X. D., Ruiz, J. R., & Castro-Pin ~ ero, J. Segura-Jime (2015). Association of sleep patterns with psychological positive health and health complaints in children and adolescents. Quality of Life Research, 24, 885e895. http://dx.doi.org/10.1007/s11136-014-0827-0. Skinner, E. A., & Belmont, M. J. (1993). Motivation in the classroom: Reciprocal effects of teacher behavior and student engagement across the school year. Journal of Educational Psychology, 85, 571e581. http://dx.doi.org/10.1037/00220663.85.4.571. Spilt, J. L., Hughes, J. N., Wu, J. Y., & Kwok, O. M. (2012). Dynamics of teacher-student relationships: Stability and change across elementary school and the influence on children's academic success. Child Development, 83, 1180e1195. http:// dx.doi.org/10.111/j.1467-8624.2012.01761.x. Swiss Coordination Centre for Research in Education. (2014). Bildungsbericht schweiz 2014 [Swiss education report 2014]. Aarau: Schweizerische Koordinationsstelle für Bildungsforschung. Swiss Federal Statistical Office. (2017). Retrieved from https://www.bfs.admin.ch/ bfs/en/home.html. Tuominen-Soini, H., & Salmela-Aro, K. (2014). Schoolwork engagement and burnout among Finnish high school students and young adults: Profiles, progressions, and educational outcomes. Developmental Psychology, 50, 649e662. http:// dx.doi.org/10.1037/a0033898. Wang, M. T., Brinkworth, M., & Eccles, J. (2013). Moderating effects of teacherestudent relationship in adolescent trajectories of emotional and behavioral adjustment. Developmental Psychology, 49, 690e705. http://dx.doi.org/10.1037/ a0027916. Weinstein, R. S., Marshall, H. H., Sharp, L., & Botkin, M. (1987). Pygmalion and the student: Age and classroom differences in children's awareness of teacher expectations. Child Development, 58, 1079e1093. http://dx.doi.org/10.2307/ 1130548. Wentzel, K. R. (1997). Student motivation in middle school: The role of perceived pedagogical caring. Journal of Educational Psychology, 89, 411e419. http:// dx.doi.org/10.1037/0022-0663.89.3.411. Wentzel, K. R. (1998). Social support and adjustment in middle school: The role of parents, teachers, and peers. Journal of Educational Psychology, 90, 202e209. http://dx.doi.org/10.1037/0022-0663.90.2.202. Wentzel, K. R. (2002). Are effective teachers like good parents? Teaching styles and student adjustment in early adolescence. Child Development, 73, 287e301. http://dx.doi.org/10.1111/1467-8624.00406. Wentzel, K. R., Battle, A., Russell, S. L., & Looney, L. B. (2010). Social supports from teachers and peers as predictors of academic and social motivation. Contemporary Educational Psychology, 35, 193e202. http://dx.doi.org/10.1016/ j.cedpsych.2010.03.002. Zee, M., Koomen, H. M., & Van der Veen, I. (2013). Studenteteacher relationship quality and academic adjustment in upper elementary school: The role of student personality. Journal of School Psychology, 51, 517e533. http://dx.doi.org/ 10.1016/j.jsp.2013.05.003.

Please cite this article in press as: Gasser, L., et al., Emotionally supportive classroom interactions and students' perceptions of their teachers as caring and just, Learning and Instruction (2017), http://dx.doi.org/10.1016/j.learninstruc.2017.08.003