Comparing students’ and teachers’ perceptions about teachers’ interpersonal behaviour in Greek secondary education

Τeachers’ interpersonal behaviour is considered important for the quality of the student–teacher relationship. The present study examined teachers’ and students’ perceptions regarding teachers’ interpersonal behaviour and the degree of their agreement at the class level. Moreover, this study investigated whether possible differences between them are associated with the teachers’ and classes’ characteristics. Participants were 1669 students and 43 secondary teachers within 85 secondary classrooms in Greece. The Greek version of the Questionnaire on Teacher Interaction was used to assess eight aspects of teachers’ interpersonal behaviour. A high degree of divergence between teachers’ self-perceptions and students’ perceptions about their teachers’ interpersonal behaviour was found. Multiple regression analyses showed that the differences among teachers’ and students’ perceptions can partially be explained by teachers’ years of teaching experience and the degree level, as well as the class educational level. Findings of this study provide specific practical implications.


Introduction
The classroom environment is a complex communication system, where two-way messaging processes create a variety of interactions between teachers and students.Both teachers' and students' behaviours mutually influence each other because each member of the interaction adapts his/her actions to the responses of the other.The way in which teachers and students react to each other -verbally and nonverbally -creates recurring patterns of interactions which lead to the development of their interpersonal relationship (Wubbels et al., 1993a).Teacher-student relationships are a multidimensional system, the quality of which is influenced by many factors such as the personality traits of the teachers and students and the perception formed by each participant about the relationship, as well as external influences from the school environment and the social context of the classroom (Pianta et al., 2003).The teacher's behaviour plays a dominant role in shaping the interpersonal relationship, as s/he is the professional who is responsible for organizing the classroom environment (Mainhard et al., 2018;Pennings & Hollenstein, 2020) and his/her actions have an impact on the students' behaviour (Suciu, 2014).
The teacher's interpersonal behaviour is considered a key factor for students' participation in the learning process.Research in the field of interpersonal relationships in the classroom supports the link between teacher behaviour and cognitive and emotional learning outcomes (den Brok et al., 2004a;Maulana et al., 2012;Wubbels & Brekelmans, 2005).Roorda et al. (2011) found a strong correlation between the encouragement and empathy of the teacher, and the emotional and cognitive results of the students.Positive and supportive teacher-student relationships are related to higher academic performance (Ma et al., 2018).In addition, the teachers' interest in their students and their willingness to help them are strong predictors of students' cognitive and affective outcomes (Roorda et al., 2011).Research findings show that teacher warmth and friendliness are associated with lower stress and greater student engagement; such evidence highlights the influence that interpersonal teacher behaviour can have on students' feelings (Mainhard et al., 2018).
To study the effect of teacher behaviour on students, the systems approach to communication (Watzlawick et al., 1967) provides the conceptual framework for understanding patterns of classroom interactions.The communicative systems approach emphasizes the repercussions of one person's actions on the other and considers every action that someone takes when another person is present as communication.It is important to explore the impact of the teachers' actions in the communicative context of the classroom because the way in which students perceive and interpret the teachers' behaviour affects their reactions, which in turn shape the patterns of interactions.Students' perceptions can differ from the teachers' actual intentions (Wubbels et al., 2006), which means that teachers' messages conveyed through teaching and interaction with students might not be perceived from all students in the same way (Mainhard, 2015).In turn, students' personal interpretations define their reactions to the teacher's behaviour and their responses to learning activities.Consequently, students' perspectives affect the quality of the relationship that they will develop with their teachers (Brinkworth et al., 2018).
Both teachers' and students' perspectives can contribute significantly when investigating the teacher behaviour.It is evident from the literature that researchers, when they explore the teachers' perceptions, focus mainly on their perceptions regarding their teaching practices and the way in which they manage their classrooms (Lüdtke et al., 2009).An initial assumption is that, as adults, teachers can easily acknowledge their thoughts and feelings that mainly define their actions (Donker et al., 2021).However, although teachers' self-perceptions constitute a significant source of information about their interpersonal behaviour, they tend to 'suffer' from potential bias, which means that some teachers are reluctant to report their weaknesses or are unaware of them (Kolar et al., 1996;Scherzinger & Wettstein, 2019).Therefore, some researchers question the effectiveness of the teachers' self-perceptions as the sole reliable source of information when examining classroom interactions (den Brok et al., 2004b;Wubbels et al., 1993b).
Moreover, examination of students' perceptions about interpersonal teacher behaviour can provide a more-comprehensive perspective of classroom interactions.In contrast with their teachers who experience much more pressure during the school day, students have more time to observe teachers' actions and practices (Scherzinger & Wettstein, 2019).In addition, students' judgments are influenced by different parameters such as the variety of school subjects, learning environments, and the different teaching styles of their teachers (den Brok et al., 2004b).Furthermore, teachers' self-perceptions are a single-source judgment, whereas the students' perceptions regarding their teacher are the result of a multisource judgment and hence can contribute significantly to the study of teachers' behaviour (den Brok et al., 2004a;Donker et al., 2021).Of course, students' assessments are also influenced by several factors, such as personal preferences, student characteristics, or the pressure of grades (Wagner et al., 2016).

Measuring interpersonal teacher behaviour
To study interpersonal teacher behaviour, Wubbels et al. (1993a) adapted Leary's (1957) interpersonal circle in the educational context to develop the Model of Interpersonal Teacher Behaviour, which in a recent version is referred to as the Teacher Interpersonal Circle (IPC-T) (Gurtman, 2011;Pennings & Hollenstein, 2020).This model comprises two independent dimensions: the Agency dimension refers to who is in control of the relationship, and the Communion dimension describes the degree of teacher-students cooperation (Fig. 1).In particular, the Agency dimension is represented by an axis with the two contrary poles of Dominance and Submission.The Communion dimension is also represented by the two contrary poles of Cooperation and Opposition.Researchers can investigate interpersonal behaviour in a variety of different positions within the above-mentioned dimensions.Furthermore, the IPC-T delineates one's behaviour in a coordinates system with two dimensions that consists of eight subdomains that develop and describe a circular pattern of eight behaviours: Directing (DC), Helpful (CD), Understanding (CS), Compliant (SC), Uncertain (SO), Dissatisfied (OS), Confrontational (OD) and Imposing (DO).Each quadrant of the coordinate structure represents two domains of behaviour (Sun et al., 2018).The Agency dimension consists of behaviours in the sectors closest to the Dominance-Submission axis, such as imposing, directing, uncertain and compliant behaviours.The sectors that make up the Communion dimension are placed closer to the Cooperation-Opposition axis: helpful, understanding, dissatisfied and confrontational behaviours.Each interpersonal behaviour is developed by the combination of Agency and Communion and defines its position in the circle.
Relying on IPC-T, Wubbels et al. (1993a) designed and validated the Questionnaire on Teacher Interaction (QTI) in the Netherlands.This measure evaluates the degree of agency and communion that a teacher communicates to the classroom.Τhe QTI was designed to assess students' perceptions of the teachers' interpersonal behaviour within a class in the secondary education (Wubbels et al., 2014).The original Dutch questionnaire consists of 77 items, an American version includes 64 items (Wubbels & Levy, 1991) and an Australian version has 48 items (Rickards et al., 2005).The QTI is the most-commonly used instrument for assessing teacher interpersonal behaviour.Therefore, there are many different versions of QTI at a worldwide level (e.g., Mainhard, 2015;Sun et al., 2018) and many studies have confirmed its reliability and validity (e.g., den Brok et al., 2006b;Telli et al., 2007).
The application of QTI in many countries using students' perceptions has contributed to understanding of interpersonal teacher behaviour in different cultural environments.Levy et al. (1993) examined average students' perceptions for a sample of Dutch, American and Australian secondary schools.Students considered teachers to have a higher degree of directing, helpful, understanding and imposing behaviours, while the lowest scores were found for uncertain and dissatisfied behaviours.It is noteworthy that American and Australian teachers were rated as more enforcing than Dutch teachers.In several surveys in Asian countries, students perceived their teachers as leading and imposing, showing a high degree of interpersonal closeness (Maulana et al., 2012;Safa & Doosti, 2017;Sivan & Chan, 2013).Despite the extensive research using the QTI in Western and Asian countries, research on interpersonal perspectives of the classroom in the Greek educational context is scarce.To the best of our knowledge, only a few studies were carried out in the Greek elementary school context, but they did not provide information about students' and teachers' evaluations of teacher behaviour (Charalampous & Kokkinos, 2018;Charalampous et al., 2016).Only one study was conducted in the secondary education, which focused on the QTI's validation (Tsigilis et al., 2021).

Comparing the perceptions of students and teachers
Researchers considering the class as the unit of analysis can explore the degree of agreement between teachers' and students' perceptions about teachers' interpersonal behaviour and identify the divergence and convergence of their perceptions (e.g., Maulana et al., 2012).Such knowledge can help teachers to better understand how effectively they communicate with their students which, in turn, can result in a higher-quality learning environment and experience for them (Brekelmans et al., 2011;Wubbels et al., 1992).Prior research with the QTI compared students' and teachers' perceptions through a variety of statistical analyses.Findings from these studies showed a discrepancy between teachers' and students' perceptions, indicating that teachers perceived themselves as more leading and supportive than their students (den Brok et al., 2002;Wubbels et al., 1993b).When Maulana et al. (2012) used absolute difference scores, difference scores and the measurement error, they found that a group of teachers reported higher scores than their students on the directing, helpful, understanding, compliant and imposing scales.In contrast, the majority of teachers reported lower scores than their students on the uncertain, dissatisfied and confrontational scales.In another study that used independent samples t-tests and effect sizes showed that the magnitude of differences between students' and the teachers' perceptions was large for confrontational, understanding, dissatisfied and compliant scales (Safa & Doosti, 2017).This could mean that students evaluated their teachers' behaviour more highly than their teachers for dissatisfied and reprimanding behaviours and lower for understanding and tolerant behaviours.When Wubbels et al. (1992) tested divergence perceptions with t-tests and absolute difference scores, some teachers perceived their behaviour more favorably than their students, while others perceived themselves more negatively.According to the researchers, the higher self-perception could reflect their desired behaviour, while the lower self-perception could reflect an attempt to protect themselves from frustration.In general, empirical data indicate that most teachers tend to overestimate -relative to the students' perceptions -behaviours that have been found to be positively related to cognitive and socio-emotional learning (e.g., Directing, Helpful), while they underestimate behaviours (e.g., Compliant, Uncertain) that are negatively related to students' progress (Brekelmans et al., 2011;Wubbels et al., 1993b).
Research into the differences between student and teacher perceptions and their association with background variables is scarce.Differences in perceptions have been associated with teaching experience, although the results are mixed.Den Brok et al. (2006a) found no significant relationship between the teachers' experience and divergence in the perceptions of the teachers' instructional behaviour.In contrast, a longitudinal analysis revealed that the discrepancy regarding teacher cooperative behaviours increases as teachers' experience increase (Brekelmans et al., 2002).
The present study explored teachers' self-perceptions and students' perceptions regarding teachers' interpersonal behaviour in Greek secondary education.We extended our examination to explore the divergence of teachers' and students' perceptions about interpersonal teacher behaviour.We then investigated whether the possible differences in their perceptions were associated with the teachers' and classes' characteristics.To this end, three research questions were formulated and tested.
1. What are teachers' and students' perceptions about teachers' interpersonal behaviour?2. Do teachers and students perceive teacher interpersonal behaviour in a similar manner?3. Do teacher-level and class-level characteristics (teaching experience, degree level, and educational level) relate to the difference between teachers' and students' perceptions of interpersonal behaviour?

Participants
The sample comprised 43 teachers and their 1669 students from 85 classes in secondary education from 18 schools of the prefecture of Thessaloniki in Northern Greece.Each teacher was teaching in two different classrooms with one exception.Of the participating teachers, 29 were female and their ages ranged from 38 to 55 years.In terms of students' gender, 896 (53.7%) were girls and 723 (43.3%) were boys, while 50 (3%) students did not report their gender.Students ranged in age between 14 and 17 years.Greek secondary education is divided into two levels: junior high school (13-15 years old students) and senior high school (16-18 years old students).655 students (39.2%) from 38 classes attended junior high school and 1014 students (60.8%) from 47 classes attended senior high school.
Table 1 presents the sample's characteristics.

Measurements: Teachers' interpersonal behaviour
The Greek version of the 36-item Questionnaire on Teacher Interaction (QTI) was implemented to examine both teachers' and students' perceptions of interpersonal teacher behaviour.Using the Greek-language version of the 64-item QTI adapted by Kyriakides (2005), Tsigilis et al. ( 2021) investigated its applicability to the Greek secondary education, by employing a cross-validation approach.This resulted in a version of 36 items based on conceptual and statistical criteria (Sun et al., 2018).Specifically, items that evaluated didactic issues, or had a student-centered rather than teacher-centered orientation, were eliminated.Also, it was observed that some items showed low item-total correlation and were discarded.The above actions resulted in a 36-item QTI (5 items each for the Directing, Helpful, Understanding and Uncertain scales and 4 items each for the Compliant, Dissatisfied, Confrontational and Imposing scales).Internal consistency values for students' perceptions were high, supporting the reliability of the 36-item Greek QTI.Moreover, when the circumplex structure of the QTI was evaluated using exploratory structural equation modeling (ESEM) and the R package CircE, which is capable of examining circular models, the construct validity of the Greek version was demonstrated.The Greek-language version of the 64-item QTI by Kyriakides (2005) that was applied to Cyprus's primary schools showed acceptable validity and reliability.Other studies examined the QTI's psychometric properties in terms of internal consistency, concurrent and factorial validity in the Greek language using Cypriot or Greek elementary students' responses (Charalampous & Kokkinos, 2018;Kokkinos et al., 2009).
In the present study, teachers' and students' responses were given on a five-point Likert scale anchored by never (1) to always (5).Following Wubbels et al. (1993a) study, each teacher (Ν = 43) who was teaching in two different classrooms was evaluated by the students of these classes.Students within a classroom evaluated only one teacher.

Procedure
Approval of this study was provided by the Greek National Educational Policy Institute.The researchers informed parents and teachers about the purpose and procedures of the current study and asked them to participate voluntarily.In addition, parents signed informed consent forms before the researchers' first visit to schools.After consent forms were obtained, students also were asked for their consent, after which they were informed about the purposes and process of this research.Hence, they were assured that their responses to the survey would be anonymous and would not be read by their teachers.Then, teachers rated their interpersonal behaviour by filling in the QTI questionnaire.Participants were selected based mainly on the teachers' availability to contribute to the study.This recruitment strategy produced a convenience sample.

Data analysis
All analyses were performed using the SPSS ver.26.MacDonald's ω coefficient was used to examine the internal consistency of QTI subscales at the class and teacher level (Hayes & Coutts, 2020).In our analysis, QTI scores were transformed from 1 to 5 to scores between 0 to 1.The transformed scores represented a proportion (of the maximum) scores, which were easier to compare (see den Brok et al., 2006b).Then, QTI scores were computed for all scales across teacher and class levels.Moreover, the Agency and Communion dimensions were estimated using the linear combinations of the eight scale scores from the QTI on the basis of their position on the interpersonal circle (e.g., Wubbels et al., 2005).
After the above preliminary analyses, descriptive statistics (means and standard deviations) were run to initially obtain teachers' and students' perceptions for each QTI scale at teacher and class level (first research question).In order to initially investigate whether there were significant differences between teachers' and students' perceptions of each QTI scale and dimension, paired samples t-tests were used.The convergence or divergence between teachers' and students' perceptions was further checked through the estimation of absolute difference scores, regular difference scores, and measurement error of the difference scores (see den Brok et al., 2006a for further details about this procedure) (second research question).In particular, difference scores were computed using student perceptions minus teacher perceptions for QTI scales and dimensions.Following den Brok et al.'s (2006a) guidelines, we then identified the frequency of classes with convergence and divergence between teachers' and students' group perceptions.
Finally, in order to explain the possible variance in the differences between teachers' and students' perceptions, two multiple regression models were estimated (third research question).The first model referred to differences between teachers' and students' perceptions on the Agency dimension and the other one the Communion dimension.In these models, teaching experience, educational level, and degree level were examined as predictors of the difference between teachers' and students' perceptions.These variables were dummy coded and then entered into the model (0 for 1-10 years, 1 for 11-20 years and 2 for < 20 years for the teaching experience variable, 0 for Bachelor, 1 for Master, and 2 for doctoral studies for the degree level, and 0 for junior high school and 1 for senior high school for the educational level).

Descriptive statistics and internal consistency of the Greek QTI
Internal consistency for the QTI scales and the two dimensions at the class and teacher levels are presented in Table 2.The ω coefficient for teachers' reports ranged from 0.44 to 0.74 at scale level, suggesting moderate levels of reliability.Low internal consistency for teachers' evaluations has been reported in several other studies (e.g., den Brok et al., 2006b;Kokkinos et al., 2009).However, internal consistency values for both dimensions indicated a high level of reliability (ω = 0.72, ω = 0.83 for the Agency and Communion dimensions, respectively).Moreover, estimation of the internal consistency of the Greek QTI showed satisfactory values at the class level.
As seen in Table 2, teachers seem to have higher ratings than their classes on both the Agency and Communion dimensions.Moreover, teachers rated themselves higher than their classes on all QTI scales with the exception of the Dissatisfied (OS) and Confrontational (OD) scales.

Comparison between teachers' and classes' perceptions of teachers' interpersonal behaviour
Paired samples t-tests were conducted to determine any significant differences between teachers' and classes' perceptions of QTI scales and dimensions.As indicated in Table 3, teachers' self-perceptions were significantly different from students' perceptions on all the scales and dimensions, except the Uncertain, Dissatisfied and Confrontational scales (p > 0.05).Cohen's d values indicate that these differences were also meaningful.
To further explore the second research question, estimations of absolute difference scores, regular difference scores, and measurement errors of difference scores were conducted (Table 4).Table 4 shows that mean absolute difference scores for all QTI scales and dimensions were larger than mean measurement errors.This suggests that, on average, a divergence exists between students' and teachers' perceptions.With regard to the magnitude of differences, three were very close to zero (Uncertain, Dissatisfied, and Confrontational), two approached 0.10 (Compliant, Imposing) and three yielded substantial differences (Directing, Helpful and Understanding).Thus, a more-negative divergence (higher teachers' reports than students' reports) was found for the Directing, Helpful and Understanding scales as well as for the Agency and Communion dimensions.It seems that some teachers only slightly differ from their students in their perceptions, while others differ considerably.
The number of classes with convergence and divergence between teachers' and students' group perceptions is presented in Table 5.In terms of the Agency and Communion dimensions, over half of the classes were characterized as divergent.Also, teachers systematically evaluated themselves higher than their students across all QTI scales and the two dimensions, with the exception of the Confrontational scale.

Background characteristics and differences between teachers' and students' perceptions on QTI dimensions
A series of multiple regression analyses were conducted to explore whether teachers' experience, educational level and degree level explain the differences between teachers' and students' perceptions on QTI dimensions.Overall, teachers' experience and educational level accounted for 25.1% of the variance in the differences between teachers' and students' perceptions for Agency (F (3, 38) = 4.25, p = 0.01).Moreover, teachers' experience and degree level accounted for 25.5% of the variance in the differences between teachers' and students' perceptions for Communion (F (3, 38) = 4.33, p = 0.01) (Table 6).These results suggest that more teacher experience, higher educational level and higher degree level of teachers lead to smaller divergence between teachers' and students' perceptions.

Discussion
The current study showed a high degree of divergence between teachers' self-perceptions and students' perceptions of interpersonal teacher behaviour, mainly for the communion dimension.Also, teaching experience, degree level and educational level were predictors of the divergence in perceptions.These findings are discussed explicitly below.
In terms of students' evaluations, Greek teachers showed a moderate degree of dominance and a very high degree of proximity.Students reported double scoring for the Communion dimension compared with the dimension of Agency, thus showing that teachers were perceived more as cooperative than dominant.The degree of teachers' influence in Greece is comparable to that of teachers in the USA (Wubbels & Levy, 1991) and in Australia (den Brok et al., 2006b), but higher than that in Netherlands (Brekelmans et al., 2011).On the other hand, for the Communion dimension, scores were one of the highest compared with the other countries.This finding is similar to results with Turkish teachers (Telli et al., 2007) and, to some extent, it can be compared with the results of Australian teachers (Brok et al., 2006b).The relatively high level of teacher control might reflect a traditional aspect of the Greek secondary education in which the use of the teacher's institutional power is linked with the pressure for high academic achievement (Koutrouba et al., 2012).It seems that Greek secondary education has similar characteristics to those of some Western countries (Wubbels & Levy, 1991), which also focus on students' cognitive outcomes.Nevertheless, the teachers' high degree of interpersonal proximity is more comparable to that identified in more collectivistic societies (e.g., Turkey).Given that Greek society has semi-collectivistic characteristics, with values such as immediacy and sharing playing a vital role (Gregoriadis et al., 2020), students' perceptions also might reflect the importance of these values.
Our investigation of the convergence between students' and teachers' perceptions of interpersonal teacher behaviour revealed a discrepancy across QTI scales and dimensions.More specifically, of the divergent classes, approximately 75% of teachers reported higher scores than their students on the Agency dimension.The percentage of teachers who evaluated themselves higher in the Proximity dimension was similar (78%).In addition, approximately 95% of teachers had higher scores than students for the directing, helpful and understanding scales, while 75% of the teachers reported higher scores on the imposing scale.However, a smaller number of teachers overestimated uncertain and dissatisfied behaviours, but this overestimation showed mean difference scores close to zero, which suggests an unimportant deviation.It seems that Greek teachers feel more leading and cooperative than their students think.Thus, a mismatch was found between the way in which teachers and their classes perceive the teacher's communication style, suggesting a miscommunication between the two parties.The strongest disagreement was mainly observed in proximity behaviours, showing that teachers overestimate the positive interactions with their students and thus the quality of their interpersonal relationship.In general, our findings are consistent with previous studies comparing the two perspectives regardless of the methodological approach followed (Brekelmans et al., 2011;Maulana et al., 2012;Safa & Doosti, 2017;Wubbels et al., 1992).To the best of our knowledge, only the study of Maulana et al. (2012) applied a similar methodology using the QTI and found a similar degree of perceptions divergence in the Agency and Communion dimensions.A possible explanation of this deviation is that the highest teachers' evaluations in their cooperative behaviours depict their desired behaviour or that their perceptions tend to be in line with socially-desirable answers.It is not clear whether students or teachers provide the most accurate judgment of the teacher behaviour.A better assessment of what is going on in a classroom could be achieved by conducting interviews and/or videotaping classroom interactions in order to understand the reasons for the disagreement.
Regarding the third research question, regression analysis showed that a significant percentage of variance in differences in perceptions of teachers' agentic behaviours is explained by teaching experience and educational level.On the other hand, the teacher's degree level predicts the divergence in perceptions of the teachers' communal behaviours.More specifically, the divergence in Agency dimension is smaller for the teachers who have more experience, indicating that more-experienced teachers show less leading and imposing behaviours.It seems that gaining experience goes along with effective classroom management, which could elicit positive student responses (Brekelmans et al., 2002).Previous studies have shown a weak or no relation between divergence and teaching experience, although the research generates ambiguous findings (Brekelmans et al., 1992(Brekelmans et al., , 2002)).However, in these studies, the difference in perceptions was not statistically tested, which does not allow comparison with the present finding.It was also found that the discrepancy in control behaviours was higher in junior high school compared with senior high school.A possible interpretation of this finding can be attributed to the fact that students in the lowersecondary classes are often unruly and therefore teachers might exhibit more confrontational behaviours (Levy et al., 1993).
With regard to the Communion dimension, smaller deviations were associated with teachers' acquisition of additional qualifications.It seems that teachers' continuing education after graduation further enhances their communication skills for engaging students in more-collaborative interactions.Unfortunately, we are not aware of prior studies that attempted to explain discrepancies between students' and teachers' perceptions, making comparison with the present findings difficult.Therefore, further research is needed to generate more reliable data about the explanatory variables that predict the discrepancy between teachers' and students' perceptions.
In conclusion, our findings show a high degree of divergence between teachers' selfperceptions and students' perceptions of teacher interpersonal behaviour mainly for closeness behaviours.Teachers assess their behaviour as friendly and supportive, evaluating their communication as extremely positive and assuming that they have developed a warm relationship with students.However, it seems that the teachers' interaction style is not perceived by students in the same way.In an attempt to account for this divergence, we found that teacher's shorter experience and the lower educational level are predictors of the perceptions' divergence in the Agency dimension.In addition, a significant percentage of variance in the perceptions' difference of teachers' communal behaviours was explained by teachers' higher degree level.
The present study offers interesting findings regarding teacher-student interaction in the classroom.To the best of our knowledge, this is the first attempt to compare teachers' self-perceptions and students' perceptions about interpersonal teacher behaviour in Greek secondary education.The Greek educational system, which functions in a semi-collectivist society (Gregoriadis et al., 2020), has been characterized as centralized because schools have few responsibilities and opportunities for initiatives.Moreover, the focus of secondary education on students' cognitive outcomes, especially in senior high school, emphasizes students' academic achievement (Koutrouba et al., 2012).Consequently, the current study adds new empirical data to the existing literature on classroom interaction in a different educational context from those of previous studies.
A strength of our study is the method by which perception convergence was examined, which contributed to determining both the overall magnitude of divergence and distinguishing between teachers with positive deviations from those with negative deviations.The detailed way of capturing teacher-student disagreement contributes to a deeper understanding of classroom interactions.Finally, the current study attempted, perhaps for the first time, to interpret differences between teachers' and students' perceptions in an educational context with special characteristics.If the results of the present study are replicated by other researchers and enriched with additional explanatory variables, they can serve as the basis for the design and implementation of professional development programs that promote positive interactions with students.
The present study had some limitations.First, it used a convenience sample with the selection of participants based mainly on teachers' availability.This could mean that teachers with positive interpersonal behaviour agreed to contribute to the study.Therefore, findings could be influenced by an overestimated aspect of the teachers' interpersonal behaviour.Future studies should involve a more-representative sample for further examining teachers' behaviour in the Greek secondary education context.A second limitation is that the findings of this study are based on the subjective perceptions of students and teachers.Future studies involving data from different sources, such as observations during class or interviews with teachers and students, could add a more-objective aspect to the interactions within classrooms.

Table 2
Reliability and descriptive statistics for QTI scales and dimensions

Table 3
Comparison between teachers' and students' perceptions of teachers' interpersonal behaviour (paired t-test) at class level * p < .05,** p < .01

Table 4
Standard error for difference scores, mean absolute difference scores and mean difference scores for QTI scales and dimensions