Emotions are multidimensional constructs that consist of (1) affective, (2) physiological, (3) cognitive, (4) expressive, and (5) motivational components (Kleinginna and Kleinginna, 1981; Scherer, 2005; Shuman and Scherer, 2014). Emotions have what Frenzel et al. (2015) call a “felt core”— the tangible experience of feeling (p. 202). When people experience emotions, the body often reacts as well. For example, the experience of fear can result in an increased heart rate or a change in breathing rate or pattern (Frenzel et al., 2015, p. 202). Emotional experiences also impact thoughts, such as when fear leads to thoughts about consequences. Emotions can be perceived by the outside world through the expressive component. For example, fear can be expressed verbally or non-verbally, such as through a worried face. Finally, the motivational component ensures that appropriate action is taken. Fear often leads to avoidance behavior.

According to Pekrun and Linnenbrink-Garcia (2012, p. 261; see also Pekrun et al., 2023), emotions can be described and differentiated according to their valence and activation. In terms of valence, a distinction can be made between positive (e.g., joy) and negative (e.g., anger). Valence in this context is related to the subjective experience of the teacher. Positive emotions are classified as those that are experienced as pleasant by the teacher, whereas negative emotions are defined as those experienced as unpleasant. Both negative and positive emotions can be functional or dysfunctional (for a critical discussion see An et al., 2017). In addition, there are physiologically activating or deactivating states. Excitement is activating, whereas relaxation is usually deactivating. These two aspects are crucial for understanding the actions that arise from emotions, as in the case of teachers who experience emotions in the classroom and act accordingly.

Teachers’ emotions have increasingly become objects of study in recent years. Frenzel (2014) proposed a reciprocal model of the causes and effects of teachers’ emotions when teaching in class. It illustrates how teachers’ emotions are triggered and influenced by and affect the teaching process. The basic assumptions of the model are based on an appraisal-theoretical understanding of emotions (Ellsworth and Scherer, 2003). Appraisal theory explains why the same external experience may not lead to the same emotional responses in all individuals; it is not the experience itself that evokes the emotion, but the subjective appraisal made by the individual (Sutton and Wheatley, 2003). For teachers, this appraisal is based on four aspects of teacher goals: (1) cognitive, (2) motivational, (3) social, and (4) relational (Frenzel et al., 2009b; Frenzel, 2014). Based on their perceptions of learners’ behavior on these four dimensions, teachers assess whether they have achieved or will achieve these goals. The outcome of this assessment process determines the teachers’ emotional response. For example, if a teacher perceives students’ engagement as high, it is likely that the teacher will experience positive emotions (e.g., enjoyment) as the students’ behavior is interpreted as goal conducive. These emotions then influence the teacher’s classroom behavior (i.e., cognitive activation, classroom management, social support). For example, teachers who experience positive emotions can build trusting relationships with their students. These instructional behavior factors affect students’ achievement, motivation, behavior in class, and relationship with the teacher. Thus, student and teacher behavior in the classroom is both the cause and effect of the teacher’s emotional experiences. This reciprocal relationship between teacher and student emotions has been empirically confirmed in a variety of studies (Frenzel et al., 2009a,b; Becker et al., 2014; Keller and Lazarides, 2021). It is expected that students’ homework behavior on the different dimensions is related to teachers’ homework-related emotions as well. It seems plausible, for example, that students who are committed to doing their homework trigger positive emotions in teachers because teachers then feel confirmed in their effectiveness and consider their goals to have been achieved. However, there are no specific empirical findings for homework practice so far.

Teachers experience emotions for a variety of reasons related to achieving or not achieving their goals (Sutton and Wheatley, 2003; Sutton, 2007; Frenzel, 2014). They experience joy in the classroom when students are motivated (Becker et al., 2015; Burić and Frenzel, 2021), engaged (Prawatt et al., 1983; Epstein and van Voorhis, 2012; Hagenauer et al., 2015; Chang, 2020), interested (Frenzel et al., 2008), disciplined (Hagenauer et al., 2015; Frenzel et al., 2020) or simply happy (Chang, 2020; Keller and Lazarides, 2021). According to Frenzel et al. (2008) and Keller and Lazarides (2021), joy is the emotion most commonly reported by teachers. When students are successful or interactive, teachers experience positive emotions (Sutton, 2005; Wu and Chen, 2018; Chang, 2020) regardless of the students’ abilities (Prawatt et al., 1983). In addition, they feel pride when a student with low abilities suddenly begins to try very hard (Prawatt et al., 1983).

However, student engagement and discipline are also significant predictors of negative emotions such as anxiety, anger, and frustration (Prawatt et al., 1983; Georgiou et al., 2002; Sutton and Wheatley, 2003; Sutton, 2005; Becker et al., 2014; Hagenauer et al., 2015; Frenzel et al., 2020). Anger is mainly evoked when students misbehave, do not participate, or are inattentive or unmotivated (Sutton, 2007; Hagenauer et al., 2015). It also arises when teachers attribute students’ academic failures to inadequate effort (Reyna and Weiner, 2001). The level of discipline has a significantly negative correlation with fear (Frenzel et al., 2008). Surprise occurs when low-ability students who make little effort nevertheless succeed or high-ability students who exert a lot of effort fail (Prawatt et al., 1983).

The relationships between students and teachers are also associated with emotions. When teachers feel connected to their students, they experience joy. If these relationships cannot be established, anger and anxiety are more likely to arise (Hagenauer et al., 2015). In addition, social relations outside the classroom—such as those with colleagues or parents—can also lead to emotional responses (Sutton and Wheatley, 2003; Sutton, 2005; Wu and Chen, 2018). Teachers feel pleasant emotions when they succeed in working with their colleagues, receive support from school leaders, or experience recognition from parents (Chen, 2019). In contrast, unpleasant emotions can result if they experience competition with their colleagues, receive little support from the administration, or interact with uncooperative parents (Sutton, 2005; Chen, 2019).

In conclusion, the main sources which trigger teachers’ emotions proposed in the model on teachers’ emotions (Frenzel, 2014) have been confirmed empirically by various studies in different countries. However, it remains an open question whether these particular sources are also at the core of teachers’ emotions related to homework.

A total of 23 secondary school teachers from the canton of Bern in Switzerland participated in this study. The conditions for participation were that they had been in the teaching profession for at least 3 years and taught German (which is the language of instruction and a primary subject in the area). The subject German was chosen as it is one of the main subjects in Swiss secondary schools. Homework and its control can be very time-consuming for teachers, as essays have to be corrected in addition to other forms of assignments. In addition, it was important for us that the teachers already had sufficient professional experience so that they could report from their broad experience.

We first contacted all secondary schools in the canton of Bern to recruit teachers who were willing to participate in interviews. There are five different school models in Bern, which differ in terms of permeability (see Figure 1). In Model 1, the students of the high-track secondary level (Sekundarschule) and the low-track secondary level (Realschule) are taught separately in different school buildings. In Model 2, the two tracks are taught separately but in the same school building (i.e., there are separate high-track and low-track classes in the same building). In Model 3a, students in the low- and high-track levels are taught separately in most subjects; however, in the main subjects (mathematics, German, French), they are grouped according to their ability levels. In Model 3b, core classes are mixed, while the three main subjects are taught in ability-level groups. In Model 4, all subjects are taught in mixed levels and classes are only differentiated internally.

Of the teachers interviewed, two teachers were from Model 1, three teachers were from Model 2, 12 teachers were from Model 3a, four teachers were from Model 3b, and two teachers were from Model 4. This distribution accurately reflects the distribution of teachers among the different models in the canton of Bern. Model 3a is the most frequently implemented (see Table 1).

Of the 23 teachers interviewed, 12 were female (52.2%) and 11 were male (47.8%). One teacher was under 30 years old, 11 were between 30 and 40 years old, eight were between 41 and 50 years old and three were over 50 years old. The teachers also differed in terms of professional experience. Two had been in the teaching profession for less than 5 years, five for 5–10 years, seven for 11–15 years, six for 16–20 years and three for over 20 years (see Table 1).

As teachers’ emotions related to homework practices are relatively unexplored, a qualitative–explorative approach was chosen to answer the proposed research questions. In addition, a short questionnaire was used to collect demographic information and the teachers’ positive and negative affect related to homework practice.

The interviews were audiotaped and transcribed verbatim. Fuss and Karbach (2019) and Kuckartz (2010) identified obligatory and pre-defined transcription rules. All transcripts conformed to these rules. Personal information provided by the participants was anonymised in the transcripts. The interviews were analysed utilizing the software MAXQDA based on the qualitative content analysis structure defined by Mayring (2017). A coding frame was developed consisting of several main categories and subcategories that structured the material. First, the interview material was coded based on Frenzel’s (2014) teachers’ emotions model. They classified the students’ behavior on cognitive, motivational, and social levels as relevant antecedents of the teachers’ emotions. All other key categories and sub-categories pertaining to the triggers of the teachers’ emotions that were part of the coding frame emerged inductively from the material. In terms of specific emotions, the emotional states from the PANAS scales were used as deductive categories. Other emotions, such as feeling insecure, emerged from the interviews, so further inductive categories were formed during the coding process. These categories and the overall coding frame were discussed several times with a second researcher. The full coding frame is available from the researchers on request. Extensive extracts from the coding frame are depicted in the results section in Tables 2–4.

For the final coding frame, each category was described and assigned a representative anchor example. In relation to the research questions, the coding scheme included 21 categories of positive emotions and 27 categories of negative emotions. A total of 116 codes in the positive emotion categories and 133 codes in the negative emotion categories were developed. The coding scheme also included 25 antecedents, 20 of which were divided into positive and negative. The exceptions were categories that were considered to be positive or negative per se (e.g., lies/excuses). There were 373 codes in the antecedent categories.

To ensure intercoder reliability, a second independent researcher who was not involved in the research project but who has expertise in the field coded a randomly selected interview using the final coding scheme. The codes were discussed with the second researcher. After a consensus was reached, the independent researcher coded four more randomly selected interviews. These were used to calculate intercoder reliability via the corrected Cohen’s Kappa coefficient, as suggested by Brennan and Prediger (1981). The intercoder reliability as a measure of the coding consistency was good, suggesting consistency in the coding process (κ = 0.78; Landis and Koch, 1977).

Findings from the PANAS scales revealed that the teachers experienced a variety of positive and negative emotions related to homework. Positive emotions dominated over negative emotions (M_{positive emotions} = 2.69; M_{negative emotions} = 1.40) (see Figure 2 and Supplementary Table S1).

To cross-validate these findings, the teachers were also asked in the interviews whether positive or negative emotions dominate from their perspective. In line with the quantitative findings, most of the teachers interviewed (n = 15) claimed that positive emotions were dominant. When prompted to elaborate, they clarified that they have a positive attitude toward homework and strive to implement high-quality homework practices. In addition, they reported that they do not receive negative feedback from students or parents on their homework practices, suggesting that they are satisfied. In contrast, negative emotions dominated among some teachers (n = 6). They argued that homework has the potential to cause negative outcomes such as conflicts with parents, stress, or students feeling overloaded. Finally, two teachers were unsure whether positive or negative emotions dominate, reflecting an ambivalent attitude toward homework.

Looking at the distinct emotions in detail, the teachers mentioned a high variation of positive and negative emotions that are triggered by their homework practice. Specifically, for the positive emotions, they reported feeling hopeful, excited, relieved, empathic, admiration, confident, determined, interested, enthusiastic, inspired, satisfied, proud, fulfilling, and relaxed. In terms of negative emotions, they reported feeling stressed, pity, sad, ineffective, overwhelmed, frustrated, guilty, including having a guilty conscience, ashamed, upset, insecure, disappointed, annoyed, scared, irritable, helpless, perplexed, hopeless, inadequate, and bored. In the following section, these distinct emotions are related to their antecedents.

Based on the interview findings, triggers of teachers’ emotions were identified and grouped into three categories: context (Section 6.2.1), teacher behavior and (inner) demands (Section 6.2.2), and student behavior (Section 6.2.3).

Based on an explorative approach, the present study has shown that teachers’ emotions can also be examined context-specifically for the teacher’s area of responsibility “homework practice.” Following the results, the model of teachers’ emotions (Frenzel, 2014), which specifies the antecedents of teachers’ emotions, can be transferred to teachers’ emotions related to homework practice, but additionally, it should be extended to include further antecedents at the teachers’ level and at the context level. Through the exploratory approach the diversity of teachers’ emotions in the homework process could be illustrated and the diverse antecedents of these emotions could be identified in depth.

Still, from an exploratory perspective, the study has some limitations as well. First, the findings represent the experiences of secondary teachers from the canton of Bern. By purposively selecting these cases, we have tried to obtain a selection that is as comprehensive as possible in terms of the school models existing in the canton of Bern. Nevertheless, further quantitative studies need to follow. These studies could test possible differences in the emotional experiences of teachers working in different school models. In our small-scale study, these group differences could not be reliably explored. In addition, further studies based on different samples in different contexts are needed, which would allow a generalization of the results beyond the Swiss (Bernese) context. If such quantitative studies are conducted, when developing measurement instruments, consideration should be given to mapping the diversity of the emotional experiences of teachers. Classical instruments, such as the Teacher Emotions Scales (TES; Frenzel et al., 2016) which were developed for teaching in general, may fall short when it comes to the specific context of homework practice. Even if anxiety, anger, and enjoyment (i.e., the core emotions of the TES) are relevant teachers’ emotions related to homework practice, other emotions, such as satisfaction, disappointment, stress, or guilt (including having a guilty conscience) should also be considered in such measurement instruments. Furthermore, the link between teachers’ emotions and homework quality needs further exploration. As a reciprocal relationship can be assumed, a complex longitudinal design needs to be applied. Second, the teachers’ emotions were measured retrospectively. It can be assumed that the teachers mainly described situations that were either very close in time or in which the emotions were experienced intensely (Heuer and Reisberg, 1992). Future studies should therefore also use situational measurements (e.g., experience sampling methods). Another limitation is that the results were based on self-reporting. This can lead to bias; for example, the teachers may have answered in a socially desirable way. We countered this effect by ensuring full anonymity and by creating a trusting environment during the interviews. It must also be mentioned that the subjective assessment of emotions is still a valid way to capture the affective core of emotions, i.e., the subjective feeling that cannot be observed. Nevertheless, if a multicomponent approach to emotions is pursued, future studies can, for example, use further data collection methods, such as physiological measures accounting for the physical arousal of emotions. Finally, it should be noted that the teachers were explicitly asked about their emotions in connection with homework. This has the advantage that teachers have purposefully reflected on their emotions and their antecedents. However, such an approach presupposes a conscious reflection on emotions by the teachers. Another pre-assumption of this study was that emotions occur in the homework process, which is why we opted for the explicit approach to explore teachers’ emotions. For future research, it would be interesting to complement these explicit approaches to capturing emotions with implicit approaches by attempting to reconstruct teachers’ emotional experiences through, for example, narrative interviews.

This study has provided a first insight into the emotional experiences of Swiss secondary teachers teaching German during the homework process and has also identified the multiple influencing conditions of these emotions. On a theoretical level, the results of this study extend the research findings on teachers’ emotions by focusing on homework practices as a specific area of action in the classroom. They enable Frenzel’s (2014) model of teachers’ emotions in the classroom to be differentiated by focusing on this specific aspect of teaching. Overall, the results clearly showed that the homework process is definitely experienced emotionally by teachers. Even though homework is done by students at home, it is still the students and their behavior that are the most emotionally relevant source for teachers’ emotions. This result is due to the fact that the homework process also includes significant teacher–student interactions in class (e.g., homework return and discussion), as well as the fact that student behavior is also visible in the quality of homework completion. Teachers, for example, are happy about the students’ learning progress that they diagnose from the homework, or they are annoyed when the students do not put in enough effort or cheat on homework. However, the demands that teachers place on themselves are also often sources of their emotions (e.g., “inner demands”), and contextual factors also influence their emotional experience (e.g., experiencing pity due to unfavorable conditions at home).

From a practical point of view, the results provide some implications for teacher education and training in Switzerland. First and foremost, pre-service teachers should acquire basic knowledge about the development of emotions and their influence on teaching and learning. In-service teachers should also be sensitized to this through professional development programs. For example, a training program developed by Carstensen et al. (2019) focusses specifically on fostering teachers’ socio-emotional competencies. If teachers develop socio-emotional competencies, they are more likely to recognize automatic patterns of action that occur due to their own emotions and thus will be better able to interrupt negative spirals that can arise from them. This could have a positive impact on the quality of the homework they assign and subsequently the behavior of the students. Teachers who can regulate their emotions appropriately (e.g., by applying cognitive reappraisal when students do not hand in their homework) are less likely to let their emotional reactions interfere with their professional teaching behavior.

These skills also have an impact on building and maintaining meaningful teacher–student relationships (Carstensen et al., 2019), which positively influence the students’ engagement and achievement (Roorda et al., 2011). Previous research has shown that cheating amongst university students is lower when the instructor is evaluated positively (Stearns, 2001) in terms of teacher–student relationships and enthusiasm (Orosz et al., 2015). Building on these findings, meaningful teacher–student relationships might decrease the triggers of negative teachers’ emotions, as students who are satisfied with their teachers are less likely to cheat on, copy, or lie about their homework and complete it more reliably.

Finally, pre-service and in-service teachers should be specifically trained in assigning high-quality homework. The results of this study demonstrate that positive emotions in teachers can be evoked by their students’ learning progress, learning outcomes, and engagement. Previous studies have shown that homework quality can promote student achievement (Rosário et al., 2018) and engagement (Trautwein et al., 2006). Assigning high-quality homework can have positive effects on both the students’ learning and the teacher’s own emotional experiences. During the training, teachers could also learn how to follow up on completed homework during class, as it was found that being unable to use homework in class leads to negative emotions.