Introduction

In everyday school life, teachers are tasked not only with preparing or giving lessons, but also with handling a wide range of student misbehaviors (Kulinna et al., 2006; Wheldall & Merrett, 1988). These situations harbor the risk of severe consequences for both teachers and students: For teachers, the extent of student misbehavior has found to be strongly linked to their well-being and health, and it has been identified as the most salient stressor related to burnout syndrome in teachers (Aloe et al., 2014; Brouwers & Tomic, 2000; McCormick & Barnett, 2011). Finding suitable and effective classroom intervention strategies for such incidents is thus a crucial challenge and major concern for teachers (Melnick & Meister, 2008). In the best case, student misbehavior can be prevented by positive and proactive classroom management approaches (Sugai & Horner, 2006). However, sometimes such misbehavior simply cannot be prevented by a well-prepared lesson (e.g., if it occurs during break). In such instances, the classroom may bear some resemblance to a courtroom, with the teacher representing the judge who is in charge of finding an appropriate response to misbehavior (Weiner, 2003).

For students, teachers’ treatment of student misbehavior is equally essential, as it is a fundamental cause of their perception of injustice in school (Fan & Chan, 1999; Israelashvili, 1997). Crucially, such injustice perceptions have a strong impact on students’ lives, both within the school (e.g., on students’ academic self-concept, motivation, and achievements; Peter et al., 2012) and beyond (e.g., on students’ attitudes toward democracy; Pretsch & Ehrhardt-Madapathi, 2018). Importantly, students’ perceptions of injustice in school are not limited to and indeed are only marginally influenced by the grading or evaluation of students’ performances. Instead, it is the treatment of student misbehavior that appears to be an important factor (Israelashvili, 1997). Research suggests that such situations are among the most frequently reported situations of students’ injustice experiences, even excluding situations of false allegations (Fan & Chan, 1999). In other words, even if the incident of student misbehavior appears to be clear (e.g., when it is obvious who the offender or victim is), students frequently feel treated unfairly by their teachers. Eventually, the “wrong” treatment of student misbehavior may cause severe negative outcomes, such as a negative classroom climate (Peter & Dalbert, 2010) or a strained relationship between most or even all students and teachers (Avtgis & Rancer, 2008; Ratcliff et al., 2010). In turn, a tense student–teacher relationship and the students’ perceptions of low support from the teacher may increase conflict and, eventually, the occurrence of further classroom misbehavior (Boyle et al., 1995; Bru et al., 2001, 2002; Ertesvåg & Vaaland, 2007).

Consequently, it is vital to understand and study teachers’ responses to student misbehavior to identify characteristics that do not help but harm learning and instruction. This is the endeavor of the present research. More precisely, we examine teachers’ approaches to respond to student misbehavior and compare these with students’ preferences for how this misbehavior should be treated. We particularly scrutinize one aspect of teachers’ decision-making process that has not been analyzed from both the teachers’ and students’ perspectives in the past: the goals teachers intend to achieve when reacting to student misbehavior. That is, when punishing students for misbehavior, teachers may pursue a variety of different goals as described in detail below. Importantly, students may agree or disagree with these goals and, consequently, may perceive the punishment as more or less appropriate and just (Gollwitzer & Okimoto, 2021).

Retribution, Special Prevention, and General Prevention as Punishment Goals

There is a considerable body of literature discussing the goals individuals generally pursue when engaging in punishment (Carlsmith et al., 2002; Cushman, 2015; Goodwin & Gromet, 2014; Twardawski et al., 2020b). On the broadest level, one can differentiate between two goals that are associated with the philosophical works by Immanuel Kant and Jeremy Bentham. According to Kant (1952) punishment should follow a deontological justice principle: An offender harms a victim, a society, and its rules, and causes an imbalance to the scales of justice. Consequently, punishment is legitimate and justified to rebalance the (moral) wrong that has been caused by the offense, paying back harm doers for their misconduct and, thus, restore justice (Gerber & Jackson, 2013). This is typically achieved by finding a proportionate punishment that “fits” the crime (Goodwin & Gromet, 2014). Accordingly, a justified punishment is primarily backward-oriented and concerned with the harm caused but not about future developments (Carlsmith et al., 2002). The punishment goal associated with this deontological justice principle is referred to as retribution (Carlsmith & Darley, 2008; Carlsmith et al., 2002).

According to Bentham (1962), on the other hand, punishment should follow a utilitarian justice principle. Correspondingly, the intrinsically damaging act of punishment is justifiable if it leads to positive future consequences—in particular, by preventing future misbehavior (McCullough et al., 2013). That is, punishment should primarily be forward-oriented and used as an instrument to facilitate compliance with social norms and reducing norm violations (Rucker et al., 2004). This utilitarian perspective on people’s punishment behavior can further be differentiated into special prevention and general prevention (Twardawski et al., 2020b). A special preventive punishment is primarily concerned with offenders themselves and intends to prevent future recidivism (Keller et al., 2010). A general preventive punishment, in turn, is primarily concerned with other members of the community that might have learned of the offense and, therefore, may imitate the misbehavior if it remains unpunished (Goodwin & Benforado, 2015).

Teachers’ and Students’ Punishment Goals

Decades of research examined laypeople’s relative endorsement of these punishment goals (for an overview, see e.g., Carlsmith & Darley, 2008). Of particular relevance for the present research, recent literature suggests that the endorsement of punishment goals is subject to power and hierarchy, that is, people differing in power differ in their preferences for specific goals (Mooijman & Graham, 2018). Particularly, after observing misbehavior, powerful people respond with distrust and increased concerns about losing their power (Mooijman et al., 2015). To prevent the loss of power, they use punishment as an instrument to deter observers from imitating the misbehavior. Consequently, general prevention (i.e., deterring observers from imitating the misbehavior) is the preferred goal of punishment among people in powerful positions. Whereas, people who do not occupy leadership positions show a preference for retribution (i.e., to even out the wrong that has been done) rather than for special or general prevention (Mooijman et al., 2015). Importantly, teachers generally occupy an inherently powerful position in school (Reeve, 2009), suggesting that their punishment may be similarly designed to assert control (over the classroom). More precisely, teachers may acknowledge punishment as an instrument to communicate behavioral norms and such communication is, by definition, central to utilitarian (i.e., general preventive, but also special preventive) punishment.

Moreover, and in line with this reasoning, teachers’ behavior is generally led by educational goals, above and beyond delivering academic curricula. More precisely, professionals in education ultimately pursue the goal of shaping learners and help them developing to empowered, independent, and righteous individuals. This translates into teachers creating a pedagogical environment that helps educating fundamental social values and norms, both during lecturing and beyond (Husu & Tirri, 2007). Importantly, teachers also follow such educational principles in the face of student misbehavior (Coverdale, 2020; Goodman, 2020; Hand, 2020; Liu, 2017). For example, it has been suggested that teachers respond in such situations “for a myriad of reasons, including but not limited to moral education of students, maintaining safety, and creating an environment conducive to learning” (Thompson et al., 2020, p. 79). Notably, all of these reasons can be subsumed under the umbrella of special and general prevention.

In sum, based on the above reasoning, it can be theorized that teachers’ endorsement of punishment in schools particularly follows utilitarian principles. Indeed, recent research suggests that teachers show a preference for general prevention (and special prevention) over retribution, at least when they attribute the misbehavior to controllable causes (Twardawski et al., 2020a). In the present research, we follow up on this research by investigating which punishment goals teachers endorse in the face of a specific misbehavior and how they evaluate punishment practices that are perceived to serve different goals. We hypothesized that teachers show greater endorsement of general prevention compared to retribution. Given the natural overlap of special prevention and general prevention as two related, yet distinct aspects of utilitarian punishment, we also expected that teachers generally support special prevention over retribution, whereas we had no hypothesis regarding potential differences between special and general prevention. It should be noted that any hypotheses concerning the role of special prevention were necessarily more speculative, as past research only rarely considered the differences between general and special prevention and, often, did not explicitly examine special prevention.

As outlined above, it is not only vital to examine and understand the relative endorsement of punishment goals among teachers. Rather, it is equally important to consider the students’ perception of teachers’ punishment to identify potential characteristics of the punishment that may foster subjective injustice (Fan & Chan, 1999; Israelashvili, 1997). Differences between teachers’ and students’ relative endorsement of punishment goals may ultimately result in such undesirable outcomes (Mooijman et al., 2017). However, there is a lack of research on students’ endorsement of punishment goals and one could therefore derive contrary hypotheses for the students’ perspective: On the one hand, students are in a relatively less powerful position in the school context (Reeve, 2009) and, thus, should show less endorsement of general prevention than of retribution (Mooijman et al., 2015). On the other hand, this reasoning largely stems from research on adults, whereas evidence on children’s endorsement of punishment goals is scarce. In fact, the few existing insights into children’s punishment goals rather suggest that children value both retribution and prevention in their own punishment (Marshall et al., 2021; Twardawski & Hilbig, 2020). Therefore, it is unclear whether students show a relatively larger endorsement of retribution as compared to special and general prevention, especially when evaluating teachers’ reactions to student misbehavior. Consequently, investigating the extent to which students share teachers’ relative endorsement of punishment goals promises a fundamental contribution to the literature.

The Present Research

The primary goal of the present research is to examine and compare teachers’ and students’ relative endorsement of punishment goals. Therefore, we provided teachers and school students with a scenario describing a specific case of student misbehavior and tested whether teachers and students show similar preferences in punishment goals (i.e., retribution, special prevention, and general prevention) when directly asked to indicate their endorsement of these goals (Mooijman et al., 2015). That is, teachers were asked to imagine being in charge of reacting to this incident of student misbehavior and to indicate the degree to which they would want to achieve either of the three punishment goals. Similarly, students indicated their endorsement of the three goals when thinking about punishment of the offender in a structurally equivalent case of student misbehavior. We refer to this approach as the direct endorsement measure of teachers’ and students’ punishment goals.

In addition to analyzing teachers’ and students’ direct endorsement of the goals punishment ought to achieve in a specific case of student misbehavior, it is important to examine teachers’ assessment of concrete punishment practices vis-à-vis students’ perceptions of these practices. Specifically, research has shown that people’s direct endorsement of punishment goals is only weakly correlated with their actual punishment of a specific case of misbehavior (Crockett et al., 2014). This also applies to teachers, as their assessment of concrete punishment practices and the goals they purportedly endorse are similarly misaligned (Twardawski et al., 2020a). For everyday school life, however, examining the perception of concrete and specific punishment practices may be equally important as the endorsement of rather abstract punishment goals—in particular if the evaluation of practices and abstract endorsement do not correspond perfectly. More specifically, teachers and students may agree on the endorsement of abstract punishment goals that may be pursued in response to a student misbehavior, but nonetheless disagree on a particular punishment practice designed to achieve these goals (or vice versa). Consequently, besides measuring teachers’ and students’ direct endorsement of abstract punishment goals in the case of a student misbehavior, we additionally measured teachers’ punishment goal preferences in a more indirect way, asking them to rate the appropriateness of three punishment practices that were perceived (in pretests) as primarily serving one of the goals. Vitally, we provided students with the same punishment practices and asked them to indicate the extent to which they evaluated these practices as fair, appropriate, and just, if shown by a teacher. We refer to this approach as the punishment practice evaluation measure of teachers’ and students’ punishment goal preferences.

Most critically, the core focus of the present research is to strictly examine whether teachers and students differ in their relative endorsement of various punishment goals and in how they evaluate corresponding punishment practices. More precisely, we examine the degree to which teachers and students show preferences with regard to the three punishment goals and compare these intra-individual preferences (i.e., the rank order of goals) between the two groups. Our hypothesis as stated above was that teachers indicate a preference for general (and special) prevention over retribution. For students’ punishment goal preferences, in turn, we had no strong a priori expectation, but deemed a preference for retribution most likely given the literature.

As recommended, we report how we determined our sample size, all data exclusions (if any), all manipulations, and all measures (Simmons et al., 2012). Furthermore, all materials (including instructions and materials of the pretest study; all materials are translated from German to English), along with all data, analyses scripts, and supplementary analyses are available on the OSF and can be accessed via the following link: https://osf.io/r5d8v/.

Method

Samples

Data were collected from both pre-service (that is university students becoming teachers) and in-service teachers in conjunction with a project on other research questions (Twardawski et al., 2020a). In-service and pre-service teachers were recruited through mailing lists, social media platforms, personal contacts, and recruitment in schools to participate in a study lasting 10–15 min. For in-service teachers, 103 participants started an online-version of the questionnaire, with n = 74 (72%) completing it. Additionally, n = 67 teachers opted for completing a paper and pencil version of the questionnaire. The total sample of in-service teachers therefore comprised N = 141 participants. Around two thirds of these participants (i.e., n = 92; 65%) were female (two participants did not indicate their gender), and ages ranged between 23 and 70 years (M = 40.77, SD = 11.13). For pre-service teachers, 160 participants started the questionnaire (all of them participated online), of which N = 119 (74%) completed it. In this final sample of pre-service teachers, ages ranged between 18 and 36 years (M = 23.38, SD = 3.09) and 84% (n = 100) of participants were female. These pre-service teachers were in their sixth semester of studies on average (M = 5.61, SD = 3.23) and mostly studied teaching on high school level (35%), teaching for primary schools (28%), or special education (22%). In total, we collected complete data sets from N = 260 pre-service and in-service teachers, of whom n = 192 were female (74%).

Data from students were collected in the fifth and sixth grade of three public German schools, again, in conjunction with projects on other research questions (Twardawski et al., 2020a). In total, N = 238 children from twelve school classes participated in the study. Around 45% of participants (i.e., n = 106) were female, most children’s mother language was German (92%), and ages ranged between 9 and 12 (M = 10.46, SD = 0.61; one child did not indicate her age).

To evaluate the sample sizes of the teachers (N = 260) and students (N = 238) with regard to the planned within-subjects comparisons of teachers’ and students’ support for the three punishment goals (retribution, special prevention, and general prevention), two separate sensitivity power analyses were conducted using G*Power (Faul et al., 2007, 2009). One of these power analyses was specified to detect within-subjects differences in participants’ assessment of the three punishment goals (i.e., as an estimate of teachers’ and students’ punishment goal preferences). Assuming a conventional α = 0.05, a nonsphericity correction of ε = 1, and a standard power criterion of 1 − β = .90, this resulted in a detectable effect size of f = .16 for the teacher and f = .15 for the student sample in a repeated measures ANOVA. We further evaluated the collective sample sizes of teachers and students (total N = 498) with regard to our main test of whether teachers and students differ in their relative endorsement of punishment goals. We therefore calculated a sensitivity power analysis to detect a within-between interaction in a mixed model with two groups (teachers and students) as between-subjects factor and the three punishment goals (retribution, special prevention, and general prevention) as within-subjects factor. Given a standard power criterion of 1 − β = .90, α = .05, number of groups = 2, number of measurements = 3, and nonsphericity correction ε = 1, this resulted in a detectable effect size of f = .11 in a mixed-model ANOVA. Thus, there was high statistical power for even relatively small effect sizes throughout.

Measures and Procedures

The Teacher Perspective

Teachers had the chance to participate in the study online or via a paper and pencil version of the questionnaire. After providing informed consent, participants received a scenario describing a student destroying a recently prepared hand drum of another student. This scenario read as follows:

Within the past lessons, you manufactured hand drums with your students that you plan to use today. You briefly turn to the board. Once you face the class again, you see how Florin causes a hole in Maxi’s drum. As a result, the drum is broken.Footnote 1

Participants were then asked to provide answers on several control variables regarding the perception of the misbehavior, starting with (i.) the stability and (ii.) controllability of the cause of the student’s misbehavior. Furthermore, participants were asked to indicate (iii.) the student’s responsibility for what occurred, how much (iv.) anger and (v.) sympathy they would feel toward the misbehaving student as well as (vi.) to what degree it is possible to influence the student’s future behavior. Each response was rated on a 6-point scale ranging from 0 = “not at all” to 5 = “completely.” Next, they answered the punishment practice evaluation measure and rated the appropriateness of three punishment practices that teachers may use as a response to the displayed student misbehavior. Punishment practices are provided in Table 1. These practices were designed based on the results of a thorough pretest (in an independent sample) and therein judged as serving predominantly one of the punishment goals.Footnote 2 Appropriateness of each practice was indicated on a 6-point scale ranging from 0 = “not at all appropriate” to 5 = “completely appropriate.” Finally, participants provided their direct endorsement of the three punishment goals in the specified situation of student misbehavior, indicating the goals they would want to accomplish if presented a chance or obliged to react to the misbehaving student. To this end, we adapted one item for each punishment goal (retribution, special prevention, and general prevention) from Orth (2003) and Weiner et al. (1997). The item measuring direct endorsement of retribution read as follows: “To what extent would you like to react to even out the wrong that Florin has done?”; the item measuring endorsement of special prevention read “To what extent would you like to react to prevent recidivism by Florin?”; and the item measuring endorsement of general prevention read “To what extent would you like to react to prevent other students of showing similar behavior in the future?” Each item was answered on a 6-point scale ranging from 0 = “not at all” to 5 = “completely,” with higher values indicating stronger endorsement of a particular punishment goal. After answering all questions, participants worked on several other tasks that pertained to a different research question (Twardawski et al., 2020a). Thus, this part of the material will not be further discussed in the present article. Finally, participants provided demographic information, before they were fully debriefed and thanked.

Table 1 Punishment practices (as derived from the pretest)
Full size table

The Student Perspective

Data from students were collected in schools. On arrival in the classroom, students provided the experimenters with the consent form signed by their parents or legal guardians. Then, students were seated individually in front of computers. Two experimenters welcomed the class and gave them verbal instructions on the subsequent tasks. The first task was on an unrelated research question (Twardawski & Hilbig, 2020) and will therefore not be further discussed in the present article. Subsequently, students provided demographic information before they were introduced to the main task. Therein, students were asked to indicate their perceptions of a misbehavior and different teacher responses to it. This misbehavior and the responses were presented in short comic strips. Using an exemplary misbehavior, students received comprehensive instructions on the procedure of the task, before working on the task individually. They were provided with a comic strip depicting one student destroying the recently prepared hand drum of the other student (i.e., a structurally equivalent scenario that teachers received). The gender of the misbehaving student and the victim were counterbalanced (i.e., it was either the boy destroying the drum of the girl or vice versa). Next, students received three comic strips depicting potential punishment responses of the teacher (either female or male; counterbalanced) in the comic—reflecting the punishment practice evaluation measure. Students were asked to rate the extent to which they perceived these practices as just, appropriate, and fair on 6-point scales ranging from 0 = “not at all” to 5 = “completely.” These three items per punishment practice were aggregated (that is, averaged) and showed high internal consistencies (retributive reaction: α = .92; special preventive reaction: α = .90; general preventive reaction: α = .93).Footnote 3 Similar to teachers, students further answered all of the control questions outlined above (i.e., the stability and controllability of the cause of the student’s misbehavior, the student’s responsibility for what occurred, how much anger and sympathy they would feel toward the misbehaving student, and to what degree it is possible to influence the student’s future behavior), as well as their direct endorsement of the three punishment goals in the specified situation. Mirroring the teachers’ materials, the item measuring direct endorsement of retribution read as follows: “Florin should be punished for their behavior to even out the wrong committed”; the item measuring endorsement of special prevention read “Florin should be punished for their behavior to prevent them from doing something like this again”; and the item measuring endorsement of general prevention read “Florin should be punished for their behavior to prevent others from imitating them.” Again, scales ranged from 0 = “not at all” to 5 = “completely.” Finally, students were fully debriefed and thanked.

Results

Preliminary Data Analyses

Before conducting our main analyses, we first ensured that we could treat pre-service and in-service teachers as one homogeneous group. Therefore, we tested whether pre-service and in-service teachers differed systematically in their endorsement of punishment goals for both the direct endorsement and the punishment practice evaluation measure. For each measure of endorsement of punishment goals, we conducted a separate mixed model ANOVA predicting the endorsement of the three punishment goals (retribution, special prevention, and general prevention) as within-subjects factor and the sample (pre-service and in-service teachers) as between-subjects factor. It turned out that there were only negligible differences between pre-service and in-service teachers (the detailed results of these tests, including test statistics, are available online at the OSF). Thus, we considered it reasonable to combine the two groups to one group of teachers.

Additionally, we considered it vital to determine whether teachers and students perceived the student misbehavior itself in a similar manner. We therefore calculated a Spearman correlation between the mean ratings of all control variables we collected on participants’ perception of the scenario (i.e., the stability and controllability of the cause of the student’s misbehavior, the student’s responsibility for what occurred, how much anger and sympathy they would feel toward the misbehaving student, and to what degree it is possible to influence the student’s future behavior). This analysis revealed a very high correlation, rs = .83. Correspondingly, differences in punishment goal preferences between teachers and students cannot be ascribed to different perceptions of the student misbehavior itself and can therefore be reasonably be attributed to different preferences for how to deal with this misbehavior.

Main Analyses

For the main analyses, we first conducted in-depth within-subjects comparisons on teachers’ and students’ ratings per punishment goal measurement approach to examine their groups’ punishment goal preferences (i.e., the intra-individual rank order in the endorsement of the three punishment goals). Subsequently, we directly compared teachers’ and students’ preferences (i.e., the rank order) of punishment goals, again separately for each punishment goal measurement approach.

Direct Endorsement of Punishment Goals

As shown in Fig. 1 (left panel), teachers indicated a generally higher direct endorsement of utilitarian punishment goals as compared to retribution. Specifically, special prevention was the most endorsed goal (M = 4.42, SD = 0.79), closely followed by general prevention (M = 4.26, SD = 0.98). Notably, in line with our hypotheses, retribution received substantially lower endorsement ratings from teachers (M = 3.29, SD = 1.30). To statistically test this pattern, we used a repeated measures ANOVA predicting the endorsement of the three punishment goals (retribution, special prevention, and general prevention) as within-subjects factor, followed by pairwise post-hoc t-tests. The analysis of variance confirmed significant differences between direct endorsement ratings of punishment goals, F(2, 518) = 109.20, p < .001, \({\hat{\upeta }}_{\mathrm{G}}^{2}\) = .19. As hypothesized, follow-up t-tests directly examining our hypothesis revealed significantly greater endorsement of general prevention compared to retribution, t(259) = 9.96, p < .001, d = 0.59. Likewise, special prevention received significantly greater endorsement than retribution, t(259) = 12.61, p < .001, d = 0.73. Interestingly, special prevention also received significantly higher ratings than general prevention, albeit yielding only a miniscule effect size, t(259) = 2.82, p = .005, d = 0.12.

Fig. 1
figure 1

Results. Comparison of teachers’ and students’ direct endorsement of the three punishment goals (left panel) and their evaluation of different punishment practices (right panel, i.e., teachers’ appropriateness ratings of the practices, and students’ evaluation of how fair, appropriate, and just the teachers’ punishment practices are). Error bars represent one standard error of the mean

Full size image

Moreover, data revealed notable differences for students’ direct endorsement of the three punishment goals. That is, students indicated highest endorsement ratings for special prevention (M = 3.70, SD = 1.44), whereas general prevention (M = 3.23, SD = 1.74) and retribution (M = 3.24, SD = 1.62) were equally supported. A repeated measures analysis of variance confirmed significant differences between endorsement ratings of punishment goals of students, F(2, 474) = 12.73, p < .001, \({\hat{\upeta }}_{\mathrm{G}}^{2}\)= .02. Follow-up t-tests revealed that special prevention was most endorsed and received slightly higher ratings than retribution, t(237) = 5.23, p < .001, d = 0.21, and general prevention, t(237) = 4.17, p < .001, d = 0.20. Differences between retribution and general prevention were negligible, t(237) = 0.07, p = .943, d = 0.003.

The core focus of the present research is a direct comparison of teachers’ and students’ punishment goal preferences (i.e., the rank order in punishment goals of the two groups). As can be seen in Fig. 1 (left panel), teachers’ and students’ direct endorsement of the three punishment goals particularly differed regarding the relative degree of endorsement of retribution. Specifically, whereas retribution received endorsement ratings descriptively comparable to special prevention and general prevention in students, it received substantially lower endorsement ratings compared to the other goals from teachers. To statistically test whether teachers and students actually differed in their preferences of punishment goals using this measurement approach, we conducted a mixed model ANOVA predicting the direct endorsement of the three punishment goals (retribution, special prevention, and general prevention) as within-subjects factor and the sample (teachers and students) as between-subjects factor. Most importantly, this analysis revealed a significant interaction of the goal to be rated and the sample, F(2, 992) = 27.88, p < .001, \({\hat{\upeta }}_{\mathrm{G}}^{2}\) = .02.

Punishment Practice Evaluation Measure of Punishment Goals

For the punishment practice evaluation measure, contrary to the direct endorsement measure of punishment goals, teachers rated the general preventive punishment practice as most appropriate (M = 2.65, SD = 1.50), closely followed by the retributive practice (M = 2.48, SD = 1.57). Importantly, the special preventive practice was rated as least appropriate (M = 1.73, SD = 1.46). These patterns are also shown in Fig. 1 (right panel). Again, we conducted a repeated measures ANOVA to compare teachers’ appropriateness ratings of the three punishment practices, followed by pairwise post-hoc t-tests. The analysis of variance confirmed significant differences between appropriateness ratings of the three punishment practices, F(2, 518) = 29.12, p < .001, \({\hat{\upeta }}_{\mathrm{G}}^{2}\) = 0.06. In contrast to the direct endorsement measure and contrary to our hypothesis, follow-up t-tests revealed no significant differences between the general preventive and retributive punishment practices, t(259) = 1.28, p = .202, d = 0.08. Interestingly, the special preventive practice was rated significantly less appropriate than both the retributive practice, t(259) = − 5.64, p < .001, d = − 0.34, and the general preventive practice, t(259) =  − 8.00, p < .001, d = − 0.43.

Similar to what we found for teachers, students’ evaluation of the three punishment practices also differed from students’ direct endorsement of punishment goals. That is, students indicated highest ratings for the retributive practice (M = 3.84, SD = 1.30), closely followed by the special preventive practice (M = 3.59, SD = 1.22). The general preventive practice received the lowest ratings (M = 2.85, SD = 1.55). Again, we conducted a repeated measures ANOVA to compare students’ perceptions of the three punishment practices, followed by pairwise post-hoc t-tests. The analysis of variance confirmed significant differences between the evaluation of the three punishment practices, F(2, 474) = 49.42, p < .001, \({\hat{\upeta }}_{\mathrm{G}}^{2}\) = .09. Follow-up t-tests revealed that the retributive practice received significantly higher ratings than the general preventive practice, t(237) = 9.31, p < .001, d = 0.46, and the special preventive practice, t(237) = 2.58, p = .010, d = 0.14. However, the effect size of the latter was rather small. Furthermore, the special preventive practice received higher ratings than the general preventive practice, t(237) = 7.01, p < .001, d = 0.36.

Again, we directly compared teachers’ and students’ evaluations of the three punishment practices. As depicted in Fig. 1 (right panel), teachers and students differed substantially regarding their evaluations of the two preventive punishment practices. That is, both teachers and students rated the retributive practice in a comparable manner (i.e., both perceived it as relatively suitable). At the same time, the special preventive practice received the lowest appropriateness ratings from teachers, whereas students perceived this practice as relatively suitable (i.e., fair, just, and appropriate). Likewise, the general preventive practice yielded differences in that it received highest appropriateness ratings from teachers, but lowest ratings from students. Again, we tested whether teachers and students differed in their preferences for the three punishment practices (i.e., the group-level rank order) using a mixed model ANOVA with the three punishment practices (retributive practice, special preventive practice, and general preventive practice) as within-subjects factor and the sample (teachers and students) as between-subjects factor. This analysis again revealed a significant interaction of the punishment practice and the sample, F(2, 992) = 52.26, p < .001, \({\hat{\upeta }}_{\mathrm{G}}^{2}\) = .05.

Discussion

Research suggests that teachers have to deal with student misbehavior on a daily basis (Kulinna et al., 2006; Wheldall & Merrett, 1988). This is not only particularly challenging for teachers (Aloe et al., 2014; Brouwers & Tomic, 2000), but also threatens to lead to perceived injustice among students (Fan & Chan, 1999; Israelashvili, 1997). Correspondingly, it is vital to investigate and analyze teachers’ classroom intervention strategies to understand the factors enhancing students’ perceptions of injustice. One key dimension of punishment behavior reflects the underlying motives and goals of the punishment (Carlsmith et al., 2002; Gromet & Darley, 2009). This aspect of the teachers’ punishment is of particular interest, given that teachers and students are likely to differ in their relative endorsement of punishment goals and these differences may ultimately result in undesirable outcomes, such as a self-perpetuating cycle of student misconduct and teacher punishment that is perceived as unjust (Mooijman & Graham, 2018).

Herein, we examined the perspectives of both teachers and school students on the goals of punishment in a specific situation of student misbehavior. Specifically, we investigated the extent to which (pre-service and in-service) teachers endorse retribution (i.e., evening out the harm caused), special prevention (i.e., preventing recidivism of the offender), and general prevention (i.e., preventing imitation of others) as goals of (their) punishment. We therefore provided teachers with a scenario describing a student destroying the belongings of another student and asked them to indicate to what extent they endorse the three punishment goals in this situation. Furthermore, we measured teachers’ relative endorsement of punishment goals in a more indirect way, asking them to rate the appropriateness of different punishment practices that were perceived as primarily achieving either of the goals (as shown in a pretest). Importantly, students were asked to indicate their endorsement of the three punishment goals as a basis for a response to a structurally equivalent student misbehavior that was used to study teachers’ perspectives. Additionally, students rated how fair, appropriate, and just they perceived teachers’ punishment practices designed to achieve different goals. In total, we thus investigated teachers’ and students’ preferences for retribution, special prevention, and general prevention as punishment goals and whether these preferences are comparable.

As hypothesized, teachers indicated a general preference for general prevention and special prevention over retribution. This was particularly true for general prevention, whereas special prevention was only preferred over retribution in the direct endorsement measure of punishment goals. Students indicated a favorable evaluation of teachers’ punishment practice that was linked to retribution, especially when compared to a more general preventive practice. That is, evaluating teacher’s punishment practices that are perceived as achieving different goals, students rated the retributive practice more favorably than the general (and, to a lesser degree, special) preventive practices. Conversely, for the direct endorsement measure, students rated special prevention as the most endorsed goal of punishment. Notably, differences in students’ endorsement of the three punishment goals were relatively rather small in general, suggesting that students have no strong punishment goal preferences when asked directly (mirroring the literature on explicit support of punishment goals in adults; Applegate et al., 1996; Carlsmith, 2008).

In sum and most importantly, the present research provides the opportunity to directly compare teachers’ and students’ approaches on punishment given the same incident of misbehavior (although materials were adapted to the age of participants, as is discussed below). Analyses comparing students’ and teachers’ relative endorsement of punishment goals indeed revealed substantial differences between students’ and teachers’ punishment goal preferences, in particular for the endorsement of general prevention. Whereas general prevention was least endorsed by students both when asked to indicate their endorsement directly and when evaluating teachers’ punishment practices linked to different goals, it consistently received high support by teachers. This may be particularly problematic, given that the pursuit of general prevention as the goal of punishment may have undesirable consequences. In fact, research in organizational psychology has shown that an authority’s punishment for general preventive purposes is perceived as a signal of distrust and actually leads to a decrease in rule compliance by subordinates (Mooijman et al., 2017). Correspondingly, it could be suggested that this decrease in rule compliance is due to differing goals of people with different hierarchical positions. However, although the present research points to this hypothesis, future research is needed to illuminate this further, in particular given that one cannot necessarily generalize from leaders reacting to subordinate misbehavior in organizational teams to teachers reacting to student misbehavior in schools.

Interesting for the general punishment goal literature, in both teachers and students, we found notable inconsistencies between the direct endorsement of punishment goals and the evaluation of practices that were perceived as achieving these goals. For example, whereas special prevention was consistently of highest preference for both teachers and students when directly asking for their endorsement of the goals, the corresponding punishment practice received relatively low ratings. Furthermore, retribution received rather low direct endorsement scores, whereas the retributive punishment practice was evaluated particularly positively. This is in line with considerable research showing that individuals endorse other goals in the abstract than they support when translated to concrete punishment practices (Applegate et al., 1996). Once more, this also emphasizes that methodological considerations are crucial for the study of people’s endorsement of punishment goals (Twardawski et al., 2020b).

The present results also have several practical implications for the treatment of misbehavior in schools. Given notable discrepancies between teachers’ endorsement of punishment goals and their support for corresponding punishment practices, one might encourage teachers and individuals involved in teacher education to reflect on the topic of punishment in the educational setting and the goals they ought to achieve when responding to misbehaving students. This is particularly important, given that teachers expressed a consistent preference for general prevention as the goal of their punishment, whereas students’ endorsement of this goal was rather low—especially for the general preventive punishment practice. However, students’ negative evaluation of the general preventive practice was arguably to be expected, as one key of actual punishment practices that are meant to prevent future misbehavior is the public display of the offender, the misbehavior, and the punishment (Carlsmith, 2006; Keller et al., 2010) and such a public reprimand has been found to be unacceptable for students (Elliott et al., 1986).

Then again, the consistencies between teachers’ and students’ punishment goal preferences when directly asked for their endorsement of the goals may be cause for optimism in that it should be possible in principle to respond to student misbehavior without triggering students’ perception of injustice—and even without giving up on the goal of general prevention in punishment. That is, agreement of teachers and students was higher when thinking about punishment (and its goals) in the abstract (i.e., the direct endorsement measure) as compared to the evaluation of concrete punishment practices ought to achieve these goals. In light of this finding, teachers may consider to explicitly discuss classroom policies in collaboration with their students to manage potentially emerging student misbehavior. Receiving students’ commitment to such policies based on an abstract support of general preventive punishment may decrease the likelihood of perceived injustice, in case the policy has to be applied to treat a concrete case of student misbehavior (however, see qualifying results from research on the Three Strikes Initiative; Applegate et al., 1996).

An alternative approach that may circumvent the problems arising from differences between students’ and teachers’ perspectives concerning rather punitive reactions to misbehavior (as examined in the present research) follows the principles of restorative justice (Bazemore, 1998; Braithwaite, 1998). This philosophy of dealing with misbehavior considers the perspectives of victims, offenders, and the community in which the offense occurred to assign a punishment. One key aspect of this approach is a face-to-face meeting involving all parties: the victim, the offender, and other community members (Wenzel et al., 2010). In this meeting, offender and victim present their perspectives on the misbehavior and, using a consensus decision-making approach, work out an appropriate punishment for the offender with participation from all parties. Potentially, such restorative justice procedures may resolve the otherwise existing differences in students’ and teachers’ views on an appropriate punishment. In fact, various schools have already introduced justice approaches inspired by restorative justice—such as peer mediation in the case of student conflict or school community conferencing—and although most programs are still in their infancy, there is first evidence for its success with decreasing rates of bullying between students and more positive teacher-student-relationships (Gregory et al., 2016).

However, such approaches also entail wide-ranging challenges (McCluskey et al., 2008). For example, the implementation of restorative justice processes requires deep changes in school climate and, therefore, takes several years to run smoothly (Gregory et al., 2016). Furthermore, not all misconducts can go through such comprehensive processes (Varnham, 2005). Therefore, it is nevertheless important to improve teachers’ ability to independently deal with student misbehavior, and to find punishments that are both appropriate for teachers and perceived as fair by students. The present findings may be helpful to achieve this.

Before concluding, potential limitations of the present research should be acknowledged. First off, the punishment practice evaluation measure, despite the additional insights if affords, does yield certain challenges. Specifically, the punishment practices we extracted from the pretest were perceived as primarily achieving one of the goals (while being equally severe). However, they still also achieved the other two goals to some extent. In fact, the practices may even differ on dimensions we did not consider (and test) in our pretest (e.g., reputational concerns among teachers). However, creating punishment practices that exclusively achieve one punishment goal but no others (while being parallel on any other dimension) may render the teacher responses somewhat artificial (at best), simply because every-day punishment often serves multiple goals at the same time (Gromet & Darley, 2009). Therefore, endorsement of these practices in our research cannot be interpreted as a direct measure of punishment goal preferences. Nonetheless, given the typically moderate correlations between endorsement of abstract punishment goals and preferred punishment practices (Crockett et al., 2014; Twardawski et al., 2020a), it seemed vital to additionally examine preferences on this specific level, too.

Moreover, we used scenarios and comic strips to investigate teachers’ and students’ perspectives on punishment, respectively, rather than observing actual behavior in schools. Indeed, there is arguably an inherent difference between situations actually occurring in class and such hypothetical scenarios (Hughes & Huby, 2004; Schoenberg & Ravdal, 2000). Of note, in the present research, extensive care was taken to ensure that the material was suitable (e.g., by consulting teachers to evaluate and improve the material) and to increase the relevance and authenticity of the student misbehavior and the concrete punishment practices of the teachers used. Additionally, similar methodological approaches have been successfully used to investigate teachers’ evaluation and decision-making in other domains (Baudson & Preckel, 2013). Nonetheless, future research may consider studying actual student misbehavior and punishment practices by teachers in school settings using field observations (Klein, 2008; Klein et al., 1993; Lipshitz et al., 2001).

Further associated with the scenario and comic strip used, teachers and students were confronted with structurally equivalent descriptions of a student misbehavior. However, presentation of the scenario was adjusted to the sample: Teachers read a verbal description of the misbehavior, whereas the misbehavior was presented as a comic strip to students (i.e., to reduce the amount of text). Such an adjustment of material is typical for developmental psychological research comparing the perspectives of adults and children (e.g., McCrink et al., 2010; Powell et al., 2012). Importantly, several control measures on the perception of the misbehavior (rather than the reactions evaluated as main variables) show that teachers and students perceived the situation very similarly, despite its adaptation to different formats. Nevertheless, differences between teachers’ and students’ perspectives may, to some extent, also be a product of this adaptation process.

Relatedly, data from students were exclusively collected in the classroom during school time. By contrast, teachers were provided with the possibility to opt for answering a paper and pencil or online version of the questionnaire and may have answered the questionnaire outside their school environment. While a mixed-mode assessments (i.e., the combination of a variety of survey modes) is becoming increasingly popular and can already be considered common practice (e.g., Dumont et al., 2019; Hübner et al., 2017; von Keyserlingk et al., 2020), the context of study participation may have influenced responses and, thus, the results yielded. In line with this reasoning, recent literature shows that unsupervised web-based study participation is not strictly equivalent to other assessment modes; although biases introduced by web-based testing were generally small (Zinn et al., 2021).

Additionally, we only used one specific instance of student misbehavior (i.e., a student destroying the belongings of another student). Therefore, the results obtained here may be subject to unknown specifics of this scenario and the punishment practices offered (Twardawski et al., 2020b). Indeed, it could be argued that teachers’ and students’ endorsement of punishment goals may be influenced by other aspects of the misbehavior not addressed herein, such as the magnitude of harm caused (Carlsmith, 2006). It is up to future research to replicate and extend the present findings to more diverse forms of student misbehavior. This research could, additionally, also make use of different forms of data collection, such as conducting interviews or collecting other more qualitative data (e.g., Penderi & Rekalidou, 2016). Likewise, we only collected student data from a very specific age group (children around the age of 10). Our theorizing herein is mostly concerned with the position in the school context (i.e., being a teacher vs. a student) rather than age. Consequently, we would not expect that age strongly determines individuals’ relative endorsement of punishment goals in school. However, this is rather speculative and past research reported age differences in evaluations of classroom intervention strategies in some domains (e.g., Bear & Fink, 1991). Hence, further research is needed to examine the role of age on students’ endorsement of punishment goals.

Lastly, it should be mentioned that, regarding the students’ view on the misbehavior and punishment, the results of the present research are limited to the role of an uninvolved observer. By contrast, in many situations of misbehavior, there are several other perspectives involved, such as from perpetrators or victims (Schmitt et al., 2005). Therefore, future work will need to examine the students’ perspective on misbehavior and a teacher’s response to it from different perspectives.

In conclusion, the present research is the first to directly compare teachers’ and students’ views on the purposes of punishment in the school context. In light of the findings and the observation that the approach, as a whole, is fruitful, other researchers are strongly encouraged to integrate all perspectives (i.e., the teachers’ and students’ views) on the psychological analysis of teaching and instruction. Finally, we hope the present findings contribute to the development of classroom intervention strategies that may reduce rather than enhance conflicts in student–teacher-interactions.