Educational leadership is a process needed to achieve common educational goals and refers to encouraging and coordinating the capacities and strengths of teachers, pupils, and parents ([16]). The leadership style of school principals can be related to teachers’ job satisfaction ([93]), teacher commitment ([35]), teaching performance ([61]), and teacher learning ([79]).

Meta-analysis ([93]) shows that ethical leadership (leaders with high moral qualities such as honesty, fairness, caring for employees, and trust) style correlates with teacher job satisfaction and employees’ well-being ([3]). Also, teachers’ commitment increases when they perceive their leaders as supportive, providing a clear school vision, setting directions, and providing instructional support to teachers ([35]). The principal’s instructional leadership affects school culture ([61]). On the other hand, poor leadership is associated with teacher burnout ([11]).

Social support can be defined as emotional, informational, and/or physical support received from other people and can be considered the opposite of social demands ([62]). Social support is an interpersonal construct and process whose main objective is to promote the thriving of an individual ([39]). Social support indicates resources coming from a social network and enhances a person’s capacity to manage stress ([28]) and is linked to health and well-being ([39]). Social support can reduce the negative impact of high-strain jobs on job satisfaction ([90]) and mediates the relationship of control on job strain ([21]). However, the results from a meta-analysis ([47]) did not find a relation between social support and different burnout dimensions, but argue that work sources of support are more strongly related to exhaustion and nonwork-related support to depersonalisation and personal accomplishment. Meta-analyses suggest that social support can reduce the negative effect of workplace bullying on work stress ([27]) and that supervisor support training has a moderate effect on reducing workplace bullying ([88]).

Research in the educational context shows that social support significantly predicts teachers’ psychological well-being ([53]) and mitigates the negative impact of workload on work-family conflict among teachers ([32]). Teachers who receive low social support are more likely to experience high levels of psychological distress ([52]).

Work roles are defined as “the total set of performance responsibilities associated with one’s employment” ([73]) and consist of many responsibilities which vary along two continuous dimensions: social embeddedness (to what extent tasks require coordination with other people) and formalization (specified tasks which are to be performed as a part of the employment relationship) ([73]). Role clarity refers to the objective availability of sufficient role-related information and the subjective perception of having as much role-relevant information as an individual wishes ([66]). Among employees in high-skilled jobs, greater role clarity is associated with higher work ability ([54]).

Work roles characteristics are expansion, ambiguity, conflict, and flexibility ([73]). Role ambiguity, role conflict, and role overload can be considered hindrance stressors as job demands and are negatively related to work affect ([113]). Role ambiguity occurs when individuals are unclear about the expectations they need to fulfill to perform their roles effectively, and role conflict occurs when an individual perceives that performing satisfactorily in one role (or role domain) hinders their ability to perform satisfactorily in another role (or domain) ([73]).

Role stress is a job demand for teachers which can lead to burnout ([45]), and role ambiguity and role conflict are the two main factors of role stress that negatively impact teachers’ job performance and work engagement ([84]; [69]). Perceived role ambiguity among teachers significantly influences their anxiety about school leadership and leadership self-efficacy (Ayyıldız et al., 2025). Meta-analytic evidence indicates that role ambiguity is positively associated with emotional exhaustion, cynicism, and diminished personal accomplishment ([6]). Higher role overload and role conflict are associated with reduced affective commitment and job satisfaction ([111]), and lecturer role conflict negatively affects teaching quality ([83]). Workers’ role conflicts predict sickness absence ([89]), and hindrance stressors have a negative impact on role-based performance ([112]).

Predictability refers to having relevant and useful information about significant upcoming events at the workplace, such as organizational changes, new technologies, or physical modifications, etc. ([76]), and is one of the psychosocial work factors in COPSOQ II measured by how well employees are informed about important decisions, changes, future plans, and provided information to work well ([80]). Predictability has been found to be correlated with quality of leadership and social support ([25]), decision authority ([77]), and employees with job strain reported lower levels of predictability ([25]). While there is limited research on predictability in educational settings, in other contexts it has been found that higher predictability significantly predicted lower absence rates among men ([76]) and lower levels of intention to leave the job ([71]). For example, considering teachers, during COVID-19, lower levels of predictability emerged in educating online compared to in-person teaching ([98]).

Therefore, there is a need to provide insight into how teachers’ and social pedagogues’ personal experiences of school- and workplace bullying behavior—either as perpetrators or victims—may shape their perceptions of interpersonal relations and leadership at work.

The aim of the study was to describe and compare teachers’ and social pedagogues’ experiences with bullying behavior and their perceptions of work-related psychosocial factors such as interpersonal relations and leadership.

According to the research aim, a research question was proposed: What are the differences between teachers and social pedagogues with different experiences of bullying behavior and their evaluations of work-related psychosocial factors such as interpersonal relations and leadership?

The data were collected between January and February 2023 using an online questionnaire. Respondents had the option to save their answers and resume filling out the questionnaire later; furthermore, they were allowed to leave any question unanswered. Data collection was anonymous, participation was voluntary, and all respondents were informed of the study’s objectives. Data were stored confidentially. To ensure the confidentiality and anonymity of the participants throughout all stages of the study, background data concerning work experience were excluded from the analysis.

The teacher sample was a purposive sample involving teachers from across Estonia. To form the sample, half of Estonia’s counties were selected, covering diverse regions (urban and rural areas, counties with high and low population density). All schools representing primary, lower-secondary, and upper-secondary education in the selected counties were identified using the Estonian Education Information System, and all teachers in these schools (N = 4324) were invited to participate via email addresses obtained from publicly available school websites. The participation rate was 10.2% among the teachers who received the invitation. The sample of teachers (n = 441) consisted of 408 women (92.5%), 30 men (6.8%), and 3 others (0.7%). The teachers were distributed into five age groups: up to 29 years old (8.5%), 30 - 39 (16.6%), 40 - 49 (25.7%), 50 - 59 (29.4%), and over 60 years old (19.8%). The sample of social pedagogues was a total population sample involving all social pedagogues across Estonia who work in schools or provide social pedagogical services to general education schools across three educational levels and other educational institutions. These institutions were identified using the Estonian Education Information System and publicly available information, and invitations were distributed via publicly available email addresses. From all social pedagogues who received the invitation (N = 462), 37% of social pedagogues participated in the study. The sample of social pedagogues (n = 171) consisted of 167 women (97.6%), 2 men (1.2%), and 2 others (1.2%). The social pedagogues were distributed into five age groups: up to 29 years old (6.6%), 30 - 39 (26.8%), 40 - 49 (32.7%), 50 - 59 (23.3%), and over 60 years old (10.6%).

Experiences with bullying.To examine teachers’ and social pedagogues’ personal history of bullying experiences (victimization and/or perpetration), a self-reported questionnaire was used, taking into account two aspects: the participant’s role in bullying—perpetrator and/or victim—and the type of bullying experience—retrospective school bullying and/or bullying at work. Retrospective reports of bullying experiences (as perpetrator and/or victim) (e.g., [106]; [67]; [97]) are considered a valid research method ([33]; [46]), as individuals’ descriptions of bullying experiences tend to remain relatively stable over time ([87]). Before the questionnaire was implemented, it was reviewed by experts and tested in a pilot study with teachers.

First, the participants were provided with a definition of bullying ([56]) as repeated, intentional aggressive behavior involving a power imbalance, followed by four questions (yes/no response options) regarding their personal history of bullying involvement: “Have you previously been a victim of bullying or other inappropriate behavior during your school years?”; “Have you previously bullied someone during your school years?”; “Have you been a victim of bullying or other inappropriate behavior by adults working at school?”; “Have you bullied other adults at work?”. Table 1 presents the responses to these self-report questions for participants belonging to groups with different experiences of behavior.

Table 1. Basis for formation of study groups: personal history of bullying behavior.

A total of eight mutually exclusive groups reflecting different bullying experiences emerged, with each participant assigned to one group. No additional categories emerged. The initial sample (n = 626) consisted of teachers (n = 451) and social pedagogues (n = 175), but since two groups (victim/bully in school and victim at work; bully at work) were too small and were excluded from the analysis, the final sample size is n = 612, teachers (n = 441) and social pedagogues (n = 171).

Perceptions of work-related psychosocial factors.To assess the perceptions of work-related psychosocial factors connected with interpersonal relations and leadership, the Copenhagen Psychosocial Questionnaire (COPSOQ II) was used ([60]; [80]). COPSOQ II has been translated into more than 25 languages and has been validated in numerous countries worldwide (e.g., [19]; [80]), also adapted in Estonia (e.g., [70]). COPSOQ II validity and reliability have been tested by several studies in different contexts (e.g., [13]; [36]; [65]; [101]).

COPSOQ II short version consists of 42 items. In this study, 8 items were used to assess interpersonal relations and leadership: predictability (2 items α = 0.798, e.g., “At your place of work, are you informed well in advance concerning, for example, important decisions, changes, or plans for the future?”), role clarity (2 items α = 0.727, e.g., “Does your work have clear objectives?”), quality of leadership (2 items α = 0.828, e.g., “To what extent would you say that your immediate superior gives high priority to job satisfaction?”), and social support from supervisor (2 items α = 0.870, e.g., “How often do you get help and support from your nearest superior?”). Previous evaluations of the psychometric properties of these subscales have confirmed the validity of internal consistency parameters ([13]). Participants responded using a 5-point Likert-type scale, assessing either intensity ranging from 1 (to a very small extent) to 5 (to a very large extent) or frequency from 1 (never/hardly ever) to 5 (always), according to the question type.

Exploratory and confirmatory factor analyses were conducted to assess the factor structure of the “Interpersonal relations and leadership” subscale using the whole sample of respondents (N = 612). For EFA, factors were extracted using the minimum residual (minres) method with oblique Promax rotation, assuming correlated factors; missing values were handled using pairwise deletion to maximize the available sample size for each variable pair. For CFA, parameters were estimated using the maximum likelihood (ML) estimator, and missing responses were handled via full information maximum likelihood (FIML).

To analyze differences among study groups with different bullying experiences, the Kruskal-Wallis test was used, followed by Dunn’s post hoc test with a Bonferroni correction for multiple comparisons; the adjusted p value is reported. For the Kruskal-Wallis and post hoc analyses, participants with missing values for the respective outcome were excluded from the corresponding analysis. Consequently, the analytical sample size varied across questionnaire items. To compare perceptions between teachers and social pedagogues, Mann-Whitney U tests were conducted. For all non-parametric analyses, participants with missing values were excluded listwise per respective item, resulting in minor variations in sample sizes across specific outcome variables.

When exploring teachers’ and social pedagogues’ experiences with bullying behavior (both retrospective in the past and in the workplace), a total of eight groups emerged: not involved in bullying behavior (25.72% of teachers and 28.57% of social pedagogues); victim at school (13.97% of teachers, 13.71% of social pedagogues); victim at work (15.52% of teachers and 12.00% of social pedagogues); victim at school and victim at work (16.85% of teachers and 20.00% of social pedagogues); victim/bully at school and victim/bully at work (17.52% of teachers and 16.00% of social pedagogues); bully at school and victim at work (8.20% of teachers and 7.43% of social pedagogues). Two groups, bully/victim at school and victim at work (n = 11), and bully at work (n = 3), were excluded from the latter analysis as the total sample size was too small. Results indicate that 73.48% of teachers and social pedagogues had a personal history with bullying behavior, and 26.52% reported not being involved in bullying. The study group not involved in bullying behavior was the highest among teachers and social pedagogues. When describing the personal history of bullying involvement, the largest proportion of teachers (17.52%) were in the group victim/bully at school and victim/bully at work, and social pedagogues (20.00%) were in the group victim at school and at work.

To specifically analyze perceptions of interpersonal relations and leadership, a separate model was derived from the original 42-item scale. This focused model comprises four factors (dimensions of work-related psychosocial factors): predictability (2 items), role clarity (2 items), quality of leadership (2 items), and social support (2 items). Exploratory factor analysis (EFA) was conducted to examine the underlying factor structure of the selected items, and EFA results indicated a good fit of the four-factor structure to the data: χ² = 1.243, df = 2, p = 0.537, CFI = 1.000, RMSEA = 0.000 (90% CI [0.000, 0.069]), TLI = 1.004, SRMR = 0.002. Factor loadings are provided in Table 2.

Table 2.Factor loadings of the exploratory factor analysis of COPSOQ II.

Subsequently, confirmatory factor analysis (CFA) was conducted, which generally supported the proposed factor structure for the sample of teachers and social pedagogues. The results were: χ² = 33.385, df = 14, p = 0.003, CFI = 0.998, RMSEA = 0.048 (90% CI [0.027, 0.069]), TLI = 0.996, SRMR = 0.021. The model plot of the CFA is presented in Figure 1.

Across four factors, the mean scores and standard deviations of the dimensions of interpersonal relations and leadership among teachers and social pedagogues with different personal histories of bullying involvement are presented in Table 3. In all factors of interpersonal relations and leadership (predictability, role clarity, quality of leadership, social support) among both samples (teachers and social pedagogues), the assessments were highest in the group not involved in bullying behavior and lowest in the group victimized at school and at work.

Figure 1. Model plot of confirmatory factor analysis of the relationships between factors. Note: Social support = Social support from supervisor.

Table 3.Interpersonal relations and leadership—mean (M) and standard deviation (SD) of teachers (T) and social pedagogues (SP).

Significant differences based on experiences of bullying behavior between study groups were found across all four work-related factors for both teachers and social pedagogues. Specifically, differences in perception of predictability (teachers: H(5) = 15.452, p < 0.01, ε² = 0.035, 95% CI [0.017, 0.085]; social pedagogues: H(5) = 22.145, p < 0.001), role clarity (teachers: H(5) = 13.325, p < 0.05, ε² = 0.030, 95% CI [0.015, 0.087]; social pedagogues: H(5) = 13.499, p < 0.05, ε² = 0.079, 95% CI [0.030, 0.199]), quality of leadership (teachers: H(5) = 12.179, p < 0.05, ε² = 0.028, 95% CI [0.013, 0.079]; social pedagogues: H(5) = 13.219, p < 0.05, ε² = 0.079, 95% CI [0.038, 0.205]) and social support (teachers: H(5) = 17.503, p < 0.005, ε² = 0.040, 95% CI [0.017, 0.090]; social pedagogues: H(5) = 11.499, p < 0.05, ε² = 0.068, 95% CI [0.024, 0.187]).

Predictability.When comparing teachers and social pedagogues with the same bullying experience, perceptions of predictability in the current workplace differed in the group not involved in bullying behavior, whereas teachers rated perceptions of predictability lower (n = 117, Mdn = 4.00, IQR = 0.50) compared to social pedagogues (n = 50, Mdn = 4.00, IQR = 0.87), p < 0.05, U = 2245.00.

Teachers not involved in bullying rated predictability higher (n = 117, Mdn = 4.00, IQR = 0.50) compared to teachers who have been victims at school and at work (n = 76, Mdn = 3.50, IQR = 1.00), p < 0.01, Z = 3.574, rrb = 0.302.

Social pedagogues not involved in bullying (n = 50, Mdn = 4.00, IQR = 0.87) reported higher levels of predictability compared to victims at school and victims at work (n = 35, Mdn = 3.50, IQR = 1.00), p < 0.001, Z = 4.513, rrb = 0.547.

Role clarity. When comparing teachers and social pedagogues, perceptions of role clarity in the current workplace differed in groups: victim/bully at school and victim/bully at work. Teachers with the same bullying experience rated role clarity higher (n = 78, Mdn = 4.00, IQR = 0.50) compared to social pedagogues (n = 28, Mdn = 4.00, IQR = 0.75), p < 0.05, U = 1414.00.

Teachers who were victims at school and victims at work (n = 76, Mdn = 4.00, IQR = 1.00) reported lower levels of role clarity compared to teachers in the group not involved in bullying (n = 116, Mdn = 4.00, IQR = 1.00), p < 0.05, Z = 3.270, rrb = 0.262, and victims at school (n = 64, Mdn = 4.25, IQR = 0.50), p < 0.05, Z = 2.999, rrb = 0.279.

Social pedagogues not involved in bullying (n = 50, Mdn = 4.00, IQR = 0.50) reported higher levels of role clarity compared to social pedagogues who were victims at school and victims at work (n = 35, Mdn = 4.00, IQR = 0.50), p < 0.05, Z = 3.373, rrb = 0.410.

Quality of leadership. When comparing perceptions of the quality of leadership among teachers and social pedagogues with the same experience with bullying behavior, there were no statistically significant differences.

Teachers as victims at school and at work (n = 75, Mdn = 4.00, IQR = 1.00) reported lower levels of quality of leadership compared to teachers not involved in bullying behavior (n = 117, Mdn = 4.00, IQR = 1.00), p < 0.05, Z = 3.261, rrb = 0.286.

Social pedagogues as victims at school and victim work (n = 34, Mdn = 3.5, IQR = 1.37) reported lower levels of quality of leadership compared to social pedagogues not involved in bullying (n = 48, Mdn = 4.00, IQR = 0.75), p < 0.05, Z = 3.322, rrb = 0.419.

Social support from supervisor. When comparing perceptions of social support among teachers and social pedagogues with the same experience with bullying behavior, there were no statistically significant differences.

Teachers as victims at school and victims at work (n = 76, Mdn = 3.75, IQR = 1.5) reported lower levels of social support compared to teachers not involved in bullying behavior (n = 116, Mdn = 4.5, IQR = 1.00), p < 0.005, Z = 3.806, rrb = 0.318.

Social pedagogues not involved in bullying (n = 50, Mdn = 5.00, IQR = 1.00) reported higher levels of social support compared to victims at school and victims at work (n = 35, Mdn = 3.50, IQR = 1.50), p < 0.05, Z = 3.141, rrb = 0.381.