School disciplinary climate and safety play a crucial role in shaping student learning, equity and well-being. Drawing on PISA 2022 data, this chapter analyses how students in Sweden experience classroom order, safety and behavioural engagement. It examines nine indices and two individual items, comparing Sweden with other OECD and Nordic countries. The analysis includes results disaggregated by gender, socio-economic status, migrant background and school location. Findings highlight Sweden’s strengths in student-teacher relationships, teacher support and sense of belonging, alongside persistent challenges related to classroom disruption, bullying and absenteeism. Importantly, positive associations between school climate and learning outcomes are not equally distributed, revealing equity gaps across the system. Policy considerations to address these challenges are presented at the end of the chapter and further developed in Chapter 5.
Education in Sweden
Abstract
Drawing on 2022 data from the OECD Programme for International Student Assessment (PISA), this chapter examines the disciplinary climate and safety conditions in Swedish schools to explore how these factors influence student learning and well-being. The analysis complements findings from the Swedish National Agency for Education’s (Skolverket) study Elevers trygghet och studiero – en fördjupning av resultat i PISA 2022 (Skolverket, 2024[1]), which investigates how aspects of the school environment affect Swedish students’ PISA outcomes. This chapter considers both students’ perceptions and school-level practices reported by principals, including classroom management, exposure to bullying, feelings of safety, and the quality of student-teacher relationships, thereby broadening the perspective offered by the report developed by Swedish authorities.
The chapter begins by outlining how PISA measures school climate and safety through student and school questionnaires, highlighting the strengths and limitations of this international dataset. It then presents a comprehensive analysis of twelve indices, comparing Sweden’s results with those of OECD and Nordic peers. Each index is examined in relation to mathematics performance and disaggregated by gender, socio-economic status, migrant background, and school location to identify equity gaps. The chapter concludes with a brief discussion of policy considerations, which will be developed further in the final chapter of this review.
Understanding the importance of disciplinary climate and safety: Research evidence and literature insights
School disciplinary climate refers to the overall classroom environment that shapes student behaviour, serving as a critical factor in influencing learning outcomes. It encompasses various elements such as classroom management, the presence of clear rules and expectations, teacher-student interactions, and the general atmosphere of respect and order. Empirical evidence points to the importance of a positive disciplinary climate in fostering environments conducive to learning and personal growth (OECD, 2020[2]; Thapa et al., 2013[3]; Wang and Degol, 2015[4]). This climate is characterised by structured, respectful interactions between students and teachers, minimal noise and disorder, few disruptions to learning, and lessons that begin with students promptly engaging in their tasks (OECD, 2020[2]).
The significance of disciplinary climate becomes even more apparent when viewed through the lens of international data. The OECD’s Programme for International Student Assessment (PISA) can serve as a valuable tool for analysing the impact of disciplinary climate and safety on student learning outcomes. The PISA offers a comprehensive, comparative view of educational environments by assessing the knowledge and skills of 15-year-old students in reading, mathematics, and science worldwide (OECD, 2023[5]). Findings from both PISA 2022 underscore a consistent, positive association between effective classroom climate and student performance. For instance, PISA 2018 showed that students who perceived their classrooms as orderly and respectful are more likely to perform better academically, even after accounting for socio-economic status and other factors (OECD, 2020[2]). Skolverket’s (2024[1]) complementary analysis of PISA 2022 data highlighted similar patterns, noting that students reporting higher levels of safety and classroom order achieved stronger learning outcomes. Earlier PISA results from 2009 similarly emphasised that fewer classroom interruptions allow teachers to devote more time to instruction rather than managing behaviour, resulting in greater student focus and engagement (OECD, 2020[2]).
Further research supports these findings by underscoring the detrimental effects of disruptive peer behaviour on individual academic achievement. A study by Blank and Shavit (2016[6]) demonstrated a negative correlation between peer disruptions and student outcomes, reinforcing the need for well-managed classrooms. Notably, the association between a positive disciplinary climate and academic performance appears stronger for students from disadvantaged or minority backgrounds. In such cases, an orderly and respectful environment can help mitigate broader social inequities by providing all students with equitable opportunities to learn (OECD, 2020[2]). Furthermore, research underscores that the quality of classroom interactions, particularly those linked to disciplinary climate, plays a significant role in shaping educational outcomes. Fauth et al. (2014[7]) found that student perceptions of teaching quality, including classroom management, were predictive of academic achievement. Their study highlights how well-managed classrooms contribute not only to order but also to more effective use of learning time. This finding is further supported in the PISA data from both 2006 and 2022, which show that while the quantity of instructional time is important, its effectiveness is largely determined by the conditions under which it occurs. Disruptions, whether due to peer misbehaviour or digital distractions, can diminish the learning potential of instructional time, underscoring the critical importance of classroom management in promoting productive learning environments (OECD, 2023[5]).
Beyond maintaining order, effective classroom management is linked to student academic motivation and subject-specific engagement. Kunter, Baumert, and Köller (2007[8]) observed that classrooms with clear routines and behavioural expectations can enhance students’ interest in the subject matter. This is especially relevant in cognitively demanding subjects such as mathematics, where teacher sensitivity and responsiveness have been shown to influence both performance and motivation (Lazarides and Buchholz, 2019[9]). Data from PISA 2022 further support this view: students who report receiving greater support from their mathematics teachers tend not only to perform better but also to report a stronger sense of belonging, lower mathematics anxiety, greater confidence in self-directed learning, and higher life satisfaction (OECD, 2023[5]). These findings reflect the multifaceted role of disciplinary climate. It serves not only as a determinant of academic outcomes but also as a foundation for emotional and motivational well-being.
In addition to classroom-level dynamics, the broader concept of school safety is integral to student learning experiences. A safe school environment protects students from physical and emotional harm, including violence and bullying, both in-person and online. It enables students to feel secure, respected, and intellectually supported, which are key elements for building trusting relationships and promoting positive engagement (Thapa et al., 2013[3]). PISA data consistently reveal that students who feel safe at school tend to achieve higher academic outcomes, as a sense of safety allows them to focus more fully on learning rather than being preoccupied with stress or fear (Bradshaw, Hoelscher and Richardson, 2006[10]; Wang and Degol, 2015[4])
The connection between school safety and student well-being is similarly well established. Students who feel secure at school are more likely to report a strong sense of belonging, greater life satisfaction, and lower levels of anxiety (OECD, 2023[5]). Conversely, those exposed to bullying or threats of harm often experience emotional distress and social exclusion. A positive school climate can serve as a buffer against these effects. For instance, Hultin et al. (2018[11]) found that supportive school environments mitigate the psychological impact of bullying. Other studies confirm that student sense of safety is closely tied to broader well-being outcomes, including emotional resilience and peer connectedness (Catalano et al., 2004[12]; Steffgen, Recchia and Viechtbauer, 2013[13])
Safety and disciplinary climate also play a vital role in shaping students’ behavioural engagement. Students who feel safe and supported are more likely to attend school regularly and arrive on time, and these are two factors strongly associated with academic progress. PISA data show that systems with stronger disciplinary environments and safety measures report higher levels of student attendance and punctuality (OECD, 2023[5]). Similarly, students in high-performing systems who report feeling secure are less likely to skip classes or arrive late, allowing them to stay on track academically. However, reducing absenteeism requires more than behavioural enforcement. Effective approaches combine consistent discipline with academic, social, and emotional supports (Germain et al., 2024[14]).
Finally, a safe and inclusive school climate supports students’ broader social development. Studies suggest that students in positive school environments are less likely to engage in risky behaviours such as smoking, drinking, or substance abuse (Catalano et al., 2004[12]; OECD, 2023[5]; LaRusso, Romer and Selman, 2007[15]). Such environments also nurture interpersonal skills and cooperative behaviour, contributing to a constructive atmosphere for both teaching and learning. Wang and Degol (2015[4]) argue that these social-emotional benefits are not merely additional but are essential to building effective and equitable education systems.
PISA collects data on disciplinary climate and school safety through both student and school questionnaires, offering a comprehensive perspective on the conditions that shape students’ ability to learn and feel secure at school. Data reported by students provide direct insights into their day-to-day experiences in the classroom, including the frequency of disruptions, perceptions of safety, and incidents of bullying or violence. In parallel, information provided by school principals on classroom management practices and institutional policies for student safety adds valuable context at the school level. By combining student perspectives with school-level information, this dual approach of the PISA enables a more nuanced and robust assessment of disciplinary climate and school safety.
The following discussion outlines the methodological strengths and limitations of using PISA data to explore these issues, highlighting the value of integrating student and institutional perspectives to better understand the learning environment.
Strengths and limitations of the PISA data to assess disciplinary climate and safety
One of the key strengths of PISA’s methodology lies in its comprehensive data collection, which integrates insights from both students and school leaders. It offers a valuable lens for examining student perceptions of disciplinary climate and safety by reflecting the first-hand experiences of those most directly affected by classroom conditions. This approach delivers an authentic and timely view of learning disruptions and the everyday realities of school life. At the same time, school questionnaires completed by school leaders offer an institutional perspective on disciplinary issues, drawing on their oversight of school-wide policies, procedures, and behavioural patterns across classrooms and student groups. Their professional experience and observational expertise enable them to identify disruptions and safety concerns that students may overlook or underreport. This multi-source approach ensures that both subjective student experiences and institutional policies are considered when evaluating school climate. Additionally, the use of standardised measures allows for consistency across different countries, making international comparisons possible and providing benchmarks for educational systems worldwide (OECD, 2023[5]).
However, there are also limitations to this approach. Since PISA relies heavily on self-reported data, student responses may be influenced by personal biases, memory recall issues, or social desirability factors, which can affect the accuracy of estimates on the prevalence of issues such as bullying or safety risks. Additionally, cultural differences in perceptions of discipline and safety may influence how students respond, potentially impacting the comparability of data across education systems. This challenge was particularly evident in the case of Sweden, Norway and New Zealand, where differing interpretations of concepts such as vandalism1 led to the adjustment of certain items from the index of school safety risks to ensure valid international comparisons. These limitations are more relevant when the data are used to assess the absolute levels of specific issues across countries. However, they are less problematic when analysing the relationships between school climate factors and student outcomes, especially when multivariate models account for socio-economic status and other contextual variables. In such cases, the analysis can still provide valuable insights into how perceived school climate is associated with learning outcomes (OECD, 2023[5]).
Another key limitation is the scope of measurement. Not all elements of school safety and disciplinary climate can be considered in the survey. It only offers a snapshot, capturing conditions as they exist at a single point in time. This cross-sectional design limits the ability to observe how school climate evolves over the course of a school year or to evaluate the sustained long-term impact of specific policy interventions. Furthermore, the data may not provide sufficient detail on the broader environmental or socio-economic conditions in which schools operate. As a result, it can be difficult to disentangle school-based disciplinary issues from external factors such as home environment, neighbourhood safety, community cohesion or social inequality that also shape student behaviour and perceptions of safety at school. Factors such as teacher-student relationships, broader school culture, and the long-term impact of school safety initiatives may require additional data sources and methodological approaches to build a more complete understanding.
In summary, PISA provides a valuable framework for assessing how school environments influence student learning and well-being. By leveraging student experiences and institutional data, PISA offers comprehensive insights into classroom order and school safety on a global scale. However, its findings must be interpreted with care. The reliance on self-reported data, the influence of cultural variations, and limited scope of measurement constrains its ability to fully capture the complexities of school safety and disciplinary climate. Recognising and understanding the strengths and limitations of this data allows policymakers to make more informed decisions when designing interventions aimed at improving school discipline and ensuring safer learning environments.
This analysis draws on data from both the PISA 2022 student and school questionnaires to examine key dimensions of disciplinary climate and safety in Swedish schools. A total of nine indices and two additional behavioural measures were analysed, covering a broad spectrum of factors that shape students’ perceptions of safety, order and support at school. Specifically, seven indices were constructed from student questionnaire data and two from school questionnaire responses, alongside two student-reported behavioural indicators related to absenteeism and lateness. These are grouped as follows:
From the PISA student questionnaire:
Index of disciplinary climate in mathematics lessons (DISCLIM)
Index of exposure to bullying (BULLIED)
Index of feeling safe (FEELSAFE)
Index of school safety risks (SCHRISK)
Index of sense of belonging (BELONG)
Index of quality of student-teacher relationships (RELATST)
Index of teacher support in mathematics (TEACHSUP)
From the PISA school questionnaire (principal-reported):
Index of student-related factors affecting school climate (STUBEHA)
Index of negative school climate (NEGSCLIM)
In addition, measures of student absenteeism and lateness, although not formal indices, were treated as index-like indicators given their relevance to school climate and behavioural engagement. They are examined in two items that belong to the PISA student questionnaire:
Student absenteeism (SKIPPING)
Student lateness (TARDYSD)
While the primary focus of this chapter is on disciplinary climate and safety, the analysis draws on a broader set of PISA 2022 indices and indicators that are conceptually and empirically related to students’ behavioural engagement and perceived learning environment. In addition to the core indices on disciplinary climate and school safety risks, the analysis includes measures such as bullying, sense of belonging, student-teacher relationships, and teacher support in mathematics. These elements were selected based on research evidence suggesting that they influence students’ experience of safety, order, and inclusion at school, and may interact with disciplinary climate in shaping learning outcomes (OECD, 2023[16]; OECD, 2023[5]). Moreover, indicators of student absenteeism and lateness, though not formal indices, were included as proxy measures of behavioural engagement, which is both a potential consequence of negative climate and a factor reinforcing it. Two school-reported indices on climate-related challenges were also incorporated to provide additional perspective from school leaders. Taken together, this broader set of indicators allows for a more holistic understanding of the school environment in Sweden and how different aspects may relate to student performance.
For each index, average results for Sweden were compared to the OECD average and to those of three Nordic peers: Finland, Norway and Denmark. To better understand how school climate conditions are associated with student learning outcomes, the analysis proceeds in several steps:
First, bivariate regressions were conducted to examine the relationship between each school climate index and mathematics performance. These models were estimated both at the overall student level and within specific subgroups (e.g. girls, boys, socio-economically advantaged and disadvantaged students, immigrant and non-immigrant students, and students in three different school locations). The bivariate analysis describes the simple association between each climate factor and student performance, without accounting for other variables.
Second, recognising that students’ family background and school socio-economic context are important influences on both performance and perceptions of school climate, a multivariate analysis was conducted. This second step accounts for students’ and schools’ socio-economic profiles, as measured by the PISA index of economic, social and cultural status (ESCS). The multivariate results provide insights into how each school climate factor relates to performance after accounting for socio-economic status. This approach helps to distinguish whether observed associations persist independently of students’ socio-economic context.
The results presented throughout the chapter reflect both the bivariate and multivariate analyses. Associations are described carefully as correlational, not causal. Moreover, all bivariate regression results are reported with statistical significance levels clearly indicated.
In addition to index-level analyses, item-level comparisons were performed to assess how Swedish students’ and principals’ responses to individual items within each index differ from OECD averages. Where relevant, bivariate regression analyses were also conducted at the item level to identify statistically significant differences in performance across the four equity dimensions (i.e. gender, socio-economic status, migrant background, and school location).
Finally, trend analyses were conducted using data from previous PISA cycles. Where feasible, index-level trends were examined across 2012, 2015, 2018 and 2022. However, given limitations in index comparability over time, trend analysis focussed on individual items that remained consistent across cycles, covering four indices in total.
The situation of disciplinary climate and safety in Swedish schools according to PISA 2022 data
According to the latest PISA data, the disciplinary and safety environment in Swedish schools presents a mixed picture with both commendable strengths and notable challenges. One significant aspect of the disciplinary climate in Sweden is the emphasis on creating a respectful and inclusive classroom atmosphere. Swedish educational policies promote democratic values and student participation, which are reflected in classroom management practices. Teachers in Sweden are encouraged to foster open communication and mutual respect, which help create a positive learning environment (OECD, 2023[17]; OECD, 2023[16]; Ministry of Education and Research, 2024[18]). Teacher-student interactions are typically characterised by a high level of trust and support. Teachers are seen not just as authority figures but also as facilitators of learning who provide guidance and support to students. In the following sections, this chapter will explore relevant items from the PISA test to explore the impact of climate and safety in schools on student learning.
Exploring disciplinary climate and safety in PISA: Student questionnaire items
Student questionnaire items in the PISA offer a crucial perspective on disciplinary climate and safety of schools in Sweden by capturing students’ direct experiences with class disruptions, perceptions of safety, and exposure to bullying or violence. However, their interpretation requires caution due to the subjective nature of self-reported data and potential cultural differences in how students respond.
Index of disciplinary climate
The index of disciplinary climate captures the extent to which classroom disturbances affect students’ ability to concentrate and engage with learning tasks. Based on responses to the PISA student questionnaire, the index includes indicators measuring student attentiveness, the frequency of lessons being disrupted by noise and disorder, and how often teachers are required to wait for students to settle down. It also reflects delays in getting started with learning, including time lost at the beginning of class and distractions caused by the use of digital devices (both by the student and their peers). A positive disciplinary climate is widely recognised as an important foundation for effective teaching and learning, enabling better classroom management, maximising instructional time, and contributing to improved student outcomes (OECD, 2023[5]; Kunter, Baumert and Köller, 2007[8]).
The mean disciplinary climate index in Sweden is -0.32, indicating a relatively negative disciplinary climate in comparison to the OECD average of 0.02. This places Sweden on the more challenging side of the disciplinary climate spectrum compared to other countries. Among Nordic countries, Denmark stands out with a positive index of 0.03, suggesting a more orderly classroom environment. Finland performs slightly better than Sweden with an index of -0.22, while Norway reports a higher score of -0.08, indicating a comparatively more positive disciplinary climate than both Sweden and Finland.
While these average scores provide a useful benchmark, it is equally important to understand what they imply for students’ learning outcomes. To explore this, bivariate regression analysis examines how disciplinary climate relates to mathematics performance. In Sweden, a one-unit increase in the index of disciplinary climate is associated with a 6-point increase in mathematics performance (Figure 4.1). It is important to note that these results describe statistical associations and do not imply causality.
Note: Statistically significant differences are shown in a blue tone. Each bar represents the coefficient from a bivariate regression of mathematics performance on the index, estimated separately for each subgroup (e.g. girls, boys, socio-economically advantaged/disadvantaged, etc.). No control variables are included in these models. Results reflect associations, not causal effects. A socio-economically disadvantaged (or advantaged) student is a student in the bottom (or top) quarter of the PISA index of economic, social and cultural status (ESCS) in Sweden. Rural areas or villages are defined as having fewer than 3000 inhabitants; town areas have between 3000 to 100 000 inhabitants; and cities have over 100 000 inhabitants. A positive value on the disciplinary climate index corresponds to more favourable student perceptions of classroom order and fewer disruptions.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Uneven experiences: How disciplinary climate varies across students
Students in Sweden do not report experiencing the disciplinary climate in the same way. When disaggregating data by gender, Sweden shows a small but significant difference: girls report a mean index of -0.35 compared to -0.28 for boys, indicating that girls perceive classroom conditions slightly more negatively. Despite this, both groups show similar patterns in the association between disciplinary climate and mathematics performance, with scores rising by 5 points for girls and 6 points for boys in response to a one-unit increase. However, two specific items within the index are associated with markedly larger score declines among boys than girls, suggesting that certain classroom conditions may be more strongly associated with boys’ academic outcomes (Table 4.1).
Changes in mathematics performance when students report the following disruptions (all values shown are statistically significant)
Note: All values are statistically significant. Values shown represent the estimated change in mathematics performance (PISA score points) associated with a one-unit increase in the reported item, based on separate bivariate regressions for girls and boys. Positive values indicate higher performance associated with reporting the behaviour. These associations should be interpreted with caution, as they do not imply causality and may reflect underlying group dynamics or response patterns.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Beyond gender, students’ family status also shapes their experiences of classroom conditions. In Sweden, socio-economically advantaged students report a less favourable disciplinary climate (-0.32) than disadvantaged students (-0.27), a pattern that contrasts with smaller gaps observed in Finland and Norway. The association with academic performance also differs. Among advantaged students, better disciplinary conditions are linked to a notable 10-point increase in mathematics scores, while disadvantaged students show a slight 2-point decline. At the item level, four indicators reveal statistically significant differences between groups, with advantaged students experiencing greater performance losses across all items (Table 4.2).
Changes in mathematics performance when students report
Note: Statistically significant values are indicated in bold. Values shown represent the estimated change in mathematics performance (PISA score points) associated with a one-unit increase in the reported item, based on separate bivariate regressions for advantaged and disadvantaged students. Positive values indicate higher performance associated with reporting the behaviour. These associations should be interpreted with caution, as they do not imply causality and may reflect underlying group dynamics or response patterns. A socio-economically disadvantaged (or advantaged) student is a student in the bottom (or top) quarter of the PISA index of economic, social and cultural status (ESCS) in Sweden.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
When considering immigrant background, Sweden exhibits a difference in the disciplinary climate between non‑immigrant and immigrant students. Non‑immigrant students report a mean index of –0.34, while immigrant students report a more positive index of –0.23. Improvements in disciplinary climate appear to be associated with different academic patterns: non‑immigrant students show a 9‑point increase in mathematics performance, while no statistically significant association is observed among immigrant students. A greater difference emerges in the item “students do not listen to what the teacher said” in every or most lessons, where the gap in mathematics performance between the two groups is 13 points: non‑immigrant students experience a decline of 3 points, whereas immigrant students show an 11‑point increase. At the item level, these results may appear counter‑intuitive: for example, when students report that they do not listen to what the teacher says, non‑immigrant students show a slight decline in scores while immigrant students report an increase. These patterns may reflect differences in how students interpret the question, or the influence of other unobserved factors such as school environment, classroom expectations or peer norms. As noted by Skolverket (2024[1]), interpretation of the Swedish results concerning this item should be handled with particular care, since the translation in PISA 2022 slightly differs from both earlier and international versions of the question (see Elevers trygghet och studiero – en fördjupning av resultat i PISA 2022). It is also possible that immigrant students are more resilient in the face of classroom disorder or attend schools where other supports mitigate its impact. As noted previously, these findings should be interpreted with caution and highlight the need for further investigation into how classroom dynamics may be experienced differently across student groups.
Geography adds yet another layer, as school location is associated with differences in disciplinary climate and academic outcomes. Students in rural and town settings report the same average disciplinary climate index of -0.35. However, a one-unit increase in the disciplinary climate index is associated with slightly higher scores in mathematics for rural students (+6 points) compared to students in city or town settings (+4 points), although the 2-point difference between the locations remains small. Students attending schools in town settings report a somewhat more favourable score in the index of disciplinary climate (- 0.27) and demonstrate the largest performance increases, with a 9-point increase in mathematics associated with better classroom conditions.
Beyond these simple associations, it is important to also examine whether the link holds once other factors, such as students’ and schools’ socio-economic profiles, are taken into account. Multivariate analysis of PISA 2022 data indicates that a more positive disciplinary climate is statistically associated with higher mathematics performance in Sweden. Specifically, a one-unit increase in the index of disciplinary climate is linked to a 6-point increase in mathematics scores, after accounting for socio-economic status (Annex Table 4.A.1). This association is statistically significant, although the index explains only a modest share of the variance, accounting for about 0.4% before accounting for socio-economic status are applied. These multivariate results align with the bivariate findings presented earlier in this section, which also show a positive relationship between disciplinary climate and mathematics performance in Sweden. However, the magnitude of the association decreases once socio-economic factors are considered, suggesting that some of the relationships observed in the bivariate analysis may reflect underlying socio-economic differences among students and schools.
Insights from international practice
These findings suggest that addressing frequent classroom disruptions could help create better learning conditions. Some countries have taken practical steps in this direction. Australia, for example, has prioritised strengthening classroom management through targeted teacher development (Box 4.1). By embedding evidence-based classroom management training into teacher education and early-career support, Australia demonstrates how clear routines, effective transitions and consistent expectations can help teachers limit disruptions, maximise learning time, and create more orderly learning environments.
Classroom noise and disorder have long been recognised as significant barriers to maximising instructional time in Australian schools. National reviews highlighted that a notable proportion of Australian teachers, particularly those early in their careers, report losing valuable learning time due to frequent interruptions, excessive noise, and challenges in settling students at the start of lessons.
In response to these challenges, Australia has progressively expanded its national focus on evidence-informed classroom management training as a core component of teacher preparation and professional development. The Australian Institute for Teaching and School Leadership (AITSL) provides structured guidance for embedding effective classroom management strategies within initial teacher education programmes and ongoing school-based induction.
Central to this approach is a clear articulation of classroom management as a learnable and improvable skill set, rather than solely a function of teacher personality or experience. Early-career teachers receive practical training on establishing routines, setting clear behavioural expectations, using non-verbal signals, and managing transitions efficiently to minimise noise and disorder during lessons. Schools are encouraged to complement this training with peer observation, mentoring, and feedback cycles, enabling novice teachers to develop confidence and adapt strategies to diverse classroom contexts.
Evaluations of these initiatives indicate promising results. Teachers who engage in structured classroom management training report higher levels of self-efficacy in maintaining orderly learning environments and spend less time addressing disruptive behaviour. Schools investing in targeted classroom management report smoother lesson transitions, fewer interruptions, and improved student engagement.
Australia’s experience underlines the importance of integrating evidence-based classroom management into both teacher education and in-service professional learning. By doing so, schools can better equip teachers to create calm, well-ordered classrooms that maximise learning time and strengthen overall disciplinary climate.
Source: OECD (2019[20]), Richardson et al. (2023[21]), Australian National Audit Office (2024[22]) and Australian Institute for Teaching and School Leadership (2025[23]).
Observed patterns in disciplinary climate
Trends in the index of disciplinary climate in Sweden show both increases and declines over time. While the full index is not directly comparable across all PISA cycles due to the addition of new items over time, five items have been consistently measured since 2012 (Figure 4.2).
Percentage of students responding that they experience disruptions “in most or every lesson”
Note: Based on student reports. Trends are shown only for the five comparable items that have been consistently measured since PISA2012. Caution is required when interpreting Swedish results for the item “Students do not listen to what the teacher says”, as the wording of this question in the Swedish PISA 2022 translation differs slightly from both earlier and international versions. This variation may partly explain observed differences in trend patterns (Skolverket, 2024[1]).
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Between PISA 2012 and PISA 2022, three items in the index of disciplinary climate show an upward trend in Sweden. The most notable change is in the share of students reporting that they “do not listen to what the teacher says” in most or every lesson, which rose from 34.4% in 2012 to 57.7% in 2022. However, this result should be interpreted with caution, as the Swedish translation of this item in PISA 2022 differs slightly from both previous national cycles and the international version. These translation differences may have contributed to the apparent increase in reported classroom disruption (Skolverket, 2024[1]). Although a similar trend is observed across the OECD, the increase was more moderate, from 22.8% to 30.4%. Sweden also saw a rise in the proportion of students reporting that they are unable to work well in class (from 25.9% to 29.9%) and a very slight increase in those reporting that they delay starting work (from 32.3% to 33.2%), although the latter change is not statistically significant.
Research shows that maximising instructional time is crucial for academic achievement. Delays in starting work can reduce overall learning time, disrupt lesson flow, diminish student engagement, and increase likelihood of further misbehaviour or distractions (Daily et al., 2020[24]; Cohen et al., 2009[25]). The frequent delays observed in Swedish classrooms may contribute to lower performance outcomes, particularly in subjects like mathematics that require sustained focus and practice (Lazarides and Buchholz, 2019[9]). In contrast, two items suggest modest improvements in classroom conditions. The share of students reporting noise and disorder declined slightly in Sweden (from 38.1% to 37.1% and statistically significant), as did the item indicating that the teacher must wait a long time for students to quiet down, which fell from 34.2% to 30.7%. These improvements align with broader OECD trends, suggesting some progress in reducing certain types of classroom disruption.
Conclusion
PISA 2022 data indicate that Sweden’s disciplinary climate remains a challenge, particularly in comparison with other Nordic countries and the OECD average. Caution is warranted when interpreting Swedish results for specific questionnaire items, such as “Students do not listen to what the teacher says”, since the PISA 2022 Swedish translation differs slightly from earlier and international versions, which may have contributed to observed differences (Skolverket, 2024[1]). Swedish students report more frequent classroom disruptions, and trend data show that disciplinary conditions have worsened in three of the five comparable PISA items since 2012. The remaining two items show modest improvements. Bivariate analysis shows that, notably, disciplinary climate is not experienced uniformly across student groups. Socio-economically advantaged students, boys, and non-immigrant students tend to report more negative disciplinary conditions and show larger performance differences when these conditions are present. In contrast, disadvantaged and immigrant students report slightly more favourable disciplinary conditions, but these do not appear to be associated with comparable academic improvements. However, after accounting for student and school socio-economic profiles, more positive disciplinary conditions are statistically associated with higher mathematics scores. However, the relationship remains modest and may reflect other unobserved factors. Overall, these results highlight that while better classroom order is linked to stronger performance, its benefits are not equally experienced by all students. This underlines the need to strengthen classroom management alongside broader supports that help students from diverse backgrounds benefit equally from improved learning conditions.
Index of exposure to bullying
The index of exposure to bullying reflects the prevalence and types of bullying behaviours experienced by students, as reported in the PISA 2022 student questionnaire. This index captures a range of physical, emotional, and relational bullying, including both in-person and online interactions. Items include statements such as: “Other students left me out of things on purpose,” “Other students made fun of me,” “I was threatened by other students,” “Other students spread nasty rumours about me,” and “I got hit or pushed around by other students.” More severe or coercive behaviours are also considered, such as “I was in a physical fight on school property,” “I gave money to someone at school because they threatened me,” and “I stayed home from school because I felt unsafe.” Higher (more positive) values on the index reflect greater exposure to bullying, while lower (more negative) values indicate less exposure.
Social exclusion, one of the relational forms of bullying captured in the index, undermines students’ psychological need for belonging, one of the cornerstones of motivation and cognitive engagement (Hessen and Kuncel, 2022[26]). Research shows that excluded students often struggle with attention regulation, persistence, and participation in classroom activities, which can lead to lower academic performance, particularly in reading and mathematics (Lazarides and Buchholz, 2019[9]; Şirin, 2022[27]; Vanderbilt and Augustyn, 2010[28]). Longitudinal studies further indicate that children who experience persistent social exclusion tend to exhibit poorer academic outcomes over time (Hessen and Kuncel, 2022[26]; Sheridan et al., 2019[29]; Thapa et al., 2013[3]). Furthermore, mocking represents a verbal form of bullying that erodes self-efficacy and amplifies stress responses, directly impeding learning processes such as problem-solving and attentional focus (Vanderbilt and Augustyn, 2010[28]). Students who experience mocking often internalise feelings of shame or humiliation, which can lead to disengagement from school activities (Woods and Wolke, 2004[30]). Research also indicates that adolescents subjected to these forms of victimisation exhibit structural changes in brain regions associated with emotional processing (Wolke and Lereya, 2015[31]; Sheridan et al., 2019[29]) which may hinder their ability to regulate emotions and focus academically. Mocked students exhibit elevated cortisol levels, impairing memory consolidation and reducing resilience in academic settings (Şirin, 2022[27]; Wolke and Lereya, 2015[31]).
In Sweden, the average index score of exposure to bullying is -0.27, close to the OECD average of -0.30 and broadly in line with Norway (-0.28). This suggests that the overall level of reported bullying in Sweden is slightly better than the OECD average but falls behind Finland (-0.39), where students report lower exposure to bullying. By contrast, Denmark stands out in the Nordic region with a considerably less favourable score of -0.04, indicating higher reported exposure to bullying. It is important to note that because higher (more positive) values reflect greater exposure to bullying, while lower (more negative) values indicate less exposure, Sweden’s index score of -0.27 signals that, on average, students in Sweden report slightly lower levels of bullying than the OECD average of -0.30. While these average values offer an overview, they do not convey how experiences of bullying may be associated with student learning. Exploring these patterns can offer insights into broader educational implications in Swedish schools.
Uneven experiences: How bullying varies across students
Exposure to bullying is strongly linked with lower academic performance. In Sweden, a one-unit increase in the index of exposure to bullying is associated with a 10-point decline in mathematics performance. This negative association is evident across student groups but differs in magnitude (Figure 4.3).
Note: All values are statistically significant. Each bar represents the coefficient from a bivariate regression of mathematics performance on the index, estimated separately for each subgroup (e.g. girls, boys, socio-economically advantaged/disadvantaged, etc.). No control variables are included in these models. Results reflect associations, not causal effects. A socio-economically disadvantaged (or advantaged) student is a student in the bottom (or top) quarter of the PISA index of economic, social and cultural status (ESCS) in Sweden. Rural areas or villages are defined as having fewer than 3000 inhabitants; town areas have between 3000 to 100 000 inhabitants; and cities have over 100 000 inhabitants.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
When results are disaggregated by gender, girls and boys report similar exposure levels (mean scores of -0.29 and -0.24, respectively), girls tend to show larger performance differences when these conditions are reported. When bullying increases, girls’ mathematics scores drop by 15 points, more than twice the 6-point decrease observed for boys. This 9-point gap is statistically significant and suggests that bullying may undermine girls’ academic engagement or confidence to a greater extent. Research suggests that gender disparities in how students experience and report bullying may be associated with distinct social behaviours and emotional response patterns. For girls, higher exposure to relational aggression is associated with elevated risks of anxiety and depressive symptoms, contributing to reduced academic engagement (Hackman et al., 2022[32]). By contrast, lower exclusion rates among boys are frequently linked to their involvement in overt bullying behaviours. While this may protect them from social rejection, it can also result in disciplinary consequences (Hessen and Kuncel, 2022[26]).
Changes in mathematics performance when students report the following incidents (all values shown are statistically significant)
Notes: All values are statistically significant. Values shown represent the estimated change in mathematics performance (PISA score points) associated with a one-unit increase in the reported item, based on separate bivariate regressions for girls and boys. Positive values indicate higher performance associated with reporting the behaviour. These associations should be interpreted with caution, as they do not imply causality and may reflect underlying group dynamics or response patterns.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Socio-economic differences also play a role, although the average exposure gap between advantaged and disadvantaged students in Sweden is small (-0.30 vs. -0.24) and not statistically significant. However, the association between bullying exposure and mathematics performance appears stronger among advantaged students. A one-unit increase in the bullying index is linked to a 13-point decline in their mathematics scores, compared to a 6-point decline among disadvantaged students. This pattern, observed in Sweden and mirrored in Norway, may reflect different expectations, coping mechanisms, or support systems across socio-economic groups. In contrast, Denmark presents the most pronounced socio-economic gap in bullying exposure in the Nordic region, with advantaged students reporting higher levels of bullying. Five out of the six items in the index reveal statistically significant differences in mathematics performance between socio-economic groups, with advantaged students tending to report greater declines across these items. The most pronounced drop (-104 points) is observed among advantaged students in response to the item “I gave money to someone at school because I was threatened,” marking the largest decline recorded for any subgroup or item in the index (Table 4.4).
Changes in mathematics performance when students report
Note: Statistically significant values are indicated in bold. Values shown represent the estimated change in mathematics performance (PISA score points) associated with a one-unit increase in the reported item, based on separate bivariate regressions for advantaged and disadvantaged students. Positive values indicate higher performance associated with reporting the behaviour. These associations should be interpreted with caution, as they do not imply causality and may reflect underlying group dynamics or response patterns. A socio-economically disadvantaged (or advantaged) student is a student in the bottom (or top) quarter of the PISA index of economic, social and cultural status (ESCS) in Sweden.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Immigrant background appears to have minimal association with overall exposure to bullying in Sweden. Non-immigrant students report a mean score of -0.29 compared to -0.25 for immigrant students, a difference that is not statistically significant. However, differences emerge in performance outcomes linked to specific bullying behaviours. When students report that others have spread nasty rumours about them, non-immigrant students show a large 32-point decline in mathematics performance, while the decline among immigrant students is 13 points. This 18-point gap between immigrant and non-immigrant students is the most significant disparity found among the items in the index, indicating that bullying experiences are often disproportional. When disaggregated by school location, the association between exposure to bullying and mathematics performance appears most pronounced in town and city schools, where scores drop by 10 points, compared to 4 points in rural areas. However, none of the individual items within the index show statistically significant differences between locations.
While these bivariate results illustrate clear patterns, multivariate analysis also shows that the negative association between exposure to bullying and mathematics performance in Sweden remains statistically significant (Annex Table 4.A.2). After accounting for student socio-economic status, a one-unit increase in the index of exposure to bullying is associated with a 9-point decline in mathematics performance. This effect is slightly smaller in magnitude but directionally consistent with the bivariate estimate presented earlier in this section (-10 points). The bullying index explains 1.2% of the variance in mathematics performance, which is above the OECD average of 1.0%. These findings reinforce the earlier observation that bullying remains an important correlate of student performance in Sweden, even after taking socio-economic differences into account, although the results do not imply causality.
Insights from international practice
Recognising that bullying is not uniformly experienced across student groups, some countries have strengthened their approaches to prevention and response through integrated well-being strategies. Sweden could draw inspiration from Ireland’s experience (Box 4.2), where anti-bullying measures are integrated into a whole-school well-being framework that helps schools tailor strategies for students most at risk.
Ireland’s school system places growing emphasis on student well-being as a foundation for learning and development. Over the past decade, there have been broad governmental efforts to embed anti-bullying measures within broader well-being frameworks at the school level.
A key milestone was the introduction of the Well-being Policy Statement and Framework for Practice 2018–2023, which positions well-being as a core element of the educational mission for all primary and post-primary schools. Under this framework, schools are required to implement a whole-school approach that addresses bullying through prevention, early detection, and consistent response mechanisms.
To translate this ambition into practice, the Department of Education provides schools with comprehensive guidance and practical tools, including mandatory annual reviews of anti-bullying policies, integration of well-being targets within School Self-Evaluation (SSE) cycles, and dedicated training for teachers and school leaders. Schools are encouraged to involve students and parents in shaping local anti-bullying strategies, with a strong focus on creating respectful peer interactions and supportive school cultures.
Evaluations by the Department of Education Inspectorate suggest that this whole-school approach helps schools sustain anti-bullying efforts over time, strengthen staff capacity to identify and address bullying behaviours, and ensure that well-being principles inform daily practice. Emerging evidence points to improvements in students’ sense of safety and connectedness where the framework is implemented robustly.
Ireland’s experience shows how an integrated, system-wide commitment to well-being can provide a strong foundation for tackling bullying effectively, helping schools create safer and more inclusive learning environments for all students.
Source: Government of Ireland Department of Education and Skills (2018[33]; 2025[34])
Observed patterns in exposure to bullying
Although the index of exposure to bullying was first introduced in PISA 2015, only six items are comparable over time due to the addition of three new items in 2022. For those six items, Sweden’s trends generally align with broader OECD patterns (Figure 4.4).
Percentage of students responding that they experience bullying a few times a month
Note: Based on student reports. Trends are shown only for the six comparable items that have been consistently measured since PISA2012.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Since PISA 2015, some improvements in the index are evident for Sweden. For instance, the proportion of students reporting being deliberately socially excluded dropped by 1.1 percentage points in 2022, which is a larger improvement than the OECD average. Similarly, the share of students reporting that others spread nasty rumours about them declined by 1.8 percentage points, slightly better than the OECD’s 1.5-point drop. However, incidents of verbal bullying became more frequent: the proportion of students reporting being made fun of increased by 2.1 percentage points, more than double the 0.8-point increase across the OECD. Other bullying behaviours have remained relatively stable over the past decade. The percentage of students who report being threatened by others declined slightly in Sweden (by 0.7 percentage points), in line with the OECD average decrease of 0.6 percentage points. Similarly, the share of students who report having their belongings taken or destroyed declined by 1.0 percentage point in Sweden, compared to a 0.8-point drop across the OECD. Notably, the proportion of students who report being physically pushed or hit did not change in Sweden, remaining at 5.4% between 2015 and 2022, while the OECD average decreased modestly by 0.3 percentage points.
Conclusion
PISA 2022 data indicate that while the overall reported bullying in Sweden is broadly similar to the OECD average, its associations with student well-being and academic performance are unevenly distributed across student groups. Girls, advantaged students, and non-immigrant students tend to show stronger negative performance patterns when higher levels of bullying are reported, while patterns for socio-economic status and immigrant background reveal item-specific disparities. Multivariate analysis indicates that these associations persist even after taking students’ and schools’ socio-economic profiles into account. Although the association is modest, it remains statistically significant and suggests that bullying may play a role in shaping student outcomes. Taken together, these findings show that reducing bullying is not only about lowering prevalence but also about addressing both the physical and emotional dimensions of school safety that can ensure inclusive and supportive learning environments for all students.
Index of feeling safe at school
The PISA 2022 index of feeling safe at school evaluates students’ perceptions of both physical and emotional safety, encompassing aspects such as the presence of bullying, the security of the school environment, and students’ sense of being free from harm. It draws on items such as feeling safe on the way to and from school, in classrooms, and in other areas such as hallways and restrooms. A safe school climate is not only widely recognised as important for student well-being, but also strongly associated with academic outcomes (Daily et al., 2020[24]; Thapa et al., 2013[3]; Catalano et al., 2004[12]). In Sweden, the average index value of student-reported safety is 0.08, slightly above the OECD average of 0.00, indicating that Swedish students perceive school as a relatively safe environment. Importantly, a one-unit increase in this index is associated with a 13-point increase in mathematics performance (highlighting the positive statistical association between students' sense of safety and their learning outcomes). However, notable variation exists when comparing Sweden to other Nordic countries. Finland (0.38) and Norway (0.30) report significantly higher mean indices, suggesting that students in these systems feel safer than their peers in Sweden. Denmark did not report data for this index, limiting regional comparability. A closer examination of the data reveals that students’ experiences of safety are not uniform but differ in meaningful ways across subgroups.
Uneven experiences: How safety perceptions vary across students
While overall results are positive, disaggregated data reveal persistent disparities. Gender differences in perceived safety are particularly marked: Swedish girls report a mean index of -0.12, whereas boys report a significantly higher 0.27, resulting in a gender gap of 0.39. This pattern is not unique to Sweden. Norway also shows a comparable gender difference of 0.44. Although there is a gender gap in perception where boys report feeling safer, the association between feeling safe and mathematics performance is similar for both boys and girls in Sweden. Girls score 13 points higher and boys 12 points higher when they report feeling safe.
Socio-economic disparities are also apparent. Advantaged students report a safety index of 0.22, compared to -0.05 for disadvantaged students, showing a significant score gap of 0.27 in perceptions of safety. Although this difference in perception is consistent with patterns seen across the OECD, its association with academic performance is more nuanced in Sweden. While the overall relationship between feelings of safety and academic outcomes is similar for both groups (showing a 9-point increase for disadvantaged students and a 10-point increase for advantaged students in mathematics), one item shows a statistically significant performance gap. When students report feeling safe in other areas at school such as in hallways or restrooms, advantaged students score 39 points higher compared to 13 points for their disadvantaged peers, resulting in a 27-point difference.
By contrast, no statistically significant differences in the association between the index of school safety and mathematics performance are observed across student backgrounds or school locations, despite perceptual gaps. Non-immigrant students in Sweden report a higher average sense of safety (0.12) than their immigrant peers (-0.03), with a notable difference of 0.15. This pattern, in which non-immigrant students tend to feel safer at school than immigrant students, is also observed in many other countries. In terms of learning outcomes, higher feelings of safety are associated with an increase in mathematics scores by 12 points for non-immigrant students and 8 points for immigrant students. Although this difference is not statistically significant, it indicates unequal experiences. It is important to note that these findings refer specifically to the relationship between perceived safety and mathematics performance, and do not reflect overall performance differences across student groups. Although overall safety levels do not differ markedly by school location, the association between feeling safe and mathematics performance does vary. Students in rural schools report the largest performance difference linked to stronger feelings of safety (15 points), followed by those in city schools (14 points) and town schools (12 points).
The results of the multivariate analysis, which account for students’ and schools’ socio-economic profiles, indicate that the positive association between students’ sense of safety at school and mathematics performance in Sweden remains statistically significant (Annex Table 4.A.3). After accounting for socio-economic factors, a one-unit increase in the index of feeling safe at school is associated with a 9-point increase in mathematics performance, slightly lower than the 13-point association found in the bivariate analysis. Students’ reported sense of safety explains 2.0% of the variation in mathematics performance in Sweden, marginally above the OECD average of 1.8%. This pattern suggests that students who feel safe at school also tend to perform better academically, even after accounting for socio-economic differences. The reduction in the strength of the association after adjusting for these factors indicates that part of the relationship may reflect broader socio-economic influences. However, the continued statistical significance highlights the potential relevance of school safety as a contextual factor linked to student learning outcomes. As with the other indices, these findings describe associations and should not be interpreted as evidence of causality.
Insights from international practice
Ensuring that students feel safe both inside the school and during their commute requires targeted measures that extend beyond general anti-bullying or well-being strategies. Portugal’s Escola Segura programme (Box 4.3) demonstrates how visible security measures, improved physical infrastructure, and community policing can directly strengthen students’ sense of safety in classrooms, hallways, and on their way to and from school.
Portugal operates a nationally coordinated framework for strengthening physical safety in and around schools through its long-standing Escola Segura (Safe School) programme. Established in the late 1990s, the programme is jointly managed by the Ministry of Internal Administration and the Ministry of Education, with operational implementation led by the Public Security Police (PSP) and the National Republican Guard (GNR). Its primary aim is to ensure that students feel safe not only within classrooms and school premises but also during their daily commutes.
A core feature of Escola Segura is the deployment of specialised officers trained to work with young people and school staff. These officers conduct regular patrols around school gates, bus stops, and common commuting routes during arrival and dismissal times to deter incidents such as bullying, harassment, vandalism, and other forms of violence. Within school grounds, officers collaborate with school leaders to identify risk areas such as poorly lit hallways, isolated corridors, and playgrounds, and to implement targeted measures to increase adult supervision and environmental security.
In addition to visible policing, the programme supports schools in carrying out safety audits of physical infrastructure, providing guidance on improvements such as fencing, lighting, and secure access points. Preventive educational activities are also integrated into the school curriculum, including classroom sessions on topics like violence prevention, conflict resolution, and safe behaviour in public spaces and online. Over the years, Escola Segura has become a key pillar in Portugal’s broader strategy for promoting student well-being and ensuring that all students can benefit from safe environments.
Source: Polícia Segurança Pública (2024[35]), Ministry of Education (2024[36]), and Eurydice (2025[37]).
Observed patterns in student perceptions of safety in Sweden
Item-level results offer additional insight into where Swedish students feel safer compared to their OECD peers (Figure 4.5). Compared to the OECD average, Swedish students report stronger feelings of safety during their commute to and from school, but slightly weaker perceptions of safety within the school premises. A higher proportion of students in Sweden feel safe on their way to school (93.3% Sweden vs. 91.8% OECD average) and on their way home (93.5% Sweden vs. 91.4% OECD). In contrast, slightly fewer Swedish students report feeling safe in classrooms (91.6% Sweden compared to 93.1% OECD) and in other areas of the school such as hallways and restrooms (88.7% Sweden vs. 89.9% OECD).
Percentage of students in Sweden and the OECD who agree/strongly agree with the following:
Note: Based on student reports. While visual differences may appear large due to the scale of the Y-axis, the actual differences between Sweden and the OECD average are relatively small (ranging between 1–2 percentage points). Index of feeling safe at school for Sweden = 0.08, OECD average = 0.00.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
As this index was introduced in PISA 2022, comparisons with other PISA cycles are not yet possible. Nonetheless, even these small differences can offer insights into how students’ perceptions of safety relate to equity and learning outcomes in Sweden.
Conclusion
Overall, Swedish students report feeling slightly safer than the OECD average, although notable differences persist across gender, socio-economic status, and migrant background. Among Nordic countries, Sweden’s mean index value lies between those of Norway and Finland, with Finland reporting the highest levels of perceived safety. Bivariate analyses show that stronger feelings of safety are linked to higher mathematics scores for all student groups, with variations in the size of this association across demographics and school locations. Multivariate results confirm that this positive relationship remains statistically significant even when accounting for students’ and schools’ socio-economic profiles. These patterns highlight that a secure and trusting environment is an essential part of an orderly and supportive school climate, reinforcing the importance of addressing physical safety alongside classroom management and belonging.
Index of school safety risks
The index of school safety risks is constructed from five yes/no questions asking students whether they have experienced or witnessed specific safety-related incidents at school in the past four weeks, including vandalism, fights resulting in injury, gang presence, threats, and weapons on school premises. While all these incidents are concerning, threats and physical aggression stand out as particularly damaging to student well-being and academic engagement. Exposure to such behaviours (whether direct or indirect) can create a pervasive climate of fear that impairs students’ ability to concentrate, retain information, and participate fully in classroom activities (Steffgen, Recchia and Viechtbauer, 2013[13]; Wolke and Lereya, 2015[31]). Heightened stress responses triggered by such experiences have been linked to difficulties with memory, problem-solving, and academic performance. In some cases, fear of harm leads to increased absenteeism, compounding the negative effects on students’ educational outcomes (Wolke and Lereya, 2015[31]; Gubbels, van der Put and Assink, 2019[38]).
In Sweden, the item on vandalism within this index required a country-specific item parameter, indicating that students’ response patterns differed significantly from those in other countries (OECD, 2023[5]). Consequently, this item was excluded from the analysis, and Sweden’s mean score of -0.20 on the index of school safety risks was calculated based on the remaining four items. This places Sweden slightly below the OECD average of 0.01. A negative score on the index suggests that Swedish students perceive their school environments as slightly safer than their peers elsewhere. While the gap is relatively modest, the findings highlight areas where targeted policy action could strengthen students’ sense of safety and well-being in school.
Among Nordic countries, Finland reports a less favourable perception of school safety risks, with a higher index score of 0.25. In contrast, Norway scores lower than Sweden at -0.43, indicating fewer perceived safety risks. Importantly, in Sweden, students’ perceptions of safety risks are negatively correlated with academic performance. A one-unit increase in perceived safety risks is associated with a 7-point decline in mathematics scores, underscoring the negative educational associations of an unsafe school environment (Figure 4.6).
Note: Statistically significant differences are shown in a blue tone. Each bar represents the coefficient from a bivariate regression of mathematics performance on the index, estimated separately for each subgroup (e.g. girls, boys, socio-economically advantaged/disadvantaged, etc.). No control variables are included in these models. Results reflect associations, not causal effects. A socio-economically disadvantaged (or advantaged) student is a student in the bottom (or top) quarter of the PISA index of economic, social and cultural status (ESCS) in Sweden. Rural areas or villages are defined as having fewer than 3000 inhabitants; town areas have between 3000 to 100 000 inhabitants; and cities have over 100 000 inhabitants.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Uneven experiences: How perceptions of safety risks vary across students
When disaggregated by student characteristics, notable disparities in perceived safety risks emerge. In Sweden, boys report higher exposure to threats than girls, reflecting a broader global trend in which boys are more likely to encounter or engage in overt forms of physical aggression and confrontations (OECD, 2020[2]; OECD, 2023[5]). This gender disparity may reflect broader socialisation processes or peer dynamics that normalise aggressive behaviours among boys. Consistent with this, the index of school safety risks shows that girls report fewer perceived risks, with an average score of -0.32 compared to -0.07 for boys. Because a more negative score indicates fewer perceived risks, these values suggest that girls feel significantly safer at school than boys. Lower perceptions of safety are strongly associated with mathematics performance: boys experience a larger decline in mathematics scores (-9 points) than girls (-6 points) in contexts where safety risks are perceived to be higher.
Negative patterns are further reflected in the item-level analysis within the index: students who report witnessing a fight that resulted in injury tend to have lower mathematics scores, with girls scoring 18 points lower on average and boys 34 points lower, resulting in a gender gap of -15 points. Notably, the psychological toll of school threats extends beyond direct victims; bystanders are also affected, with heightened stress and a broader climate of heightened safety risks that can undermine students’ cognitive functioning and engagement (Wolke and Lereya, 2015[31]; Steffgen, Recchia and Viechtbauer, 2013[13]).
When disaggregated by socio-economic status, disadvantaged students report fewer perceived safety risks, with an average index score of –0.23, compared to –0.17 among advantaged students. This 0.06-point difference suggests that advantaged students feel marginally less safe at school. Despite these perceptions of risk, disadvantaged students show a smaller decline in mathematics performance (–4 points) compared to advantaged students (–12 points) when safety concerns are present. Two items in the index show statistically significant differences in mathematics performance between these two groups (Table 4.5).
Changes in mathematics performance when students report witnessing the following incidents (all values shown are statistically significant)
Note: All values shown are statistically significant. Values shown represent the estimated change in mathematics performance (PISA score points) associated with a one-unit increase in the reported item, based on separate bivariate regressions for advantaged and disadvantaged students. Positive values indicate higher performance associated with reporting the behaviour. These associations should be interpreted with caution, as they do not imply causality and may reflect underlying group dynamics or response patterns. A socio-economically disadvantaged (or advantaged) student is a student in the bottom (or top) quarter of the PISA index of economic, social and cultural status (ESCS) in Sweden.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
A similar pattern does not appear with immigrant background. Both immigrant and non-immigrant students report the same average score on the school safety risk index (–0.21), suggesting no overall difference in perceived safety risks between the two groups. However, in terms of academic outcomes, non-immigrant students experience a slightly greater decline in mathematics scores (–7 points) than immigrant students (–4 points) in response to increased safety risks. At the item level, there are no statistically significant differences between immigrant and non-immigrant students in mathematics performance.
School location does not appear to play a role. Mathematics performance is 5 points lower for students in rural areas, compared to a 7-point difference for those in town settings, and a 9-point difference for students in city schools. However, none of these differences are statistically significant (either at the item level or in the overall index) suggesting that school location is not associated with academic performance in this context.
While these patterns shed light on visible group differences, it is also important to examine whether the observed associations persist after accounting for students’ and schools’ socio-economic characteristics. The results of the multivariate analysis, which takes into account the PISA index of economic, social and cultural status (ESCS) at both student and school levels, confirm that higher levels of perceived school safety risks in Sweden are statistically associated with lower mathematics performance (Annex Table 4.A.4). Specifically, after accounting for socio-economic factors, a one-unit increase in the school safety risk index is associated with a 6-point decrease in mathematics performance. This effect is only slightly smaller than the 7-point decrease observed in the bivariate analysis, suggesting that the relationship remains robust even after adjusting for socio-economic context. The share of variance in mathematics performance explained by the school safety risk index is modest in Sweden (0.8%), slightly below the OECD average (1.8%). These findings indicate that students who report higher levels of safety risks at school also tend to have lower academic outcomes, even after accounting for socio-economic background. While these results describe statistically significant associations, they should not be interpreted as evidence of causal relationships.
Insights from international practice
Many education systems have adopted whole-school approaches to address safety risks in a comprehensive way. Reducing risks related to physical fights, threats, and weapons on school grounds requires more than general safety awareness. Ontario, Canada (Box 4.4) demonstrates how a comprehensive, whole-school approach that combines clear policies, mandatory reporting, progressive discipline and community partnerships, can effectively prevent and address violent incidents and create safer, more supportive learning environments.
Ontario’s education system has adopted a whole-school approach to prevent and address violence, vandalism and weapon-related incidents in schools through its Safe and Accepting Schools initiative. In response to growing concerns about bullying, physical fights and the presence of gangs or weapons on school property, the Accepting Schools Act (2012) established clear requirements for prevention, incident reporting and intervention.
School boards are mandated to develop and implement comprehensive policies to create safe, inclusive and accepting learning environments. These policies include explicit procedures for reporting and responding to violent incidents such as fights, threats, and possession of weapons. To ensure consistent practice, all serious incidents are formally reported by staff, with progressive discipline guidelines that promote fair and constructive responses to student behaviour.
Ontario’s model also strengthens collaboration across the school community. Schools establish Safe School Teams that include students, staff and parents, providing peer leadership in promoting a positive school climate. In high-risk areas, many schools partner with local police through School Resource Officer (SRO) programmes to prevent gang activity and violence near school grounds.
Professional development is a key pillar of the approach, requiring regular training for staff on de-escalation strategies, inclusive practices and early intervention. Evaluations suggest that this integrated model has contributed to declines in suspension rates for violent behaviour and improved students’ perceptions of safety and belonging in school.
Source: Ontario Ministry of Education (2025[39])and Government of Ontario (2012[40]).
Observed patterns in student perceptions of school safety risks in Sweden
In terms of specific risks, 40.6% of Swedish students report that their school had been vandalised, compared to an OECD average of 19.8% (Figure 4.7). In Sweden, the English term “vandalism” is typically translated as skadegörelse, which broadly means “damage” or “destruction of property.”
While this translation conveys the general meaning of vandalism, its everyday interpretation in Swedish may differ in both scope and perceived severity. Skadegörelse is often associated with less serious incidents (such as graffiti, minor defacement or damage) rather than serious criminal acts of destruction. These linguistic and cultural difference in interpretation most likely contributed to the substantial discrepancy between Sweden and other countries, and ultimately led to the exclusion of this item from Sweden’s overall calculation of the index of school safety risks (OECD, 2024[41]).
Percentage of students reporting the following incidents in the past four weeks:
Note: Based on student reports. Index of safety risks for Sweden = -0.02, OECD average = 0.01.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
In Sweden, 18.8% of students report witnessing a fight on school property that resulted in someone getting hurt, slightly above the OECD average of 17%. Victims of physical violence often suffer both visible injuries and emotional trauma, including anxiety and depression, which undermine their ability to learn effectively (Gubbels, van der Put and Assink, 2019[38]; Vanderbilt and Augustyn, 2010[28]). These incidents not only disrupt the learning environment but also erode trust between students and staff, contributing to a hostile school climate (Steffgen, Recchia and Viechtbauer, 2013[13]; Turanovic and Siennick, 2022[42]; Way, Reddy and Rhodes, 2007[43]). A greater proportion of Swedish students also report hearing a peer threaten to harm another student (21.5%) compared to the OECD average (20.2%). However, fewer Swedish students report seeing gangs at school (10.6%) than the OECD average (12.1%), and the proportion of students who have seen weapons at school is also lower in Sweden (9.9%) compared to the OECD average of 10.5%.
Conclusion
The index of school safety risks provides valuable insights into how Swedish students perceive safety in their learning environments and how these perceptions may be linked to patterns in academic outcomes. While Sweden’s average score suggests only a modest deviation from the OECD norm, disaggregated findings highlight important differences by gender, socio-economic status and migrant background, with boys and disadvantaged students tending to perceive greater safety risks and showing more pronounced associations with lower performance. The exclusion of the vandalism item, due to cultural and linguistic interpretation, underscores the importance of contextualising international measures to reflect local realities. Importantly, multivariate analysis confirms that even after accounting for students’ and schools’ socio-economic profiles, higher reported safety risks remain statistically associated with lower mathematics performance. Taken together with related findings on school climate, these results highlight that perceptions of safety risks remain relevant for students’ learning experiences. While the overall association is modest, the findings underline the value of continuing efforts to ensure that all students feel secure and supported at school.
Index of sense of belonging
The index of sense of belonging captures students’ feelings of acceptance, inclusion and connectedness within their school environments. Based on student responses in the PISA 2022 questionnaire, this index reflects the extent to which students feel part of the school community, enjoy positive peer relationships, and perceive themselves as being accepted by others. The index is composed of positive items such as: “I feel like I belong at school,” “I make friends easily at school,” and “Other students seem to like me,” where agreement shows sense of belonging. There are also three negative indicators such as “I feel like an outsider (or left out of things),” “I feel awkward and out of place in my school,” and “I feel lonely at school” where higher levels of disagreement signal a stronger sense of belonging. A strong sense of belonging is widely recognised as an important factor to student well-being, engagement and academic success that can contribute to fostering positive learning environments, disciplinary climate and safety in schools (OECD, 2023[5]; Hoge, Smit and Hanson, 1990[44]; MacNeil, Prater and Busch, 2009[45]).
In Sweden, students report a slightly positive average score on this index (0.09), placing the country above the OECD average (-0.02) and close to its Nordic neighbours Denmark (0.11) and Finland (0.10). Sweden’s index, however, is well below Norway’s (0.23). This suggests that, on the whole, students in Sweden feel moderately connected to their school communities. Importantly, a stronger sense of belonging is associated with better academic outcomes. In Sweden, a one-unit increase in the index of student sense of belonging is associated with a higher score in mathematics performance by an average of 6 points (Figure 4.8). This positive association holds across all demographic groups, although the strength of the relationship with academic outcomes varies by gender, socio-economic status, migrant background, and school location.
Note: Statistically significant differences are shown in a blue tone. Each bar represents the coefficient from a bivariate regression of mathematics performance on the index, estimated separately for each subgroup (e.g. girls, boys, socio-economically advantaged/disadvantaged, etc.). No control variables are included in these models. Results reflect associations, not causal effects. A socio-economically disadvantaged (or advantaged) student is a student in the bottom (or top) quarter of the PISA index of economic, social and cultural status (ESCS) in Sweden. Rural areas or villages are defined as having fewer than 3000 inhabitants; town areas have between 3000 to 100 000 inhabitants; and cities have over 100 000 inhabitants.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Uneven experiences: How sense of belonging varies across students
Sweden exhibits one of the narrowest gender gaps in the Nordic region on this index; however, notable differences in experiences persist. Boys report a mean index score of 0.21, while girls report a negative score of -0.03, indicating lower levels of emotional connection and sense of belonging at school. Both girls and boys tend to achieve higher and comparable mathematics scores when they report a stronger sense of belonging, with an associated increase of 9 points for boys and 7 points for girls.
Two items within the index reveal meaningful gender differences in performance (Table 4.6). In both cases, girls show a stronger association with positive academic outcomes when they feel more included, highlighting the importance of addressing emotional well-being and social integration for girls in Sweden.
Changes in mathematics performance when students to the following statements (all values shown are statistically significant)
Notes: All values are statistically significant. Values shown represent the estimated change in mathematics performance (PISA score points) associated with a one-unit increase in the reported item, based on separate bivariate regressions for girls and boys. Positive values indicate higher performance associated with reporting the behaviour. These associations should be interpreted with caution, as they do not imply causality and may reflect underlying group dynamics or response patterns.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Socio-economic status is also an important factor in students’ sense of belonging. Advantaged students in Sweden report a mean index of 0.22, while disadvantaged students report an average index of -0.05. This gap is the second largest in the Nordic region, following Norway. Despite this disparity, both advantaged and disadvantaged students show positive associations between academic outcomes and a stronger sense of belonging. Mathematics scores increase by 3 points for advantaged students and 4 points for disadvantaged students. There are no items in the index that show a statistically significant difference between socio-economic groups, suggesting that the overall experience of belonging may be more consistent than the overall index suggests. Nevertheless, the overall index gap signals deeper structural inequities in how school communities can foster inclusion.
Sense of belonging also differs by immigrant background. Non-immigrant students report a higher average sense of belonging (0.14) than their immigrant peers (-0.04). The association between sense of belonging and mathematics performance is also stronger for non-immigrant students: a one-unit increase in the index corresponds to a 7-point increase in mathematics scores for non-immigrant students, compared to a 4-point increase for immigrant students. Among the six items that make up the index, one shows a statistically significant performance difference between these groups: when students agree or strongly agree that they feel they belong at school, non-immigrant students score 27 points higher, while immigrant students score 11 points higher. This 16-point gap indicates that, even when immigrant students report a sense of belonging, the associated academic benefits may not be as pronounced.
While average levels of belonging are similar across school locations, the academic benefits of feeling included are particularly strong in rural settings. A one-unit increase in belonging correlates with a 15-point increase in mathematics performance among rural students, compared to 8 points in town areas and 6 points in city schools. One finding stands out: among all subgroups, students in town settings who disagree with the statement that they “feel like an outsider” or are “left out of things at school” are associated with a difference in mathematics performance of 33 points, compared to their peers who agree.
The results of the multivariate analysis, which takes into account the students’ and schools’ socio-economic profiles, show that a stronger sense of belonging at school remains statistically associated with higher mathematics performance in Sweden (Annex Table 4.A.5). After accounting for socio-economic factors, a one-unit increase in the index of sense of belonging is associated with a 5-point increase in mathematics performance. This association is somewhat smaller than in the bivariate analysis (8 points), but it remains statistically significant. The share of variance in mathematics performance explained by the sense of belonging index is 0.9%, slightly above the OECD average of 0.7%. These findings suggest that, after adjusting for socio-economic status, students who report feeling more connected and included in their school community tend to perform better academically in mathematics.
Insights from international practice
Fostering a stronger sense of belonging requires more than recognising its link to academic outcomes. Sweden already performs above the OECD average on this index, but trends suggest that targeted actions are needed to reverse declines and reduce persistent gaps for girls, disadvantaged students and those with an immigrant background. Sweden can draw on the experience of Poland (Box 4.5), which has taken a whole-school approach to strengthening students’ sense of belonging and emotional well-being. Through initiatives like the School of Positive Thinking programme, Poland shows how equipping schools with structured tools and support for social-emotional learning can help reduce loneliness, promote positive peer relations and build more inclusive school communities.
To address rising concerns about student loneliness and social disconnection, Poland has developed targeted initiatives that promote positive peer relationships and a stronger sense of belonging within schools. One such example is the Szkoła Myślenia Pozytywnego (School of Positive Thinking) programme, implemented by the Empowering Children Foundation (Fundacja Dajemy Dzieciom Siłę) with support from the Ministry of National Education.
This nationwide programme adopts a whole-school approach to fostering inclusion and emotional well-being. Schools participating in the initiative receive structured guidance, resources and training to help educators and school leaders recognise and respond to students’ social-emotional needs. Key components include professional development for teachers to build skills in supporting student mental health, classroom activities that strengthen trust and cooperation among peers, and mechanisms for engaging students in decision making to increase their sense of agency and connectedness.
Schools have dedicated well-being teams composed of teachers, counsellors and students to monitor school climate and to design actions that respond to identified needs. Evaluations of the programme have indicated positive outcomes, including improved classroom climate, lower reports of student loneliness, and higher levels of student engagement and participation. This initiative illustrates Poland’s broader commitment to integrating mental health promotion and social-emotional learning into education policy frameworks to support students’ well-being and academic success.
Source: Ministry of National Education (2023[46]), Fundacja IEP (2023[47]) and Fundacja Dajemy Dzieciom Siłę (2021[48]).
Observed patterns in students’ sense of belonging
While Sweden’s overall index score for sense of belonging is modestly positive compared to the OECD average, when disaggregated at the item level, the data point to a consistent downward trend. Direct comparison of the full index across PISA 2012 and PISA 2022 is limited due to changes in its composition, but six items are comparable as they have been consistently measured over the past decade (Figure 4.9).
Percentage of students responding agree/strongly agree
Note: Based on student reports. Trends are shown only for the six comparable items that have been consistently measured since PISA2012.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Trends seen in this index suggest a more fragmented and ambivalent experience of belonging at school. Between 2012 and 2022, more students in Sweden report feelings of loneliness at school, with the share agreeing or strongly agreeing with this sentiment rising by 5.5 percentage points (from 9.5% to 15.0%). While concerning, this increase is slightly below the OECD average, where reports of loneliness grew by 7.4 percentage points over the same period. In contrast, students in both Sweden and the OECD report identical increases in feelings of awkwardness or being out of place by 8 percentage points and in feeling like an outsider (5 percentage points), suggesting a broader international trend of growing social discomfort among students.
At the same time, indicators of peer connectedness and social affirmation also show downward trends. In Sweden, the share of students who agreed that they feel like they belong at school falls to 70.0%, slightly below the OECD average of 70.2%. However, the drop in Sweden (from 78.6%) is steeper than the OECD average decline, which went from 81.3% to 74.6%. The proportion of students who report making friends easily drops by 7 percentage points in Sweden compared to a 10-point decrease across the OECD. Similarly, agreement with the statement “other students seem to like me” falls by 7 percentage points in both Sweden and the OECD.
Conclusion
Strengthening students’ sense of belonging can play a key role in reinforcing positive disciplinary climate and safety, helping students feel connected, respected, and ready to learn. The PISA 2022 data suggest that while Swedish students report a modestly positive average sense of belonging, which is slightly above the OECD average, underlying trends point to a more complex picture. Item-level responses show general declines over the past decade, with more students reporting feelings of loneliness, exclusion, and social discomfort. However, these patterns are not unique to Sweden; rather, they reflect a broader international trend observed across many OECD countries, suggesting that growing social disconnection among students is a shared challenge. Additionally, in Sweden, sense of belonging is not experienced uniformly across all student groups. Boys, socio-economically advantaged students, and non-immigrant students report higher levels of belonging and tend to show stronger associations with academic performance when belonging increases. In contrast, girls, disadvantaged students, and immigrant students report lower levels of belonging and exhibit weaker associations with academic outcomes. These positive associations remain significant even after accounting for socio-economic factors and reflect broader structural differences. Targeted efforts are therefore needed to address the deeper emotional, cultural and relational barriers that may be hindering full inclusion for all students.
Index of quality of student-teacher relationships
The index of quality of student-teacher relationships reflects the extent to which students feel supported, respected, and understood by their teachers. It is based on eight items that capture both positive and negative aspects of these relationships, offering a comprehensive view of students’ perceptions. Positive items include whether teachers show respect, express genuine interest in students' well-being, demonstrate concern when students are upset, and would be excited to see students return in the future. Negative indicators capture feelings of intimidation or perceptions that teachers are unfriendly or mean. Together, these items provide a holistic view of the quality of student-teacher relationships.
PISA 2022 data show a strong and positive relationship between student performance and supportive teachers across most countries and economies (OECD, 2023[5]). In consultations conducted by the OECD with stakeholders in Sweden, several groups observed that schools with high quality teaching and teachers who are actively engaged with students tend to experience fewer behavioural issues and nurture more positive learning climates. Importantly, the benefits of positive student-teacher relationships extend beyond academic achievement. Students who perceive higher levels of teacher support are more likely to show greater engagement in learning, including increased proactiveness in mathematics, stronger critical thinking skills, better self-regulation, and a deeper enjoyment of learning (Yu and Singh, 2016[49]).
Sweden’s mean index for student-teacher relationships is 0.19, slightly above the OECD average of 0.00, suggesting that Swedish students generally perceive their relationships with teachers more positively than the typical student across OECD countries. Importantly, improvements in the quality of student-teacher relationships are positively associated with academic achievement: in Sweden, a one-unit increase on the index of quality of student-teacher relationships correlates with a 17-point increase in mathematics performance, highlighting the potential academic benefits of strong, supportive teacher-student connections (Figure 4.10).
Note: All values show statistically significant differences. Each bar represents the coefficient from a bivariate regression of mathematics performance on the index, estimated separately for each subgroup (e.g. girls, boys, socio-economically advantaged/disadvantaged, etc.). No control variables are included in these models. Results reflect associations, not causal effects. A socio-economically disadvantaged (or advantaged) student is a student in the bottom (or top) quarter of the PISA index of economic, social and cultural status (ESCS) in Sweden. Rural area or villages are defined as having fewer than 3000 inhabitants; town areas have between 3000 to 100 000 inhabitants; and cities have over 100 000 inhabitants.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Comparing Sweden to other Nordic countries shows interesting variations. In Finland, the index is slightly negative at -0.10, suggesting Finnish students, on average, report somewhat less positive student-teacher relationships compared to their Swedish peers. In contrast, Norway’s index is positive at 0.10, but still lower than Sweden’s. Denmark stands out with a notably high index score of 0.36, indicating particularly strong student-teacher relationships. However, as Denmark did not meet PISA’s sampling requirements, this result should be interpreted with caution.
Uneven experiences: How quality of student-teacher relationships vary across students
When disaggregated by gender, the quality of student-teacher relationships in Sweden appears largely equitable: girls score 16 points higher and boys 18 points higher in mathematics for every one-unit increase in the index. No statistically significant difference is observed between girls and boys, suggesting that both groups show similar associations between strong teacher-student relationships and academic performance. However, at the item level, students who report feeling intimidated by their teachers show notable gender differences in academic outcomes: boys score 49 points lower on average, compared to a 24-point difference for girls, resulting in a gender gap of 25 points.
When looking at socio-economic status, larger disparities emerge. Advantaged students in Sweden show a 17-point difference in mathematics performance associated with stronger student-teacher relationships, compared to a 9-point difference for disadvantaged students. This is a statistically significant gap of 8 points between advantaged and disadvantaged students. This performance gap widens further for the item asking whether teachers would be excited to see students return in the future: advantaged students score 39 points higher, while disadvantaged students score 16 points higher, resulting in a statistically significant difference of 23 points. These findings suggest that advantaged students may be better positioned to benefit academically from positive teacher interactions, or that the quality and nature of their relationships with teachers differ in ways not fully captured by the index.
A similar trend is observed when looking at immigrant background. Among non-immigrant students, a one-unit increase in the index of student-teacher relationship quality is associated with an 18-point difference in mathematics, compared to a 10-point difference for immigrant students. This 7-point gap between the two groups is statistically significant. This difference highlights potential challenges immigrant students may face in building equally strong relationships with teachers, which are often linked to factors such as language barriers, cultural differences, or experiences of bias (Archambault et al., 2024[50]). Across the index, there are two items that show significant differences in mathematics performance between immigrant and non-immigrant students (Table 4.7).
Changes in mathematics performance when students report
Notes: Statistically significant values are indicated in bold. Values shown represent the estimated change in mathematics performance (PISA score points) associated with a one-unit increase in the reported item, based on separate bivariate regressions for immigrant and non-immigrant students. Positive values indicate higher performance associated with reporting the behaviour. These associations should be interpreted with caution, as they do not imply causality and may reflect underlying group dynamics or response patterns.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
School location also seems to shape the relationship between student-teacher interactions and academic performance. Students in rural schools show the largest difference in mathematics scores, with a 26-point increase associated with better student-teacher relationships, compared to 17 points in town schools and 13 points in city schools. The 12-point gap between rural and city schools is statistically significant, suggesting that smaller, more close-knit rural communities may foster stronger, more supportive educational relationships. At the item level, students in rural settings who feel their teachers are more engaged in individual conversations score 54 points higher, compared to 27 points in city settings. This suggests that stronger teacher-student interactions may be particularly associated with higher performance in rural schools.
Multivariate analysis further confirms the positive association between the quality of student-teacher relationships and mathematics performance in Sweden, even after accounting for students’ and schools’ socio-economic profiles (Annex Table 4.A.6). After accounting for socio-economic status, a one-unit increase in the index of quality of student-teacher relationships is associated with a 14-point increase in mathematics performance, which is somewhat lower than the 17-point association observed in the bivariate analysis. The proportion of variance in mathematics performance explained by this index is 3.5%, noticeably higher than the OECD average 1.7%. This suggests that, even after adjusting for socio-economic differences, Swedish students who report stronger, more positive relationships with their teachers tend to perform better academically. The size and persistence of this association highlights the potential importance of nurturing positive teacher-student relationships, while recognising that this remains an observed association rather than evidence of a causal effect.
Insights from international practice
While Sweden’s performance on the index of student-teacher relationships is slightly above the OECD average, the data highlight areas where more deliberate action could help close persistent gaps. Strengthening relational practices in schools can play an important role in this regard. Estonia’s experience (Box 4.6) shows how dedicating time for regular one-on-one conversations between teachers and students, supported by multi-professional teams, can help teachers build trust, respond to individual needs and support student engagement more effectively. Drawing lessons from such practices could help Swedish schools ensure that all students feel known, supported and connected.
Estonia places strong emphasis on fostering high-quality student-teacher relationships as part of its broader commitment to equity and excellence in education. Building on a robust tradition of teacher professionalism, Estonia’s national education strategy prioritises student-centred pedagogy and the development of relational competencies as integral to both initial teacher education and continuous professional development.
To further strengthen student-teacher interactions, Estonia has expanded the use of individual mentoring and guidance practices in lower secondary education. Teachers are given dedicated time to hold regular one-on-one conversations with students to discuss learning progress, well-being and any emerging challenges. These structured meetings aim to build trust and ensure that all students feel known and supported, with particular attention to those at risk of disengagement.
In addition, Estonian schools benefit from well-developed multi-professional support teams, including school psychologists, counsellors, special education educators, who work alongside teachers to address students’ academic and socio-emotional needs in a coordinated manner. This integrated approach contributes to high levels of student well-being and engagement.
Source: OECD (2020[51]); Republic of Estonia Ministry of Education and Research (2024[52]).
Observed patterns in the quality of student-teacher relationships in Sweden
Data from PISA reveal that Swedish students consistently report more positive relationships with their teachers than students in other OECD countries.
Across every positive indicator of student-teacher relationships, Sweden outperforms the OECD average, suggesting a school culture marked by respect, friendliness, and genuine care. An overwhelming 91.1% of Swedish students say their teachers are friendly, compared to 87.2% across the OECD. Similarly, nearly 9 in 10 students (89.3%) feel that their teachers show them respect, which is also above the OECD average of 85.9%. Swedish students are also more likely to report that teachers are engaged in conversations with them (73.4% vs. 67.4%) and genuinely interested in their well-being (78.8% vs. 74.9%). The most striking difference appears in the sense of long-term connection and care: 81.3% of Swedish students believe their teachers would be excited to see them return in the future, compared to just 69.7% in the OECD, resulting in a gap of 11.6 percentage points. This item captures more than momentary kindness; it reflects a lasting emotional bond and a school climate that values relationships beyond the classroom. Importantly, Swedish students also report fewer negative experiences with their teachers. Only 7.1% say they feel intimidated by teachers, which is less than half the OECD average of 16%. Similarly, 7.1% perceive their teachers as mean, compared to 10.4% across the OECD. These results suggest that Sweden has cultivated a school environment where students feel seen, respected, and supported by their teachers.
Conclusion
Overall, the findings highlight that strong, respectful student-teacher relationships are an important pillar of a positive disciplinary climate, reinforcing trust, reducing disruptions and helping all students engage more fully in learning. The index of quality of student-teacher relationships stands out as a relative strength for Sweden, both in international comparison and in its positive association with student outcomes. Swedish students consistently report more respectful, friendly, and supportive interactions with teachers than their OECD peers, with fewer reports of negative experiences such as intimidation. Bivariate and multivariate analyses show that stronger student-teacher relationships are statistically associated with higher mathematics performance, even after accounting for socio-economic status, with the strength of this association notably above the OECD average. While this pattern holds for both girls and boys, notable differences persist across socio-economic status, immigrant background, and school location. Advantaged, non-immigrant, and rural students tend to report stronger links between supportive teacher relationships and academic performance, suggesting that some groups may benefit more from these interactions or experience them differently. These findings highlight the value of reinforcing relational practices that ensure all students but particularly disadvantaged, immigrant, and town students can build trusting, supportive connections with their teachers.
Index of teacher support in mathematics
The index of teacher support in mathematics measures the extent to which students perceive their mathematics teachers as supportive, attentive, and committed to their learning. It is based on students’ responses to items that reflect whether their teacher shows interest in each student’s progress, provides additional help when needed, assists with learning challenges, and continues instruction until students understand the material. These elements collectively capture not only the availability of instructional support but also the emotional and pedagogical responsiveness of the teacher. Research highlights that teacher support plays a crucial role in fostering a positive learning environment and is associated with stronger academic outcomes, increased student motivation, and enhanced self-efficacy in mathematics.
In Sweden, students report comparatively high levels of teacher support, with an average index score of 0.19, well above the OECD average of -0.03. This indicates that Swedish students, on average, feel more supported in mathematics classrooms than their peers in many other OECD countries. Research shows that higher levels of perceived teacher support are often associated with more positive learning climates, stronger student engagement, and greater persistence in learning particularly in demanding subjects such as mathematics (OECD, 2023[5]; Gase et al., 2017[53]; Lazarides and Buchholz, 2019[9]).
Compared to other Nordic countries, Sweden stands out in a notably more positive light. Both Finland and Denmark report index scores of 0.00, while Norway’s index sits at -0.17, indicating that students in these countries perceive less support from their mathematics teachers. In contrast, Sweden’s score of 0.19 underscores the relative strength of teacher-student relationships in mathematics classrooms, suggesting a more supportive and responsive instructional environment.
Positive perceptions of teacher support in mathematics are also reflected in student academic outcomes (Figure 4.11). In Sweden, a one-unit increase in the index of teacher support in mathematics is associated with a 10-point increase in mathematics performance, indicating the potential beneficial impact that supportive teaching can have on learning. This positive association is consistent across student groups.
Note: Statistically significant differences are shown in a blue tone. Each bar represents the coefficient from a bivariate regression of mathematics performance on the index, estimated separately for each subgroup (e.g. girls, boys, socio-economically advantaged/disadvantaged, etc.). No control variables are included in these models. Results reflect associations, not causal effects. A socio-economically disadvantaged (or advantaged) student is a student in the bottom (or top) quarter of the PISA index of economic, social and cultural status (ESCS) in Sweden. Rural areas or villages are defined as having fewer than 3000 inhabitants; town areas have between 3000 to 100 000 inhabitants; and cities have over 100 000 inhabitants.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Uneven experiences: How perceptions of teacher support vary across students
Gender differences on this index are negligible. Both girls and boys score 10 points higher in mathematics performance for each one-unit increase in perceived teacher support, with no statistically significant difference between them. This consistency suggests that teacher support in Sweden is perceived relatively equitably across gender. No significant gender differences are observed at the item level either.
Socio-economic status, however, reveals a more nuanced picture. Advantaged students show a stronger association between increased teacher support and mathematics performance, with a 12-point difference compared to a 5-point difference for disadvantaged students, a statistically significant gap of 8 points. While this suggests that teacher support is positively associated with academic outcomes, advantaged students appear to be better positioned to translate this support into academic improvement. Moreover, three items in the index show statistically significant disparities in performance increases between socio-economically advantaged and disadvantaged students, with advantaged students consistently showing stronger associations across these measures (Table 4.8). Notably, disadvantaged students seem to demonstrate the largest academic improvements when teachers actively help them with their learning, whereas advantaged students show the most progress when teachers provide extra help beyond regular instruction.
Changes in mathematics performance when students report
Note: Statistically significant values are indicated in bold. Values shown represent the estimated change in mathematics performance (PISA score points) associated with a one-unit increase in the reported item, based on separate bivariate regressions for advantaged and disadvantaged students. Positive values indicate higher performance associated with reporting the behaviour. These associations should be interpreted with caution, as they do not imply causality and may reflect underlying group dynamics or response patterns. A socio-economically disadvantaged (or advantaged) student is a student in the bottom (or top) quarter of the PISA index of economic, social and cultural status (ESCS) in Sweden.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
When comparing students with immigrant and non-immigrant backgrounds, both groups show positive associations with academic performance (7 points and 11 points, respectively) with no statistically significant difference between them. Furthermore, there are no significant differences between immigrant and non-immigrant students at the item level.
Finally, the association of teacher support on mathematics and student academic outcomes is consistently positive across geographical regions. Students in rural (11 points), town (11 points), and city (10 points) schools all show higher scores in mathematics, with no statistically significant differences across school locations. This points to a fairly even distribution of supportive teaching practices throughout the country, regardless of school setting. Furthermore, there are no significant differences in mathematics performance among students from different school locations at the item level.
The multivariate analysis, which takes into account the students’ and schools’ socio-economic profiles, indicates that teacher support in mathematics remains positively associated with mathematics performance in Sweden (Annex Table 4.A.7). After adjusting for socio-economic status, a one-unit increase in the index of teacher support is associated with a 10-point increase in mathematics performance. This is consistent with the effect size observed in the bivariate analysis prior to accounting for socio-economic factors, suggesting that the association between perceived teacher support and performance in Sweden remains relatively robust even when accounting for students’ socio-economic profiles. The explained variance in mathematics performance linked to teacher support is 1.4%, which is higher than the OECD average of 0.7%. This means that, in relative terms, the index of teacher support explains a larger share of performance variation in Sweden compared to many other countries. The persistence of this association after accounting for socio-economic differences suggests that students who perceive higher levels of teacher support in mathematics tend to have better academic outcomes, although the results should be interpreted as correlational rather than causal.
Insights from international practice
Maintaining high levels of perceived teacher support requires deliberate and sustained investment in teacher capacity and professional culture. While Sweden outperforms the OECD average on this index, modest declines over time and persistent socio-economic gaps highlight the need for continued attention to sustaining supportive instructional practices for all students. In this regard, Sweden can draw on the experience of Singapore (Box 4.7), which has embedded strong teacher support into its system through their Teacher Growth Model. By framing teacher development as a career-long process and fostering collaborative professional cultures, Singapore demonstrates how structured investments in teacher capacity and peer learning can help ensure that students consistently perceive their teachers as supportive, attentive, and committed to their progress.
Singapore’s education system places strong emphasis on building teacher capacity and fostering a professional culture of support for student learning. Central to this approach is the Teacher Growth Model, a comprehensive framework introduced by the Ministry of Education (MOE) to guide professional development throughout a teacher’s career. The model emphasises student-centric, values-driven teaching by framing teacher development as an ongoing, career-long process focussed on three core values: learner-centredness, teacher identity, and service to the profession and community, encouraging educators to adopt practices that are responsive to students’ academic and socio-emotional needs.
Teachers receive systematic training through the National Institute of Education (NIE) and benefit from structured mentoring and peer collaboration once in schools. Schools are given autonomy to customise Professional Learning Communities (PLCs) where teachers work collaboratively to improve instructional practices based on student data and feedback. These structures ensure that teachers are equipped not only with subject expertise but also with skills to provide tailored support, including identifying learning gaps and adjusting pedagogy accordingly.
To ensure continued alignment with student outcomes, the Enhanced Performance Management System incorporates multiple indicators, including student engagement and feedback, in evaluating teacher performance. This accountability framework reinforces a culture where teachers are supported and expected to help every student achieve their full potential.
Singapore’s holistic and integrated approach to teacher development ensures that instructional support is not incidental, but a deliberate and sustained feature of classroom practice, contributing to consistently high levels of student performance and engagement.
These structures help to ensure that teachers show interest in students’ progress, provide timely help, persist in clarifying material until students understand, and address individual learning challenges. By embedding these expectations within training, appraisal, and school culture, Singapore aims to sustain a high level of perceived teacher support in classrooms, contributing to strong student motivation and academic outcomes.
Source: Singapore Ministry of Education (2025[54]); Singapore National Institute of Education (2025[55]).
Observed patterns in teacher support in mathematics
The data reveal that Sweden consistently outperforms the OECD average across all four items measuring perceived teacher support in mathematics, both in PISA 2012 and PISA 2022. This indicates that Swedish students, on average, feel more supported by their mathematics teachers than their peers in many other countries. However, when examining trends over time, a more nuanced picture emerges (Figure 4.12).
Percentage of students responding they receive the following teacher support “in most or every lesson”
Note: Based on student reports. Trends are shown only for the four comparable items that have been consistently measured since PISA2012.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
While Sweden has maintained relatively high levels of teacher support, two items show a decline between 2012 and 2022. The share of students who report that their teacher helps them with their learning falls from 84.5% in 2012 to 79.1% in 2022, by 5.5 percentage points, and those who say their teacher continues teaching until students understand drops from 74.9% in 2012 to 72.8% in 2022. These declines over time, although moderate, suggest a slight weakening in students’ perceptions of more sustained or in-depth instructional support.
In contrast, the OECD averages for all four items declines more sharply over the same period, pointing to a broader international trend of diminishing perceived teacher support. What is particularly noteworthy is that Sweden’s relative position compared to the OECD average improves in most areas. This is because perceptions of teacher support across the OECD declined more sharply than in Sweden. For example, the OECD average for students who felt their teacher continues teaching until they understand dropped by 4.4 percentage points (from 68.0% in 2012 to 63.6% in 2022) compared to a smaller decline of 2.1 points in Sweden. Similar downward shifts can be observed in the remaining items, indicating that students in many OECD countries feel less supported by their mathematics teachers than they did a decade ago. As a result, Swedish students now report comparatively stronger support from their mathematics teachers than they did ten years ago, not necessarily because student perceptions have improved domestically, but because they seem to have deteriorated more significantly in many other countries in the OECD.
Conclusion
Taken together, the results underscore how subject-specific teacher support contributes not only to higher achievement but also to a more supportive and orderly learning environment overall. PISA 2022 highlights Sweden’s comparatively strong performance in fostering supportive mathematics classrooms. Swedish students consistently report higher levels of teacher support than their OECD peers, and this support is positively associated with mathematics performance across gender, immigrant background, and school location. After accounting for socio-economic status, this association remains statistically significant, underscoring its consistency across contexts. While differences between girls and boys are negligible, notable socio-economic gaps persist. Advantaged students show stronger associations between perceived teacher support and performance than disadvantaged students, suggesting that some groups may be better positioned to benefit from supportive teaching practices. Although student perceptions of teacher support in mathematics have declined slightly since 2012, these declines have been more moderate than the sharper drops observed in many other OECD countries, allowing Sweden to maintain its relative lead in this area. These results underline the importance of continued investment in building teacher capacity and sustaining supportive, responsive instructional practices.
Student absenteeism: Skipping an entire day and some classes
The PISA 2022 data on student absenteeism provides insights into the prevalence and patterns of students missing school without a valid reason, drawing on students’ self-reported data regarding their attendance in the two weeks prior to the assessment. Absenteeism is a critical indicator of student engagement and school climate and is associated with a range of academic and social outcomes (OECD, 2023[5]; Sobba, 2018[56]). Regular attendance is fundamental to learning, while frequent absences are linked to lower academic achievement, disengagement, and increased risk of early school leaving. The PISA captures this dimension by asking students how many times they skipped a whole day of school or partial day of school (i.e. missing some classes). These behaviours not only disrupt individual learning trajectories but may also reflect deeper issues related to student motivation, school connectedness, and broader systemic factors (Germain et al., 2024[14]). Students who frequently skip classes or days of school are also more likely to drop out entirely, limiting their future educational and career prospects (Gubbels, van der Put and Assink, 2019[38]). Understanding the patterns of absenteeism is therefore essential for informing policies aimed at fostering inclusive, supportive, and effective learning environments.
In Sweden, a relatively low percentage of students report skipping an entire school day, with 9.5% of students indicating they had skipped at least one day during the two full weeks prior to the PISA test. This figure is notably lower than the OECD average of 19.8%, which suggests that Swedish students are less likely to skip school compared to the average across OECD countries. Students were more likely to report skipping some classes in the last two full weeks of school (19% of all students), although still lower than the OECD average of 22.5%. Among the Nordic countries, Sweden's percentage of full-day absenteeism is in the lower range. For instance, Denmark reports 20.4%, and Finland shows 14.2%, both of which are higher than Sweden's. In contrast, Norway reports a similar figure of 13.9%. In terms of students reporting skipping some classes, Sweden sits between Denmark and Finland. Denmark has a higher partial absenteeism rate of 24.9%, while Finland shows a slightly lower rate of 18.3%.
The data on mathematics performance penalties associated with absenteeism reveals important disparities across gender, socio-economic status, migrant background, and school location. The relationship between absenteeism and academic performance is also well-documented in the literature, with consistent findings that students who miss more than ten days of school annually perform significantly worse on standardised tests across all subjects (OECD, 2024[57]). Frequent absenteeism is also associated with weaker social bonds at school and an increased risk of dropping out altogether (Gubbels, van der Put and Assink, 2019[38]). PISA 2022 data reinforce these findings, showing that both skipping a whole day of school and skipping some classes are associated with lower mathematics performance (Figure 4.13).
In the last two full weeks of school, I skipped a whole day, or I skipped some classes, at least once
Note: All values are statistically significant. Each bar represents the coefficient from a bivariate regression of mathematics performance on the index, estimated separately for each subgroup (e.g. girls, boys, socio-economically advantaged/disadvantaged, etc.). No control variables are included in these models. Results reflect associations, not causal effects. A socio-economically disadvantaged (or advantaged) student is a student in the bottom (or top) quarter of the PISA index of economic, social and cultural status (ESCS) in Sweden. Rural areas or villages are defined as having fewer than 3000 inhabitants; town areas have between 3000 to 100 000 inhabitants; and cities have over 100 000 inhabitants.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Uneven experiences: How absenteeism varies across students
When disaggregating by gender, girls in Sweden have a slightly higher tendency to skip school (10.8%) than boys (8.3%), a trend seen across various countries. Swedish girls (21.2%) also report a higher rate of partial school skipping than boys (16.7%), resulting in a gender gap of 4.5 percentage points. This trend of girls having higher absenteeism rates than boys is common across many countries, including in Norway and Denmark, where boys also show lower absenteeism rates. However, gender differences in the data show that boys tend to have a stronger negative association with absenteeism than girls. Boys who report skipping a whole day of school score, on average, 63 points lower in mathematics, compared to a 46-point difference seen among girls, resulting in a 17-point gender gap. A similar, although smaller, pattern is seen for skipping classes, where boys score 53 points lower and girls 43 points lower, yielding a 10-point gap. This suggests that absenteeism may exacerbate existing gender differences in academic resilience or engagement.
Socio-economic status also plays a significant role in school absenteeism, with disadvantaged students in Sweden showing a higher rate of absenteeism (13.3%) compared to their more advantaged counterparts (6.1%). This mirrors broader OECD patterns, where disadvantaged students are more likely to skip school partially or for an entire day. When it comes to the relationship between absenteeism and mathematics performance, advantaged students score 67 points lower when they skip a whole day and 48 points lower when they skip some classes, compared to 26 and 32 points respectively among disadvantaged students. This 42-point gap associated with skipping an entire day is noteworthy and statistically significant.
In Sweden, migrant background has a notable influence on absenteeism patterns. Immigrant students are more likely to skip a whole day of school, with 13.8% reporting this behaviour compared to just 8.0% of non-immigrant students. However, the relationship between absenteeism and mathematics performance appears more pronounced among non-immigrant students: those who skip a full day score 50 points lower in mathematics, compared to a 30-point drop among their immigrant peers. Partial absenteeism follows a similar trend. A significantly higher proportion of immigrant students (25.9%) report skipping some classes, compared to 16.6% of non-immigrant students, creating a gap of 9.3 percentage points. This disparity may reflect the unique challenges immigrant students face in integration and maintaining engagement with school. Yet again, non-immigrant students see a larger negative association between skipping some classes and mathematics performance, with a drop in 41 points compared to a 37-point difference for immigrant students. The 20-point difference when students report skipping a whole day may point to differences in learning environments, support systems, or levels of academic resilience. These findings underscore that absenteeism does not affect all students uniformly and highlight the need for more targeted and equitable support strategies.
Finally, some variation is observed by school location. Students in city schools show the largest negative association between skipping a whole day of school and mathematics performance (-63 points), compared to -51 points in town areas and -43 points in rural settings. However, the differences narrow when considering partial absenteeism, with only a 3-point gap between city and rural students. The greater decline associated with full-day absences in cities may be linked to more academically demanding environments or more structured learning time being lost.
Multivariate analysis results provide further insight into the association between school absenteeism and student performance in mathematics in Sweden. Both forms of absenteeism examined, skipping at least one whole school day and skipping at least some classes, are negatively associated with mathematics performance even after accounting for students’ and schools’ socio-economic profile. For students who reported skipping at least one whole school day in the two weeks prior to the PISA 2022 assessment, the analysis shows a 42-point lower mathematics performance in Sweden, after accounting for socio-economic factors. This is a statistically significant decline and larger than the OECD average of 30 points (Annex Table 4.A.8a). For skipping at least some classes, the corresponding performance decline in Sweden is 35 points, again above the OECD average of 28 points (Annex Table 4.A.8b). These results confirm that the negative associations between absenteeism and mathematics performance observed in the earlier bivariate analyses remain robust after accounting for socio-economic profiles.
The strength of these associations suggests that school absenteeism may be an important factor linked to lower mathematics outcomes in Sweden. While these findings do not imply causality, they highlight that students who are absent from school, whether for entire days or for specific classes, tend to have lower mathematics scores, even after adjusting for their socio-economic context. These patterns reinforce the importance of ongoing efforts to address absenteeism as part of broader strategies to improve student engagement and learning outcomes.
Insights from international practice
While Sweden’s absenteeism rates are close to the OECD average, persistent absenteeism patterns observed in Sweden suggest that more targeted, coordinated efforts are needed. Portugal’s Programa de Promoção do Sucesso Escolar (Programme for the Promotion of School Success) offers a useful example of how schools can tackle the root causes of truancy by identifying students at risk early on, providing individualised support, and building stronger, more trusting relationships between teachers and students to sustain regular attendance (Box 4.8).
Recognising the persistent challenges of student absenteeism and early school leaving, Portugal launched the Programme for the Promotion of School Success (Programa de Promoção do Sucesso Escolar, PPSE) in 2016 as part of a broader strategy to improve student outcomes and equity. This initiative provides additional resources and guidance for schools and municipalities to develop context-specific action plans targeting low attendance, low achievement and disengagement.
Under the PPSE, schools analyse local data on student attendance, lateness, performance and risk factors to identify students in need of targeted support. Schools then design and implement flexible measures, which can include individual tutoring, personalised learning plans, strengthened monitoring of attendance and punctuality, and closer cooperation with families and local social services. By giving schools autonomy to tailor strategies to their community’s needs, the programme encourages schools to address both whole-day absenteeism and persistent lateness as interconnected elements of student engagement and school climate.
Municipalities and school clusters are supported through regional co-ordination, sharing of good practices and professional development opportunities for teachers and school leaders. This collaborative approach has coincided with a steady decline in Portugal’s early school leaving rate since 2016 and has helped strengthen schools’ capacity to support student motivation and regular attendance, particularly for students at greater risk of disengagement. These efforts have also been reflected in Portugal’s steady improvement in PISA outcomes.
Source: Ministério da Educação (2017[58]); Direção-Geral da Educação (2019[59]); Programa Nacional de Promoção de Sucesso Escolar (2023[60]).
Observed patterns in student absenteeism
The data on student absenteeism from PISA 2015, 2018, and 2022 highlight both stability and emerging concerns in Sweden, particularly in comparison to OECD averages (Figure 4.14).
Students skipping a whole day of school or some classes, once or more, in the last two weeks of school
Note: Based on student reports.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
When it comes to skipping a whole day of school, Sweden demonstrates consistency: 9.0% in 2015, a slight increase of 9.9% in 2018, and a slight decrease of 9.5% in 2022. In contrast, the OECD average remains much higher over the same period: 21.2% in 2015, 21.9% in 2018, and declining to 19.8% in 2022. This suggests that while Sweden has maintained relatively low levels of full-day truancy, other OECD countries may have made progress in reducing it more recently.
A more dynamic trend appears in the rates of skipping some classes. In Sweden, this form of absenteeism has steadily increased: from 16.3% in 2015 to 17.0% in 2018, and further to 19.0% in 2022. Conversely, the OECD average remained stable between 27.8% in 2015, 27.9% in 2018, before falling significantly to 22.5% in 2022. This points to a gradual increase in partial absenteeism in Sweden, even as rates have been falling elsewhere across the OECD. Overall, the data show that although Sweden consistently reports lower absenteeism than the OECD average, the rising trend in skipping classes may indicate emerging risks related to student engagement, classroom climate, or broader well-being issues in Sweden.
Conclusion
This evidence suggests that tackling absenteeism requires addressing underlying disciplinary and safety conditions that can ensure that all students feel motivated, included and safe enough to attend regularly. The PISA 2022 findings indicate that Sweden continues to report comparatively low levels of student absenteeism relative to the OECD average, yet important disparities persist across student groups and forms of absence. While full-day truancy has remained relatively stable, partial absenteeism has gradually increased, diverging from the decline seen in many other OECD countries. The observed associations between absenteeism and lower mathematics performance vary notably by gender, socio-economic status, migrant background and school location, and these links remain significant even when accounting for socio-economic factors in multivariate analysis. This suggests that missing school is rarely an isolated individual choice but is often connected to wider influences such as student well-being, the quality of learning environments and the strength of school engagement. These patterns underscore the importance of sustaining policies and practices that not only monitor attendance but also foster inclusive, supportive and engaging school climates so that all students can remain connected to their education.
Late arrivals to school
The PISA 2022 data on student late arrivals to school offers valuable insights into behavioural patterns that can subtly but significantly affect the learning environment and academic outcomes. Based on students’ self-reports, the data measure the frequency with which students arrived late for school in the two weeks preceding the assessment. While not as severe as absenteeism, habitual lateness can still signal issues related to student engagement, time management, and school climate (OECD, 2023[5]; James et al., 2023[61]). Late arrivals may interrupt classroom routines, reduce instructional time, and contribute to a less orderly learning environment. Moreover, frequent tardiness is often correlated with lower academic performance and may be indicative of broader personal or contextual challenges such as difficulties at home, long commutes, or limited motivation (OECD, 2023[5]).
In Sweden, 44.0% of students report never arriving late to school in the two weeks prior to the PISA test, which is slightly below the OECD average of 54.7%. While this suggests that a significant proportion of students in Sweden are on time, a notable percentage of students still arrive late at least once in a while. Specifically, 31.4% of Swedish students report being late to school once or twice in the two weeks leading up to the test, which is fairly consistent with the OECD average of 28.9%. However, Sweden stands out when it comes to more frequent tardiness. A combined total of 24.6% of students in Sweden report arriving late three or more times (12% three to four times and 12.6% five or more times), well above the OECD combined total of 16.4%. This indicates that a larger proportion of students in Sweden experience more persistent tardiness compared to their peers in the OECD. Moreover, PISA 2022 data reveal that arriving late to school is consistently associated with lower academic achievement among students in Sweden, with students who report being late once or more scoring 24-point lower in mathematics (Figure 4.15). Compared to other Nordic countries, Sweden sits in the middle. Finland, with its culture emphasising punctuality, has a notably higher percentage of students arriving on time (60.2%). Meanwhile, Denmark and Norway show comparable patterns to Sweden, with only about half of students reporting on-time arrivals (50.6% in Denmark and 49.0% in Norway) and relatively high rates of frequent lateness.
In the last two full weeks of school, I arrived late for school once or more.
Note: All values show statistically significant differences. Each bar represents the coefficient from a bivariate regression of mathematics performance on the index, estimated separately for each subgroup (e.g. girls, boys, socio-economically advantaged/disadvantaged, etc.). No control variables are included in these models. Results reflect associations, not causal effects. A socio-economically disadvantaged (or advantaged) student is a student in the bottom (or top) quarter of the PISA index of economic, social and cultural status (ESCS) in Sweden. Rural areas or villages are defined as having fewer than 3000 inhabitants; town areas have between 3000 to 100 000 inhabitants; and cities have over 100 000 inhabitants.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Uneven experiences: How late arrivals vary across students
Gender differences in Sweden indicate that boys are slightly more likely to arrive late to school than girls. According to PISA 2022 data, 59.3% of boys report being late at least once in the two weeks prior to the assessment, compared to 52.7% of girls, resulting in a statistically significant gap of 6.6 percentage points. However, the academic consequences of tardiness appear more severe for girls. Girls who report arriving late score 30 points lower in mathematics, compared to an 18-point decline for boys, which is a gap of 13 points in performance loss. This suggests that lateness may be particularly detrimental to girls’ academic outcomes, or that it coincides with other compounding challenges. This pattern is consistent with findings from other Nordic countries such as Norway and Denmark, where boys also tend to report higher rates of being late than girls.
Differences in punctuality are mirrored in academic outcomes: disadvantaged students who are late score 22 points lower in mathematics, compared to a 17-point drop among their advantaged peers. Although this five-point difference is modest, it suggests that tardiness may further exacerbate existing educational inequalities. The gap in punctuality between socio-economic groups in Sweden reflects a broader pattern observed across many countries, including in neighbouring Nordic countries, where disadvantaged students are more likely to face barriers to regular and timely school attendance.
Student immigrant background also influences patterns of punctuality. A higher proportion of immigrant students (66.3%) report arriving late to school at least once in the two weeks prior to the PISA assessment, compared to 52.9% of non-immigrant students. In terms of academic outcomes, declines in mathematics performance are associated with lateness for both groups: 21 points for immigrant students and 17 points for non-immigrant students, indicating a broadly similar academic impact. While the difference is relatively small, immigrant students show a slightly larger drop in scores.
Geographically, mathematics performance associated with lateness is most pronounced among students in town areas, who score 30 points less. This is notably larger than the 19-point drop among students in rural areas and the 22-point difference among those in city schools, suggesting that frequent tardiness may be more strongly associated with larger performance gaps in town settings. These differences could reflect variations in school environments, transport logistics, or socio-cultural expectations related to punctuality across different types of communities.
Multivariate analysis offers some additional nuances (Annex Table 4.A.9). A one-unit increase in the index of late arrivals is associated with a decrease of 20 score points in mathematics performance after accounting for socio-economic factors. The proportion of variance in mathematics performance explained by late arrivals in Sweden was relatively small (1.6%). These results are broadly consistent with the bivariate analysis, which also show a negative and statistically significant association between late arrivals and mathematics performance in Sweden. Although the analysis does not imply causality, the persistence of this association after accounting for socio-economic factors suggests that frequent late arrivals may be a marker of student disengagement or other school-related challenges that could affect learning outcomes.
Insights from international practice
Despite some improvements in related areas, Sweden’s lateness rates point to a need for more targeted and community-focussed approaches to support punctuality, especially for disadvantaged students. Ireland’s School Completion Programme (Box 4.9) provides a relevant example of how schools can address lateness as part of a broader effort to strengthen attendance and engagement. By combining practical supports such as breakfast clubs and transport assistance with targeted family liaison and mentoring, Ireland’s programme aims to tackle the underlying factors contributing to repeated lateness and create a more supportive school climate that encourages punctuality.
Ireland’s School Completion Programme (SCP) is a core component of the national Delivering Equality of Opportunity in Schools (DEIS) policy framework, which targets educational disadvantage by providing additional supports to students most at risk of absenteeism and early school leaving. Introduced in the early 2000s, the SCP continues to operate in DEIS schools, which serve communities with higher levels of socio-economic disadvantage.
The primary objective of the SCP is to keep students engaged in education by improving attendance, addressing persistent lateness, and reducing drop-out rates. A distinctive feature of the programme is its flexible, locally tailored design: schools and community-based project teams are given the autonomy to develop and implement interventions that respond directly to the specific barriers faced by their students and families.
To tackle both absenteeism and lateness, which is recognised as an early indicator of disengagement, the SCP combines practical supports with strong family and community engagement. Schools implement a range of measures adapted to their local context:
Breakfast clubs: Many participating schools operate breakfast clubs, providing students with a nutritious meal and a positive start to the school day, while encouraging timely arrival and supporting families with morning routines.
Transport supports: In communities where transport challenges contribute to lateness, the SCP may fund transport assistance to help students reach school on time and regularly.
Daily check-ins: Some schools run daily check-ins or mentoring sessions at the start of the day to help students develop punctuality habits and strengthen their connection to school.
Family liaison: Home-School Community Liaison Officers work closely with families to address the underlying causes of frequent lateness and absences (such as housing instability, caring responsibilities or family routines) and to build supportive relationships that promote student attendance.
Community collaboration: Many SCP projects collaborate with local youth services and community organisations to offer additional supports such as homework clubs, mentoring and out-of-school activities, all of which contribute to sustaining student engagement and regular attendance.
Evaluations of the DEIS framework, including the most recent national review in 2022, highlight the positive impact of the SCP on improving daily attendance and punctuality, particularly for students at greater risk of disengagement. The widespread use of breakfast clubs, in particular, has been shown to support punctuality, concentration and students’ readiness to learn.
By combining targeted practical measures with sustained family and community partnerships, Ireland’s School Completion Programme demonstrates how locally responsive supports can address both absenteeism and lateness as part of a coherent strategy to reduce educational disadvantage and promote student success.
Source: Weir et al. (2017[62]); Government of Ireland Department of Education and Youth (2025[63])
Observed patterns in late arrivals
Across all three PISA cycles, a consistently higher proportion of students in Sweden report arriving late to school at least once in the two weeks prior to the assessment than the OECD average (Figure 4.16). In PISA 2022, 56% of Swedish students report being late, compared to 45.3% across the OECD, resulting in a gap of more than 10 percentage points.
Students arriving late to school, once or more, in the last two weeks of school
Note: Based on student reports.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Over time, the trend in Sweden remains relatively stable, with minor fluctuations. In 2015, an average of 54.5% of students reported being late, dipping slightly to 53% in 2018 before rising again to 56% in 2022. In contrast, the OECD average shows a slightly different trajectory. From 2015 to 2018, the proportion of students reporting being late to school increased from 44.5% to 47.5%, followed by a decline to 45.3% in 2022.
The persistent gap between Sweden and the OECD average points to possible systemic or cultural factors affecting punctuality in Swedish schools. The increase in Sweden’s 2022 figure, set against a small decline at the OECD level, may also signal a need for renewed attention to punctuality and student engagement in Swedish education policy. While the year-to-year variations are not dramatic, the consistently elevated rates of tardiness in Sweden underscore its relevance as a policy issue, particularly given the established links between punctuality, academic performance, and school climate.
Conclusion
In summary, the PISA 2022 data on student punctuality in Sweden highlight an important, yet often overlooked, aspect of school engagement and equity. While many students continue to arrive on time, Sweden consistently records a higher share of students who are late, particularly those reporting frequent tardiness, than the OECD average. This pattern has remained stable across recent PISA cycles. Notably, lateness is linked with lower mathematics performance across all student groups, with the steepest performance differences observed among girls, students from disadvantaged backgrounds, and those attending schools in town areas. Although the gaps are not uniformly large, they point to ways in which frequent lateness can compound existing learning challenges and contribute to persistent inequalities. Multivariate analyses also confirm that this negative association between lateness and mathematics outcomes remains statistically significant even after accounting for socio-economic factors. In line with absenteeism, reducing late arrivals depends on creating orderly, welcoming classrooms where students feel engaged and supported to begin lessons on time. Student lateness may reflect deeper structural, cultural or individual barriers to full engagement with school. As such, supporting punctuality should not be viewed in isolation but as part of a wider strategy to strengthen school climate, address barriers that different student groups face, and promote equitable learning conditions for all students in Sweden.
Exploring disciplinary climate and safety in PISA: School questionnaire items
In addition to student perspectives, PISA collects crucial information on disciplinary climate and school safety through the school questionnaire, which is completed by school principals. This questionnaire provides an institutional perspective on classroom management, student behaviour, and the measures schools implement to ensure a safe and orderly school climate. (OECD, 2023[5]).
School principals report on various factors influencing the learning environment, including policies on discipline, strategies to manage classroom behaviour, and measures taken to prevent disruptions. Their responses contribute to the indices of disciplinary climate and school safety which, alongside student-reported data, provides a comprehensive assessment of how conducive classrooms are to learning (OECD, 2023[5]). While this school-reported data offers valuable insight into institutional frameworks and policy intent, it also has limitations. Principals may describe policies as they are designed rather than as they are enacted in daily practice, and such institutional perspectives may not fully reflect the lived experiences of students (OECD, 2023[5]).
To address this, PISA integrates both student and school questionnaire data, facilitating a more nuanced understanding of school security and identifying gaps between student experiences and institutional policies (OECD, 2023[5]). This combined approach helps identify potential gaps between institutional policy and student experience, supporting policymakers in designing targeted interventions that ensure alignment between intended practices and actual conditions, and to foster safe, orderly, and effective schools.
Index of student-related factors affecting school climate
The index of student-related factors affecting school climate is constructed from principals’ reports on several behavioural issues that may undermine the learning environment. These include truancy, bullying, substance abuse, disrespect for teachers, and overall classroom disorder. Positive values on the index indicate that such behaviours are reported as more frequent or serious, while negative values suggest fewer student-related disciplinary issues and a more positive school climate. In Sweden, the mean score on this index is 0.07, slightly above the OECD average of 0.04, indicating that principals perceive student-related challenges to be slightly more frequent than in other OECD countries.
Among the Nordic countries, Denmark reports the lowest index score at -0.58, suggesting that school leaders perceive fewer student-related behavioural problems, and thus a more favourable school climate than their peers elsewhere. By contrast, Finland has a higher mean score of 0.24, indicating more frequent reports of student-related challenges. Norway, with a score of 0.36, reports the highest level of student-related issues among the Nordic countries.
In Sweden, this index is negatively associated with mathematics performance. A one-unit increase in reported student-related challenges corresponds to a 15-point decrease in students’ mathematics scores on average (Figure 4.17). This suggests that schools where principals report more frequent disciplinary issues tend to also have lower academic outcomes, though this association does not imply causality.
Note: Statistically significant differences are shown in a blue tone. Each bar represents the coefficient from a bivariate regression of mathematics performance on the index, estimated separately for each subgroup (e.g. girls, boys, socio-economically advantaged/disadvantaged, etc.). No control variables are included in these models. Results reflect associations, not causal effects. A socio-economically disadvantaged (or advantaged) student is a student in the bottom (or top) quarter of the PISA index of economic, social and cultural status (ESCS) in Sweden. Rural areas or villages are defined as having fewer than 3000 inhabitants; town areas have between 3000 to 100 000 inhabitants; and cities have over 100 000 inhabitants.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Uneven experiences: How student-related factors affect school climate differently across student groups
When disaggregated by gender, mathematics performance in Sweden is negatively associated with the index of student-related factors affecting school climate for both girls and boys. A one-unit increase in the index corresponds to a 14-point decrease in mathematics scores for girls and a 16-point decrease for boys. However, this difference is not statistically significant, suggesting that the relationship between disruptive student behaviours and performance is broadly similar across genders.
At the item level, no statistically significant gender differences are observed in mathematics performance related to this index. By contrast, socio-economic status reveals sharper disparities. Among socio-economically advantaged students, a one-unit increase in the index is associated with a 16-point drop in mathematics performance, compared to a 2-point drop among disadvantaged students. This 14-point difference is statistically significant and suggests that advantaged students may be more academically affected by the presence of student-related disciplinary issues in schools. Three items in the index reveal statistically significant differences in performance between socio-economic groups (Table 4.9). In all three cases, advantaged students show stronger negative associations with mathematics performance than disadvantaged students. Notably, for the item “students not being attentive,” disadvantaged students even show a positive association with performance, which may reflect reporting or contextual differences across school types.
Changes in mathematics performance when principals report learning of students is hindered by
Note: Statistically significant values are indicated in bold. Values shown represent the estimated change in mathematics performance (PISA score points) associated with a one-unit increase in the reported item, based on separate bivariate regressions for advantaged and disadvantaged students. Positive values indicate higher performance associated with reporting the behaviour. These associations should be interpreted with caution, as they do not imply causality and may reflect underlying group dynamics or response patterns. A socio-economically disadvantaged (or advantaged) student is a student in the bottom (or top) quarter of the PISA index of economic, social and cultural status (ESCS) in Sweden.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Differences by immigration background are comparatively small. In Sweden, a one-unit increase in the index of student-related factors affecting school climate is associated with a 12-point decline in mathematics scores for non-immigrant students and a 9-point decline for immigrant students. This 3-point difference is not statistically significant, suggesting that the association between deteriorating school climate and academic performance is relatively consistent across immigration status. At the item level, no statistically significant differences are observed between immigrant and non-immigrant student groups. By contrast, school location is associated with more notable disparities in school climate and its academic impact. While city schools report the most favourable conditions on average (mean index of 0.03, indicating fewer student-related behavioural issues), followed by town schools (0.06), rural schools show the most frequent disciplinary concerns (0.39). These results suggest that student-related disruptions are perceived to be most common in rural areas and least common in cities, although the overall differences across locations remain modest.
Despite reporting fewer student-related behavioural issues overall, the relationship between school climate and mathematics performance is more pronounced in city schools. A one-unit increase in negative student-related behaviours is associated with a 24-point drop in mathematics performance among students in cities, compared to an 8-point decline in town schools and a 1-point decline in rural schools. These findings suggest that students in cities may be more vulnerable to the academic consequences of disruptive behaviour, even if such behaviours are less frequently reported. Item-level analysis reinforces this pattern: in the three items where statistically significant differences are observed, students in city schools consistently show the largest declines in mathematics performance (Table 4.10).
Changes in mathematics performance when the principal reported learning of students hindered by
Note: Statistically significant values are indicated in bold. Values shown represent the estimated change in mathematics performance (PISA score points) associated with a one-unit increase in the reported item, based on separate bivariate regressions for students in various geographical locations. Positive values indicate higher performance associated with reporting the behaviour. These associations should be interpreted with caution, as they do not imply causality and may reflect underlying group dynamics or response patterns. Rural areas are defined as having fewer than 3000 inhabitants; town areas have between 3000 to 100 000 inhabitants; and cities have over 100 000 inhabitants.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Multivariate analysis, which takes into account the students’ and schools’ socio-economic characteristics, confirms a more modest and statistically non-significant association: a one-unit increase in the index is linked to a 2-point decline in mathematics performance (Annex Table 4.A.10). This drop is smaller than the 14-point decrease observed in the bivariate analysis, and the loss of statistical significance suggests that socio-economic differences across schools explain much of the observed association. Nevertheless, the remaining association could indicate that schools with more frequent reports of student behavioural problems may also struggle to support academic achievement, even after accounting for socio-economic context. These results highlight the importance of considering school context and composition when interpreting the academic consequences of disruptive behaviour.
Insights from international practice
Sweden’s overall index suggests only moderately higher levels of disruptive student behaviour than the OECD average; however, the prevalence of truancy, disrespect towards teachers and bullying highlights the need for more comprehensive, school-wide responses. Singapore’s experience (Box 4.10) illustrates the potential of aligning discipline policies, early interventions and teacher development to create classrooms that remain respectful, safe and conducive to learning.
In Singapore, the Ministry of Education (MOE) implements a Holistic Student Management Framework that combines clear, school-wide discipline policies with coordinated support for student well-being and teacher professional development to maintain a positive and safe school climate. The framework directly targets students’ disruptive behaviours and negative attitudes, some of which include truancy, bullying, disrespect for teachers, and general classroom disorder that can undermine learning environments.
Under this approach, all schools develop and enforce whole-school discipline plans aligned with MOE’s guidelines, blending clear behavioural expectations with supportive measures that build respect, responsibility, and resilience. Dedicated Student Management Teams, which typically comprise of discipline leads, counsellors, and educators, work collaboratively to identify students at risk of disengagement or disruptive behaviour, intervene early, and coordinate support with families and community services.
Teachers receive ongoing training in classroom management, conflict resolution, and strategies to strengthen respectful teacher-student interactions. This integrated system of prevention, intervention, and capacity-building contributes to comparatively low levels of student misconduct and strong classroom order in Singapore’s schools. The framework illustrates how consistent behavioural standards, proactive support and well-equipped teachers can together sustain a safe and orderly learning environment.
Source: Singapore Ministry of Education (2022[64]; 2024[65]); Yuen and Yen (1993[66]).
Observed patterns in school leader perceptions of student-related factors affecting school climate
The data reveal several notable differences between Sweden and the OECD average in terms of student-related behaviours that may hinder learning (Figure 4.18).
Percentage of principals reporting that student learning is hindered by the following to some extent or a lot
Note: Based on principal reports
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
In Sweden, principals report higher levels of student truancy (44.3%) and skipping classes (50.2%) compared to the OECD averages of 41.7% and 36.7% respectively. A significantly larger share of Swedish principals also indicates that a lack of respect for teachers is an issue in their schools, with 37.3% reporting it as a hindrance to learning, compared to just 24% across the OECD. Similarly, 16.3% of Swedish principals report issues related to intimidation or bullying of other students, above the OECD average of 12.6%. In contrast, the use of alcohol or illegal drugs appears to be less of a concern in Swedish schools, with only 3.3% of principals identifying it as a problem, well below the OECD average of 9%. In addition, Swedish principals are less likely to cite the lack of student attentiveness as a barrier to learning, with 49.9% reporting it as an issue compared to 60.1% across the OECD. These figures suggest that while Sweden faces particular challenges in areas such as attendance, classroom discipline, and student-teacher relations, there appears to be fewer problems related to substance use and inattentiveness during lessons.
Conclusion
Overall, findings reinforce that broader student-related challenges such as absenteeism, misbehaviour, and disengagement can shape the school climate in meaningful ways, underscoring the need for whole-school strategies that combine effective classroom management with wider supports for students’ academic and social development. Sweden’s overall score on the index of student-related factors affecting school climate indicates that principals perceive slightly more frequent behavioural issues than their peers in other OECD countries. A more detailed analysis reveals persistent differences: Swedish principals report comparatively high levels of truancy, classroom skipping, and lack of respect for teachers, all of which are consistently associated with lower mathematics performance.
Bivariate analyses show that these associations are especially pronounced for socio-economically advantaged students and for those attending schools in cities, where reported behavioural issues are linked to larger performance declines, despite these schools not having the highest overall index scores. In contrast, differences by gender and immigration background are smaller and not statistically significant. However, results from the multivariate analysis (after accounting for schools’ and students’ socio-economic profiles) show that the strength of the relationship between student-related behaviour and mathematics performance weakens considerably, with only a modest and statistically non-significant association remaining. This suggests that much of the observed variation in performance may reflect underlying socio-economic conditions and local school contexts rather than the direct effects of disruptive behaviour alone. Taken together, these findings underline the importance of tackling socio-economic and geographical inequalities in tandem with efforts to improve student engagement, reduce absenteeism, and foster respectful learning environments.
Index of negative school climate
The index of negative school climate, based on the PISA 2022 school questionnaire completed by principals, captures systemic issues that compromise the quality of the learning environment and overall school functioning. Rather than isolated incidents, the index reflects broader organisational dysfunctions through a composite of school-level indicators. These include the frequency and prevalence of profanity, vandalism, theft, and intimidation or verbal abuse among students (whether in-person or through digital means) as well as incidents of verbal or physical aggression directed at teachers and staff. It also accounts for serious behavioural issues that may lead to student transfers. Together, these indicators provide a window into the extent to which schools struggle with maintaining order, safety, and respect. Positive values on the index of negative school climate indicate that principals report more frequent or severe problems within their schools, while negative values suggest that such issues are less prevalent or less serious compared to the OECD average. Empirical studies point to ways that negative school climates can erode trust, disrupt teaching and learning, and contribute to both student disengagement and teacher dissatisfaction (OECD, 2023[5]; La Salle-Finley et al., 2024[67]; Steffgen, Recchia and Viechtbauer, 2013[13]). By focussing on these institutional dimensions, the index offers critical insights into the structural conditions that shape students’ educational experiences
The results for Sweden show interesting patterns when compared to other Nordic countries and the OECD average. Sweden’s mean score on the index is 0.54, notably higher than the OECD average of 0.09, indicating that Swedish students are, on average, exposed to a more negative school environment than their peers in other countries. When viewed alongside other Nordic countries, Sweden's elevated index of 0.54 contrasts with more favourable values seen in Finland (0.12) and Norway (0.42). Denmark, by comparison, reports a negative index of -0.22, indicating the most positive school climate among countries in the Nordic region.
Bivariate regression analysis reveals that mathematics performance also appears sensitive to school climate conditions: a one-unit increase in the index of negative school climate is associated with a 19-point drop in mathematics scores on average (Figure 4.19).
Note: Statistically significant differences are shown in a blue tone. Each bar represents the coefficient from a bivariate regression of mathematics performance on the index, estimated separately for each subgroup (e.g. girls, boys, socio-economically advantaged/disadvantaged, etc.). No control variables are included in these models. Results reflect associations, not causal effects. A socio-economically disadvantaged (or advantaged) student is a student in the bottom (or top) quarter of the PISA index of economic, social and cultural status (ESCS) in Sweden. Rural areas or villages are defined as having fewer than 3000 inhabitants; town areas have between 3000 to 100 000 inhabitants; and cities have over 100 000 inhabitants.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Uneven experiences: How negative school climate varies across students
Gender differences in Sweden show that boys are slightly more affected by negative school climate than girls. A one-unit increase in the index is associated with a 22-point decrease in boys’ mathematics scores and a 16-point decrease in girls’ mathematics scores. However, the six-point difference between the two is not statistically significant, indicating that the relationship between negative school climate and academic performance is similar across both genders. Similarly, there are no significant differences seen between girls and boys at the item-level.
In contrast, when broken down by socio-economic profile, the disparities are more striking. Declines in mathematics performance were strongly associated with a one-unit increase in the index. Advantaged students who show a 16-point drop in scores, compared to just a 2-point drop among disadvantaged students. This 14-point statistically significant difference suggests that advantaged students may be more vulnerable to the disruptions caused by a deteriorating school climate. Two items reveal significant differences in mathematics performance between advantaged and disadvantaged students, with advantaged students showing more negative associations of academic performance than shown by disadvantaged students (Table 4.11):
Changes in mathematics performance when principals report, to moderate or large extent
Note: Statistically significant values are indicated in bold. Values shown represent the estimated change in mathematics performance (PISA score points) associated with a one-unit increase in the reported item, based on separate bivariate regressions for advantaged and disadvantaged students. Positive values indicate higher performance associated with reporting the behaviour. These associations should be interpreted with caution, as they do not imply causality and may reflect underlying group dynamics or response patterns. A socio-economically disadvantaged (or advantaged) student is a student in the bottom (or top) quarter of the PISA index of economic, social and cultural status (ESCS) in Sweden.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Immigration background also reveals a divergence, although not statistically significant. In Sweden a one-unit increase in negative school climate is associated with a 17-point decrease in mathematics performance among non-immigrant students, compared to an 8-point drop among immigrant students. The smaller score difference among immigrant students may reflect different expectations or resilience factors, but further research is needed to interpret this pattern. At the item level, there are two items in the index that show significant differences in mathematics performance between immigrant and non-immigrant students (Table 4.12).
Changes in mathematics performance when principals report, to moderate or large extent
Note: Statistically significant values are indicated in bold. Values shown represent the estimated change in mathematics performance (PISA score points) associated with a one-unit increase in the reported item, based on separate bivariate regressions for immigrant and non-immigrant students. Positive values indicate higher performance associated with reporting the behaviour. These associations should be interpreted with caution, as they do not imply causality and may reflect underlying group dynamics or response patterns.
Source: OECD (2023[19]), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html.
Geographically, the results indicate that the rural areas or villages in Sweden report a mean index of 0.89, suggesting that students in these areas face more negative school climates than those in town areas or cities, where the mean index values are 0.63 and 0.32, respectively. However, the academic impact of these conditions tells a different story. Students in city schools show the largest negative association between school climate and mathematics performance, with a 25-point decrease for each one-unit increase in negative school climate, compared to 10 points in town areas and just 3 points in rural schools. The 22-point gap between city and rural schools is statistically significant, suggesting that while rural schools report more challenging climates overall, the relationship between school climate and academic performance appears more pronounced in city schools, possibly reflecting greater sensitivity to disruptions in these environments.
Multivariate analysis shows that after accounting for students’ and schools’ socio-economic profiles, a one-unit increase in the index of negative school climate is associated with a 4-point decrease in mathematics performance, although this relationship is not statistically significant (Annex Table 4.A.11). The proportion of variance in mathematics performance explained by this index is 1.8%, suggesting a modest association between school climate as perceived by principals and student outcomes in mathematics. Compared to the bivariate analysis, where the association is stronger (a statistically significant 18-point decrease before accounting for socio-economic profiles), the strength of the association in Sweden is substantially reduced after accounting for socio-economic factors. This pattern may suggest that part of the initial relationship observed in the bivariate results reflects differences in school composition and socio-economic context, rather than school climate factors alone.
Insights from international practice
Although Sweden reports fewer incidents of theft and physical violence than the OECD average, the comparatively high prevalence of profanity, intimidation, and verbal abuse suggests that schools may benefit from whole-school approaches that address these behaviours more systematically. The United Kingdom’s Behaviour Hubs Programme (Box 4.11) provides a relevant example of how schools can strengthen their capacity to reduce the frequency and severity of disruptive and aggressive behaviours by implementing clear codes of conduct, improving classroom management, and supporting staff to respond consistently to intimidation, vandalism, and verbal or physical aggression.
The United Kingdom’s school system operates within a decentralised framework, with individual schools and multi-academy trusts (MATs) responsible for ensuring safe, orderly, and respectful learning environments in line with national behaviour standards. To tackle persistent problems that compromise the quality of school climate such as profanity, vandalism, intimidation, and aggression directed at both students and staff, the Department for Education (DfE) launched the Behaviour Hubs Programme in 2021.
Through this initiative, schools with exemplary behaviour cultures are designated as Lead Schools or Lead MATs and partnered with schools facing challenges related to frequent disruptive behaviour and widespread misconduct. The programme provides tailored coaching and practical resources to help school leaders establish clear, robust codes of conduct that explicitly address the use of abusive language, incidents of vandalism or theft, and intimidation or harassment, including cyberbullying.
Participating schools receive professional development for teachers and staff in effective classroom management, de-escalation strategies, and consistent responses to verbal or physical aggression. Schools are also supported in building whole-school approaches that actively monitor and reduce negative behaviours, drawing on evidence from student incident reports and behaviour logs to guide targeted interventions.
A key feature of the programme is its emphasis on shifting school-wide culture towards positive norms and mutual respect, reducing the conditions that allow persistent disrespect, bullying, and aggression that undermine safe teaching and learning. By strengthening leadership capacity and peer collaboration, the Behaviour Hubs Programme aims to prevent the escalation of serious behavioural issues that often result in student disengagement, staff dissatisfaction, or transfers.
Since its launch, the programme has expanded across England, supporting the government’s broader priority of raising standards of discipline and improving the climate for learning in schools where systemic behavioural issues persist.
Source: United Kingdom Department for Education (2024[68]; 2024[69]); Education Endowment Foundation (2021[70]).
Observed patterns in school leader perceptions of negative school climate
Data from the PISA 2022 school questionnaire reveal notable differences between Sweden and the OECD average in principals’ reports of negative school climate indicators. In several areas, Swedish principals report a markedly higher prevalence of problematic behaviours.
In PISA 2022, over half of Swedish school leaders (50.7%) indicate that profanity occurs in school to a moderate or large extent, compared to just 29.6% across the OECD. Similarly, intimidation or verbal abuse among students is reported by 50.5% of Swedish principals which is double the OECD average of 25.3%. Intimidation and verbal abuse are prevalent forms of psychological aggression that erode students’ sense of safety and belonging within schools. Students exposed to frequent verbal abuse often develop avoidance behaviours that hinder their participation in group activities or collaborative learning tasks (Wolke and Lereya, 2015[31]; Woods and Wolke, 2004[30]). Prolonged exposure can lead to chronic stress responses that impair cognitive functions such as memory retention and problem-solving abilities (Hackman et al., 2022[32]; Hessen and Kuncel, 2022[26]). Vandalism is also more frequently reported in Sweden, with 24.5% of principals citing it as a concern compared with 14.7% across OECD countries. However, comparisons should be interpreted with caution because, as noted by several countries including Sweden, Norway and New Zealand, differences in national interpretations of what constitutes “vandalism” may influence responses (Skolverket, 2024[1]). Verbal abuse directed at teachers or non-teaching staff is reported by 9.9% of Swedish school leaders, which is somewhat higher than the OECD average of 6.5%.
In contrast, Sweden reports fewer incidents of theft and physical violence. Only 4.3% of Swedish principals cite theft as a moderate or serious issue, below the OECD average of 6.1%, and just 3.2% report physical injury caused by students, compared to 5.9% across the OECD. The most striking difference is seen in the reported likelihood of transferring students due to behavioural problems: only 6.5% of Swedish principals view this as likely, in stark contrast to 35.5% across the OECD. This pattern may reflect a greater reliance on in-school approaches among Swedish schools, such as counselling, behavioural interventions, or restorative practices rather than exclusionary measures. Overall, the data present a complex picture of school climate in Sweden. While verbal misconduct and peer conflict appear to be widespread, more serious incidents of physical violence or school transfers are reported less frequently. This may point to cultural or systemic differences in behaviour management, with Swedish schools potentially favouring more inclusive responses to behavioural issues within the school environment.
Conclusion
Taken together, the findings from PISA 2022 highlight a multifaceted picture of school climate in Sweden. While principals report relatively high levels of verbal misconduct and peer-related tension, more severe behavioural incidents such as theft, physical violence, or student transfers, remain less prevalent than the OECD average. The relationships observed between negative school climate and mathematics performance suggest that students’ academic outcomes can be sensitive to these conditions, but that this sensitivity varies by context and student group. For example, students in city schools and those from socio-economically advantaged backgrounds appear particularly exposed to the academic impacts of deteriorating school climate.
Importantly, multivariate analysis indicates that once students’ and schools’ socio-economic contexts are taken into account, the association between negative school climate and mathematics performance is modest and no longer statistically significant, explaining only a small share of the variation in scores. This suggests that some of the apparent links between school climate and performance reflect broader contextual or compositional factors rather than school climate alone. Taken together with other results, these patterns suggest that negative aspects of the broader school climate can spill over into classrooms, undermining efforts to maintain order, safety and student engagement. Addressing such challenges requires strong leadership and a collaborative school culture, with early, systemic action tailored to the specific contexts in which they occur and the students they most affect.
Taken together, the findings from PISA 2022 reveal a nuanced picture of disciplinary climate and safety in Swedish schools. While national averages on key indices indicate that Sweden faces some challenges, particularly related to classroom disruption, absenteeism, and student perceptions of safety, these issues are not evenly distributed across the education system. The data suggest that certain student groups and schools are more exposed to negative school climate conditions, with town schools reporting higher levels of disruption and safety risks. Moreover, the analysis indicates that some students, particularly those from more advantaged socio-economic status, may be more academically vulnerable to these conditions when they occur.
When disaggregated by gender and migration background, the results are more mixed. Girls and boys report similar levels of exposure, and while boys sometimes show slightly larger academic declines, the differences are not statistically significant. Similarly, immigrant and non-immigrant students experience comparable school climate conditions, but non-immigrant students appear somewhat more negatively affected in academic terms, particularly in relation to verbal abuse and vandalism.
Item-level analysis reinforces this picture, showing high reported rates of behaviours such as profanity, disrespect for teachers, and intimidation among students, even when overall index values are moderate. Multivariate analysis results suggest that part of the observed relationship between school climate and academic outcomes reflects underlying school composition and contextual factors. These findings underscore the importance of a targeted and context-sensitive policy response, focussed on supporting schools and students facing the greatest behavioural challenges, while continuing to promote safe, inclusive, and respectful learning environments across the system.
Based on these findings, this chapter concludes with a set of four policy considerations to address the challenges identified. Finally, it should be noted that, the policy considerations outlined below, will be further developed in the Chapter 5 of this report.
Policy consideration 1: Strengthening professional learning for teachers and school leaders considering a focus on behavioural management in high-need contexts
Teachers and school leaders in Sweden would benefit from enhanced support in addressing student disengagement, classroom disorder, and disrespectful behaviour, particularly in schools where these problems are more frequent or have greater academic impact. TALIS 2018 shows that teacher preparedness and access to professional learning are strongly related to school climate quality (OECD, 2019[20]). For example, this chapter highlights that the Singapore Holistic Student Management Framework (Box 4.10) offers a compelling example of how professional development in classroom management, inclusive discipline, and respectful teacher-student interactions can help reduce behavioural incidents system-wide. In the Swedish context, this suggests the need for context-sensitive and sustained professional learning opportunities, especially in schools serving mixed socio-economic student populations. Special emphasis could be placed on preventive strategies, restorative practices, and whole-school discipline planning, ensuring that staff are equipped to manage behavioural complexity without resorting to exclusionary practices.
Policy consideration 2: Supporting school leaders to address challenges in school climate and safety in their schools
The school principal plays a pivotal role in shaping and maintaining the school climate (OECD, 2019[20]). Swedish school leaders, especially in town settings where challenges concentrate, would benefit from dedicated training, peer exchange, and strategic guidance to support school-wide behavioural goals. This chapter illustrates how the UK Behaviour Hubs Programme (Box 4.11) supports leaders in implementing consistent conduct policies, building a positive school culture, and responding effectively to verbal abuse, intimidation, and other forms of disruption. Sweden could adapt similar mechanisms by identifying and networking schools with strong climate practices and pairing them with schools seeking to improve. Supporting leadership development around data-informed decision making, staff co-ordination, and early intervention could empower Swedish principals to lead improvement efforts aligned with both local realities and national values.
Policy consideration 3: Reinforcing partnerships with families and external support services in high-need schools
Improving levels of parental engagement in the school life, particularly when interacting with teachers and other members of the school community, might be an important policy lever to consider. According to PISA 2022, in Sweden, the proportion of parents who engaged proactively with schools declined between 2018 and 2022: principals reported that in 2022, only 20% of students attended schools where at least half of families discussed their child’s progress on their own initiative, compared with 41% in 2018 (OECD, 2023[71]). This point is relevant because schools experiencing high levels of student absenteeism or disrespectful behaviour often operate in socially complex environments where families might play a crucial role to support schools. Stronger and more proactive engagement with families, through clear communication channels, shared behavioural norms, and opportunities for involvement, can help reinforce school expectations at home. This chapter highlights that inclusive partnerships with external support services (like mental health support), youth workers, and community organisations are important for tackling the underlying causes of behavioural issues, particularly in schools with high concentrations of disadvantaged students. Sweden could benefit from pilot initiatives or guidelines that formalise and strengthen cooperation between schools and these external actors.
Policy consideration 4: Promoting system-level coherence in behavioural and safety expectations, while respecting local autonomy
Sweden’s decentralised education governance allows schools and municipalities autonomy in defining behavioural standards, safety policies, and student support frameworks. However, the PISA 2022 data presented in this chapter suggest that this flexibility may also lead to variability in implementation, with some schools facing persistent behavioural problems while others operate in relatively calm and safe environments. Sweden could consider developing a set of shared behavioural expectations or frameworks that can be adapted locally but ensure baseline consistency across the system. The examples of national frameworks in Singapore and the UK, as mentioned before, show how systems can promote consistent behavioural norms without undermining school-level innovation. Peer learning platforms and model policies could help schools learn from each other while addressing challenges specific to their contexts.
References
[50] Archambault, I. et al. (2024), “Relationships with teachers and sense of belonging in school: Inter- and intra-individual variations according to adolescent immigrant background”, Learning and Instruction, Vol. 92, p. 101930, https://doi.org/10.1016/j.learninstruc.2024.101930.
[23] Australian Institute for Teaching and School Leadership (2025), Australian Institute for Teaching and School Leadership, http://www.aitsl.edu.au (accessed on 14 October 2025).
[22] Australian National Audit Office (2024), Australian Institute for Teaching and School Leadership’s Administration of National Standards and Frameworks, https://www.anao.gov.au/work/performance-audit/aitsl-administration-national-standards-and-frameworks (accessed on 10 September 2025).
[6] Blank, C. and Y. Shavit (2016), “The Association Between Student Reports of Classmates’ Disruptive Behavior and Student Achievement”, AERA Open, Vol. 2/3, p. 233285841665392, https://doi.org/10.1177/2332858416653921.
[10] Bradshaw, J., P. Hoelscher and D. Richardson (2006), “An Index of Child Well-being in the European Union”, Social Indicators Research, Vol. 80/1, pp. 133-177, https://doi.org/10.1007/s11205-006-9024-z.
[12] Catalano, R. et al. (2004), “The Importance of Bonding to School for Healthy Development: Findings from the Social Development Research Group”, Journal of School Health, Vol. 74/7, pp. 252-261, https://doi.org/10.1111/j.1746-1561.2004.tb08281.x.
[25] Cohen, J. et al. (2009), “School Climate: Research, Policy, Practice, and Teacher Education”, Teachers College Record: The Voice of Scholarship in Education, Vol. 111/1, pp. 180-213, https://doi.org/10.1177/016146810911100108.
[24] Daily, S. et al. (2020), “School Climate as an Intervention to Reduce Academic Failure and Educate the Whole Child: A Longitudinal Study”, Journal of School Health, Vol. 90/3, pp. 182-193, https://doi.org/10.1111/josh.12863.
[69] Department for Education (2024), Behaviour in Schools: Advice for headteachers and school staff, https://assets.publishing.service.gov.uk/media/65ce3721e1bdec001a3221fe/Behaviour_in_schools_-_advice_for_headteachers_and_school_staff_Feb_2024.pdf (accessed on 25 November 2025).
[68] Department for Education (2024), Guidance: Behaviour Hubs, https://www.gov.uk/guidance/behaviour-hubs (accessed on 14 OCtober 2025).
[34] Department of Education and Skills (2025), Anti-bullying procedures for primary and post-primary schools, https://www.gov.ie/en/department-of-education/publications/anti-bullying-procedures-for-primary-and-post-primary-schools/ (accessed on 25 November 2025).
[33] Department of Education and Skills (2018), Wellbeing Policy Statement and Framework for Practice, https://assets.gov.ie/24725/07cc07626f6a426eb6eab4c523fb2ee2.pdf (accessed on 25 November 2025).
[63] Department of Education and Youth (2025), DEIS Delivering Equality of Opportunity In Schools, https://www.gov.ie/en/department-of-education/policy-information/deis-delivering-equality-of-opportunity-in-schools/ (accessed on 31 October 2025).
[59] Direção-Geral da Educação (2019), Programa Nacional de Promoção do Sucesso Escolar, https://www.dge.mec.pt/programa-nacional-de-promocao-do-sucesso-escolar (accessed on 25 November 2025).
[70] Education Endowment Foundation (2021), Improving Behaviour in Schools: Make informed decisions about behaviour strategies, https://educationendowmentfoundation.org.uk/education-evidence/guidance-reports/behaviour (accessed on 10 September 2025).
[37] Eurydice (2025), Portugal: Organisation and governance, https://eurydice.eacea.ec.europa.eu/eurypedia/portugal/organisation-and-governance (accessed on 25 November 2025).
[7] Fauth, B. et al. (2014), “Student ratings of teaching quality in primary school: Dimensions and prediction of student outcomes”, Learning and Instruction, Vol. 29, pp. 1-9, https://doi.org/10.1016/j.learninstruc.2013.07.001.
[48] Fundacja Dajemy Dzieciom Siłę (2021), Fundacja Dajemy Dzieciom Siłę, https://fdds.pl/ (accessed on 10 September 2025).
[47] Fundacja IEP (2023), Program Szkoła Myślenia Pozytywnego, https://www.gov.pl/web/plastyknaleczow/szkola-pozytywnego-myslenia-w-naszej-szkole (accessed on 25 November 2025).
[53] Gase, L. et al. (2017), “Relationships Among Student, Staff, and Administrative Measures of School Climate and Student Health and Academic Outcomes”, Journal of School Health, Vol. 87/5, pp. 319-328, https://doi.org/10.1111/josh.12501.
[14] Germain, E. et al. (2024), Reducing chronic absenteeism: Lessons from community schools, Learning Policy Institute, https://doi.org/10.54300/510.597.
[40] Government of Ontario (2012), Accepting Schools Act, 2012, S.O. 2012, c. 5 - Bill 13, https://www.ontario.ca/laws/statute/s12005 (accessed on 25 November 2025).
[38] Gubbels, J., C. van der Put and M. Assink (2019), “Risk Factors for School Absenteeism and Dropout: A Meta-Analytic Review”, Journal of Youth and Adolescence, Vol. 48/9, pp. 1637-1667, https://doi.org/10.1007/s10964-019-01072-5.
[32] Hackman, D. et al. (2022), “School Climate, Cortical Structure, and Socioemotional Functioning: Associations across Family Income Levels”, Journal of Cognitive Neuroscience, Vol. 34/10, pp. 1842-1865, https://doi.org/10.1162/jocn_a_01833.
[26] Hessen, P. and N. Kuncel (2022), “Beyond grades: A meta-analysis of personality predictors of academic behavior in middle school and high school”, Personality and Individual Differences, Vol. 199, p. 111809, https://doi.org/10.1016/j.paid.2022.111809.
[44] Hoge, D., E. Smit and S. Hanson (1990), “School experiences predicting changes in self-esteem of sixth- and seventh-grade students.”, Journal of Educational Psychology, Vol. 82/1, pp. 117-127, https://doi.org/10.1037/0022-0663.82.1.117.
[11] Hultin, H. et al. (2018), “Pedagogical and Social School Climate: Psychometric Evaluation and Validation of the Student Edition of PESOC”, Scandinavian Journal of Educational Research, Vol. 63/4, pp. 534-550, https://doi.org/10.1080/00313831.2017.1415962.
[61] James, S. et al. (2023), “School start time delays and high school educational outcomes: Evidence from the START/LEARN study”, Journal of Adolescence, Vol. 95/4, pp. 751-763, https://doi.org/10.1002/jad.12151.
[8] Kunter, M., J. Baumert and O. Köller (2007), “Effective classroom management and the development of subject-related interest”, Learning and Instruction, Vol. 17/5, pp. 494-509, https://doi.org/10.1016/j.learninstruc.2007.09.002.
[67] La Salle-Finley, T. et al. (2024), “Understanding and Promoting School Climate, Bullying, and Social-Emotional Learning: Transdisciplinary and Transnational Science Advancing Positive Youth Outcomes”, School Psychology Review, Vol. 53/5, pp. 417-424, https://doi.org/10.1080/2372966x.2024.2386235.
[15] LaRusso, M., D. Romer and R. Selman (2007), “Teachers as Builders of Respectful School Climates: Implications for Adolescent Drug Use Norms and Depressive Symptoms in High School”, Journal of Youth and Adolescence, Vol. 37/4, pp. 386-398, https://doi.org/10.1007/s10964-007-9212-4.
[9] Lazarides, R. and J. Buchholz (2019), “Student-perceived teaching quality: How is it related to different achievement emotions in mathematics classrooms?”, Learning and Instruction, Vol. 61, pp. 45-59, https://doi.org/10.1016/j.learninstruc.2019.01.001.
[45] MacNeil, A., D. Prater and S. Busch (2009), “The effects of school culture and climate on student achievement”, International Journal of Leadership in Education, Vol. 12/1, pp. 73-84, https://doi.org/10.1080/13603120701576241.
[36] Ministério da Educação (2024), Programa Escola Segura, https://www.sembullyingsemviolencia.edu.gov.pt/ (accessed on 25 November 2025).
[58] Ministério da Educação (2017), Ministry of Education, https://www.portugal.gov.pt/pt/gc21/area-de-governo/educacao (accessed on 10 September 2025).
[54] Ministry of Education (2025), Ministry of Education: Moulding the future of our nation, https://www.moe.gov.sg/ (accessed on 25 November 2025).
[65] Ministry of Education (2024), Counselling and Student Welfare, https://www.moe.gov.sg/education-in-sg/our-programmes/counselling-and-student-welfare (accessed on 16 November 2025).
[64] Ministry of Education (2022), Discipline, https://www.moe.gov.sg/education-in-sg/our-programmes/discipline (accessed on 19 November 2025).
[52] Ministry of Education and Research (2024), Republic of Estonia: Ministry of Education and Research, https://www.hm.ee/en (accessed on 25 November 2025).
[18] Ministry of Education and Research (2024), The political and reformative qualities of the Sweden education system 2018-2024.
[46] Ministry of National Education (2023), Szkoła Myślenia Pozytywnego w naszej Szkole, https://www.gov.pl/web/plastyknaleczow/szkola-pozytywnego-myslenia-w-naszej-szkole (accessed on 25 November 2025).
[55] National Institute of Education (2025), Nanyang Technological University: National Institute of Education Professional Programmes, https://www.ntu.edu.sg/nie/admissions/professional-programmes (accessed on 25 November 2025).
[57] OECD (2024), “Evaluating post-pandemic education policies and combatting student absenteeism beyond COVID-19”, OECD Education Policy Perspectives, No. 101, OECD Publishing, Paris, https://doi.org/10.1787/a38f74b2-en.
[41] OECD (2024), PISA 2022 Technical Report, PISA, OECD Publishing, Paris, https://doi.org/10.1787/01820d6d-en.
[17] OECD (2023), Education at a Glance 2023: OECD Indicators, OECD Publishing, Paris, https://doi.org/10.1787/e13bef63-en.
[19] OECD (2023), PISA 2022 Database, https://www.oecd.org/en/data/datasets/pisa-2022-database.html (accessed on 29 January 2024).
[71] OECD (2023), PISA 2022 Results (Volume I and II). Country Notes: Sweden, OECD Publishing, https://www.oecd.org/content/dam/oecd/en/publications/reports/2023/11/pisa-2022-results-volume-i-and-ii-country-notes_2fca04b9/sweden_3c417884/de351d24-en.pdf (accessed on 10 September 2025).
[16] OECD (2023), PISA 2022 Results (Volume I): The State of Learning and Equity in Education, PISA, OECD Publishing, Paris, https://doi.org/10.1787/53f23881-en.
[5] OECD (2023), PISA 2022 Results (Volume II): Learning During – and From – Disruption, PISA, OECD Publishing, Paris, https://doi.org/10.1787/a97db61c-en.
[2] OECD (2020), “Disciplinary climate”, in PISA 2018 Results (Volume III): What School Life Means for Students’ Lives, OECD Publishing, Paris, https://doi.org/10.1787/f05bb3ee-en.
[51] OECD (2020), Education Policy Outlook in Estonia, https://www.oecd.org/content/dam/oecd/en/publications/reports/2020/12/education-policy-outlook-in-estonia_ddbaa07f/9d472195-en.pdf (accessed on 15 November 2025).
[20] OECD (2019), TALIS 2018 Results (Volume I): Teachers and School Leaders as Lifelong Learners, TALIS, OECD Publishing, Paris, https://doi.org/10.1787/1d0bc92a-en.
[39] Ontario Ministry of Education (2025), Creating Safe and Accepting Schools, https://www.ontario.ca/page/creating-safe-and-accepting-schools (accessed on 18 November 2025).
[60] PNPSE (2023), Programa Nacional de Promoção de Sucesso Escolar: Nota de Apresentação 2023, https://pnpse.min-educ.pt/programa (accessed on 25 November 2025).
[35] Polícia Segurança Pública (2024), Escola Segura, https://www.psp.pt/Pages/atividades/programa-escola-segura.aspx (accessed on 14 october 2025).
[21] Richardson, S. et al. (2023), Effectively managing classrooms to create safe and supportive learning environments, https://www.edresearch.edu.au/research/discussion-papers/effectively-managing-classrooms-create-safe-and-supportive-learning-environments (accessed on 14 October 2025).
[29] Sheridan, S. et al. (2019), “A Meta-Analysis of Family-School Interventions and Children’s Social-Emotional Functioning: Moderators and Components of Efficacy”, Review of Educational Research, Vol. 89/2, pp. 296-332, https://doi.org/10.3102/0034654318825437.
[27] Şirin, H. (2022), “Bullying in Schools”, in Advances in Social Networking and Online Communities, Handbook of Research on Bullying in Media and Beyond, IGI Global, https://doi.org/10.4018/978-1-6684-5426-8.ch020.
[1] Skolverket (2024), Elevers trygghet och studiero – en fördjupning av resultat i PISA 2022, Skolverket, https://www.skolverket.se/getFile?file=10392 (accessed on 25 November 2025).
[56] Sobba, K. (2018), “Correlates and buffers of school avoidance: a review of school avoidance literature and applying social capital as a potential safeguard”, International Journal of Adolescence and Youth, Vol. 24/3, pp. 380-394, https://doi.org/10.1080/02673843.2018.1524772.
[13] Steffgen, G., S. Recchia and W. Viechtbauer (2013), “The link between school climate and violence in school: A meta-analytic review”, Aggression and Violent Behavior, Vol. 18/2, pp. 300-309, https://doi.org/10.1016/j.avb.2012.12.001.
[3] Thapa, A. et al. (2013), “A Review of School Climate Research”, Review of Educational Research, Vol. 83/3, pp. 357-385, https://doi.org/10.3102/0034654313483907.
[42] Turanovic, J. and S. Siennick (2022), The Causes and Consequences of School Violence: A Review, https://www.ojp.gov/library/publications/causes-and-consequences-school-violence-review (accessed on 10 September 2025).
[28] Vanderbilt, D. and M. Augustyn (2010), “The effects of bullying”, Paediatrics and Child Health, Vol. 20/7, pp. 315-320, https://doi.org/10.1016/j.paed.2010.03.008.
[4] Wang, M. and J. Degol (2015), “School Climate: a Review of the Construct, Measurement, and Impact on Student Outcomes”, Educational Psychology Review, Vol. 28/2, pp. 315-352, https://doi.org/10.1007/s10648-015-9319-1.
[43] Way, N., R. Reddy and J. Rhodes (2007), “Students’ Perceptions of School Climate During the Middle School Years: Associations with Trajectories of Psychological and Behavioral Adjustment”, American Journal of Community Psychology, Vol. 40/3-4, pp. 194-213, https://doi.org/10.1007/s10464-007-9143-y.
[62] Weir, S. et al. (2017), Addressing educational disadvantage. A review of evidence from the international literature and of strategy in Ireland: an update since 2005., https://www.erc.ie/wp-content/uploads/2018/01/Addressing-Educational-Disadvantage-2017.pdf (accessed on 25 November 2025).
[31] Wolke, D. and S. Lereya (2015), “Long-term effects of bullying”, Archives of Disease in Childhood, Vol. 100/9, pp. 879-885, https://doi.org/10.1136/archdischild-2014-306667.
[30] Woods, S. and D. Wolke (2004), “Direct and relational bullying among primary school children and academic achievement”, Journal of School Psychology, Vol. 42/2, pp. 135-155, https://doi.org/10.1016/j.jsp.2003.12.002.
[66] Yuen, C. and O. Yen (1993), “Pastoral Care and Career Guidance in Singapore Schools”, Australian Journal of Career Development, Vol. 2/2, pp. 34-37, https://doi.org/10.1177/103841629300200210.
[49] Yu, R. and K. Singh (2016), “Teacher support, instructional practices, student motivation, and mathematics achievement in high school”, The Journal of Educational Research, Vol. 111/1, pp. 81-94, https://doi.org/10.1080/00220671.2016.1204260.
← 1. While the word “skadegörelse” is a legal and linguistic equivalent of “vandalism” in English, its connotation in Swedish is often milder and is more closely associated with graffiti or superficial damage rather than serious property destruction. This semantic nuance may have contributed to measurement challenges in cross-country comparisons, as observed in PISA 2022, and led to the exclusion of this item from the calculation of the index in Sweden. However, this interpretation is not definitive and may warrant further investigation.