This chapter presents teachers’ results on the general pedagogical knowledge assessment and how teachers with different levels of general pedagogical knowledge are distributed within countries and across schools. The chapter also explores how general pedagogical knowledge relates to a range of teachers’ professional outcomes.
Results from the Teacher Knowledge Survey
What Teachers Know About General Pedagogy
Abstract
Results from the Teacher Knowledge Survey (TKS) provide evidence that general pedagogical knowledge (GPK) may contribute to improving both students’ and teachers’ outcomes. In countries where teachers scored higher on GPK, 15-year-old students tend to perform better in mathematics and reading as assessed by the Programme for International Student Assessment (PISA). At the individual level, teachers with higher GPK report lower levels of work-related stress, suggesting that they can better cope with a wide range of job demands. Possibly thanks to better classroom management practices, they spend less time keeping order in the classroom and more time on actual teaching and learning.
Teachers with more general pedagogical knowledge are selective with their teaching practices. They are more likely to consider the specific contexts in which they are teaching and adapt their practices, for example, by adjusting the difficulty of practice tasks to support student learning.
Teachers in Portugal achieved the highest scores in the GPK assessment: 274 score points on average, compared to 266 across the four OECD countries participating in the survey (Chile, Poland, Portugal and the United States). Teachers in Poland, Croatia and the United States scored slightly below 270 points, while teachers in Chile achieved 254 score points on average. Average scores in the other participating countries were lower: 225 score points in South Africa, 224 in Morocco and 218 in Saudi Arabia.
Teachers in the United States show considerable variation in GPK, with 175 score points separating the top and bottom 10% of teachers. In the other participating countries, this same gap ranges between 50 points in Morocco and 86 points in South Africa.
Variation in GPK within schools is larger than variation in GPK across schools. As a result, students across different schools tend to have access to teachers with similar GPK, on average. However, teachers with the highest GPK scores are not evenly distributed across schools. This unevenness is largest in South Africa and smallest in Portugal. In most countries, no systematic differences in teachers’ GPK are observed according to school characteristics. In South Africa, teachers in schools with a higher intake of disadvantaged students tend to have lower GPK. The opposite is true in Morocco.
The Teacher Knowledge Survey shows that assessing teachers’ general pedagogical knowledge is feasible and meaningful. These results provide the first international benchmark of teachers’ knowledge base in general pedagogy and highlight the need for policy interventions to strengthen this important component of teachers’ professional competencies.
Teachers in Portugal achieved the highest results (274 score points, on a scale where the average score across the eight participating countries equals 250 score points and the standard deviation equals 50 score points). Average scores of teachers in Portugal are eight points above the average across the four OECD countries that participated in the Teacher Knowledge Survey (Chile, Poland, Portugal and the United States – henceforth referred to as “the average” or “the international average”; Table 1.1 and Table E.1.1).
In Poland, Croatia and the United States, average scores are also above the average of the four participating OECD countries, although only in Poland is this difference statistically significant (Table 1.1). Demonstrated GPK is lower than the OECD average in Chile (254 score points) and, by a larger amount, in South Africa, Morocco and Saudi Arabia (225, 224 and 218 score points, respectively).
While average scores are a useful metric for international benchmarking, they do not describe the extent to which teachers within a given country differ in GPK proficiency. Ultimately, students may be confronted with actual teachers who are very different from the “average” teacher. It is therefore important to look at the entire distribution of GPK scores.
Source: OECD, TALIS TKS 2024 Database, Table E.1.1.
Figure 1.1 shows that countries with similar average scores can be characterised internally by very different distributions. Compare, for example, Poland and the United States. The average score on the GPK assessment is similar in the two countries, but while in Poland most teachers score close to the national average, in the United States, a large number of teachers achieve either very high or very low scores. The top quarter of teachers in the United States score above 312 points, much higher than the cut point above which the top quarter of teachers in Poland sit.1 At the other end of the distribution, however, the cut point below which the bottom quarter of teachers sit is 255 points in both countries. The bottom 10% of teachers in the United States score below 155 points, while the bottom 10% of teachers in Poland score below 237 points (Table E.1.1).
Note: Countries are sorted according to the mean GPK scores of their teachers.
Source: OECD, TALIS TKS 2024 Database, Table E.1.1.
Differences between teachers within the same country are often larger than cross-country differences in average scores. On average among the four OECD countries participating in TKS (henceforth, “on average”), 39 points separate teachers at the top and bottom quarter (interquartile range), and 96 points separate the top and bottom 10% of teachers (interdecile range; Table E.1.1).
The interquartile range is as high as 57 points in the United States and as low as 26 points in Morocco and lies between 30 and 40 points in most other countries; the interdecile range reaches 175 points in the United States, and only 51 points in Morocco, ranging between 60 and 86 points in the other countries.
Consistent with the large dispersion in the distribution of GPK, teachers in the United States have the highest share scoring at Level 3 (51%), but also a sizeable portion scoring at Level 1 (20%; Figure 1.2 and Table E.1.2; see “What is the Teacher Knowledge Survey?” and OECD (forthcoming[1]) for a description of the proficiency levels). In Chile, a similar share of teachers score at Level 1 (22%), but a much smaller share (17%) score at Level 3.
Note: Countries are sorted in descending order according to the share of teachers at Level 3 (Advanced). For a description of the knowledge possessed by teachers at different levels and a discussion of how the proficiency levels were established, see Table 1 in “What is the Teacher Knowledge Survey?” and OECD (forthcoming[1])
Source: OECD, TALIS TKS 2024 Database, Table E.1.2.
Portugal stands out as a system that equips the vast majority of its teachers with a good level of GPK. It has the second-highest share of teachers scoring at Level 3 (40%) and the lowest share at Level 1 (8%). In Poland, only 11% of teachers scored at the lowest level of GPK, but the share scoring at Level 3 is below average at 25%. In Croatia, a higher share of teachers scored at Level 3 (34%), but also a higher share scored at Level 1 (14%). In South Africa, Saudi Arabia and Morocco, the majority of teachers scored at Level 1. In Morocco and Saudi Arabia, fewer than 1% of teachers scored at Level 3, whereas 4% did so in South Africa.
Theory positions general pedagogical knowledge as a core element of teachers’ professional competences, which underpin effective teaching (Guerriero, 2017[2]; OECD, 2025[3]). A growing body of empirical research has begun to test this claim, examining whether teachers who score higher on GPK assessments also teach more effectively, produce better outcomes for their students and enjoy better professional outcomes themselves. A systematic review of international evidence found a moderate positive effect of GPK on various indicators of teaching quality, as well as on student outcomes (Ulferts, 2019[4]). These findings suggest that GPK matters not only for what teachers do in the classroom — including how they structure lessons, support student learning and manage the classroom environment — but also for what students ultimately learn.
Beyond teaching and learning, more recent work has extended this line of inquiry to teachers' own professional well-being. Lauermann and König (2016[5]) demonstrated that GPK can function as a protective factor against burnout, both directly and indirectly through its positive association with teaching self-efficacy, suggesting that stronger pedagogical knowledge helps teachers cope with the demands of the profession.
This section draws on the results of the GPK assessment and the contextual information contained in the TKS questionnaire to assess the relationship between teachers’ demonstrated pedagogical knowledge and students’ outcomes, teaching practices and teachers’ professional outcomes.
General pedagogical knowledge and students’ outcomes
The Teacher Knowledge Survey does not link individual teachers and students, which prevents an investigation of the relationship between teachers’ characteristics (including their general pedagogical knowledge) and students’ outcomes. However, that relationship can be evaluated across countries. Figure 1.3 shows that in countries where teachers’ average GPK scores are higher, 15-year-old students achieved better results in mathematics in the PISA 2022 survey, and a similar relationship holds for students’ results in reading.
This correlation is based on only seven countries that participated in both PISA and TKS, and may therefore not be very robust. Moreover, it should not be interpreted as evidence of a causal link between teachers’ general pedagogical knowledge and students’ outcomes. It does, however, provide strong suggestive evidence that teachers’ knowledge may have a concrete impact on students’ learning, and suggests that investments in teacher professionalism can be vital.
Correlation between average scores in GPK in TALIS TKS 2024 and average mathematics scores of 15-year-old students in PISA 2022
Note: r is the Pearson correlation coefficient between average PISA mathematics scores and average GPK scores; R2 is the R-squared from a linear regression of average GPK scores on average PISA mathematics scores.
Source: OECD, TALIS TKS 2024 Database, Table E.1.1; OECD, PISA 2022 Database, Table I.B1.2.1.
General pedagogical knowledge and classroom management
Classroom management is a key aspect of instruction as it supports the engagement of all students in appropriately paced learning (OECD, 2025[3]). It requires teachers to be aware of all classroom activities and to handle the development of a range of (possibly simultaneous) events to maintain students’ attention on relevant tasks. Teachers may rely on a range of techniques and strategies in this endeavour, including, but not limited to, clarifying classroom expectations, establishing routines, using consistent consequence systems, or organising seating charts.
Teachers with more general pedagogical knowledge tend to report spending less time on classroom management. Holding constant a range of teachers and class characteristics (teachers’ gender, age, years of experience, as well as class size and class composition), an increase of one-standard-deviation on the GPK scale is negatively associated with the reported share of class time spent on keeping order and maintaining discipline in all countries except Poland and South Africa (Table E.1.3 and Figure 1.4). Differences are most pronounced in Saudi Arabia, where this increase in GPK is associated with a decrease of almost 8 percentage points (p.p.) in the share of class time reportedly spent on keeping order. To put this number in perspective, teachers in Saudi Arabia report spending, on average 19% of class time keeping order and maintaining discipline (17% across OECD countries participating in TKS; Table E.1.4).
Change in the average share of time teachers report typically spending on keeping order (maintaining discipline) in a target class1 associated with a one-standard-deviation increase on the general pedagogical knowledge scale2
Notes: Countries are sorted in descending order of the size of the coefficient.
Statistically significant coefficients are highlighted with filled circles (see Annex D). Filled circles above 0 indicate a positive association between teachers’ GPK and the share of class time spent on keeping order in a typical class, while those below 0 reflect a negative relationship.
1. These data refer to a class randomly selected from teachers' current weekly timetable during the week preceding the survey.
2. Results based on linear regression analysis, showing the change in the outcome variable associated with a one-standard-deviation increase in the explanatory variable. The regressions control for teacher characteristics (gender, age and years of teaching experience) and class characteristics (class size and the shares of students in the class that teachers report having difficulties understanding the language of instruction, being low academic achievers and having special education needs).
Source: OECD, TALIS TKS 2024 Database, Table E.1.3.
The lower amount of time spent on maintaining order by more knowledgeable teachers is reflected in the reported frequency of certain behaviour management practices. In most countries, teachers with higher GPK (henceforth, “knowledgeable teachers”) are less likely to report that they “frequently” or “always” tell students to listen to what they say (Table E.1.5). Similarly, having higher GPK tends to be negatively associated with frequently telling students to quiet down at the start of a lesson, telling students to follow classroom rules, and calming disruptive students.
It’s likely that teachers with higher GPK have implemented effective classroom management systems so they do not have to constantly discipline students. An alternative explanation could be that teachers with higher GPK are systematically assigned classes that are easier to teach, or students who naturally tend to behave better. The former interpretation is arguably more likely, though, given that the analysis controls for classroom composition and that teachers are asked to report on a randomly chosen class they teach.2
Teachers with higher GPK also tend to report spending less of their class time on administrative tasks: a one-standard-deviation increase in GPK is associated with a reduction in the share of class time spent on administrative tasks ranging between 0.6 p.p. in Poland and 3.6 p.p in Saudi Arabia (Table E.1.3). On average, teachers report spending about 10% of their class time on administrative tasks (Table E.1.4).
Spending less time on maintaining discipline and on performing administrative tasks leaves more time for actual teaching and learning. In all countries but Poland, higher GPK is positively associated with the share of class time reportedly spent on actual teaching and learning (Table E.1.3). This could contribute to the observed positive association between GPK and students’ outcomes. On average, a one-standard-deviation increase in GPK is associated with a 2 p.p increase in the share of class time devoted to actual teaching and learning, with larger coefficients estimated for Morocco (9 p.p.) and Saudi Arabia (11 p.p.). Considering that, on average, teachers report spending 75% of their class time on actual teaching and learning in Morocco, and 69% in Saudi Arabia, these would translate to increases of 12% (in Morocco) and of 16% (in Saudi Arabia) in actual time spent on teaching and learning.
Change in the average share of time teachers report spending on actual teaching and learning in a typical lesson1 associated with a one-standard-deviation increase on the general pedagogical knowledge scale2
Note: Countries are sorted in descending order of coefficient size.
Statistically significant coefficients are highlighted with filled circles (see Annex D). Filled circles above 0 indicate a positive association between teachers’ GPK and the share of class time spent on keeping order in a typical class, while those below 0 reflect a negative relationship.
1. These data refer to a class randomly selected from teachers' current weekly timetable during the week preceding the survey.
2. Results based on linear regression analysis, showing the change in the outcome variable associated with a one-standard-deviation increase in the explanatory variable. The regressions control for teacher characteristics (gender, age and years of teaching experience) and class characteristics (class size and the shares of students in the class that teachers report having difficulties understanding the language of instruction, being low academic achievers and having special education needs).
Source: OECD, TALIS TKS 2024 Database, Table E.1.3.
Teachers with more general pedagogical knowledge tailor their teaching practices
Teaching is a complex activity that requires professional judgement to select the best approach in a continuously evolving environment. In preparing lessons, teachers must make decisions about relevant teaching strategies, as well as the content, tasks and materials needed. In the classroom, teachers need to make ongoing decisions about how to interact with students and whether planned activities need to be adapted in response to classroom events. GPK can help to guide teachers’ choices when adapting their teaching practices to the demands of specific classroom situations.
The TKS contextual questionnaire asked teachers to report how often they engage in certain teaching practices. Some of these practices explicitly focus on adapting teaching to the concrete context teachers face (“adaptive practices”). Other questions asked about practices focused on students’ cognitive activation and others about assessment practices.
TKS results suggest that knowledgeable teachers strongly believe in tailoring their teaching to specific students’ needs. In the majority of countries participating in TKS, the odds that teachers report “always” considering students’ prior knowledge and needs when planning a lesson increase when teachers have higher GPK (Figure 1.6). Moreover, GPK can help teachers tailor their teaching within a lesson. In half of the participating countries, teachers with higher GPK are more likely to report “always” changing their approach to explaining when a student has difficulties understanding a topic or task (Figure 1.6).
Note: Statistically significant coefficients are highlighted with filled circles and country labels (see Annex D).
1. Binary variables: the reference category refers to teachers reporting that they “never or almost never”, “occasionally”, or “frequently” engage in the different practices.
2. Results based on 5 separate binary logistic regressions. The estimated odds ratios from each regression are displayed on a different line. An odds ratio indicates the degree to which an explanatory variable is associated with a categorical outcome variable. An odds ratio below 1 denotes a negative association; an odds ratio above 1 indicates a positive association; and an odds ratio of 1 means that there is no association. The regressions control for teacher characteristics (gender, age and years of teaching experience) and class characteristics (class size and the shares of students in the class that teachers report to having difficulties understanding the language of instruction, being low academic achievers and to having special education needs).
Source: OECD (2024), TALIS TKS 2024 Database, Table E.1.6.
However, tailoring teaching may not always be feasible or possible because of a lack of time, resources or an excessively heterogeneous composition of the class. In half the countries studied, more knowledgeable teachers are less likely to report “always” referring students to different materials for learning depending on their needs. This could be because teachers with higher GPK select other adaptive teaching strategies that are less time-consuming to prepare and complex to implement than differentiated teaching materials (Gaitas and Alves Martins, 2016[6]). Also, few consistent relationships emerge between GPK and teachers reporting that they “always” adapt their teaching methods to students’ needs. This could be because teachers account for students’ prior knowledge and needs by adapting lesson content or their explanations, rather than by adapting their teaching methods – especially if they view differentiation strategies as being part of their regular practice.
When examining other teaching practices, a broad pattern emerges: teachers with higher GPK tend to be less likely to “always” adopt a given practice. This could again be evidence of the tendency of teachers with higher GPK to pick the tool appropriate to a specific situation – rather than blindly following prescribed lesson plans or teaching methods – or to appreciate the importance of diversifying teaching practices and the complementarities among them.
Having higher GPK, for example, is linked to a lower likelihood that teachers report “always” letting students practise similar tasks until every student has understood the subject matter (Table E.1.7). This relationship is statistically significant in Croatia, Poland, Portugal and South Africa. One possible explanation is that teachers with greater pedagogical knowledge may be more effective at supporting student progress, as they spend less time on tasks pitched at a similar level rather than increasingly challenging ones.
A consistent pattern emerges: teachers with higher GPK tend to report “always” performing only a limited number of teaching practices. This could reflect a consensus, among more knowledgeable teachers, on the effectiveness of these approaches. This may be the case, for example, for selecting tasks for student practice that gradually increase in difficulty. In Croatia, Morocco and Poland, GPK is positively associated with teachers “always” reporting doing this (Table E.1.7). Because more knowledgeable teachers report taking students’ growing abilities into account when setting tasks, this finding aligns with the results above, demonstrating a positive association between GPK and teachers’ consideration of students’ prior knowledge in lesson planning. Possibly for a similar reason, teachers with higher GPK tend to report that they do not “always” let students practise similar tasks.
Teachers with higher GPK tend to be particularly discerning in their use of teaching practices for cognitive activation that aim to foster deep conceptual understanding and problem-solving skills. A one-standard-deviation increase in GPK is consistently associated with a decrease in the odds that teachers report “always” using a range of related practices (Table E.1.8, Figure 1.7). In all countries except Chile and Morocco, teachers with higher GPK are less likely to report “always” giving tasks that require students to think critically. Negative relationships are similarly found in five of the eight participating countries for presenting tasks for which there is no obvious solution, having students work in small groups to come up with a joint solution to a problem, and asking students to decide on their own procedures for solving complex tasks.
These findings are consistent with the positive relationships between teachers’ GPK and their consideration of students’ prior knowledge in lesson planning, as well as their selection of tasks that gradually increase in difficulty. Teaching practices for cognitive activation engage students in higher-level thinking, which requires them to work from a solid knowledge base – without which a task may become too challenging.
Note: Statistically significant coefficients are highlighted with filled circles and country labels (see Annex D).
1. Binary variables: the reference category refers to teachers reporting that they “never or almost never”, “occasionally”, or “frequently” engage in the different practices.
2. Results based on 5 separate binary logistic regressions. The estimated odds ratios from each regression are displayed on a different line. An odds ratio indicates the degree to which an explanatory variable is associated with a categorical outcome variable. An odds ratio below 1 denotes a negative association; an odds ratio above 1 indicates a positive association; and an odds ratio of 1 means that there is no association. The regressions control for teacher characteristics (gender, age and years of teaching experience) and class characteristics (class size and the shares of students in the class that teachers report having difficulties understanding the language of instruction, being low academic achievers and having special education needs).
Source: OECD (2024), TALIS TKS 2024 Database, Table E.1.8.
When looking at assessment practices, the use of students’ self-assessment stands out as something that teachers with higher GPK are consistently less likely to “always” do. This is the case in all countries except Chile and Croatia, where the relationship, although negative, is surrounded by too large a margin of uncertainty and cannot therefore be considered statistically different from zero (Table E.1.9, Figure 1.8). This suggests that more knowledgeable teachers are cautious about over-reliance on students’ self-assessments, although they may incorporate them in their teaching from time to time as part of their wider repertoire of assessment strategies, depending on their goals for a particular lesson.
Note: Statistically significant coefficients are highlighted with filled circles and country labels (see Annex D).
1. Binary variables: the reference category refers to teachers reporting that they “never or almost never”, “occasionally”, or “frequently” engage in the different practices.
2. Results based on separate binary logistic regressions. The estimated odds ratios from each regression are displayed on a different line. An odds ratio indicates the degree to which an explanatory variable is associated with a categorical outcome variable. An odds ratio below 1 denotes a negative association; an odds ratio above 1 indicates a positive association; and an odds ratio of 1 means that there is no association. The regressions control for teacher characteristics (gender, age and years of teaching experience) and class characteristics (class size and the shares of students in the class that teachers report having difficulties understanding the language of instruction, being low academic achievers and having special education needs).
Source: OECD (2024), TALIS TKS 2024 Database, Table E.1.9.
Administering an assessment at the end of a unit or a block of lessons, on the other hand, appears to be a practice favoured by knowledgeable teachers, with those in Croatia, Morocco and Poland more likely to report “always” doing this (Table E.1.9).
General pedagogical knowledge and teachers’ professional outcomes
General pedagogical knowledge is an important resource that teachers can draw upon when coping with the demands of their job. It can potentially affect not only students’ outcomes (through the adoption of more effective teaching practices) but also teachers’ professional outcomes, such as their well-being. For example, GPK tends to be negatively related to teachers’ reports of experiencing stress from students’ behaviour. After accounting for a range of teacher and school characteristics, in five of the eight countries participating in TKS (Croatia, Morocco, Portugal, Saudi Arabia and South Africa), an increase of one-standard-deviation on the GPK scale is associated with a reduction in the scale of student behaviour stress (Table E.1.10 and Figure 1.9).3 The relationship is particularly strong in Saudi Arabia, where a one-standard-deviation increase in GPK is associated with a reduction in the scale of student behaviour stress of about 45% of a standard deviation.4
Note: Statistically significant coefficients are highlighted with filled circles (see Annex D). Filled circles above 0 indicate a positive association between general pedagogical knowledge and the scale of student behaviour stress, while those below 0 reflect a negative relationship.
1. Standardised scale scores with a standard deviation of 2 and the value of 10 corresponding to the item mid-point value of the response scale.
For more information on the scales, see Annex D.
2. Results based on linear regression analysis, showing the change in the outcome variable associated with a one-standard-deviation increase in the explanatory variable. The regressions control for teacher characteristics (gender, age and years of teaching experience) and school characteristics (school location, school governance type, school intake of students from socio-economically disadvantaged homes, school intake of students who have difficulties understanding the language(s) of instruction and school intake of students with special education needs).
Source: OECD (2024), TALIS TKS 2024 Database, Table E.1.10.
Another set of questionnaire items asks about workload-related stress. In Morocco, Portugal, Saudi Arabia and South Africa, teachers with higher GPK report lower levels of workload-related stress. The opposite is true, however, in Poland (Table E.1.11 and Figure 1.10).
Note: Statistically significant coefficients are highlighted with filled circles (see Annex D). Filled circles above 0 indicate a positive association between general pedagogical knowledge and the scale of student behaviour stress, while those below 0 reflect a negative relationship.
1. Standardised scale scores with a standard deviation of 2 and the value of 10 corresponding to the item mid-point value of the response scale.
For more information on the scales, see Annex D.
2. Results based on linear regression analysis, showing the change in the outcome variable associated with a one-standard-deviation increase in the explanatory variable. The regressions control for teacher characteristics (gender, age and years of teaching experience) and school characteristics (school location, school governance type, school intake of students from socio-economically disadvantaged homes, school intake of students who have difficulties understanding the language(s) of instruction and school intake of students with special education needs).
Source: OECD (2024), TALIS TKS 2024 Database, Table E.1.11.
Looking at the specific sources of workload-related stress, the relationship observed in Poland appears to be driven by teachers with higher GPK, who are more likely to report stress from excessive marking (Table E.1.12). In Croatia, Portugal, Saudi Arabia and South Africa, teachers with higher GPK scores are less likely to report having too many lessons to teach as a source of stress (Table E.1.13).
Figure 1.11 summarises the relationship between GPK and the likelihood of reporting stress from numerous other sources, including accountability, diverse learning needs, keeping up with reforms and professional growth. Overall, there is a consistent negative relationship between GPK and stress from many of these sources in most countries, supporting the idea that general pedagogical knowledge helps (directly or indirectly) teachers cope with a broad range of job demands.
Note: Statistically significant coefficients are highlighted with filled circles and country labels (see Annex D).
1. Binary variables: the reference category refers to “not at all” or “to some extent”.
2. Results based on separate binary logistic regressions. An odds ratio indicates the degree to which an explanatory variable is associated
with a categorical outcome variable. An odds ratio below 1 denotes a negative association; an odds ratio above 1 indicates a positive association;
and an odds ratio of 1 means that there is no association. The regressions control for teacher characteristics (gender, age and years of teaching experience) and school characteristics (school location, school governance type, school intake of students from socio-economically disadvantaged homes, school intake of students who have difficulties understanding the language(s) of instruction and school intake of students with special education needs).
Source: OECD (2024), TALIS TKS 2024 Database, Table E.1.14.
Despite being less likely to report stress at work, teachers with more general pedagogical knowledge did not necessarily score higher on the composite scale of enjoyment of teaching: after accounting for teachers and school characteristics, the relationship between GPK and enjoyment of teaching is positive and statistically significant only in Portugal and Saudi Arabia (Table E.1.15 and Figure 1.12).
Note: Statistically significant coefficients are highlighted with filled circles (see Annex D). Filled circles above 0 indicate a positive association between general pedagogical knowledge and the scale of student behaviour stress, while those below 0 reflect a negative relationship.
1. Standardised scale scores with a standard deviation of 2 and the value of 10 corresponding to the item mid-point value of the response scale.
For more information on the scales, see Annex D.
2. Results based on linear regression analysis, showing the change in the outcome variable associated with a one-standard-deviation increase in the explanatory variable. The regressions control for teacher characteristics (gender, age and years of teaching experience) and school characteristics (school location, school governance type, school intake of students from socio-economically disadvantaged homes, school intake of students who have difficulties understanding the language(s) of instruction and school intake of students with special education needs).
Source: OECD (2024), TALIS TKS 2024 Database, Table E.1.15.
Despite the broad positive association between general pedagogical knowledge and teachers’ professional outcomes, having high GPK is not strongly associated with a higher likelihood of reporting an intention to continue working as a teacher for more than five years. After controlling for teacher and school characteristics, only in Saudi Arabia, Morocco and Portugal, teachers with more general pedagogical knowledge report being less likely to plan to quit the profession in the next five years (Table E.1.16, Table E.1.17 and Figure 1.13). This is likely because teachers’ career intentions are influenced by a wide range of factors going beyond those analysed in this report. Findings from TALIS 2024, for example, point to the importance of intrinsic motivation and a broad range of employment conditions, such as material benefits, opportunities for career progression and work schedules (OECD, 2025[7]). The fact that more knowledgeable teachers are equally likely to intend to stay in the profession should encourage education authorities to find ways to better value and recognise the professional contribution they provide.
Note: Statistically significant coefficients are highlighted with filled circles (see Annex D).
1. Binary variables: the reference category refers to teachers wanting to continue to work as teachers for more than 5 years.
2. Results based on binary logistic regressions. An odds ratio indicates the degree to which an explanatory variable is associated
with a categorical outcome variable. An odds ratio below 1 denotes a negative association; an odds ratio above 1 indicates a positive association;
and an odds ratio of 1 means that there is no association. The regressions control for teacher characteristics (gender, age and years of teaching experience) and school characteristics (school location, school governance type, school intake of students from socio-economically disadvantaged homes, school intake of students who have difficulties understanding the language(s) of instruction and school intake of students with special education needs).
Source: OECD (2024), TALIS TKS 2024 Database, Table E.1.16.
An important objective of education policy is to reduce social and educational inequalities and to ensure equal access to publicly funded resources. The mechanisms governing the allocation of teachers to schools, and therefore the matching of teachers and students, are among the levers that education authorities can use to attain these objectives (OECD, 2022[8]). Allocating teachers is even more relevant given the large variation observed in teachers’ GPK in some countries (
Figure 1.1).
Equal access of students to high-GPK teachers implies randomly allocating teachers to schools. This would ensure that all schools would have a similar distribution of teachers’ characteristics (including GPK). Statistics that capture departures from a random allocation can be used to assess how close education systems are to achieving equality of access, at least at the school level.
Importantly, equality of access at the school level may not be sufficient to guarantee equal chances for all students, as different groups of students can then be assigned to different teachers within the school. Unfortunately, TKS data do not allow analysis at this more granular level.
The distribution of knowledgeable teachers across schools
This report analyses the distribution of teachers via two methods, decomposition and the dissimilarity index. Decomposition partitions the overall variance in GPK scores into two components: within-school and between-school. If a large portion of the variance lies between schools, it means that schools differ significantly in their teachers’ average GPK. If most of the variance lies instead within schools, it means that most schools replicate the variation of GPK observed at the national level.
On average, only about 8% of the variance in GPK lies between schools (Table E.1.18 and Figure 1.14). The share of between-school variance is, however, as high as 23% in South Africa, and exceeds 10% in Chile, Saudi Arabia and the United States; it is smallest in Croatia (1%), Portugal (2%) and Poland (3%).
The dissimilarity index measures departure from an even allocation of high-GPK teachers (those in the top quarter of the national distribution of GPK). When the index equals 1, all high-GPK teachers (those in the top quarter of the national distribution of GPK) are concentrated in a single school. When the index equals 0, all high-GPK teachers are allocated equally across schools.
The dissimilarity index for teachers in the top quarter of the national distribution of GPK is highest in South Africa (0.65) and is lowest in Portugal (0.29; Table E.1.18 and Figure 1.14). This implies that 24% of teachers in South Africa would need to change schools in order to achieve an even allocation of high-GPK teachers, while in Portugal, only 11% of teachers would need to be involved in such transfers.5
Results from decomposition and the dissimilarity index provide different insights into how teachers are allocated. The decomposition approach suggests that teachers’ GPK varies more within a school than across schools. However, decomposition results can be strongly influenced by teachers with very high GPK. For example, one high-performing teacher will have a greater effect on their school’s within-school variance than on their country’s between-school variance. The dissimilarity index, on the other hand, groups teachers by thresholds (e.g. the top quartile). According to this approach, there is considerable clustering of high-GPK teachers in schools. However, this approach ignores variation in GPK within those groups (e.g. the highest-performing teachers).
Share of variance in GPK between schools and dissimilarity index for teachers in the top quarter of the national distribution of GPK
Note: Countries are ranked in descending order according to the share of variance in GPK scores between schools. More details on the computation of the dissimilarity index and its interpretation can be found in Annex D.
Source: OECD, TALIS TKS 2024 Database, Table E.1.18.
Which schools have more teachers with greater general pedagogical knowledge
In TKS, it is possible to compare schools according to their location (rural area, town or city), whether they are privately or publicly managed and according to the composition of the student body, in particular the intake of students from socio-economically disadvantaged homes, or who have difficulties understanding the language of instruction, or who have special education needs.
In California (United States), the Governor’s Teaching Fellowship programme and the Talent Transfer Initiative provided financial bonuses to talented novice teachers who transfer to low-performing schools. The incentive amounted to USD 20 000 and was allocated competitively to talented novice teachers, who accepted to stay in the new school for at least four years.
Steele, Murnane and Willett (2010[9]) estimate that the programme successfully attracted teachers to low-performing schools – teachers who would not have chosen to work in these schools otherwise. Glazerman et al. (2013[10]) found that the programme had a positive impact on test scores (math and reading) in primary schools, but not in middle schools. The incentive also had a positive impact on teacher retention rates.
The Accelerating Campus Excellence (ACE) program in Dallas, Texas is another example of the effectiveness of financial incentives to attract high-performing teachers to low-performing schools (Morgan et al., 2023[11]). The reduction of the incentives, however, caused many high-performing teachers to leave and led to a drop in test scores. This suggests that policymakers should carefully consider the long-term financial sustainability of these types of programmes and perform rigorous cost-benefit analysis before introducing them.
South Africa is the only country where average GPK differs across schools along all these dimensions. In South Africa, more advantaged schools (those with a lower intake of disadvantaged students) and privately managed and urban schools systematically have teachers with higher GPK scores, on average. The gaps are relatively large (about 20 points) along all these dimensions, except for intake of students with special education needs, where the difference between advantaged and disadvantaged schools is 12 score points (Table E.1.19 and Figure 1.15).
Notes: A rural school is located in a rural area or village of up to 3 000 people; an urban school is located in a city of over 100 000 people.
A public school is a school whose principal reported that it is managed by a public education authority, government agency, municipality, or governing board appointed by the government or elected by public franchise. The question does not refer to the source of the school’s funding, which is reported in the preceding question.
A private school is a school whose principal reported that it is managed by a non-governmental organisation (e.g. a church, trade union, business or other private institution). The question does not refer to the source of the school’s funding, which is reported in the preceding question. In some countries, the private school category includes schools that receive significant government funding (government-dependent private schools).
Schools with many disadvantaged students have at least 30% of students coming from socio-economically disadvantaged homes, while schools with few disadvantaged students have fewer than 10% of students from such homes. Socio-economically disadvantaged homes are those that lack the basic necessities or advantages of life, such as adequate housing, nutrition or medical care.
Schools with many SEN students have at least 30% of students with special education needs, while schools with few SEN students have fewer than 10%.
Students with special education needs are those for whom a special education need has been formally identified because they are mentally, physically or emotionally disadvantaged.
Schools where many students have language difficulties have at least 10% of students who have difficulty understanding the language(s) of instruction, while schools where few students have language difficulties have no students who have difficulty understanding the language(s) of instruction.
Source: OECD, TALIS TKS 2024 Database, Table E.1.19.
Morocco is the only other country in which differences between schools are detected across most of these dimensions (the exception being differences by the intake of students with special education needs). In the case of Morocco, however, the differences are reversed (and smaller in magnitude), with more disadvantaged schools, as well as publicly managed and rural schools, having teachers with higher GPK scores on average.
In Croatia, Poland, Portugal and the United States, no differences in average GPK emerge across the school characteristics analysed. In Chile and Saudi Arabia, teachers in privately managed schools tend to have higher GPK than those in publicly managed schools, by 11 and 13 points, respectively (Table E.1.19 and Figure 1.15).
References
[6] Gaitas, S. and M. Alves Martins (2016), “Teacher perceived difficulty in implementing differentiated instructional strategies in primary school”, International Journal of Inclusive Education, Vol. 21/5, pp. 544-556, https://doi.org/10.1080/13603116.2016.1223180.
[10] Glazerman, S. et al. (2013), Transfer Incentives for High Performing Teachers: Final Results from a Multisite Experiment (NCEE 2014-4003), National Center for Education Evaluation and Regional Assistance, Institute of Education Sciences, U.S. Department of Education, https://files.eric.ed.gov/fulltext/ED544269.pdf (accessed on 16 December 2025).
[2] Guerriero, S. (ed.) (2017), Pedagogical Knowledge and the Changing Nature of the Teaching Profession, Educational Research and Innovation, OECD Publishing, Paris, https://doi.org/10.1787/9789264270695-en.
[5] Lauermann, F. and J. König (2016), “Teachers’ professional competence and wellbeing: Understanding the links between general pedagogical knowledge, self-efficacy and burnout”, Learning and Instruction, Vol. 45, pp. 9-19, https://doi.org/10.1016/j.learninstruc.2016.06.006.
[11] Morgan, A. et al. (2023), Attracting and Retaining Highly Effective Educators in Hard-to-Staff Schools, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w31051.
[7] OECD (2025), Results from TALIS 2024: The State of Teaching, TALIS, OECD Publishing, Paris, https://doi.org/10.1787/90df6235-en.
[3] OECD (2025), TALIS Teacher Knowledge Survey 2024 Conceptual and Assessment Framework, TALIS, OECD Publishing, Paris, https://doi.org/10.1787/65903902-en.
[8] OECD (2022), Mending the Education Divide: Getting Strong Teachers to the Schools That Need Them Most, TALIS, OECD Publishing, Paris, https://doi.org/10.1787/92b75874-en.
[1] OECD (forthcoming), TALIS 2024 Technical Report, TALIS, OECD Publishing, Paris.
[9] Steele, J., R. Murnane and J. Willett (2010), “Do financial incentives help low-performing schools attract and keep academically talented teachers? Evidence from California”, Journal of Policy Analysis and Management, Vol. 29/3, pp. 451-478, http://www.jstor.org/stable/40802084.
[4] Ulferts, H. (2019), “The relevance of general pedagogical knowledge for successful teaching: Systematic review and meta-analysis of the international evidence from primary to tertiary education”, OECD Education Working Papers, No. 212, OECD Publishing, Paris, https://doi.org/10.1787/ede8feb6-en.
← 1. This cut point is also called the 75th percentile of the distribution. Similarly, the cut point below which one quarter of teachers are located is called the 25th percentile.
← 2. More precisely, teachers are asked to report on the first ISCED Level 2 class that they taught after 11 a.m. the last Tuesday preceding the survey. If they had not taught any class on that Tuesday, they should focus on the first class taught following the last Tuesday.
← 3. The scale of Student Behaviour Stress – as all others in TALIS and TKS – is constructed by combining information from teachers’ answers to a battery of questions and is standardised to have a standard deviation of two across all education systems participating in TALIS and so that the value 10 corresponds to the item mid-point value of the response scale. For more information on the scales, see Annex D.
← 4. While the annex tables present regression coefficients, the text mostly expresses the estimated association in terms of standard deviation changes (this is done by simply dividing the estimated coefficient by 2). This is meant to facilitate interpretation, as all associations are expressed on the same metric. For more information on the scales, see Annex D.
← 5. The dissimilarity index is directly related to the share of teachers from different groups (e.g. those with low and high GPK) that would need to change schools in order to achieve an equal distribution across schools, while maintaining school size constant. With only two groups (teachers who are or not in the top quarter of the national distribution of GPK), such movements would necessarily entail the swapping of teachers from the two groups, if school size needs to stay constant. In this setting, the dissimilarity index equals the sum of the shares of teachers from the two groups that would need to swap places to achieve an even allocation. As the relative size of the two groups in the population is known (by construction, 25% of teachers are in the top quarter of the distribution of GPK, and 75% are not), multiplying the dissimilarity index by 2×0.75×0.25=0.375 gives the percentage of teachers in the overall teacher population that needs to change school to restore an even allocation. See Annex D for more details on the dissimilarity index and its interpretation.