This chapter examines whether there is a link between inequalities in access to effective teachers and learning divides between socio‑economically advantaged and disadvantaged students. It also tries to understand if system‑level policies can promote greater equity in students’ access to effective teachers. More specifically, the chapter discusses system‑level relationships between the sorting of effective teachers and teaching practices, including those related to the use of digital tools, and socio‑economic inequalities in student performance. It also highlights associations between the allocation of effective teachers and teaching practices, and measures of school competition and autonomy, thus hinting at the potential of policy intervention.
Mending the Education Divide
4. Teacher allocation and learning divides
Abstract
Highlights
Countries and territories participating in the Teaching and Learning International Survey (TALIS) characterised by a more uneven distribution of experienced teachers tend to have lower average scores in the Programme for International Student Assessment (PISA) 2018 reading assessment. Likewise, the uneven distribution of teachers who have undergone comprehensive initial training is negatively correlated with students’ mean reading score in PISA at the system level. Both points are especially true for socio‑economically disadvantaged students.
Students tend to perform worse in reading in education systems where teachers who spend a larger share of class time on actual teaching are unevenly distributed across schools. This relationship also holds for socio‑economically disadvantaged students. It is important to note, however, that the findings in this report cannot be interpreted as causal. Advantaged schools might have fewer disciplinary problems, which allows teachers to spend more time on actual teaching and less on classroom management.
Disadvantaged students tend to have just as much or more opportunity to learn digital literacy skills (such as detecting if the information read is subjective or biased) in education systems where teachers with high self‑efficacy in information and communications technology (ICT) and teachers who regularly teach using ICT are more evenly distributed across schools.
Experienced teachers are distributed more evenly across schools in countries where a higher share of principals report that their school has autonomy in appointing, hiring, dismissing or suspending teachers. In systems with more school autonomy, staffing decisions seem to take into account a wider range of factors, reducing the relative importance of seniority.
Introduction
When education systems provide equal learning opportunities to all students, differences in students’ outcomes are no longer driven by factors that are outside of the control of any individual, such as socio‑economic background, gender and disabilities (OECD, 2018[1]). This also implies that students who are most in need – for instance, those from socio‑economically disadvantaged backgrounds – are exposed to good teachers and effective teaching practices. However, in many education systems, schools with a large portion of socio‑economically disadvantaged students are precisely those that find it difficult to attract experienced teachers, who are often more effective than their junior colleagues (OECD, 2006[2]).
In today’s knowledge‑based economies, a poor education can have more punishing consequences: poor skills limit access to better‑paying and more rewarding jobs. More generally, it limits access to better health and living conditions, and hinders social and political participation (Hanushek et al., 2015[3]; OECD, 2016[4]).
Despite significant efforts to narrow disparities in students’ outcomes in the recent past, students’ socio‑economic background remains strongly correlated with their academic performance (OECD, 2019[5]; OECD, 2018[1]). Analyses based on PISA data show consistently that while many socio‑economically disadvantaged students succeed at school, students from socio‑economically advantaged family backgrounds tend to outperform their disadvantaged peers in all subjects (OECD, 2019[5]). Moreover, results from Chapters 2 and 3 of this report show that effective teachers are not distributed randomly across schools and can be concentrated in certain schools depending on school characteristics such as socio‑economic profile and location. It is important to highlight here that throughout this report terms such as quality teachers, effective teachers and good teachers are used as synonyms: they all refer to teacher characteristics and teaching practices that are robustly associated with higher student proficiency, as discussed in Chapters 2 and 3.
Hence, the following questions arise: How is the sorting of good teachers and effective practices related to inequalities in student outcomes? Do socio‑economic inequalities in student outcomes tend to be more moderate in education systems where good teachers are allocated more evenly across schools? Can education systems address inequalities in student outcomes by reallocating good teachers to more disadvantaged schools?
This chapter aims to answer these questions by correlating indicators of teacher allocation (from Chapters 2 and 3 of this report) with measures of inequalities in learning outcomes (as assessed by PISA).1 The TALIS indicators of teacher allocation include the dissimilarity index, which is a commonly used measure to analyse deviation from evenness (see Box 2.1 in Chapter 2 for more detail) and the difference between disadvantaged and advantaged schools in the share of effective teachers and teachers who use effective teaching practices. The dissimilarity index assesses departure from random allocation and is, therefore, useful for seeing if teachers with certain characteristics tend to be concentrated in a restricted number of schools. However, even if the dissimilarity index shows that teachers are allocated unevenly across schools, this would not necessarily mean that a school system is inequitable. Equitable education systems may deliberately allocate more resources (including effective teachers) to disadvantaged schools – those attended by students from socio‑economically disadvantaged backgrounds – as a way to compensate for less affluent parental background and hence to provide equal opportunities (OECD, 2019[5]). For this reason, the report also looks at average differences across schools in students’ exposure to effective teachers and teaching practices. Strikingly, based on findings from Chapters 2 and 3, whenever there is evidence for systematic sorting of effective teachers across schools, more often than not, it goes to the benefit of schools with a high share of students from a socio‑economically advantaged background.
The two angles, equality and equity, are complementary. Although the analysis looking at equality in students’ access to effective teachers and teaching practices disregards the characteristics of the students, it can still identify teacher characteristics and practices along which teachers tend to sort across schools. The dissimilarity index can highlight overall imbalances in teacher allocation. On the other hand, analysis focusing on equity draws a more detailed picture of teacher allocation. Notably, it examines how teachers with certain characteristics and practices are distributed across different types of schools.
PISA‑based measures of inequalities in student performance included in the analysis are the share of variance in reading performance explained by students’ socio‑economic profile; the difference between advantaged and disadvantaged students in reading; and the mean reading performance of students at the bottom quarter of students’ socio‑economic profile. This chapter highlights reading because it was the focus domain in the 2018 round of PISA, which means it was tested in more detail than the other two domains, which are mathematics and science. In addition to these indicators, TALIS measures of digital divides are also correlated with PISA‑based measures of socio‑economic inequalities in students’ digital literacy skills.
For the analyses presented in this chapter, it is important to highlight how PISA and TALIS measure the socio‑economic status of students. PISA relies on the index of economic, social and cultural status (ESCS), a composite measure that combines the financial, social, cultural and human‑capital resources available to students into a single score (OECD, 2019[5]). TALIS measures students’ socio‑economic status by asking principals to report the share of students in their school coming from socio‑economically disadvantaged homes. Throughout this report, "advantaged” schools are those in which 10% or less of the student body are reported to be socio‑economically disadvantaged and “disadvantaged” schools refer to those with more than 30% of the students from socio‑economically disadvantaged homes.
In addition, it should be noted that the system‑level correlational analyses presented in this chapter are somewhat limited due to the potential mismatch between the lower secondary teacher population covered by TALIS and the 15‑year‑old student population sampled by PISA. While TALIS covers lower‑secondary teachers, PISA collects information on 15‑year‑old students; yet, 15‑year‑olds are not taught by lower secondary teachers in every education system as upper secondary teachers sometimes fill this role. Moreover, it is also important to note that the findings of the system‑level correlational analyses cannot be interpreted as causal.
This chapter also tries to understand if there are system‑level policies that are associated with a more even and equitable sorting of teachers across schools. Namely, the chapter explores if school competition and school autonomy in hiring and dismissing teachers or determining teachers’ salaries can be an effective policy lever to address inequities in teacher sorting. Although the empirical evidence on whether school competition for students is beneficial for student achievement and equity in education is mixed (Boeskens, 2016[6]; OECD, 2020[7]; OECD, 2019[8]; Urquiola, 2016[9]), competition may provide incentives for schools to improve their instructional quality, including by competing for quality teachers. Thus, among other factors, the effectiveness of school competition on student performance and equity may depend on the level of autonomy schools have to hire, dismiss and remunerate their teachers. Past research also shows that school characteristics tend to be linked to teacher quality as schools with certain characteristics may find it difficult to hire and retain high‑quality teachers (Echazarra and Radinger, 2019[10]; OECD, 2018[11]). In general, schools with more challenging work environments or a higher share of socio‑economically and otherwise disadvantaged students are thought to face this issue (Allen and Sims, 2018[12]; Goldhaber, Lavery and Theobald, 2015[13]; Guarino, Santibañez and Daley, 2006[14]; Johnson, Kraft and Papay, 2012[15]; Loeb, Kalogrides and Horng, 2010[16]) as do rural schools far from urban centres (Beesley and Clark, 2015[17]; Brasche and Harrington, 2012[18]; Cowen et al., 2012[19]; Downes, 2018[20]; Fowles et al., 2013[21]).
The chapter is organised as follows. First, it looks at how TALIS measures of teacher allocation relate at the system level to socio‑economic inequality in student performance. The first section also explores the system‑level relationships between teacher allocation in relation to digital learning and socio‑economic inequality in students’ digital skills. Then, the chapter examines how TALIS measures of teacher allocation relate at the system level to school competition and autonomy.
How access to effective teachers is related to socio‑economic inequality in student performance
This section describes the system‑level relationships between differences in students’ access to effective teachers and socio‑economic inequality in student performance. More specifically, it looks at the relationship between PISA‑based measures of socio‑economic inequalities in student performance and imbalances in the allocation of effective teachers as defined with regard to teacher characteristics and practices also discussed in Chapter 2. The distinction between teachers’ characteristics and practices is that the former are somewhat fixed, portable assets that teachers always possess irrespective of the schools where they are employed. Practices are instead (at least partly) an explicit choice made by teachers teaching in a given context, and nothing ensures that they would adopt the same practices in a different school (or even with different students in the same school).
Well‑trained and experienced teachers, and reading scores
At the system level, the mean reading score in PISA tends to be negatively associated with the dissimilarity index for experienced teachers (i.e. teachers with more than ten years of teaching experience) across countries and territories (linear correlation coefficient (r) = -0.44) (Table 4.1). That is, at the system level, the uneven (non‑random) distribution of experienced teachers is associated with lower average reading scores. This suggests that experienced teachers are not directed to the schools that need them the most. There is probably scope to increase average scores by relocating experienced teachers from schools where they are in surplus to schools that do not have enough of them.
As highlighted in Chapter 2, experienced teachers are more likely to work in schools with few socio‑economically disadvantaged students (10% or less of the student body) than in schools where disadvantaged students constitute more than 30% of the student population in many of the countries participating in TALIS. The system‑level correlation also shows that an uneven distribution of experienced teachers is negatively associated (linear correlation coefficient (r) = -0.42) with the PISA reading score of the most disadvantaged students in the country (i.e. those in the bottom quarter of the distribution of the ESCS index in the country) (Figure 4.1). This reveals a tendency for lower reading scores among disadvantaged students when experienced teachers are not evenly distributed and clustered in schools that are predominantly socio‑economically advantaged.
Table 4.1. System‑level relationships between TALIS measures of teacher allocation and equity in student performance
System‑level correlation coefficients
Notes: System‑level correlation coefficients are calculated by correlating country‑level indicators that are based on TALIS and PISA data. TALIS indicators for Alberta (Canada), Ciudad Autónoma de Buenos Aires (hereafter CABA [Argentina]), England (UK) and Shanghai (China) are correlated with PISA‑based measures for Canada, Argentina, the United Kingdom and the four PISA-participating provinces/municipalities of China: Beijing, Shanghai, Jiangsu and Zhejiang, respectively.
Correlation coefficients that are equal to, or lower than -0.35, or else equal to, or higher than +0.35, are highlighted.
Correlation coefficients range from -1.00 (i.e. a perfect negative linear association) to +1.00 (i.e. a perfect positive linear association). When a correlation coefficient is 0, there is no linear relationship between the two measures.
1. A socio‑economically advantaged (disadvantaged) student is a student in the top (bottom) quarter of ESCS in his or her own country/territory.
2. ESCS refers to the PISA index of economic, social and cultural status.
3. Comprehensive formal education or training includes: content, pedagogy, classroom practice, cross‑curricular skills, teaching in mixed ability setting and classroom management (i.e. items a, b, c, d, e, g and i of Question 6 of the TALIS 2018 Teacher Questionnaire).
4. Restricted to countries and territories where the overall share of teachers with the specific characteristic analysed is 75% or less.
Source: TALIS 2018 database, Tables 2.3, 2.5, 2.6, 2.10, 2.8, 2.12; and OECD (2019[5]), PISA 2018 Results (Volume II): Where All Students Can Succeed, https://doi.org/10.1787/b5fd1b8f-en, Tables II.1 and II.B1.2.3.
Figure 4.1. Experienced teachers and students’ reading performance in PISA in the national bottom quarter of socio‑economic status
Notes: Linear correlation coefficient (R) =-0.42. The PISA‑based indicator for Alberta (Canada), CABA (Argentina), England (UK) and Shanghai (China) refers to the values for Canada, Argentina, the United Kingdom and the four PISA-participating provinces/municipalities of China: Beijing, Shanghai, Jiangsu and Zhejiang, respectively.
The dissimilarity index measures if the allocation of teachers in a country’s schools resembles the teacher population of the country. This index ranges from 0 (no segregation) to 1 (full segregation). Countries and territories where the overall share of experienced teachers is above 75% are excluded.
ESCS refers to the PISA index of economic, social and cultural status.
Source: OECD, TALIS 2018 Database, Table 2.3 and OECD (2019[5]), PISA 2018 Results (Volume II): Where All Students Can Succeed, https://doi.org/10.1787/b5fd1b8f-en, Table II.B1.2.3.
While a teacher’s experience may be assumed to affect the quality of teaching they provide, the nature and scope of their training also influences what teaching they can provide and what practices they choose to adopt (OECD, 2009[22]). There is a negative association (linear correlation coefficient (r) = -0.36) between the dissimilarity index for teachers who had a comprehensive formal education or training (including pedagogy, classroom practice, cross‑curricular skills, teaching in mixed ability setting and classroom management) and the mean reading score as measured by PISA 2018 (Table 4.1). In other words, the more concentrated teachers with comprehensive initial training are in certain schools, the worse students tend to perform on the PISA reading test. Moreover, the system‑level relationship between the allocation of teachers with comprehensive training across schools and reading performance is especially strong for the most socio‑economically disadvantaged students (i.e. those at the bottom quarter by socio‑economic background) (linear correlation coefficient (r) = -0.40) (Figure 4.2).
Policies can be put in place, though, to improve teaching quality in disadvantaged schools and areas. Box 4.1 provides three examples of such policies, and Box 4.2 suggests how legal reforms are a possible way to reduce inequity by evening out how good teachers are distributed, in this case in the United States.
Figure 4.2. Teachers with a comprehensive formal education and students’ reading performance in the national bottom quarter of socio‑economic status
Notes: Linear correlation coefficient (R) =-0.40. The PISA‑based indicator for Alberta (Canada), CABA (Argentina), England (UK) and Shanghai (China) refers to the values for Canada, Argentina, the United Kingdom and the four PISA-participating provinces/municipalities of China: Beijing, Shanghai, Jiangsu and Zhejiang, respectively.
The dissimilarity index measures if the allocation of teachers in a country’s schools resembles the teacher population of the country. This index ranges from 0 (no segregation) to 1 (full segregation). Countries and territories where the overall share of teachers who had a comprehensive formal education or training is above 75% are excluded.
Teachers who had a comprehensive formal education or training refer to teachers for whom content, pedagogy, classroom practice, cross‑curricular skills, teaching in mixed ability setting and classroom management were included in their formal education or training.
ESCS refers to the PISA index of economic, social and cultural status.
Source: OECD, TALIS 2018 Database, Table 2.5 and OECD (2019[5]), PISA 2018 Results (Volume II): Where All Students Can Succeed, PISA, https://doi.org/10.1787/b5fd1b8f-en, Table II.B1.2.3.
Box 4.1. Working to ensure high teacher quality in disadvantaged schools and areas
England (United Kingdom), Finland and Korea are three countries that combine high PISA reading scores with a low dissimilarity index for both experienced teachers and teachers who had comprehensive formal education (Figure 4.1 and Figure 4.2). This box gives a short overview of some recent policy initiatives in these countries that might have contributed to the even distribution of high‑quality teachers.
In the United Kingdom, the Department for Education has identified improving the quality of teaching in challenging areas and schools as a key to more equitable education. Various monetary incentives were introduced in 2010 to train, attract and retain strong teachers in the most challenging areas. Examples include: reimbursement of student loans for teachers teaching certain subjects in short supply; retention payments for mathematics teachers in challenging areas; and a greater focus of the Teach First programme in challenging areas. In addition, GBP 30 million was dedicated to supporting disadvantaged schools with considerable recruitment and retention issues. A series of grants and premiums were also introduced to support professional development available in challenging areas, emphasising that it should be evidence‑based training (Department for Education, 2017[23]).
Finland ran a large‑scale project called New Comprehensive Education in 2016‑19, which also addressed teacher education. As part of this, the Teacher Education Forum was created in 2016 under the auspices of the Ministry of Education and Culture to develop teacher education through co‑configurative collaboration with universities and other stakeholders. The forum formulated the Teacher Education Development Programme as a plan for teachers’ pre- and in‑service training, involving nearly 100 representatives from teacher education, the teachers’ union, local governments, researchers, principals and teachers. EUR 28 million was allocated to projects for starting the implementation of this programme. One example of a 2017 research‑oriented project was the development of pedagogical studies for teachers to enhance their’ competence base, especially in subject education, at the University of Helsinki. This was done through the creation of a digital database with co‑operation models for pedagogical studies (Ministry of Education and Culture, Finland, 2021[24]).
An interesting aspect of teacher training in Korea is the high prevalence of online professional development. This development started in 2000 and online teacher training is currently offered by more than 20 government‑authorised public and private training centres. Quality checks and co‑ordination are conducted by the Korean Education and Research Information Service (KERIS), which belongs to the Ministry of Education (Minea-Pic, 2020[25]). It is possible that this wide availability of online resources and training opportunities together with central co‑ordination has contributed to competence‑building across the board since it minimises the negative impact of physical distances and makes it easier to interact with colleagues in other areas. By bridging differences in professional development, an even distribution of well‑trained teachers is achieved.
Box 4.2. Legal reform for student equity in the United States
The United States is one of several countries that combines a high mean reading score for socio‑economically disadvantaged students with low dissimilarity index value for teachers who had comprehensive training (Figure 4.2).
The federal government has carried out a series of legal reforms aimed to reduce inequity, including the No Child Left Behind Act of 2001 and the Every Student Succeeds Act of 2015. The latter is concerned with four categories of disadvantaged students: students in poverty, minorities, students who receive special education and students with limited English language skills. It stipulates for the first time the requirement for all students across the United States to be taught to the same high academic standards, with associated assessments to measure student progress toward those standards. At the same time, the law recognises the differences in circumstances between states and includes local flexibility provisions that can be applied as long as the states in question formulate comprehensive plans for achieving the law’s equity goals. As such, states are granted more autonomy in designing and implementing general education standards. It also requires states to hold schools accountable in terms of achievement, including in digital learning. Importantly, it tasks states with identifying schools that struggle with achievement and formulating concrete plans to assist those schools (US Department of Education, 2015[26]).
It is not only thought to have contributed to further professionalisation of teachers but also to have distributed more high‑quality teachers to disadvantaged schools and areas via financial incentive (Boyd et al., 2008[27]). That said, the impact of legal reform on the federal level will necessarily be modulated by state‑level legislation relating to the distribution of educational opportunities (Knight, 2019[28]). Nonetheless, this highlights the importance of legal reform to support policy implementation.
Time spent on actual teaching and reading scores
A common complaint of teachers around the world is that they have a heavy administrative burden and spend time on maintaining order in the classroom. Both hinder them from conducting actual teaching (as opposed to other work‑related activities, such as lesson preparation and marking) (OECD, 2019[29]). This is important when considering that differences in time spent on actual teaching in the classroom can explain differences in student achievement in mathematics, science and reading between countries (Lavy, 2015[30]; OECD, 2021[31])
Teachers’ ability to maximise instruction time is a key component of classroom management (Ainley and Carstens, 2018[32]; Kane et al., 2010[33]; Stronge et al., 2007[34]). Its effect on students’ achievement, on the other hand, depends on the classroom climate, which can be partly outside of the teachers’ control. On average, higher instruction time is more beneficial in classrooms with a better climate (Rivkin and Schiman, 2015[35]). Regardless of whether this ability is constrained by the teacher’s own skills or by school characteristics, it is instructive to investigate how distributions of such teachers correlate with inequitable student performance.
Figure 4.3. Time spent on actual teaching and students’ reading performance in PISA in the national bottom quarter of socio‑economic status
Notes: Linear correlation coefficient (R) =-0.36. The PISA‑based indicator for Alberta (Canada), CABA (Argentina), England (UK) and Shanghai (China) refers to the values for Canada, Argentina, the United Kingdom and the four PISA-participating provinces/municipalities of China: Beijing, Shanghai, Jiangsu and Zhejiang, respectively.
The dissimilarity index measures if the allocation of teachers in a country’s schools resembles the teacher population of the country. This index ranges from 0 (no segregation) to 1 (full segregation).
ESCS refers to the PISA index of economic, social and cultural status.
Source: OECD, TALIS 2018 Database, Table 2.12 and OECD (2019[5]), PISA 2018 Results (Volume II): Where All Students Can Succeed, PISA, https://doi.org/10.1787/b5fd1b8f-en, Table II.B1.2.3.
In school systems where teachers who spend more class time on actual teaching are concentrated in a limited number of schools, the mean reading score of students tends to be lower, especially for the least privileged students (Table 4.1): the dissimilarity index for teachers who are in the top quarter based on class time spent on actual teaching and learning is negatively correlated with the mean reading score of students in the bottom quarter of socio‑economic status (linear correlation coefficient (r) = -0.36) (Figure 4.3). Based on the findings of Chapter 2, large and systematic differences are observed between different types of schools in the share of teachers who spend a high share of class time on instruction. Notably, teachers that spend a larger share of class time on actual teaching are more likely to work in advantaged schools as well as private schools. However, the system‑level relationship does not necessarily mean that exposing disadvantaged students to such teachers will improve their performance. There might be other factors that play a part; for example, advantaged schools might have fewer disciplinary problems in the classroom overall, which allows teachers to spend more time on actual teaching instead of classroom management.
Teachers’ digital self‑efficacy and ICT use, and students’ digital skills
TALIS measures of teacher allocation in relation to digital learning refer to differences in students’ access to teachers who are trained and have high self‑efficacy in ICT use and those who use ICT2 for teaching on a regular basis. Research suggests that teaching using ICT has the potential to improve student outcomes in various ways. It can allow self‑paced and individualised instruction; access to information and specialised materials well beyond what textbooks can offer; better tools for collaborative work; and project‑based and inquiry‑based pedagogies. ICT tools also increase students’ engagement (Bulman and Fairlie, 2016[36]; OECD, 2015[37]).
Yet, evidence on how ICT use at school affects student outcomes is mixed. Past research shows that the use of ICT at school does not automatically lead to better academic results (Bulman and Fairlie, 2016[36]; OECD, 2019[38]; OECD, 2015[37]). While moderate use of ICT in schools can be beneficial (OECD, 2015[37]), frequent use of technology can have the opposite effect and may be associated with lower student performance, whether in science, mathematics or reading (OECD, 2019[38]). A recent study by Borgonovi and Pokropek (2021[39]) also shows that students with either low or high levels of ICT use tend to have lower levels of reading achievement than those with moderate use of digital technology.
Nevertheless, ICT use at school can help students acquire digital skills (Bulman and Fairlie, 2016[36]). This includes fundamental skills such as understanding basic ICT concepts; being able to manage computer files; using keyboards or touch‑screen devices; using work‑related software; creating online content; evaluating online risks; and distinguishing fact from opinion (OECD, 2019[38]). Analyses based on PISA data show a positive association between students’ access to digital learning at school and their acquisition of digital skills (OECD, 2021[40]; OECD, 2015[37]). Importantly, those with poor acquisition of basic digital skills will find it difficult to navigate a digital world that is increasingly becoming central in everyday and work life (OECD, 2015[37]). This hits socio‑economically disadvantaged students harder as they consistently appear to have lower levels of digital literacy (Karpiński, 2021[41]). Yet, past studies have also shown that access to technology is not enough to improve student learning: effective integration of technology into teaching and learning requires teachers who are well‑trained and able to use digital tools for instruction (Fraillon et al., 2019[42]; OECD, 2021[40]; OECD, 2019[38]; OECD, 2015[37]). This section looks at system‑level patterns of teacher distribution that relate to students’ acquisition of digital skills.
Digital devices, especially those that are connected to the Internet, tend to offer more textual information and for a broader range of purposes but from different sources (OECD, 2021, p. 36[40]). Thus, reading in digital environments often requires navigating through multiple sources of text, selecting relevant information and assessing the quality of information (OECD, 2021, p. 36[40]). Multiple‑source items in the PISA 2018 computer‑based reading assessment, which are defined as having different authors, being published at different times or bearing different titles or reference numbers, provide a proxy measure for students’ digital skills when it comes to reading in a digital environment. PISA 2018 data show that being taught in school how to detect whether information is subjective or biased is positively associated with the estimated percentage correct in the item that focuses on distinguishing facts from opinions in the PISA reading assessment (OECD, 2021[40]).
The results of the analysis presented in Table 4.2 show that in school systems where teachers in disadvantaged schools are just as likely, or even more likely, to participate in professional development in the use of ICT for teaching, students who have access at home and in school to digital learning tools have a greater advantage over those who do not (linear correlation coefficient (r) = -0.49). This could be due to the fact that teachers’ proficiency in ICT yields higher returns among disadvantaged students (who are presumably less supported at home), but only to the extent that students have access to appropriate digital tools. Box 4.3 provides an example of how professional development in ICT can be organised to benefit schools across the nation.
Table 4.2. System‑level relationships between TALIS measures of digital divides and equity in students’ digital skills
System‑level correlation coefficients
Notes: System‑level correlation coefficients are calculated by correlating country‑level indicators that are based on TALIS and PISA data. TALIS indicators for Alberta (Canada), CABA (Argentina), England (UK) and Shanghai (China) are correlated with PISA‑based measures for Canada, Argentina, the United Kingdom and the four PISA-participating provinces/municipalities of China: Beijing, Shanghai, Jiangsu and Zhejiang, respectively.
Correlation coefficients that are equal to, or lower than -0.35, or else equal to, or higher than +0.35, are highlighted.
Correlation coefficients range from -1.00 (i.e. a perfect negative linear association) to +1.00 (i.e. a perfect positive linear association). When a correlation coefficient is 0, there is no linear relationship between the two measures.
1. ESCS refers to the PISA index of economic, social and cultural status.
2. A socio‑economically advantaged (disadvantaged) student is a student in the top (bottom) quarter of ESCS in his or her own country/territory.
3. Restricted to countries and territories where the overall share of teachers with the specific characteristic analysed is 75% or less.
Source: TALIS 2018 database, Tables 3.5, 3.7, 3.12, 3.15; and OECD (2021[40]), 21st‑Century Readers: Developing Literacy Skills in a Digital World, https://doi.org/10.1787/a83d84cb-en, Tables B.2.5 and B.2.6.
Another teacher characteristic that is worth examining is digital self‑efficacy. Self‑efficacy is the perception of one’s own capacity to perform a specific task well. It is separate from self‑confidence, which is a more general trait (Ainley and Carstens, 2018[32]). In this case, it is the confidence to use ICT in teaching. At the system level, teacher allocation across schools based on their self‑efficacy in ICT and their actual use of digital technology in the classroom are weakly correlated with students’ skills in navigating through multiple‑source text, selecting relevant information and assessing the quality of information (Table 4.2). Yet, the dissimilarity index for teachers with high self‑efficacy in the use of ICT is positively correlated with the difference between the most and the least advantaged students in terms of opportunity to learn digital literacy skills at school. Detecting if information is subjective or biased (linear correlation coefficients (r) = 0.49) is one aspect of these skills (Figure 4.4). As highlighted in Chapter 3, the share of teachers who feel they can support student learning through the use of digital technology and also of those who use ICT for instruction on a regular basis is larger in private than in public schools in almost a quarter of countries and territories participating in TALIS.
Distributing teachers who have high ICT self‑efficacy more evenly among schools may help provide disadvantaged students with the same opportunity to learn digital literacy skills as their peers from socio‑economically advantaged families. Teachers who feel that they can support student learning through the use of digital technology tend to be more motivated and able to offer such opportunity. Together with the general availability of adequate digital infrastructure, access to teachers with high ICT self‑efficacy can increase engagement in digital learning of both advantaged and disadvantaged students.
Figure 4.4. Teachers’ self‑efficacy in the use of ICT and students’ opportunity to learn digital literacy skills at school
Notes: Linear correlation coefficient (R) =-0.49. The PISA‑based indicator for CABA (Argentina), England (UK) and Shanghai (China) refers to the values for Argentina, the United Kingdom and the four PISA-participating provinces/municipalities of China: Beijing, Shanghai, Jiangsu and Zhejiang, respectively.
The dissimilarity index measures if the allocation of teachers in a country’s schools resembles the teacher population of the country. This index ranges from 0 (no segregation) to 1 (full segregation). Countries and territories where the overall share of teachers who feel they can support student learning through the use of digital technology “quite a bit” or “a lot” is above 75% are excluded.
ESCS refers to the PISA index of economic, social and cultural status.
ICT refers to information and communication technology.
Source: OECD, TALIS 2018 Database, Table 3.12 and OECD (2021[40]), 21st‑Century Readers: Developing Literacy Skills in a Digital World, https://doi.org/10.1787/a83d84cb-en, Table B.2.5.
A more even allocation of teachers who use ICT for teaching on a regular basis is positively associated with the difference between the most and the least advantaged students in having the opportunity to learn how to detect whether information is subjective or biased (linear correlation coefficients (r) = 0.45) (Table 4.2). That is, disadvantaged students tend to have just as much or more opportunity to learn such digital skills in countries where teachers who “frequently” or “always” use ICT for instruction are more evenly distributed across schools.
Box 4.3. Teacher professional development in ICT in Estonia
Estonia achieves high PISA reading scores and, as shown above, stands out as the only country where students who report having little or no access to digital learning perform better in reading multiple‑source texts. Estonia employs comprehensive national strategies for ICT use in schools, for example, as part of the Estonia Lifelong Learning Strategy 2020. This was launched in 2014 and its central goal is to incorporate “digital culture” at every school level (Ministry of Education and Research, Republic of Estonia, 2015[43]). According to the European Commission's 2nd Survey of Schools: ICT in Education, Estonia is one of Europe’s top countries when it comes to teacher involvement in ICT learning, both during their own time and as provided by their schools, ranking first (90%) and second (79%) respectively at lower secondary level (data.europa.eu, 2019[44]). In particular, ICT training is largely provided to teachers through HITSA (Hariduse Infotehnoloogia Sihtasutus), which is a non‑profit organisation jointly established by the Estonian national government, the University of Tartu and Tallinn University. It aims to foster educational leaders in a digital age by helping teachers to conduct goal‑oriented teaching using ICT. It also formulates visions and action plans that incorporate digital tools (European Schoolnet, 2015[45]).
How access to effective teachers is related to school competition and autonomy
This section is concerned with the relationship between system‑level characteristics and policies, and the allocation of effective teachers and teaching practices. In particular, the focus will be on the degree of autonomy that schools enjoy, and on the degree of competition they face. Both aspects are likely to have an impact on the management of human resources in schools, in particular the hiring, training and firing of teachers.
The TALIS 2018 Principal Questionnaire includes questions on how many schools are competing for the same students and to what degree principals feel that their school has autonomy over the appointment or hiring of teachers; dismissal or suspension of teachers; and determining teachers’ salaries. In TALIS 2018, “autonomous” means that significant responsibility is taken solely by a principal, other members of the school management team or teachers not part of the school management team (OECD, 2020[46]).
There is no theoretical reason to expect that the relationship between equity and school competition goes in one direction or another. Policies that increase school choices may allow disadvantaged students to opt for high‑performing schools outside of their immediate neighbourhoods. But it can also allow schools to select only the best students (Boeskens, 2016[6]; OECD, 2020[7]; OECD, 2019[8]; Urquiola, 2016[9]). At the same time, competition among schools has the potential to incentivise them to hire and retain better teachers as long as they have the autonomy to do so. Policy reviews have noted that school autonomy for staff‑related tasks can help avoid misallocations and can better match staff profiles to the needs of the school. However, an increase in autonomy entails recruitment and management costs that may lead to greater disparities in staff qualifications among schools (OECD, 2020[46]).
Importantly, the reduction in misallocation can occur in ways that are difficult to observe, such as the ability of teachers to work with a particular type of students. Reducing misallocation along these unobservable – but relevant – dimensions can go hand in hand with an increase in imbalances in other dimensions that are easier to observe but maybe less relevant. As such, while it is reasonable to expect less misallocation in systems with greater school autonomy, that same autonomy may lead to more imbalances between schools as teachers opt for schools with certain characteristics more than others. However, countries that increase autonomy can also, at the same time, put in place compensatory mechanisms to help high‑need schools attract and retain good teachers: this is one way to interpret results from PISA showing that increased autonomy in staffing practices is not necessarily associated with more inequality (OECD, 2018[11]). Moreover, past findings also suggest that school autonomy is positively associated with greater equity in student performance if it is accompanied by higher levels of accountability (OECD, 2018[11]; OECD, 2016[47]; Torres, 2021[48]).
This section explores associations between school competition and autonomy, and students’ access to effective teachers and teaching practices to see if relevant school management policy measures could potentially facilitate more equitable learning on the system level.
According to system‑level correlational analysis, the association between school competition and school autonomy in hiring, dismissing and determining teachers’ salaries and TALIS measures of teacher allocation is weak. As shown in Table 4.3, in general, there is hardly any correlation between indicators of school competition and autonomy, and the allocation of teachers, based on neither their characteristics nor the practices they use. The only exception to this pattern is the sorting of experienced teachers across schools. One possible reason for this exception is that, in systems that are very centralised and where schools have little autonomy, experience is one of the main criteria used to allocate teachers (seniority‑based system). In these systems, more experienced teachers are able, after many years of career, to move to more desirable schools such as advantaged and urban schools. On the other hand, systems with more school autonomy might have more diversity in terms of teacher characteristics because a wider range of criteria is taken into account in the hiring process. The decentralisation of hiring can allow assessment of a wider range of candidates’ characteristics, thus reducing the relative importance of other elements like experience. This is precisely one way more autonomy can reduce misallocations. This would explain why experienced teachers are distributed differently between systems with high and low school autonomy.
Across TALIS participants, the higher the share of principals within a country who report that their school has autonomy in appointing or hiring teachers, the more evenly experienced teachers tend to be distributed across schools (linear correlation coefficient (r) = -0.51) (Figure 4.5). Differences in the share of principals within a country who report that their school has autonomy in appointing or hiring teachers account for 26% of the differences in the dissimilarity index for experienced teachers. Similarly, the higher the share of principals within a country who report that their school has autonomy in dismissing or suspending teachers from employment, the more evenly experienced teachers tend to be distributed across schools (linear correlation coefficient (r) = -0.47) (Table 4.3). These findings suggest that higher school autonomy in staffing practices can result in a more equal distribution of teachers across schools. Past research has found that higher levels of school autonomy in managing teachers tend to produce a more equitable sorting of teachers across schools (OECD, 2018[11]). Yet, disadvantaged schools may need monetary or other support to be able to attract and retain the teachers they want (OECD, 2018[11]).
Box 4.4 points to the examples of Brazil, China, Turkey and Japan to illustrate how various incentives as well as mandatory teacher allocation policies can be used to achieve different outcomes in teacher distribution.
Table 4.3. System‑level relationships between TALIS measures of teacher allocation, and school competition and autonomy
System‑level correlation coefficients
Notes: System‑level correlation coefficients are calculated by correlating country‑level indicators that are based on TALIS and PISA data. Correlation coefficients that are equal to, or lower than -0.35, or else equal to, or higher than +0.35, are highlighted.
Correlation coefficients range from -1.00 (i.e. a perfect negative linear association) to +1.00 (i.e. a perfect positive linear association). When a correlation coefficient is 0, there is no linear relationship between the two measures.
1. Restricted to countries and territories where the overall share of teachers with the specific characteristic analysed is 75% or less.
Source: TALIS 2018 database, Tables 2.3, 2.5, 2.6, 2.10, 2.8, 2.12; and OECD (2020[46]), TALIS 2018 Results (Volume II): Teachers and School Leaders as Valued Professionals, https://doi.org/10.1787/19cf08df-en, Table II.5.1.
Figure 4.5. Experienced teachers and school autonomy in appointing or hiring teachers
Notes: Linear correlation coefficient (R) =-0.51.
The dissimilarity index measures if the allocation of teachers in a country’s schools resembles the teacher population of the country. This index ranges from 0 (no segregation) to 1 (full segregation). Countries and territories where the overall share of experienced teachers is above 75% are excluded.
Source: OECD, TALIS 2018 Database, Table 2.3 and OECD (2020[46]), TALIS 2018 Results (Volume II): Teachers and School Leaders as Valued Professionals, https://doi.org/10.1787/19cf08df-en, Table II.5.1.
Box 4.4. Incentives to attract quality teachers and mandatory teacher transfers
Incentive‑based policies
In a fairly unregulated teacher labour market, it is possible to introduce various incentives to try to attract good teachers to disadvantaged schools that find it difficult to hire and retain such teachers due to factors like disruptive learning environments, negative reputation, and geographical remoteness. The following are a few examples of incentive‑based policies.
In São Paulo state, Brazil, the government has implemented what is known as the ALE (Adicional por Local de Exercício) programme since 2008 (Secretaria de Orçamento e Gestão, Governo do Estado de São Paulo, 2018[49]). As part of this, salary premiums between 24% and 36% of the base salary are offered to teachers working in disadvantaged schools. Disadvantaged schools are categorised as being located either in rural areas or in socially vulnerable areas peripheral to a large urban centre. This measure has been found to reduce teacher turnover by 8.3 percentage points in public schools in São Paulo (Camelo and Ponczek, 2021[50]).
In China, career‑related incentives are used to attract teachers to remote areas. The Special Teaching Post Plan for Rural Schools was initiated in 2006 and is based on the recruitment of university graduates to remote areas in central and western China with large minority populations and socio‑economic disadvantage. The contract is for three years, after which the teachers are asked to take a test. Those who qualify are then given the opportunity to stay and take up a tenure track position. In 2015, around 90% of the teachers who finished the three‑year period stayed in their schools (OECD, 2016[51]).
Mandatory teacher transfers
An alternative approach to the above is to control teacher allocation and teacher transfers centrally. Although this space is not sufficient to properly discuss the topic, involuntary transfers of teachers between schools that have different characteristics has the potential of serving as a tool with the specific aim of reducing inequities. Here, involuntary simply means that there are institutional mechanisms in place that either assign teachers to workplaces or prevent them from remaining in one school for a longer period of time. Although data are scarce, there are indications such as those from an American county that if principals in disadvantaged schools are allowed to guide transfers of low‑performing teachers to more advantaged schools, the teachers that replace them will often perform better when measuring factors like work absence and value added to students’ achievement in mathematics and reading (Grissom, 2014[52]). Still, more research is needed before definite conclusions can be drawn. Nonetheless, it is useful to note this potential when considering the examples of Turkey as well as Japan below, contrasting them with approaches that depend more on the forces of an open teacher labour market.
Interestingly, while both Turkey and Japan have one of the lowest percentages of principals who reported autonomy in teacher appointments, Turkey displays a high dissimilarity index for experienced teachers, unlike Japan that does not (Figure 4.6).
The Turkish school system is highly centralised in terms of decision making and is markedly bigger in size than other similarly centralised systems in Europe like Greece and Luxembourg. Likewise, Turkey has a highly centralised system of teacher allocation based on initial assignments of new teachers and seniority‑based transfers (Kitchen et al., 2019[53]). This means that teachers are assigned to schools by the Turkish Ministry of National Education (MoNe) early in their career but then gain increasing freedom to transfer to schools as they wish when they accrue points in a seniority score system. In addition to this, the system employs various incentives to make more remote and undeveloped regions attractive such as higher gains in seniority score by working in those areas (Ozoglu, 2015[54]).
In 2015, Turkey introduced a new probation appraisal and induction programme for trainee teachers. As part of this, they are assigned to a school where they conduct practical work and have a supervisor who mentors them during this period (Kitchen et al., 2019[53]). This is meant to give them better preparation before they become certified teachers as their initial assignments by MoNe will often be rural, disadvantaged schools that can be challenging for novice teachers.
Japan is an OECD high‑achiever in PISA reading. It also has fairly high mean reading performance for most disadvantaged students combined with a low dissimilarity index of experienced teachers (Figure 4.1). Among the countries with a low percentage who reported school competition over students, Japan has the highest difference between disadvantaged and advantaged schools in how many self‑efficacy teachers they have (Figure 4.6).
Somewhat similar to Turkey, Japan uses a mandatory mobility scheme where teachers hired on the prefectural level are regularly assigned to new schools across municipalities in the same prefecture, which has the effect of regular high turnover. The stated aims of this policy include balancing attributes like age and gender in the teacher populations of schools, giving teachers varied professional experience and achieving a more equal spread of educational quality. The system is not uniform across Japan, however, as prefectural and municipal boards of education work together and allocate responsibilities in a variety of ways with different rules as to how often teachers should be transferred and according to what criteria. For example, in Iwate prefecture teachers are transferred after having worked in a remote area for three years, and in Osaka prefecture principals can initiate a process to retain teachers for more than ten years (Numano, 2017[55]).
In a system such as this, high turnover can have an equalising effect if transfers are random in the sense that the distribution of various teacher characteristics can become even. If, however, the transfers are guided by set criteria, the same measures can enhance equity by ensuring that useful teacher characteristics benefit schools and areas that need them the most. Of course, this is based on the assumption that criteria and transfer mechanisms are fine‑tuned enough to achieve the desired outcomes.
Other associations of note are those between schools’ autonomy in determining teachers’ salary increases and dissimilarity indices for teachers who frequently use practices relating to clarity of instruction (linear correlation coefficient (r) = 0.37) (Table 4.3). The pattern here is, in a way, opposite to the above finding about the distribution of experienced teachers. But, as discussed previously, a high dissimilarity index can also reflect a deliberate policy of supporting schools that are most in need. Clarity of instruction is also a typical example of something that is difficult to observe. Only schools that have higher autonomy would be able to identify teachers that engage in these practices and reward them. So, school autonomy with regard to salary setting can act as a powerful tool for disadvantaged schools, who can attract teachers who frequently use these practices. However, such policy would likely require monetary support for disadvantaged schools so that they can pay higher salaries.
In general, there is no strong correlation between system‑level policies such as school competition and school autonomy in hiring; dismissing and determining teachers’ salaries; and the sorting of teachers who are trained in and feel capable of using ICT or who teach using digital technology. Yet, the system‑level correlational analysis suggests that the higher the share of schools within a country that has autonomy in appointing and hiring teachers, the greater the concentration of teachers whose professional development included ICT skills for teaching (linear correlation coefficient (r) = 0.41) (Table 4.4). The analysis also shows that differences between disadvantaged and advantaged schools in terms of the share of teachers whose professional development included ICT skills for teaching is negatively correlated with the share of principals who reported having the autonomy to set teachers’ starting salaries and to increase teacher salaries (linear correlation coefficients (r) = -0.46 and -0.37, respectively) (Table 4.4). These correlations indicate that education systems where schools have more autonomy in hiring may have a more uneven distribution of teachers who receive in‑service training in ICT use, with these teachers going to more advantaged schools – at least when the autonomy has to do with salary setting.
Table 4.4. System‑level relationships between TALIS measures of digital divides, and school competition and autonomy
System‑level correlation coefficients
Notes: System‑level correlation coefficients are calculated by correlating country‑level indicators that are based on TALIS and PISA data.
Correlation coefficients that are equal to, or lower than -0.35, or else equal to, or higher than +0.35, are highlighted.
Correlation coefficients range from -1.00 (i.e. a perfect negative linear association) to +1.00 (i.e. a perfect positive linear association). When a correlation coefficient is 0, there is no linear relationship between the two measures.
1. Restricted to countries and territories where the overall share of teachers with the specific characteristic analysed is 75% or less.
Source: TALIS 2018 database, Tables 3.5, 3.7, 3.12, 3.15; and OECD (2020[46]), TALIS 2018 Results (Volume II): Teachers and School Leaders as Valued Professionals, https://doi.org/10.1787/19cf08df-en, Table II.5.1.
The system‑level analysis also shows that differences between disadvantaged and advantaged schools in terms of the share of teachers with high self‑efficacy in ICT use is negatively correlated with the share of principals who reported that two or more schools in their district were competing for their students (linear correlation coefficient (r) = -0.40) (Figure 4.6). Thus, in education systems where there is more competition for students among schools, teachers who are self‑efficacious in the use of digital technologies tend to sort into advantaged schools. The empirical evidence on the effect of school competition on teacher quality is mixed. There are studies showing that “more competition can increase teacher quality, particularly for schools serving predominantly lower‑income students” (Hanushek and Rivkin, 2003[56]). This may be the case if competition enhances the productivity of disadvantaged schools more than it benefits advantaged schools. Competition can provide incentives for considerable improvements in disadvantaged schools’ hiring, retention, monitoring and other teacher management practices. However, increased competition across schools can also result in more disparities in teacher quality, in favour of socio‑economically advantaged schools. In general, these schools are assumed to be more effective in attracting and retaining good teachers. Yet, as with all other findings presented in this chapter, one should be cautious in interpreting the results, which are only correlational and not causal. The observed system‑level correlation between school competition and the differences in the share of teachers with high self‑efficacy in ICT use between disadvantaged and advantaged schools may be a result of mediating factors. For example, in education systems where school competition is common, the gap in the quality of ICT infrastructure between advantaged and disadvantaged schools may be larger, which, in turn, is related to differences in teachers’ self‑efficacy in ICT use between disadvantaged and advantaged schools. Results from Chapter 3 show that, on average across the OECD, the share of teachers with high self‑efficacy in ICT use is higher in schools where the quality of instruction is not hindered by a shortage or inadequacy of digital technology, or insufficient Internet access (Table 3.1).
Figure 4.6. Teachers’ self‑efficacy in the use of ICT and school competition for students
Notes: Linear correlation coefficient (R) =-0.40.
Disadvantaged schools refer to schools with more than 30% of students from socio‑economically disadvantaged homes. Advantaged schools refer to schools with less than or equal 10% of students from socio‑economically disadvantaged homes.
ICT refers to information and communication technology.
Source: OECD, TALIS 2018 Database, Table 3.12.
References
[32] Ainley, J. and R. Carstens (2018), “Teaching and Learning International Survey (TALIS) 2018 Conceptual Framework”, OECD Education Working Papers, No. 187, OECD Publishing, Paris, https://dx.doi.org/10.1787/799337c2-en.
[12] Allen, R. and S. Sims (2018), “Do pupils from low-income families get low-quality teachers? Indirect evidence from English schools”, Oxford Review of Education, Vol. 44/4, pp. 441-458, https://doi.org/10.1080/03054985.2017.1421152.
[17] Beesley, A. and T. Clark (2015), “How rural and nonrural principals differ in High Plains U.S. states”, Peabody Journal of Education, Vol. 90/2, pp. 242-249, https://doi.org/10.1080/0161956X.2015.1022114.
[6] Boeskens, L. (2016), “Regulating Publicly Funded Private Schools: A Literature Review on Equity and Effectiveness”, OECD Education Working Papers, No. 147, OECD Publishing, Paris, https://dx.doi.org/10.1787/5jln6jcg80r4-en.
[39] Borgonovi, F. and M. Pokropek (2021), “The evolution of the association between ICT use and reading achievement in 28 countries”, Computers and Education Open, Vol. 2, p. 100047, https://doi.org/10.1016/J.CAEO.2021.100047.
[27] Boyd, D. et al. (2008), “The narrowing gap in New York city teacher qualifications and its implications for student achievement in high-poverty schools”, Journal of Policy Analysis and Management, Vol. 27/4, pp. 793-818, https://doi.org/10.1002/pam.20377.
[18] Brasche, I. and I. Harrington (2012), “Promoting teacher quality and continuity:Tackling the disadvantages of remote indigenous schools in the northern territory”, Australian Journal of Education, Vol. 56/2, pp. 110-125, https://doi.org/10.1177/000494411205600202.
[36] Bulman, G. and R. Fairlie (2016), “Chapter 5 - Technology and education: Computers, software, and the Internet”, in Hanushek, E. (ed.), Handbook of the Economics of Education, Elsevier, Amsterdam, https://doi.org/10.1016/B978-0-444-63459-7.00005-1.
[50] Camelo, R. and V. Ponczek (2021), “Teacher turnover and financial incentives in underprivileged schools: Evidence from a compensation policy in a developing country”, Economics of Education Review, Vol. 80, https://doi.org/10.1016/j.econedurev.2020.102067.
[19] Cowen, J. et al. (2012), “Teacher retention in Appalachian schools: Evidence from Kentucky”, Economics of Education Review, Vol. 31/4, pp. 431-441, https://doi.org/10.1016/J.ECONEDUREV.2011.12.005.
[44] data.europa.eu (2019), 2nd Survey of Schools: ICT in Education, Directorate-General for Communications Networks, Content and Technology, https://data.europa.eu/data/datasets/2nd-survey-of-schools-ict-in-education?locale=en (accessed on 19 October 2021).
[23] Department for Education (2017), Unlocking Talent, Fulfilling Potential: A Plan for Improving Social Mobility, Department for Education, UK, ref. Cm 9541, https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/667690/Social_Mobility_Action_Plan_-_for_printing.pdf.
[20] Downes, N. (2018), “Revisiting the schoolhouse: A literature review on staffing rural, remote and isolated schools in Australia 2004-2016”, Australian and International Journal of Rural Education, Vol. 28/1, pp. 31-54, https://journal.spera.asn.au/index.php/AIJRE/article/view/112.
[10] Echazarra, A. and T. Radinger (2019), “Learning in rural schools: Insights from PISA, TALIS and the literature”, OECD Education Working Papers, No. 196, OECD Publishing, Paris, https://dx.doi.org/10.1787/8b1a5cb9-en.
[45] European Schoolnet (2015), Estonia: Country Report on ICT in Education, European Schoolnet (EUN), Brussels, http://www.eun.org/documents/411753/828792/Country+Report+Estonia+2015.pdf/c7822e64-16be-43f4-bd97-d0c296b4d7d3.
[21] Fowles, J. et al. (2013), “Public employee quality in a geographic context: A study of rural teachers”, The American Review of Public Administration, Vol. 44/5, pp. 503-521, https://doi.org/10.1177/0275074012474714.
[42] Fraillon, J. et al. (2019), Preparing for Life in a Digital World: IEA International Computer and Information Literacy Study 2018 International Report,, Springer Nature, Cham, https://doi.org/10.1007/978-3-030-38781-5.
[13] Goldhaber, D., L. Lavery and R. Theobald (2015), “Uneven playing field? Assessing the teacher quality gap between advantaged and disadvantaged students”, Educational Researcher, Vol. 44/5, pp. 293-307, https://doi.org/10.3102/0013189X15592622.
[52] Grissom, L. (2014), “Strategic involuntary teacher transfers and teacher performance: Examining equity and efficiency”, Journal of Policy Analysis and Management, Vol. 33/1, pp. 112-140, https://www.jstor.org/stable/24033298.
[14] Guarino, C., L. Santibañez and G. Daley (2006), “Teacher recruitment and retention: A review of the recent empirical literature”, Review of Educational Research, Vol. 76/2, pp. 173-208, https://doi.org/10.3102/00346543076002173.
[56] Hanushek, E. and S. Rivkin (2003), “Does public school competition affect teacher quality?”, in Hoxby, C. (ed.), The Economics of School Choice, University of Chicago Press, Chicago, IL, http://www.nber.org/chapters/c10084.
[3] Hanushek, E. et al. (2015), “Returns to skills around the world: Evidence from PIAAC”, European Economic Review, Vol. 73, pp. 103-130, https://doi.org/10.1016/j.euroecorev.2014.10.006.
[15] Johnson, S., M. Kraft and J. Papay (2012), “How context matters in high-need schools: The effects of teachers’ working conditions on their professional satisfaction and their students’ achievement”, Teachers College Record, Vol. 114/10, pp. 1-39, https://www.tcrecord.org/content.asp?contentid=16685.
[33] Kane, T. et al. (2010), “Identifying Effective Classroom Practices Using Student Achievement Data”, NBER Working Paper Series, No. 15803, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/W15803.
[41] Karpiński, D. (2021), “Computational Thinking, Socioeconomic Gaps, and Policy Implications”, IEA Compass: Briefs in Education, No. 12, IEA, Amsterdam, https://www.iea.nl/sites/default/files/2021-01/Compass%2012-Computational%20thinking%2C%20socioeconomic%20gaps%20and%20policy%20implications_0.pdf.
[53] Kitchen, H. et al. (2019), OECD Reviews of Evaluation and Assessment in Education: Student Assessment in Turkey, OECD Reviews of Evaluation and Assessment in Education, OECD Publishing, Paris, https://dx.doi.org/10.1787/5edc0abe-en.
[28] Knight, D. (2019), “Are school districts allocating resources equitably? The Every Student Succeeds Act, teacher experience gaps, and equitable resource allocation”, Educational Policy, Vol. 33/4, pp. 615-649, https://doi.org/10.1177/0895904817719523.
[30] Lavy, V. (2015), “Do differences in schools’ instruction time explain international achievement gaps? Evidence from developed and developing countries”, The Economic Journal, Vol. 125/588, pp. F397-F424, https://doi.org/10.1111/ECOJ.12233.
[16] Loeb, S., D. Kalogrides and E. Horng (2010), “Principal preferences and the uneven distribution of principals across schools”, Educational Evaluation and Policy Analysis, Vol. 32/2, pp. 205-229, https://doi.org/10.3102/0162373710369833.
[25] Minea-Pic, A. (2020), “Innovating teachers’ professional learning through digital technologies”, OECD Education Working Papers, No. 237, OECD Publishing, Paris, https://dx.doi.org/10.1787/3329fae9-en.
[24] Ministry of Education and Culture, Finland (2021), New comprehensive education, https://okm.fi/en/new-comprehensive-education (accessed on 1 December 2021).
[43] Ministry of Education and Research, Republic of Estonia (2015), OECD Review of Policies to Improve the Effectiveness of Resource Use in Schools: Country Background Report, Estonia, Ministry of Education and Research, Republic of Estonia, http://www.oecd.org/education/school/EST_Country_background_report_final_30.11.15_Version2.pdf.
[55] Numano, T. (2017), Teacher Transfer and Appointment System in Japan: Reference from Personnel Transfer Policies and Guidelines, https://www.nier.go.jp/English/educationjapan/pdf/201703TTASJ.pdf.
[40] OECD (2021), 21st-Century Readers: Developing Literacy Skills in a Digital World, PISA, OECD Publishing, Paris, https://dx.doi.org/10.1787/a83d84cb-en.
[31] OECD (2021), Positive, High-achieving Students?: What Schools and Teachers Can Do, TALIS, OECD Publishing, Paris, https://dx.doi.org/10.1787/3b9551db-en.
[7] OECD (2020), PISA 2018 Results (Volume V): Effective Policies, Successful Schools, PISA, OECD Publishing, Paris, https://dx.doi.org/10.1787/ca768d40-en.
[46] OECD (2020), TALIS 2018 Results (Volume II): Teachers and School Leaders as Valued Professionals, TALIS, OECD Publishing, Paris, https://dx.doi.org/10.1787/19cf08df-en.
[8] OECD (2019), Balancing School Choice and Equity: An International Perspective Based on Pisa, PISA, OECD Publishing, Paris, https://dx.doi.org/10.1787/2592c974-en.
[38] OECD (2019), OECD Skills Outlook 2019: Thriving in a Digital World, OECD Publishing, Paris, https://dx.doi.org/10.1787/df80bc12-en.
[5] OECD (2019), PISA 2018 Results (Volume II): Where All Students Can Succeed, PISA, OECD Publishing, Paris, https://dx.doi.org/10.1787/b5fd1b8f-en.
[29] OECD (2019), TALIS 2018 Results (Volume I): Teachers and School Leaders as Lifelong Learners, TALIS, OECD Publishing, Paris, https://dx.doi.org/10.1787/1d0bc92a-en.
[11] OECD (2018), Effective Teacher Policies: Insights from PISA, PISA, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264301603-en.
[1] OECD (2018), Equity in Education: Breaking Down Barriers to Social Mobility, PISA, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264073234-en.
[51] OECD (2016), Education in China: A Snapshot, OECD, Paris, https://www.oecd.org/china/Education-in-China-a-snapshot.pdf.
[47] OECD (2016), PISA 2015 Results (Volume II): Policies and Practices for Successful Schools, PISA, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264267510-en.
[4] OECD (2016), Skills Matter: Further Results from the Survey of Adult Skills, OECD Skills Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264258051-en.
[37] OECD (2015), Students, Computers and Learning: Making the Connection, PISA, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264239555-en.
[22] OECD (2009), Creating Effective Teaching and Learning Environments: First Results from TALIS, TALIS, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264068780-en.
[2] OECD (2006), “Recruiting, Selecting and Employing Teachers”, in Teachers Matter: Attracting, Developing and Retaining Effective Teachers, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264018044-6-en.
[54] Ozoglu, M. (2015), “Teacher allocation policies and the unbalanced distribution of novice and senior teachers across regions in Turkey”, The Australian Journal of Teacher Education, Vol. 40/10, https://doi.org/10.14221/ajte.2015v40n10.2.
[35] Rivkin, S. and J. Schiman (2015), “Instruction time, classroom quality, and academic achievement”, The Economic Journal, Vol. 125/588, pp. F425-F448, https://doi.org/10.1111/ECOJ.12315.
[49] Secretaria de Orçamento e Gestão, Governo do Estado de São Paulo (2018), Adicional de Local de Exercício - Quadro do Magistério [Allowance for Teaching Location – Teaching Career], http://vclipping.planejamento.sp.gov.br/Vclipping1/index.php/Adicional_de_Local_de_Exerc%C3%ADcio_-_Quadro_do_Magist%C3%A9rio (accessed on 2 December 2021).
[34] Stronge, J. et al. (2007), “What is the Relationship Between Teacher Quality and Student Achievement? An Exploratory Study”, Journal of Personnel Evaluation in Education, Vol. 20/3-4, pp. 165-184, https://doi.org/10.1007/s11092-008-9053-z.
[48] Torres, R. (2021), “Does test-based school accountability have an impact on student achievement and equity in education?: A panel approach using PISA”, OECD Education Working Papers, No. 250, OECD Publishing, Paris, https://dx.doi.org/10.1787/0798600f-en.
[9] Urquiola, M. (2016), “Chapter 4 - Competition among schools: Traditional public and private schools”, in Hanushek, E., S. Machin and L. Woessmann (eds.), Handbook of the Economics of Education, Elsevier, Amsterdam, https://doi.org/10.1016/B978-0-444-63459-7.00004-X.
[26] US Department of Education (2015), Every Student Succeeds Act (ESSA), https://www.ed.gov/essa?src=rn (accessed on 19 October 2021).
Notes
← 1. The system-level correlational analyses presented in this chapter include correlations of country-level indicators that are based on TALIS and PISA data.
← 2. ICT refers to tools that can be used for projects or class work as defined in the TALIS 2018 Teacher Questionnaire, which makes it a broad concept.