Interplay of factors that contribute to ESL at school level

Thursday 23 July 2015, by Maša Vidmar

Although ESL is impacted by the composition of a school (e.g. mean SES) and its structure (e.g. size), school practices, especially the way curriculum is delivered in practice as well as caring, supportive and respectful teachers (and other school staff) who believe in students’ ability to succeed, seem particularly important for ensuring (potential) ESLers remain in school.

In recent decades, a growing body of literature has identified factors at the school level related to ESL. The aim of this article is to review the empirical research in the field and show a variety of school factors that may contribute to becoming an ESLer. Research indicates that schools can make a difference in students’ ability to persist in school. The specific characteristics of school composition (e.g. mean SES, share of students with a minority/migrant background) and school structure (e.g. size, private/public) were found to impact ESL, but these effects are likely confounded with other school characteristics, particularly school practices.

With regard to school practices, ESLers (or those at risk) often report experiencing the curriculum as too complicated, too academic, and disconnected from real life. Innovative provision of the curriculum is called for – different pedagogic approaches are proposed (e.g. recognition of informal learning, the use of technology, collaborative approaches through outdoor activities, teaching through arts and sports). Moreover, within the context of the school climate, caring, supportive and respectful teachers (along with other school staff) who believe in students’ ability to academically succeed have a significant impact on the lives of students at-risk for ESL and their decision to remain at school. The opposite effect was found for splits in the communication between students and teachers. Students encounter important support or discouragement already in everyday school activities and interactions, thus financial resources or time outside the classroom is not required. Interestingly, a climate of respect and caring is often intended or assumed by the adults; however, it is not necessarily experienced by the student in this way. This issue indicates the need to strengthen teachers’ initial and continuous education in this respect.


As described in the two articles on ESL factors associated with: (1) an individual, their family and social background; and (2) to system characteristics, the phenomenon and process of ESL can be viewed from the perspective of Bronfenbrenner’s (1996) ecological systems theory of human development. Bronfenbrenner recognised that many different levels of environmental influence exist and impact one’s development. He identified five environmental systems with which individuals interact: the microsystem (i.e. people and institutions in the immediate surroundings – family, peers, school), the mesosystem (i.e. interconnections between settings in which the child participates, e.g. home-school relations), the exosystem (processes and settings with an indirect influence on the individual, e.g. the neighbourhood-community context, education system, social welfare), the macrosystem (i.e. cultural influences) and the cronosystem (i.e. events and changes over time) (see Figure 1). This multidimensional framework presents the basis for understanding risk and protective ESL factors and the multifaceted nature of ESL. In the present article, we look in particular at one specific setting within the microsystem – the school – and review the scientific literature examining its impact on ESL.

JPEG - 88.9 kb
Figure 1. Bronfenbrenner’s model of human development (adapted from Pinterest). The black rectangle constitutes the focus of the article.

Early studies on ESL mostly considered risk or predictive ESL factors within the individual and their family (see e.g. Ensminger & Slusarcick, 1992; Barclay & Doll, 2001 for studies in the period 1950–1970). Later studies revealed ESL to be a process without any one individual cause and may already start in primary school (e.g. Alexander, Entwisle, & Horsey, 1997; McGarr, 2010) or even as early as the toddlers years (e.g. Jimerson, Egeland, Sroufe, & Carlson, 2000). A considerable amount was known about how the individual’s characteristics and their specific background affected the ESL process, yet much less was known about the influences of the school. But as understanding grew that ESL is a complex phenomenon entailing predictors from multiple levels of the individual’s ecology, in the last few decades empirical and review studies have expanded to also include other levels and other settings from the microsystem level, for instance the school (i.e. school or institutional factors; for example, see Audas & Willms, 2001; Janosz, LeBlanc, Boulerice, & Tremblay, 1997; Knesting, 2008; Lee & Burkham, 2003). Several factors within schools were recognised as playing a role in ESL, with some 20% of the variability in student outcomes explained by the characteristics of the schools students attend (Rumberger & Lim, 2008). Many studies are based on multi-level datasets (i.e. samples of students within schools within countries/districts) that enable researchers to disentangle student-level and school-level effects. However, establishing a causal relationship between the variable of interest and ESL remains problematic. School factors belong to different levels of the school environment (the student in the classroom, the school buildings, the school district etc.) and the ways these levels interact also impact a student’s development (Eccles, 2004).

Different classifications of school (education institution) factors exist:

  • school composition (e.g. SES intake); structure (e.g. size, sector and location); resources (e.g. physical, financial and human resources); and practices (e.g. instructional practices) (Rumberger & Lim, 2008);
  • school structure and resources; school practices (Lyche, 2010);
  • structural characteristics (e.g. composition of children/students, staff characteristics) and mediating factors (e.g. educators’ beliefs and expectations, relationships) (Hasselhorn, Andresen, Becker, Betz, Leuzinger-Bohleber, & Schmid, 2015).

In this article, we consider the first of these classifications and aim to review and demonstrate the variety of classroom/school factors that may contribute to becoming an ESLer. The article reviews empirical research in the field.


To conduct the scientific literature review, relevant publications were identified by using computerised searches in the Arizona State University Library search engine (including databases such as e.g. PsycINFO, Academic Search Premier (EBSCOhost), ERIC (Proquest), JSTOR Arts and Sciences, ProQuest, SAGE Premier, Science Direct) and other online resources (e.g. ResearchGate, institution or project webpages). We used the following keywords “early school leaving”, “drop out” AND “school factors/precursors/determinants”, “institutional factors”, “curriculum”, “pedagogy”, “school composition”, “school resources”, “school practices”, “teacher characteristics”, “metaanalysis” etc. We then examined references cited in the articles (i.e., “backward search” procedures). Original scientific articles and monographs together with reports by or for the European Commission and the OECD are considered.

The impact of school composition on ESL

School composition or the characteristics of the student body are one of the most examined school factors concerning ESL and achievement. This means the characteristics of students as a whole can influence a student over and above the effect of their individual characteristics (Rumberger & Lim, 2008). In the literature review conducted by these authors, the following indicators of school demographics are identified as having a direct effect on ESL: mean SES, the share of at-risk students (e.g. poor grades, truancy, discipline problems, grade retention), the share of racial or linguistic minorities, the share of high-mobility students (those changing schools or residences) and the share of students from single-parent families (or other family composition alternatives). Traag and van der Velden (2008) also found a large percentage of students from ethnic minorities in school increase the risk for some groups of ESLers (i.e. without a lower secondary diploma) after controlling for the individual’s minority background; in general, a 10% drop in the share of students with a minority background decreased the risk of ESL by 13%. For other groups of ESLers (i.e. with a lower secondary diploma; in an apprenticeship programme), the share of minority students from ethnic minorities in a school had no effect on the risk of ESL. McNeal (1997) also found the percentage of minorities affecting the chances of ESL. A study in Ireland found the effect of the school’s socio-economic composition – schools with more students from disadvantaged backgrounds had higher ESL rates, over and above individual student characteristics (Houses of the Oireachtas, 2010).

It is important to note that some studies indicate that, after accounting for several other school characteristics (e.g. school size, climate, resources, student-teacher relations), mean SES and other composition variables no longer have a direct effect on ESL (see, for example, Lee & Burkam, 2003; Rumberger & Palardy, 2005). This shows that the effect of school composition is mediated through other variables. Moreover, different structural and organisational characteristics of a school are likely to appear together (Lee & Burkham, 2003). On a similar note, Palardy (2013) controlled for differences in student background, peer influences, and school inputs and practices and found no differences in high-school graduation probability between low and high SES schools.

The impact of school structure on ESL

In this section, we examine the effects of the school characteristic of size, sector (public or private), location (urban or rural) and tracks provided (academic and other lower levels). The sector (public or private) of the school was examined in Rumberger and Rim’s literature review (2008). Of the 63 analyses treated, no relationship was found between public-private (in some cases only Catholic schools were considered) school and ESL rates in middle school, but in high school ESL rates were generally lower in Catholic schools (in some studies the effect could be attributed to school practices). As the authors highlight, it is common for students from private schools to transfer to a public school instead of or before leaving school, thereby ’inflating’ ESL rates in public schools. In another study, the private/public aspect of a school was found to be unrelated to ESL rates after controlling for students’ background and behaviour (Lee & Burkham, 2003). This finding is also supported by the OECD (2011) which examined the issue of private and public schools relative to achievement in the PISA study. They concluded that students in private schools do indeed perform better in the PISA assessment than students in public schools, yet when the socioeconomic context is similar students in public and private schools tend to do equally well. Moreover, ’’countries with a larger share of private schools do not perform better in PISA’’ (p. 1).

In relation to the location of the school, the relationship between the urbanisation of the municipality within which the school is located and ESL is not straightforward. For some groups of ESLers (i.e. those without a lower secondary diploma), an above-average level of urbanisation (high or very high) increased the risk for ESL, and in lowly urbanised areas the chances of ESL dropped, while in other groups of ESLers the effects were not as pronounced (Traag & van der Velden, 2008). Rumberger and Lim (2008) identified 12 analyses examining this issue, and the results were mixed; in some studies, being in an urban school increased the risk of ESL, while in others it decreased the risk or the effects were non-significant.

As for school size, the same authors (Rumberger & Lim, 2008) also found mixed results in their literature review. Out of 12 analyses, 6 found no significant effects, while in 3 students were more likely to drop out of large schools and in 3 other studies ESL was less likely in large schools. The relationship might not be non-linear – a smaller (but not too small) size is generally better (Lee & Burkham, 2003). As noted by Lee & Burkham (2003), school size per se is unlikely to directly affect the risk for ESL; it is more likely that other school, student and staff characteristics play a role. For example, a small school may have better student-teacher relations, better organisational trust, commitment to common goals, contacts (ibid., namely a better social climate). On the other hand, large schools may offer more curriculum, programme, classes and extracurricular options and benefit students’ learning in this respect (Rumberger & Palardy, 2005).

Although not a typical structural characteristic, the issue of different tracks (in terms of levels, i.e. academic or non-academic) provided in schools has also been examined. Traag (2012) and Traag and van der Velden (2008) found that schools’ heterogeneity in terms of the tracks they provided was a significant predictor of ESL. Students in lower secondary education who attended more heterogeneous schools (i.e. also providing a higher track of secondary education) were less at risk of ESL than their counterparts who attended a school which only provided lower secondary tracks. The risk of ESL decreased by 25%. The authors argued this may be a result of a more academic climate. This finding is supported by Lee and Burkham (2003); after accounting for student characteristics, schools with more challenging courses, fewer remedial or non-academic courses were more likely to keep students in school compared to their less ‘constrained’ counterparts. This means that offering a number of undemanding courses in high school does not keep students in schools.


In the literature review by Rumberger and Rim (2008), the effects of different indicators for physical, financial and human resources were observed. These included mean expenditures per pupil, mean teacher salaries, the student-teacher ratio, and the share of teachers with advanced degrees. Overall, relatively few studies found significant effects on ESL. Something similar was observed for students’ performance in PISA, TIMSS and PIRLS – namely, shortages of educational resources (instructional materials, computers etc.) did not affect students’ average performances greatly (Haahr, Nielsen, Hansen, & Jakobsen, 2005). Regarding class size, Finn, Gerber and Boyd-Zaharias (2005) found that 4 years of elementary school in a small class (13–17 students) increased the odds of finishing high school by 80% compared to their full-size-class (22–26 students) counterparts, where the effect was especially prominent among students from low-income homes. Small classes effect academic achievement and a student’s social behaviour (Finn, Pannozzo, & Achilles, 2003). Perhaps this effect mainly holds for Western societies because, in relation to class size, the OECD (2012) concluded that reducing class sizes is on its own an insufficient policy measure to improve the performance of education systems (after having examined 55 different countries or economies from around the globe).

Impact of school practices

School processes and practices are seen as the most promising for understanding and improving school performance and tackling ESL (e.g. teaching practices, a school climate conducive to student engagement and learning, social relationships among students, parents, teachers, and administrators). The studies differed in which specific practices were examined, and how they were measured. We separated the findings into those more related to instruction, curriculum and those more related to social and emotional aspects (i.e. climate).

Instructional practices, curriculum, assessment

Instruction is a key element of the teaching process that takes two main forms: direct instruction (built around problems with clear, correct answers the teacher provides) and student-centred instruction (teacher facilitating students’ own inquiry; see European Commission, 2014). Metaanalysis on learner-centred approaches found significant correlations of all person-centred teacher variables (defined as non-directivity, encouraging learning, encouragement of higher order thinking, adapting to differences, empathy, warmth, genuineness and learner-centred beliefs) with each affective or behavioural student outcome, including ESL (Cornelius-White, 2011), indicating the importance of student- or learner-centred approaches for ESL and ESLers. Finn (1989) also emphasised the quality of instruction that supports student involvement (e.g. classroom discourse as opposed to lecturing; also see McGarr, 2010). On the contrary, Van Klaveren (2011) found the proportion of time teachers spend lecturing in front of the class (compared to a more personal approach) did not (negatively) impact students’ achievement.

In a literature review for the European Commission (2014), it was recognised that a flexible and relevant curriculum and its provision are important for keeping students at school. However, too much theory, a complicated, rigid and boring curriculum, programmes based on memorising were seen as the biggest issues in this respect (also see Lee & Miu-Ling Ip, 2003; Rumberger, 2011). Indeed, almost 50% of the respondents reported that a key reason for ESL was that classes were not interesting and they were bored (Bridgeland, DiIulio, & Morison, 2006). The ESLers in the interviews mention the need for a more balanced, practical and real-life-based provision of the curriculum, with more physical activities and hands-on approaches (Houses of the Oireachtas, 2010; Lee & Miu-Ling Ip, 2003; McGarr, 2010). Another important aspect is formative assessment (European Commission, 2014) that evaluates a student’s progress and provides feedback to allow for improvement (OECD, 2008).

Moreover, the institutional providing of and encouragement to participate in extracurricular activities may contribute to a student’s sense of identification with the school, belonging to the school and may be a primary source of attachment to the school for students who are not as successful academically – in this sense, they may contribute to lower ESL (Finn, 1989).

School climate and disciplinary practices

Most studies examine the effects of different aspects of the school academic and disciplinary climate, including student-teacher relationships. A literature review (Rumberger & Lim, 2008) found that the general school climate (e.g. school loyalty and a low level of problem-student behaviour (i.e., fighting, cutting class)) and a strong academic climate (i.e. more students in academic tracks) reduced the likelihood of ESL, while a poor disciplinary climate (e.g. student reports of student disruptions in class or discipline problems in the school) and feeling unsafe increased the chances of ESL (also see Houses of the Oireachtas, 2010). McNeal did not find academic climate (a strong emphasis on academic achievement) as impacting ESL (1997). Positive student-teacher relations also reduced the risk of ESL (Byrne & Smyth, 2010; Hymel & Ford, 2014; Lee & Burkham, 2003). The authoritativeness of teachers and their fairness were identified as important in interviews with ESLers (Lee & Miu-Ling Ip, 2003). On a similar note, the literature review by Audas and Willms (2001) highlights that ESLers perceived the teacher to be less interested in the students and that perceived disciplinary practices were ineffective, unclear and unfairly applied. Downes (2013a) also identified school expulsion and suspension (usually both happen for non-academic reasons) as problematic in terms of ESL; instead, he proposed a multiplicity of intervention approaches to different ESL prevention issues. Moreover, a climate of teasing and bullying (as reported by students and teachers) increased the percentage of ESLers after controlling for the effects of other predictors (e.g. school composition variables, community crime, performance in standardised achievement testing; Cornell, Gregory, Huang, & Fan, 2013; also see Houses of the Oireachtas. 2010). A lack of emotional support services available to students is also important for the ESL issue (Downes, 2011). Brock (2011) found that teachers’ expectations concerning students’ ultimate education level and their report on self-efficacy did not predict students’ ESL. Based on a review of 800 studies, Hattie (2009) concluded the optimal classroom climate for learning is one based on trust and in which errors are allowed.

Knesting (2008) conducted a qualitative study interviewing students at risk for ESL, their teachers who were identified by the same students as either supportive or unsupportive, and certain other school staff. She also observed the teachers so identified in their classrooms. Those teachers who tried to understand the students’ behaviour believed that all students could succeed at school and communicated this (i.e. high expectations and respect), and accepted the students ‘as is’, were seen by students as supportive and helping them to remain at school. Moreover, every student in the classroom was important to these teachers and they interacted similarly with all students. Teachers who, according to the students, simply do not care (e.g. they did not mind when students left the room, did not ask about homework, did not seem excited about the material they were teaching) were listed as a reason for not liking school. On the school level, the students also mentioned an overemphasis on discipline, control, educational conformity and creating a caring and supportive environment with high expectations for everyone’s success (also see Hattie, 2009). The sense that some students were valued more than others (e.g. students with higher grades, successful sportsmen or cheerleaders) was also mentioned (also see Downes, 2013a, b). To sum up, students at risk for ESL need their voices to be heard and valued and require caring and respectful communication with teachers, administrators and other school staff. Interestingly, this climate of respect and caring is often intended or assumed by the adults – they assume they are providing it; yet students’ reports show that students do not experience it in this way –the question thus arises of how adults express care or respect in order for students to experience it as such. On a similar note, Downes (2013a) found splits in communication between students and teachers (e.g. a fear of asking the teacher questions, grading approaches that may lack transparency, access to the toilet, perceived snobbery) and advocated that students’ voices be listened to. Rogers (2016) found the following characteristics of teacher are important when working with disengaged students, namely one who: listens to students and involves them in decisions about their learning, has respectful interactions, has high expectations, takes time to get to know a student and their background, interest and strengths, and rarely displays irritation towards a student.

The findings indicate a great need for the professional development of teachers (and school staff) in the areas of conflict resolution and diversity skills so as to enable a classroom disciplinary climate to be established that does not include authoritarian teaching, autocratic and rigid behaviour management, unjustified behaviour undermining a pupil’s agency (e.g. being picked on by a teacher), offensive language – all behaviours of teachers reported by students (Downes, 2013a, b). Such experiences are an attitudinal precursor for the risk of ESL, thus a positive classroom and school climate are seen as key protective factors against ESL (Downes, 2013a). Based on analyses of data from PISA, TIMSS and PIRLS, it was concluded that while a positive school climate may be a precondition, it is insufficient for ensuring a strong academic performance and participation among students (Haar et al., 2005).


Based on the literature reviews in the three articles on ESL factors (individual/family factors and system factors), it is obvious that a variety of factors and influences impact a young person’s educational pathway, including the decision to leave school. Explanations that rely solely on the characteristics of students and their families are incomplete; there is mounting evidence that schools exhibit important effects on students’ leaving or staying in school (Lee & Burkham, 2003). Schools and their characteristics, including their teachers and other school staff, can make a real difference in students’ ability to remain at school. Concerning school-level factors, relationships with the families’ microsystem are also important (i.e. the mesosystem). (Potential) ESLers often describe having experienced contradictions between cultural and familial narratives (home culture) and the institutional narratives (i.e. stories of the school – school culture and structure) as contributing to ESL (Clandinini, 2010; Patterson, Hale, & Stessman, 2007). However, guidance to schools on how to provide the necessary multi-tiered support frameworks addressing multiple ESL risk factors (rather than interventions targeting single components) and early intervention is needed (Freeman & Simonsen, 2015).

To sum up, the specific characteristics of school composition (e.g. mean SES, share of students with a minority/migrant background) were found to impact ESL, but these effects are likely confounded with other school characteristics (e.g. school practices). Something similar could be established for school size and public/private sector. The findings regarding school location (urban/rural) were inconsistent. As stated by Audas and Willms (2001), the effects of school composition and structure are important because they are the main areas attracting governmental attempts to tackle ESL; yet, understanding the impact of schools’ policies and practices on the risk of ESL is important – as these have holding power over individuals.

In relation to school practices, a mismatch between the (provision of the) curriculum and a student’s interest and aspirations was a primary factor identified by ESLers (or those at risk). Innovative delivery of the curriculum is warranted – different pedagogic approaches, including a personalised curriculum, the recognition of informal learning, the use of technology, collaborative approaches via outdoor activities, as well as teaching through arts and sports (Rogers, 2016). Embedded (into the more real-life context of other subjects) approaches to teaching Maths and the mother tongue are suggested to increase the relevance and authenticity of the subjects (ibid.). All of these characteristics were also identified in Hattie’s (2009) metaanalysis as being relevant for enhancing student learning (e.g. teachers know how to introduce new content knowledge in a way that integrates it with students’ prior knowledge, they can relate the current lesson to other subject areas, and they can adapt the lessons according to students’ needs). The findings indicate the need to strengthen teacher’s initial and continuous education in this respect.

With regard to school climate, caring, supportive and respectful teachers (and other school staff) who believed in the students’ ability to succeed at school have a significant impact on students and them staying in school. Often this does not require financial resources or time outside of the classroom; instead, students found important support or discouragement in everyday school activities and interactions. In order to better understand the roles played by schools and teachers in ESL, it is necessary to listen to the voices of the youngsters themselves. Yet, as Downes (2013a) states, such “dialogue with students arguably comes too late in the process and needs systematic expression at a range of earlier stages as part of prevention focus” (p. 346) at the European level (Downes, 2013b).

, , & (1997). From first grade forward: Early foundations of high school dropout. Sociology of Education, 70(2), 87–107.

, & (2001). Engagement and dropping out of school: A life course perspective. Working Paper for the Applied Research Branch, Strategic Policy, Human Resources Development.

, & (2001). Early prospective studies of the high school dropout. School Psychology Quarterly, 16(4), 357–369.

, , & (2006). The silent epidemic: Perspectives of high school dropouts. Civic Enterprises, LLC, Washington.

(2011). Who is accountable for high school dropout? A study of the personal, parental, and teacher related factors of elementary students as predictors of high school dropout. Ottawa, Canada: Library and Archives Canada. Retrieved from

(1996). The ecology of human development experiments by nature and design. Cambridge, MA: Harvard University Press.

, & (2010). No way back? The dynamics of early school leaving. Dublin, Ireland: The Economic and Social Research Institute, The Liffey Press.

, , , , , , … (2010). Composing lives: A narrative account into the experiences of youth who left school early. Retrieved from

(2007). Learner-centered teacher-student relationships are effective: A meta-analysis. Review of Educational Research, 77(1), 113–143.

, , , & (2013). Perceived prevalence of teasing and bullying predicts high school dropout rates. Journal of Educational Psychology, 105(1), 138–149.

(2011). The neglected shadow: European perspectives on emotional supports for early school leaving prevention. The International Journal of Emotional Education, 3(2), 3–36.

(2013a). Developing a framework and agenda for students’ voices in the school system across Europe: From diametric to concentric relational spaces for early school leaving prevention. European Journal of Education, 48(3), 346–362.

(2013b). Holistic Approach to Early School Leaving Prevention in Europe: Key Strategic Priorities for System Level Development. Keynote presentation, European Commission, EUNEC (European Network of Education Councils), Lithuanian EU Presidency Conference on Early School Leaving, Seimas (Parliament) of the Republic of Lithuania, 17–19 November.

(2004). Schools, academic motivation, and stage-environment fit. In R. Lerner, & L. Steinberg (Eds), Handbook of adolescent psychology (2nd ed., pp. 125–153). Hoboken, NJ: John Wiley & Sons.

, & (1992). Paths to high school graduation or dropout: A longitudinal study of a first-grade cohort. Sociology of Education, 65(2), 95–113.

. (2014). Study on the effective use of early childhood education and care in preventing early school leaving: Annex 1. Literature review (N° EAC/17/2012). Luxembourg: Publications Office of the European Union.

(1989). Withdrawing from school. Review of Educational Research, 59(2), 117–142.

, , & (2005). Small classes in the early grades, academic achievement, and graduating from high school. Journal of Educational Psychology, 97, 214–223.

, , & (2003). The “why’s” of class size: Student behavior in small classes. Review of Educational Research, 73(3), 321–368.

, & (2015). Examining the impact of policy and practice interventions on high school dropout and school completion rates: A systematic review of the literature. Review of Educational Research, 85(2), 205–248.

, , , & (2005). Explaining Student Performance Evidence from the international PISA, TIMSS and PIRLS surveys. Final report. Taastrup: Danish Technological Institute. Retrieved from

, , , , , & (2015). Children at risk of poor educational outcomes: In search of a transdisciplinary theoretical framework. Child Indicators Research, 8(2), 425–438.

(2009). Visible learning: A synthesis of 800+ meta-analyses on achievement. Oxford, UK: Routledge.

. (2010). Staying in Education: A New Way Forward School and Out-of-School Factors Protecting Against Early School Leaving. Joint Committee on Education and Skills, First Report. Retrieved from

, & (2014). School completion and academic success: The impact of early social-emotional competence. In: R. E. Tremblay, M. Boivin, R. Peters (eds.), F. Vitaro F. (topic ed.) Encyclopedia on Early Childhood Development [online]. Retrieved from

, , , & (1997). Disentangling the weight of school dropout predictors: A test on two longitudinal samples. Journal of Youth and Adolescence, 26(6), 733–762.

, , , & (2000). A prospective longitudinal study of high school dropouts examining multiple predictors across development. Journal of School Psychology, 38(6), 525–549.

(2008). Students at risk for school dropout: Supporting their persistence. Preventing School Failure: Alternative Education for Children and Youth, 52(4), 3–10.

, & (2003). Young school dropouts: Levels of influence of different systems. Journal of Youth Studies, 6(1), 89–110.

, & (2003). Dropping out of high school: the role of school organization and structure. American Educational Research Journal, 40(2), 353–393.

(2010). Taking on the completion challenge: A literature review on policies to prevent dropout and early school leaving. OECD Education Working Papers, 53, OECD Publishing.

(2010). Early school leaving: An exploration of the factors contributing to school non-completion. Master’s thesis, Dublin Institute of Technology.

(1997). High school dropouts: A closer examination of school effects. Social Science Quarterly 78(1), 209–222.

. (2008). Assessment for learning: Formative assessment, OECD/CERI International Conference Learning in the 21st century: Research, Innovation and Policy, the Centre for Education Research and Innovation. Paris: CERI.

.(2011). Private schools: Who benefits? PISA in Focus, 7, 1–4.

. (2012). How does class size vary around the world? Education Indicators in Focus, 9, 1–4.

(2013). High school socioeconomic segregation and student attainment. American Educational Research Journal, 50(4), 714–754.

, , & (2007). Cultural contradictions and school leaving: A case study of an urban high school. The High School Journal, 91(2), 1–15.

(2016). Disengagement from education. London: Institute of Education Press.

(2011). Dropping out: Why students drop out of high school and what can be done about it. Cambridge, MA, USA: Harvard University Press.

, & (2005). Test scores, dropout rates, and transfer rates as alternative indicators of high school performance. American Educational Research Journal, 42(1), 3–42.

, & (2008). Why students drop out of school: A review of 25 years of research, California Dropout Research Project Report #5. Retrieved from

(2012). Early school-leaving in the Netherlands. A multidisciplinary study of risk and protective factors explaining early school-leaving. The Hague: Statistics Netherlands.

, & (2008). Early school-leaving in the Netherlands. The role of student-, family- and school factors for early school-leaving in lower secondary education. Maastricht: Research Centre for Education and the Labour Market.

(2011). Lecturing style teaching and student performance. Economics of Education Review, 30(4), 729–739.