Early between school tracking (i.e. tracking) is an institutional set-up that has strong implications for students and their future outcomes. While there is increasing consensus that tracking does not increase average levels of achievement but average levels of inequality (e.g., Terrin & Triventi 2022, meta-analysis), we know little about the underlying mechanisms. The focus of this study is a potential mechanism that is as prominent as it is understudied: the role of unequal school resources in tracked systems. In other words, while higher tracks (i.e. academic tracks) benefit from above-average resources, lower tracks (i.e. vocational tracks) are disadvantaged. We call such a situation input-stratification, because the total input (resources) assigned to the education system is stratified across tracks (c.p. Esser 2016). 

Input stratification has been discussed as a potential mechanism throughout the literature (e.g. Betts 2011; van de Werfhorst 2021, Terrin & Triventi 2022), but there are hardly any studies providing evidence on the unequal distribution of resources across tracks. For instance, van de Werfhorst  (2021; 1214)  and Terrin & Triventi (2020, meta-analysis; 7) still cite the prominent paper of Brunello & Checchi (2007). While Brunello & Checchi (2007; 795) only provide student-teacher ratios in 2004 for a handful of countries, ostensibly because resources are not the focus of their study, they apparently provide the best evidence on input-stratification to this date. The scarcity of evidence may be explained by the inherent difficulty of estimating (causal) effects of school resources (e.g. Gibbons & McNally 2013) combined with the scarcity of high-quality data that allows for comparison across countries. Thus, a consequence is that we do not know whether a) input stratification (unequal resources) actually exists in tracked systems and b) whether it drives unequal levels of achievement in different tracks. Although answering b) is outside the scope of this study, we argue that it is still valuable to know about a) since b) presupposes a). Our approach is driven by two goals. 

First, we want to provide a thorough discussion on the role of school resources in tracked systems. While the topic of school resources is often mentioned in the literature, it is hardly spelled out with its theoretical implications. The unequal distribution of resources across tracks could be a mechanism that explains why tracked systems fail to show increased (average) achievement but increased social inequalities (i.e. Terrin & Triventi 2022). Accordingly, students on higher (lower) tracks benefit (are disadvantaged) through resources above (below) average. But since tracks are segregated by social status (Strello et al. 2022), high (low) status students are more (less) likely to benefit from above-average resources, explaining the increased social inequality in tracked systems. 

Given the limited space, we merely note that we incorporate the existing literature on school resources in our study (e.g. Krüger 2003 vs. Hanushek 2003). We distinguish between explicit resources (e.g. official government funding) and implicit resources which are indirect consequences of the institutional set-up (e.g. self-selection of more capable or motivated teachers). Further we discuss whether differences in resource allocation across tracks should be seen as a bug or a feature of tracked systems (i.e. vocational vs. academic training; c.p. Esser 2021). 

Second, we want to assemble data sources that are either informative about resource levels 1) across tracks within a country or 2) across tracked and untracked countries. Noting the inherent difficulties of estimating the effects of school resources, we will restrict our analysis to a descriptive analysis in order to answer the question whether the existence of input-stratification is plausible given the existing evidence.

Research plan: In general we want to restrict our analysis to a thorough descriptive analysis of tracked (and untracked) education systems. Our analysis comprises two parts: First, using administrative data from the German speaking tracking countries (Germany, Austria, Switzerland, Luxembourg) we will discuss educational spending per student, teacher-student ratios, teaching hours and teacher qualifications across tracks in a case-study like fashion. Second, for our main analysis we will use large-scale assessment data (LSA) from the last 25 years. Unfortunately, measures of resources in PIRLS & TIMSS 4 (primary school) and PISA (secondary school) are largely incomparable which hampers efforts to compare change over time from primary to secondary school. However, it is possible to compute teacher-student ratios from LSA data (e.g. Woessmann & West 2006), which is a resource indicator commonly used in the school resource literature (e.g. Gibbons & McNally 2013). This allows us to track the change in the variance of teacher-student ratios in tracked and untracked countries as they move from primary (no country is tracked) to secondary school (some countries have administered tracking). Lastly we will use PISA data to compute a broader set of resource indicators which resonates with our idea of explicit and implicit resources. Using PISA has the advantage that we can draw on a broad set of variables and that we can directly identify the school track (as compared to TIMSS 8). Unfortunately, however, PISA has only administered (short) teacher questionnaires since 2015. To remedy this shortcoming, we aim to match PISA with TALIS data, which provides in-depth data on the teacher and school principal level. Taken together, this allows us to compute different indicators of explicit and implicit school resources (e.g. material resources, student-teacher ratios, teacher qualifications, teacher motivation, parental support) across track types.

We are still in the process of data collection (cleaning & assembling). The limited evidence on school resources indicates large differences in student-teacher ratios in five tracked systems (Brunello & Checchi, 2007). However, Brunello & Checchi also report administrative data from Austria that shows that spending per student is higher in vocational tracks. Overall, we expect more mixed evidence when it comes to explicit resources (i.e. official allocation of resources through the government) as compared to implicit resources (i.e. differences in resources as a consequence of the institutional set-up). Tracking is often understood as a form of stratification, inducing an implicit “better” or “worse” into the system. Further, it is theoretically plausible (e.g. Boudon 1974) and empirically validated (Strello et al 2022) that tracked systems are segregated by social status. We argue that this could lead to differences in implicit resources because involved actors take this stratification and segregation into account. More able or motivated teachers could self-select into higher tracks, parental support via booster clubs is likely to depend on the average social status of parents at the school and so on.

Betts, J. R. (2011). The economics of tracking in education. In Handbook of the Economics of Education (Vol. 3, pp. 341-381). Elsevier. Boudon, R. (1974). Education, opportunity, and social inequality: Changing prospects in western society. Brunello, G., & Checchi, D. (2007). Does school tracking affect equality of opportunity? New international evidence. Economic policy, 22(52), 782-861. Esser, H. (2016). Bildungssysteme und ethnische Bildungsungleichheiten. Ethnische Ungleichheiten im Bildungsverlauf: Mechanismen, Befunde, Debatten, 331-396. [English: “Education systems and ethnic educational inequalities” in “Ethnic inequality along the educational pathway: Mechanisms, Results, Debates”] Esser, H. (2021). » Wie kaum in einem anderen Land...«?: Die Differenzierung der Bildungswege nach Fähigkeiten und Leistungen und ihre Wirkung auf Bildungserfolg,-ungleichheit und-gerechtigkeit. Band 1: Theoretische Grundlagen. Campus Verlag. [English: “‘Hardly any other country…’?: Differentiation of educational pathways according to aptitude and performance and their consequences for educational attainment, inequality and justice. Volume one: Theoretical foundations” ] Gibbons, S., & McNally, S. (2013). The effects of resources across school phases: A summary of recent evidence. Hanushek, E. A. (2003). The failure of input‐based schooling policies. The economic journal, 113(485), F64-F98. Krueger, A. B. (2003). Economic considerations and class size. The economic journal, 113(485), F34-F63. Strello, A., Strietholt, R., & Steinmann, I. (2022). Does tracking increase segregation? International evidence on the effects of between-school tracking on social segregation across schools. Research in Social Stratification and Mobility, 78, 100689. Terrin, E., & Triventi, M. (2022). The effect of school tracking on student achievement and inequality: A meta-analysis. Review of Educational Research, 00346543221100850. van de Werfhorst, H. G. (2021). Sorting or mixing? Multi‐track and single‐track schools and social inequalities in a differentiated educational system. British Educational Research Journal, 47(5), 1209-1236. Woessmann, L., & West, M. (2006). Class-size effects in school systems around the world: Evidence from between-grade variation in TIMSS. European Economic Review, 50(3), 695-736.