Keywords

1 Introduction

Countries differ widely in the organization of their educational systems. Considering the stage of secondary education, these systems range from comprehensive forms without ability grouping until the end of compulsory schooling (e.g., Nordic countries) to models in which students are assigned to different and clearly distinguishable educational programmes early on in secondary education (e.g., Austria, Germany, the Netherlands, and Switzerland; cf. H.-P. Blossfeld et al., 2016). There is an ongoing controversy regarding how ability grouping contributes to educational success and to the inequality of educational opportunities (Esser, 2016; OECD, 2016). For Germany, the discussion on the advantages and disadvantages of early selection into different types of school has been going on for generations and has recently been fuelled by findings from international comparative studies (e.g., PISA). Currently, most of the 16 German federal states are working on refining their secondary education systems by means of manifold structural reforms. At present, it seems that, in the long run, some form of a two-tier system with a differentiation between the highest programme (the traditional Gymnasium as the highest form of secondary education) and various combinations of the other programmes may become dominant. Discussions on the consequences of ability grouping and on external differentiation into different types of school are based on questions such as how much this change would affect social inequalities and whether current pathways in education are actually equally open to everyone. We contribute to the current state of research in this area by examining mobility between the types of schools available during secondary schooling in Germany.

2 Secondary Education in Germany

In the Federal Republic of Germany, the individual states are responsible for the organization of education. Because of this federalism, there is not one common German education system, but rather, sixteen distinct education systems that exist in parallel. However, many states share certain characteristics regarding educational structures.Footnote 1 In general, education in Germany can be characterized as being highly stratified, particularly when considering the field of secondary schooling (Allmendinger, 1989; Buchholz et al., 2016). After jointly attending primary school for 4 years (apart from in the federal states of Berlin and Brandenburg where children regularly attend primary school for 6 years), stratification begins, and children are allocated to different types of school. Traditionally, three distinct types of schools are available. The only type of school that can be found in all German federal states is the Gymnasium (i.e., upper secondary schooling; henceforth abbreviated USS). This type of school ends after Class 12 or 13 with a final examination (Abitur) that functions as a university entrance qualification. In addition to the Gymnasium, the most common types of school are the Realschule (middle secondary schooling; henceforth MSS) and the Hauptschule (lower secondary schooling; henceforth LSS), in which middle and low school-leaving certificates can be obtained, respectively. These certificates are not sufficient for accessing higher education; rather, they serve as certificates of entry into vocational training and they regulate access to further education. In some federal states, comprehensive schools are also available, whereas other states offer types of school that integrate LSS and MSS. We address all types of school that offer more than one type of school-leaving certificate under the label ‘schools offering several courses of education’ (SCE).

Independently of the available types of school, the transition from primary to secondary education can be regarded as an important milestone of the individual educational trajectory in all German federal states. However, the federal states have quite different regulations for this transition. In many states, a certain grade point average in selected subjects is a necessary precondition to choosing a higher programme of secondary education. Some states additionally require a recommendation by the primary school teacher, which is usually based on marks in core subjects and/or additional criteria such as learning behaviour. If parents prefer a higher type of school than that recommended by the teacher, some states require an extensive entrance exam. In the majority of states, however, parents can freely choose the type of secondary school for their children; nevertheless, schools and teachers are supposed to offer parents counselling regarding the ‘right’ choice of school type for their child.

After the initial decision on a type of secondary school has been taken, children attend the different types of school. Many federal states have an orientation phase (Orientierungsstufe) that typically lasts 2 years. During this time, teachers systematically monitor and evaluate the children’s progress in order to make sure that they are at the appropriate type of school. At the end of the school year, the initial decision can be either confirmed or revised, with a revision leading to a change in the initial type of school. Such changes can occur in both directions—that is, upwards to a type of school with a more ambitious curriculum (and a higher school-leaving certificate) or downwards to a type of school with a less demanding curriculum (and a lower school-leaving certificate). Furthermore, transitions to, from, and within comprehensive schools are also feasible.

Altogether, the highly stratified secondary education system in Germany results in a great variety of possible pathways between initial enrolment and final graduation. Our article aims to shed greater light on the social selectivity that originates in transitions during the phase of secondary education. We concentrate on mobility between the traditional types of school and do not include transitions to comprehensive schools or to schools that offer several courses of education that would be difficult to classify as either upward or downward moves.

3 State of Research

3.1 Social Selectivity in Secondary Education in Germany

International comparative studies such as PISA, PIRLS, and TIMSS consistently identify Germany as a country with pronounced social inequalities in education. Among the explanations given for the persistence of educational disparities, the early timing of allocating students to different educational programmes is often considered a major factor (Brunello & Checci, 2006). As a consequence, the bulk of research has investigated social selectivity during the transition from primary to secondary schooling (e.g., Pietsch & Stubbe, 2007). Following Boudon’s (1974) theoretical distinction between primary and secondary effects of stratification, various studies have investigated the impact of students’ social origins on their academic achievement (‘primary effect’) and on the choice of which type of secondary school they enter (‘secondary effect’) after primary schooling (Ditton, 2010, 2013b; Dollmann, 2016; Müller-Benedict, 2007; Schimpl-Neimanns, 2000). Based on Boudon’s ideas, several formalized models of ‘rational choice’ have been developed with which to investigate the influence of social origin on educational attainment and achievement (e.g., Breen & Goldthorpe, 1997; Erikson & Jonsson, 1996; Esser, 1999). These models share the basic assumption that ‘rational’ actors take a decision by consciously considering the benefits, costs, and the probability of success for each option. Findings on the decomposition of primary and secondary effects vary considerably depending on the characteristics of the database being utilized (particularly the regions/federal states included), the methodological approach, the dependent variable (marks, test results, recommendations, types of schools, etc.), and the independent variables included. However, there is a general consensus regarding the fundamental importance of this early transition in terms of the emergence of social inequalities.

In addition to the rational choice tradition, another well-established research tradition has developed on the basis of Bourdieu’s (1977) theory of cultural and social reproduction. Bourdieu assumes that the educational system reproduces the unequal distribution of economic and cultural capital that families in different social positions maintain. Families from different social classes equip their children differently with skills and attitudes before they enter school, and these differences persist throughout schooling and result in unequal educational outcomes. The key component in Bourdieu’s theoretical framework is what he calls habitus—a ‘system of dispositions which acts as a mediation between structures and practice’ (Bourdieu, 1977, p. 487). The basic idea of habitus is that people internalize their surrounding social structures through daily practice and interactions with others. As a result of this habitualization, people produce thoughts, perceptions, expressions, and actions that tend to reproduce these structures. Bourdieu refers to habitus as a ‘generative grammar’ (Bourdieu, 2002, p. 30) that engenders thoughts and actions within the limits of the particular conditions of its production. Hence, habitus can be considered ‘a sense of one’s (and others’) place’ (Hillier & Rooksby, 2002, p. 5). In contrast to the assumptions of rational choice models, Bourdieu’s conception of habitus includes both conscious and unconscious aspects of behaviour. Current research on educational inequalities has further developed Bourdieu’s ideas by investigating the fit between children’s academic habitus and the demands of distinct school cultures (see e.g., Kramer, 2013, 2017). Depending on this fit, children develop a feeling of being either ‘in the right place’ or ‘out of place’ and obtain educational outcomes accordingly. In a similar fashion but with a particular focus on the perspective of the different ways in which parents structure their children’s everyday life, Lareau (2011) asserts that children develop either a ‘sense of entitlement’ or a ‘sense of constraint’ in their interactions with institutions.

Based on these two approaches, numerous studies have examined social selectivity during the stage of primary schooling and the allocation of students to different types of secondary school. However, empirical findings on the transitions that occur during secondary schooling remain relatively scarce. According to official school statistics, downward transitions by far outweigh upward transitions between Classes 7 and 9 (Authoring Group Educational Reporting, 2008). In a meritocratic system, academic performance would be expected to be the main driving force behind school changes. However, it remains unclear exactly how the openness of the educational system affects social selectivity. Some scholars assume that increased openness leads to a greater equality of chances (e.g., Bellenberg & Forell, 2012), whereas others are less optimistic about the categorical advantage of open educational pathways. Buchholz et al. (2016, p. 90) find that ‘existing social inequalities are deepened further because it is already privileged social groups that profit most from atypical paths to higher secondary education.’ According to their analyses, the parental level of education plays a significant role in educational mobility, even after controlling for differences in performance. Similarly, for students from Bavaria or Saxony, Ditton (2013a) has observed that upgrades from LSS are clearly associated with various other indicators of social background such as parents’ rating of the accessibility of higher school-leaving certificates, cultural capital, and parental support for learning. Furthermore, academic performance in primary school turns out to be an important predictor for both upward and downward transitions. In an analysis of the development of academic competencies after entering USS in the federal states of Hesse or Bavaria, Pfost et al. (2018) observed significant differences between students who had versus had not received a recommendation for USS. By the end of Class 7, the latter group had attained considerably lower marks in German and mathematics, and they even performed slightly worse in comparison with students from LSS and MSS. Roeder and Schmitz (1995) found effects of migration status on downward transitions from USS in a study of students from Hamburg. In another study in Hamburg, Stubbe (2009) reported significant effects on the decision to leave USS for academic performance (tests and marks) and social background (ISEI) as well as for migration status and gender. With data from the adult cohort of the National Educational Panel Study, Blossfeld (2018) found effects for parental education and gender on both downward and upward mobility in West Germany.

3.2 Research Questions

Only a small number of studies have examined mobility in secondary school thus far. Because the share of students with a change in type of school varies considerably between the federal states (Bellenberg, 2012), the results of the few available studies are further limited in terms of their comparability and generalizability. Hence, there is still a lack of studies that investigate inequalities in educational transitions during secondary school, and there is a particular need for analyses that utilize comprehensive data from more than just a few selected regions in Germany. This article seeks to contribute to research in this area by describing and analysing educational trajectories across Germany as a whole and focusing on changes in the type of school between Classes 5 and 9. Our main interest lies in determining the extent to which such transitions in secondary schooling exhibit patterns of social selectivity. Which factors are important for educational trajectories and educational success? In line with the literature, we assume that in addition to academic performance, socio-economic background plays an important role in educational mobility. We expect that children from less advantaged social backgrounds more often move downwards to less demanding education programmes, whereas children from more privileged families should be less prone to leaving USS before graduation. Furthermore, we consider migration status and the child’s gender as influences on school changes during secondary schooling in order to ascertain whether these factors play a role in the social selectivity of transitions.

4 Data and Methods

4.1 Sample

To examine the development of educational trajectories during secondary schooling and answer our research questions, we used data from Starting Cohort 4 of the German National Educational Panel Study (NEPS; see H.-P. Blossfeld et al., 2011).Footnote 2 The sample of 16,425 students attending Class 9 was established in a stratified multistage sampling design with a selection of schools during the first stage and a selection of two classes within each school during the second stage (Skopek et al., 2013). Because this article focused on changes between the various regular types of school, we restricted our analyses to students attending regular schools in Germany and excluded a subsample consisting of children attending a school for students with special educational needs. All types of school that offer more than one kind of school-leaving certificate or in which a decision is not necessarily taken from the beginning (e.g., due to an orientation phase) were subsumed under the label ‘schools offering several courses of education’ (SCE). We put specific emphasis on describing and analysing transitions between the three regular types of school (LSS, MSS, and USS; i.e., Hauptschule, Realschule, and Gymnasium).

Retrospective data on the children’s educational trajectory form the basis of this examination. We reconstructed the educational trajectory until the students attended Class 9 during the 2010/2011 school year with manifold information on the school biography collected from parents and children over all available panel waves. We excluded cases with incomplete or inconsistent information (e.g., missing school-entry or -leaving dates, missing information on the type of school, gaps or overlaps in the educational trajectory, etc.) from our analyses. Furthermore, we excluded cases with episodes in a Waldorf school or in a school for students with special educational needs prior to Class 9, because we argue that these cases’ educational trajectories contain particular features that inevitably differ from ‘typical’ school careers in Germany.

Finally, even though the information from later panel waves would theoretically cover a longer time span, we restricted our analyses to the phase of secondary schooling. Therefore, our analyses covered the educational trajectory between the end of primary schooling and the time the first panel wave was conducted. After all exclusions, data on 12,434 cases remained for analyses.

4.2 Variables and Data Handling

Educational Trajectory

The central dependent variable in this examination is the development of students’ educational careers during the stage of secondary schooling. In order to analyse and compare diverging trajectories, we disassembled the educational pathways into their most basic components. We distinguished between students who remain at their initial type of secondary school without a transition and students with a school change during the stage of secondary schooling. Among those who did change between the LSS, MSS, and USS types of school, we furthermore differentiated between upward transitions to a more prestigious type of school and downward transitions to a less prestigious type of school. Because the possible directions of a transition depend on the type of school attended, we always specified the initial type of secondary school. Because transitions to and out of SCE may also occur without aiming for a higher or lower school-leaving certificate, school changes involving an SCE were simply considered changes without further differentiation of direction. Because our main interest lies in changes between types of school, changes within comprehensive schools are not reported here.

Academic Performance

Considering academic performance is crucial when assessing educational trajectories and the social selectivity processes involved in them. Unfortunately, the retrospective data utilized here do not contain any direct performance measures such as test results or marks in the students’ educational trajectory prior to Class 9. Therefore, evaluating the adequacy of the initial choice of secondary school or the necessity of school changes during secondary schooling is a challenging task. However, we do possess information from the parent interviews on primary school teachers’ recommendations for the most suitable type of school. This variable correlates strongly with academic performance at the end of primary schooling. Furthermore, in some federal states, this recommendation serves as ‘gatekeeper’ that keeps low-performing students out of the more demanding types of school. Thus, we used the primary school teachers’ recommendations as an indicator for students’ academic performance when entering the phase of secondary schooling. Consistent with the type of school information, we coded the recommendation as a nominal variable with four possible values (i.e., SCE, LSS, MSS, or USS).

As to academic performance at the end of secondary schooling, we could draw on results of competence tests in the domains of reading (for detailed information, see Haberkorn et al., 2012), mathematics (Duchhardt & Gerdes, 2013), and science (Schöps & Saß, 2013) and could thereby rely on a more elaborated measure. These tests were designed specifically to allow for systematic comparisons of students in different educational programmes (Weinert et al., 2011) and were captured in Class 9. We utilized standardized weighted maximum likelihood estimates as point estimates of the individual competence scores in all three domains (Pohl & Carstensen, 2012).

Socio-economic Status

As a measure of the family’s socio-economic status, we use the parental level of education in our analyses. Respondents provided information on both their own and their partner’s education, and we combined these into a single variable indicating the highest level of parental education. The categories are LSS or lower, MSS, and USS or higher.

Migration Status

To test whether children from families with migrant backgrounds display different mobility in secondary school than children from native German families, we considered whether their parents had been born in Germany or abroad. The variable employed in our analyses was coded as a dummy variable with 0 for Germans and 1 if at least one parent had been born abroad.

Gender

We considered the gender of the child as a dummy variable with 0 for girls and 1 for boys.

4.3 Analytical Strategy and Limitations

Missing Values

A typical problem with survey data is missing values (Little, 1988; Rubin, 1987). Within the dataset utilized here, apart from the educational trajectory (for which missing information led to the elimination of a case), all independent variables displayed at least a small degree of missing values. We estimated these missing values by using the ‘multiple imputation of chained equations’ technique (van Buuren & Oudshoorn, 1999, 2000; White et al., 2011). The reported findings are averaged results over all of the 10 complete datasets that were generated.

Analytical Strategy

We began our analyses with an overview of how students in our sample were initially distributed over the different types of secondary school. Furthermore, we examined the connection between registered types of school and primary teachers’ recommendations while considering socio-economic status, migration status, and gender. In a next step, we quantified the number of school changes and determined the ratio between stable students and school leavers for each type of school. To identify patterns of social selectivity, we compared the various subgroups in relation to each covariate separately. Because there is an ongoing debate over the feasibility of pooling results from Chi2 tests across multiply imputed datasets, we calculated Chi2 test results separately for each imputed dataset and reported them on a significance level of p < .05. In a final step, we merged the results of all independent variables in a multivariate analysis in which we used a logistic regression analysis to test their effects on the likelihood of experiencing a downward transition from USS. In this way, we aimed to improve our understanding of mobility in the German education system and of the mechanisms of social selectivity at work.

Limitations

Describing and comparing educational trajectories within the diversified educational landscape in the 16 federal states of Germany is possible only within certain boundaries. In addition to considering only the most common types of school and restricting the educational trajectory to the phase of secondary schooling, we had to accept some further limitations. First, we accounted exclusively for the first change of type of school. Some students, however, changed their type of school more than once, and there may even have been a few cases in which a student made more than two changes. As interesting as such uncommon pathways are for research on educational mobility, due to space limits, they cannot be considered in detail here. Accordingly, we report only results on the first transition. Second, our analyses ignored the individual timing of the transition within secondary schooling. For research interested in the development of cognitive competencies or non-cognitive factors (such as academic self-concept), it would be important to consider whether a school change occurred early or late in the phase of secondary schooling. However, with this chapter, our focus lies on gathering knowledge about which cases experienced a school change rather than analysing when the school change happened. This is why we did not consider any effects of transition timing.

5 Results

5.1 Initial Distribution of Students in Secondary Schools

The educational trajectories of all available cases include at least an initial episode in a primary school and a subsequent episode in a secondary school. By the end of their time in primary school, students are sorted into the different types of school in the secondary school system. In all, 82.7% of the students in the sample changed to one of the three major types of school (LSS, MSS, or USS), whereas 17.3% changed to SCE. The largest share of students in the sample began their phase of secondary schooling in a USS (n = 4968, 40.0%). About one quarter of the students (n = 3022, 24.3%) attended an MSS, and 18.4% (n = 2289) began at an LSS.

To evaluate the adequacy of the initial choice of secondary school, we took the primary school teachers’ recommendations for the most appropriate type of school into account. Table 8.1 presents the distribution of registered types of school by recommended type of school. Children were most likely to attend the recommended type of school within all recommendation groups; however, there were differences regarding the share of families who followed the recommendations. Of the children who had received a recommendation for LSS, about two thirds (69.0%) attended LSS subsequent to primary school. Among the children with a recommendation for MSS, about one half (54.6%) initially registered at an MSS. Of the children who had received a recommendation for a USS, 82.0% subsequently attended a USS. Recommendations for SCE were followed in 61.6% of all cases.

Table 8.1 Registered type of school by primary school teachers’ recommendation (row percentages)
Full size table

Altogether, 69.5% of the children attended their recommended type of school after primary school. The remaining 30.5% began their time in secondary school in a type of school that deviated from their primary teachers’ professional counsel. Considering socio-economic status, the distribution of children in the different secondary schools reveals the well-known stratified pattern of social origin (cf. Appendix, Table 8.A1). Children in an LSS came most frequently from a family in which the parents had obtained a school-leaving certificate from an LSS (45.8%) or an MSS (40.5%). At the MSS, the majority of children had parents who had also graduated from an MSS (50.6%); and at the USS, the greatest share of parents (65.5%) held a school-leaving certificate from a USS. For migration status, there were only minor differences across the different types of school (cf. Appendix, Table 8.A2). The share of children from migrant families was highest at the LSS (11.4%) and lowest at the USS (8.2%). The gender distribution also varied only moderately across types of school (cf. Appendix, Table 8.A3). At the LSS, there were more boys than girls (56.0%), whereas at the USS, the girls were the largest group (54.0%). At the MSS and SCE, gender was distributed evenly.

5.2 Quantifying Changes Between Different Types of Secondary School

The type of school that students attended at the beginning of their secondary education was not always identical with the type of school at which they completed their compulsory education. In total, 1658 cases (i.e., 13.3% of all cases in the sample) experienced at least one change in type of school during their time in secondary education. Table 8.2 reveals the volume of transitions between the different types of secondary school. The denoted percentages refer to the children who initially began their phase of secondary schooling in the respective learning environment.

Table 8.2 Transitions between different types of secondary school
Full size table

Subsequent to primary school, the largest share of students in the sample switched to USS. From there, 427 children (8.6%) descended to MSS, and 13 children (0.3%) even dropped to LSS. Among the children beginning secondary schooling in MSS, 365 students (12.1%) changed downwards to an LSS, whereas only 81 students (2.7%) moved upwards to a USS. From LSS, 131 children (5.7%) experienced an upward transition to an MSS. Nineteen children (0.8%) even managed to take a big step forward and change from an LSS directly to a USS.

Comparing the different student outflows, mobility turned out to be greatest in the MSS. In sum, 82.4% of students remained there steadily by Class 9, whereas 17.6% of the children who had begun in MSS left to attend another type of school. Furthermore, the number of children arriving from other types of school was highest in MSS, and the total student number thus grew noticeably during the time of secondary schooling. The largest share of stable students could be found in USS at 88.7%. However, because the number of children entering was smaller than the number of children leaving, student numbers at USS diminished over time. This finding is in contrast with the comparatively high entry quota at the beginning of secondary schooling. Altogether, downward mobility was considerably higher than upward mobility during the phase of secondary schooling.

5.3 Identifying Selectivity in Upward and Downward Transitions

To learn more about how the educational trajectory develops and which factors participate in shaping it, we examined a set of variables that we expected to contribute significantly to educational inequalities. In this section, we merely distinguish between downward and upward transitions without further differentiating between the different possible types of targeted school.Footnote 3

Teachers’ Recommendations

Teachers’ recommendations played an important role in determining which children would go to the which type of school at the beginning of their time in secondary education. Additionally, we utilized the teachers’ recommendations as an indicator as to how successful a child would be in progressing through schooling in the new learning environment. We would expect to find differential shares of children remaining in one type of school or leaving it before graduation depending on the fit between teachers’ recommendations and the initial registration at a type of school. Children with a lower recommendation than their registration should be more likely to experience a downward transition, whereas children with an equal or higher recommendation than their registration should be more likely to remain stable or even achieve an upward transition. Table 8.3 displays the results of comparing the information regarding whether a student registered at the recommended type of school with her or his educational trajectory. For each type of secondary school, there is a row displaying whether the registration at this type of school was higher, equal to, or lower than the level recommended by the primary school teacher. Because there is no possible recommendation lower than registration in LSS and there is likewise no possible recommendation higher than registration in USS, these rows remain empty. For the same reason, the columns of downward transition to LSS and upward transition to USS are not available. The table displays row percentages that can be interpreted as a student’s chances of undergoing a transition depending on the particular recommendation and registration.

Table 8.3 Transitions by teachers’ recommendations and initial registration at a type of secondary school (row percentages)
Full size table

Among the children who began secondary schooling in LSS and who received the corresponding recommendation, 94.0% remained there without transition, whereas 6.0% attained a more prestigious type of school before graduation. The group of children with a recommendation for a more ambitious type of school had a slightly higher chance of upgrading (10.4%). Chi2 was significant in nine out of the 10 imputed datasets.

For children attending MSS, transitions were possible in both directions. In this educational environment, the interplay of recommendation and observable mobility prospects revealed a very clear pattern. A large portion of students with a lower recommendation changed downwards (38.7%), whereas students who had registered there in accordance with their recommendation had a significantly lower risk of a downward transition (10.7%). The group of students who had registered at an MSS regardless of their higher recommendation displayed the lowest downward transition rate. This group also displayed the highest share of steady students (89.8%) and the highest share of upward transitions (6.2%). These numbers additionally illustrate the great discrepancy between upward and downward transitions: chances were considerably higher for dropping down than for moving up. Chi2 was significant in all 10 datasets.

In USS, the risk of dropping out was about four times higher for children who had not received the corresponding recommendation (27.7%) compared with those who had (7.3%). However, it is important to emphasize that almost three quarters (72.4%) of the children who had begun attending USS without having received the respective recommendation remained there without dropping out until at least Class 9. The transition rate for this group (27.7%) was thus considerably lower than that for its equivalent in MSS (38.7%). Chi2 was significant in all 10 datasets.

Overall, findings confirmed our expectations regarding the connection between teachers’ recommendations, the initial type of secondary school, and upward and downward mobility during secondary schooling. The recommendation was clearly associated with both the chances of remaining in and the risk of dropping out of a particular learning environment. This observation may be taken as a sign for the high prognostic validity of primary teachers’ recommendations and thus confirms their pedagogical value.

Parental Level of Education

As a second important factor influencing educational success and mobility in secondary schooling, we considered the association between changes in the type of school and the parents’ level of education. Figure 8.1 illustrates this relationship graphically. White bars indicate upward transitions; black bars, downward transitions. The given percentages relate to the share of children who experienced a transition within each type of school and the parental level of education. All Chi2 tests resulted in significant differences.

Fig. 8.1
A bar chart of transitions by the parental level of education. The two types of transition are downward and upward transition. The levels of education are of three categories lower, middle, and upper secondary schooling.

Transitions by parental level of education. (Source: Own calculations based on data from NEPS Starting Cohort 4 (release 9.1.0))

Full size image

The bar graph reveals very clear trends for the association between parental education and children’s educational trajectory. In LSS, only 4.7% of children from families with a low level of education accomplished an upward transition, whereas 8.1% of children from families with a middle level of education and 11.7% of children from families with a high level of education managed to upgrade their initial type of secondary school. A similar albeit less pronounced pattern emerged for MSS in which upward transition rates covaried with parental education and ranged from 2.1% to 3.3%.

With downward transitions, an inverted trend could be found for both MSS and USS. The risk of dropping down was substantially higher for children from low-educated families and lowest for children from higher-educated families.

Altogether, the higher the parental level of education, the lower the risk for a downward transition and the higher the chance for an upward transition. This pattern could be found in all three learning environments, and it appeared to be stronger for downward than for upward transitions.

Migration Status

Another traditional factor in research on education inequalities is the migration status of the family. Various disadvantages for foreign children have been well-documented in the German education system (Diefenbach, 2010). Considering the transition from primary to secondary school, children from migrant families have a lower chance of receiving a recommendation for USS and a higher risk of changing to LSS (Kristen, 2002). In our sample, the initial share of children with a migrant background was 11.4% in LSS, 10.4% in MSS, and 8.2% in USS, which is in line with the literature and our expectations. Table 8.4 displays the share of students with and without transitions both for children from German families and for children from families with a migrant background.

Table 8.4 Transitions by migration status (row percentages)
Full size table

In LSS, 6.7% of the German children changed to a higher type of school, whereas 9.6% of children from families with a migration background managed to improve their learning environment. The slightly lower upward transition ratio for German children was a surprising finding, because we would have expected another disadvantage for migrant children here. However, differences (Chi2) were significant in only four of the 10 imputed datasets.

In MSS, we monitored a similar pattern, although the difference was even less pronounced (Chi2 was significant in only two imputed datasets). Whereas 2.7% of German children moved upwards from MSS, the rate for children from families with a migration background was slightly higher at 3.1%. For downgrades, however, the trend was reversed, with 12.1% of German children and 15.1% of children from migrant families changing to a lower type of school.

Among students beginning secondary education in USS, 8.9% of those from German families experienced a downward transition, whereas 10.9% of those from families with a migration background changed schools to attend a less ambitious type of school. Chi2 was significant in four out of the 10 imputed datasets.

Gender

In research on inequalities in education, testing for gender differences has a long tradition. As with the dimensions detailed above, we compared the proportions of upward and downward transitions for boys and girls separately over all initial types of school (see Table 8.5). In LSS, 8.7% of the female students changed to a more prestigious type of school, whereas the share of male students with such a transition was slightly lower (5.7%). A similar picture was visible in MSS in which 3.7% of the girls and only 1.9% of the boys experienced an upward transition. In contrast, the girls’ share in downward transitions (11.3%) was lower than the boys’ (13.5%). From USS, 7.6% of the female students and 10.8% of the male students moved downwards to a less demanding type of school. Even though the difference in transition rates was comparatively small, girls clearly performed better than boys across all three types of school. All Chi2 tests resulted in significant differences.

Table 8.5 Transitions by gender (row percentages)
Full size table

Academic Performance

Finally, we considered academic performance as an important dimension of mobility during secondary education. In our retrospective data, only performance test results for Class 9 were available to us. This meant that we could not consider academic performance prior to the transition in our analyses. However, assuming that academic performance at a certain point in the educational trajectory is determined largely by previous performance in the respective domain, we expected to find differential competence profiles for varying transition types. In the domain of reading, the students achieved a mean test score of x̄ = 0.131. In mathematics, the mean test score was x̄ = 0.221, and in science, it was x̄ = 0.140 (cf. Appendix, Table 8.A4). Because test scores were standardized over the initial sample, these results suggest that our analysis sample was somewhat positively biased in terms of academic performance.

When differentiating performance by initial type of school, we found the expected pattern. For all three competence domains, students beginning secondary education in LSS achieved the lowest test scores (see Table 8.6); students beginning secondary education in MSS achieved test scores that were a bit lower than the overall average scores; and children in USS achieved the highest test scores.

Table 8.6 Competence scores by initial type of school
Full size table

So far, these results confirm the common view that different types of school constitute different learning environments. After primary school, the performance-based preselection of students into different educational programmes is reinforced by differences in curricula and style of instruction, and this results in different learning paces and consequently in diverging competencies. For students with a change in their type of school during secondary schooling, two distinct environments influence competence development. Therefore, we would expect such cases to feature a level of competence that lies between the typical levels of either of the two types of school. Figure 8.2 illustrates the differential competence averages in reading, mathematics, and science for stable students within each type of school and for students with an upward or downward transition (for additional information on standard errors and confidence intervals, see Appendix, Table 8.A5). Changes of the initial type of school indeed resulted in quite different levels of competencies compared with students who remained continuously in one type of school. All differences in competence scores between stable students and those who left their initial type of school (in either direction) were significant without exception. Cases of students with an upward transition from LSS had significantly higher skill levels on average in all three competence domains than stable LSS students. However, the group of former LSS students did not attain the average competence level of stable MSS students in the fields of reading or science and therefore assumed a position among the weaker students in their new learning environment. In contrast, even though their competence levels were significantly lower than those of their former schoolmates, students with a downward transition from MSS outperformed their new peers in reading and science and therefore assumed a leading position in the new learning environment. The same patterns relating to competence differences in all three domains appeared for students with an upward change from MSS and for students with a downward change from USS.

Fig. 8.2
A bar chart with negative and positive values on the Y axis represents competence scores by transition type. The transition types include reading, mathematics, and science. The x-axis reads L S S stable, M S S down, L S S up, M S S stable, U S S down, M S S up, and U S S stable.

Competence scores by transition type. (Source: Own calculations based on data from NEPS Starting Cohort 4 (Release 9.1.0))

Full size image

In total, upward transitions in education seemed to come at the expense of a reduced competence development relative to the new reference group, whereas a downward transition turned out to be of advantage in terms of a leading position among the new peers. As we know from studies investigating reference-group effects such as the big-fish-little-pond effect (see e.g., Marsh & Hau, 2003), a change in the learning environment is closely connected to the development of the academic self-concept, effort, and motivational dimensions. We hence conclude that considering upward and downward transitions as either mere advantages or disadvantages respectively means taking a one-sided perspective that ignores the other side of the coin. Our findings suggest that students considering an upward transition have to deliberate about whether they prefer achieving a higher school-leaving certificate that is nonetheless below average or achieving a lower school-leaving certificate that is nonetheless above average.

5.4 Multivariate Findings: Downward Transitions from USS

As a final step of our analyses, we merged all variables into a single model with a logistic regression analysis. The dependent variable was downward transition from USS. Table 8.7 displays the results (y-standardized coefficients). We found that a recommendation of the primary school teacher for LSS increased the downward transition risk considerably (.783). A recommendation for MSS also led to a higher chance of dropping out (.456), and only a recommendation for SCE had no different effect than a USS recommendation. The parental level of education had a substantial impact on downgrade transitions for both parents with an LSS school-leaving certificate (.482) and for parents with an MSS school-leaving certificate (.334). The migration status had no effect on downgrades. For gender, we found that boys had a considerably higher risk of dropping out of USS by Class 9 (.332) than girls did. Competence had a negative effect on transitions, with mathematics asserting the strongest influence (−.308), followed by reading (−.137). After controlling for the other variables, science literacy did not additionally affect the transition risk.

Table 8.7 Downward transition from USS (logistic regression models)
Full size table

Apart from migration status and science literacy, all the factors considered had a significant effect on the risk of leaving USS by Class 9. In total, the influences lead to a pseudo R2 of .336, which can be seen as a reasonable fit for a model that does not consider the particular transition timing and marks.

6 Discussion

International comparisons reveal striking differences in educational systems, especially according to the forms of ability grouping that are applied. Moreover, there are ongoing disputes about the advantages and disadvantages of ability grouping and its consequences for educational success and for inequalities in educational opportunity (OECD, 2016). We analysed mobility between different types of school in the German secondary school system with reference to rational choice and cultural reproduction theories. We assumed that the processes behind mobility were comparable with those that lie behind entering a type of school for the first time. The dataset we used was a nationwide German sample from the National Educational Panel Study (NEPS, Starting Cohort 4). Our findings reveal that the most common occurrence is remaining in the initial type of school, and that moving down is generally much more likely than moving up. There is little evidence to suggest that the openness of secondary schooling in Germany offers any great opportunity to catch up to the level of a higher type of school if this level has not been reached immediately after primary schooling. At the initial transition from primary to secondary schooling, most parents opt for the type of school recommended by their child’s teacher. If they decide to deviate from this recommendation, the probability of moving up or down later on is considerably higher. Both performance measures that we used (teachers’ recommendations and competence tests) influenced the risk of leaving USS by Class 9 comparatively strongly. This indicates that the openness of the education system leads to a continued sorting of children according to their ability during the phase of secondary schooling. However, of particular relevance to educational inequalities is the observation that even after controlling for performance, the parental level of education still has a significant effect on the risk of experiencing a downward transition. Considering the observation that students from higher-educated parents have a greater chance of entering USS, this result reveals that children from low-educated parents are sorted out even further during secondary schooling. Hence, the social composition of children at USS becomes more and more selective over time. Given that upward transitions from USS and LSS are also more likely with increasing levels of parental education, the German educational system appears to stratify children continuously according to their ability and their socio-economic status. This finding implies that the gap between more and less privileged groups is hardly narrowed but rather widened over time by opportunities to move between types of schools, which indicates a continuing process of ‘cumulative advantage’ (cf. DiPrete & Eirich, 2006). Furthermore, our analyses reveal a persistent disadvantage for boys in the selection processes during secondary schooling.

Overall, mobility in secondary education remains at a moderate level and represents more of an exception than the rule. Although not all students undergo a transition during secondary schooling, our findings suggest that selectivity during this stage is quite similar to the mechanisms that are at play during the transition from primary to secondary schooling. In addition to academic achievement, the effects of social origin on educational trajectories persist or even increase during the stage of secondary schooling. In terms of both rational choice and cultural reproduction theories, these results raise several implications. Because upward moves are not obligatory, secondary effects as defined by Boudon may be of particular relevance for students who exhibit the necessary performance. Downward moves, on the other hand, can occur either as a free decision or out of necessity, which indicates a potentially higher relevance of primary effects for low-performing students in MSS and USS. Unfortunately, we have no measure indicating the voluntariness of a downgrade; thus, clear statements on the proportion of primary and secondary effects of social origin remain a task for future research. Considering Bourdieu’s theoretical framework, downward moves can be regarded as a consequence of a poor fit between a student’s academic habitus and the overwhelming demands of the initial school culture. After the transition, an improvement of this fit and hence an improvement in school outcomes were to be expected. In contrast, upward moves should be more likely to hinder the apparently extraordinarily good fit between a student’s habitus and the former school environment. In such cases, the chances of obtaining a higher school-leaving certificate would come at the expense of giving up a top-ranking position among former classmates and having to adjust to the demands of a new environment and reference group. From this perspective, deciding against a transition and remaining in a relatively advantageous position may even be regarded as a worthwhile option. Follow-up analyses could aim to confront the consequences of these vastly different strategies and assess which one turns out to be more successful in terms of later social position or occupational success.

Finally, there are some limitations to the present analyses. The performance measures from Class 9 were gathered after a transition had occurred between different secondary schools. Although this situation is not ideal, we consider these measures to be feasible proxies. Ideally, instead of using retrospective information on education, concurrent longitudinal follow ups would have been better. However, using data from other sources or another cohort in the German NEPS (SC2; Starting Cohort 2) would have implied other and even more severe shortcomings (dropout rates for school leavers who have been followed up individually, sample selectivity, missing data). Moreover, we ignored transitions concerning different types of comprehensive schools in Germany, because it would have made no sense to classify them as upward or downward moves. The highly stratified system of secondary schooling in Germany makes it very difficult to come to an exhaustive valuation of mobility during school years. With our analyses, we concentrated on so called horizontal mobility (within an education stage). However, this needs to be complemented by reflected vertical mobility (across stages) in order to create a comprehensive picture. In fact, an increasing number of students continue schooling after initially obtaining a low or middle school-leaving certificate. However, there is strong evidence suggesting that the decision to continue education after the time of compulsory schooling is based on similar factors to those examined here, indicating that it is again privileged social groups that benefit most from these options (Buchholz & Schier, 2015).

School systems within Europe apply different forms of ability grouping or differentiation, especially in secondary schools. Selection into different types of school is not the most common form applied—at least, not in an early stage. From our results, it seems that external differentiation is not an ideal method of avoiding social selectivity. Our next step will be to investigate which school-leaving certificates students earn by staying in or leaving their initial type of school. Concerning research in the future, it would also be of interest to compare the different systems within Europe in respect to openness, levels of achievement, (social) selectivity, and the school-leaving certificates obtained.