Track prejudice in Belgian secondary schools: examining the influence of social-psychological and structural school features

ABSTRACT

While considerable research in education has established objective and subjective status differences between tracks and focused on the outcomes of ability grouping on students’ educational and broader outcomes, there is virtually no research that explains students’ variability in track valuation. This study relies on theoretical insights from social psychology, ethnic studies and school effects research to develop hypotheses about the influence of individual and school level features on students’ track valuation. Data from The School, Identity and Society survey, involving 4,540 adolescents from 64 Belgian schools is utilised, using multilevel modelling. The findings show the relevance of social identity theory and social norms in students judging all tracks; and track chauvinism, patriotism and cross-track friendships in explaining variability in students’ prejudice towards other tracks. However, these relationships vary according to the track position of the student. The conclusions discuss the implications of this study for future research and social policy.

KEYWORDS:

• Educational tracking
• secondary education
• quantitative research
• patriotism
• chauvinism
• prejudice

1. Introduction

Educational systems widely apply homogeneous ability grouping to improve educational outcomes (Hallinan et al., 2003). More specifically, tracks (Belgium/US) or streams (UK) divide students into separate groups wherein they are taught different curricula and prepared for disparate futures (Gamoran, 1992). Research, however, has shown that tracking creates/enhances inequality for outcomes like achievement, deviant behaviour and sense of futility (Hanushek & Wößmann, 2006; Van Houtte & Stevens, 2008, 2015).

In society, different statuses are assigned to these tracks due to the different employment opportunities they create (Andersen & Van de Werfhorst, 2010). Consequently, tracking can cause a ‘cascade effect’ wherein secondary education students ‘aim high’, by trying the academic tracks first and eventually ‘go down’ to ‘lower’, more practical, technical tracks (Boone et al., 2018). Students experience judgement based on track position, with vocational and technical students feeling ‘looked down upon’ (Spruyt et al., 2015).

Educational tracking can also lead to prejudice towards students, because people connect objectionable group qualities to certain tracks, leading to hostile or negative attitudes towards students solely based on these group qualities (Allport, 1954). Both teachers and students believe vocational students are not willing or able to fulfil educational expectations (Stevens & Vermeersch, 2010). Vocational students are, among other things, perceived by teachers as less confident (Ecclestone, 2004), more disinterested and disruptive (Van Houtte, 2017). Socially, these students are considered less refined (Stevens & Vermeersch, 2010). Based on these negative characteristics, society values vocational track membership negatively, which can lead to negative attitudes and prejudice towards vocational education.

Studying track prejudice is important, as psychological research shows that experiencing prejudice has both physical and psychological consequences (McCoy & Major, 2003). Threats against the ingroup result in lower self-esteem and depressed feelings, depending on how strongly someone identifies with the ingroup (McCoy & Major, 2003). Prejudice can also influence educational outcomes, as track stigma-consciousness leads to a higher sense of futility in students (Spruyt et al., 2015).

Despite the importance of track prejudice, there is virtually no research that tries to explain variability in students’ track prejudice. Although track prejudice differences exist between countries (Andersen & Van de Werfhorst, 2010), there are no studies that aim to establish if students show prejudice towards tracks. In addition, there is no research that looks at individual and school factors that might inform students’ track prejudice. In sharp contrast, there is a wealth of research that explains variability in ethnic prejudice in schools, looking at both individual (e.g. Stevens et al., 2014) and school level features (e.g. Vervaet et al., 2018). This study builds on this educational research by investigating the influence of individual and school-level features on students’ track prejudice in the Belgian context.

2. Theoretical framework

2.1. Individual determinants of track prejudice

According to Social Identity Theory (SIT; Tajfel et al., 1979) people show the most affinity with the social group to which they belong (‘ingroup’), regardless of context. This ingroup favouritism is not automatically linked to people’s (negative) evaluation of outgroups or the status distinctions among them (Brewer, 2001). The empirical evidence for ingroup favoritism is ample, for example between national and religious groups (e.g. Cairns et al., 2006). Therefore, we hypothesise that, in general, students will show the most positive feelings towards their own track, regardless of its place in the societal status hierarchy.

Nationality and identity research shows that there are at least two ways in which people can positively perceive their ingroup, by being either chauvinistic or patriotic. Chauvinism is a positive ingroup feeling based on a sense of ingroup homogeneousness and superiority, leading to exclusionary and hostile attitudes towards outgroups (Raijman et al., 2008). Chauvinism is usually studied in relation to feelings of threat, namely a competing (ethnic) identity threatening the majority identity or the perceived threat to welfare by migrants (e.g. Coenders et al., 2004; Van der Waal et al., 2010). In the context of educational tracking, the threatened status of vocational education students is a reflection of the lesser valued blue collar jobs they are trained for (e.g. Andersen & Van de Werfhorst, 2010). Patriotism is a positive ingroup evaluation based on ingroup pride and does not elicit hostile outgroup attitudes (Raijman et al., 2008). Patriotism is often connected to mental wellbeing, for example higher patriotism lowers the impact of racist experiences (Bynum et al., 2008) and heightens self-esteem (Davis et al., 2017). Ethnic and national identity research shows that chauvinistic attitudes increase prejudice and feelings of antagonism towards outgroups, while patriotic attitudes decrease antagonism and might even stimulate feelings of unity towards outgroups (e.g. Carter & Pérez, 2016; Huddy & Del Ponte, 2019; Stevens et al., 2014). In times of conflict or perceived threat, chauvinism can also act as a defence mechanism by stimulating a ‘my group first’ mentality and employing outgroup derogation to fight back against the outgroup (Huddy & Del Ponte, 2019).

Chauvinism is generally studied from the viewpoint of the dominant group. For example, studies on nationalism and ethnicity often only include the perception of dominant national citizens vis-à-vis vulnerable outgroups like immigrants (e.g. Raijman et al., 2008). An exception to this is Carter and Pérez (2016), who studied pride and national chauvinism in both Whites and non-Whites in their opinions on immigrants. Their findings correspond with the general tendency of chauvinism causing feelings of hostility and patriotism leading to positive or at least neutral feelings towards outgroups for non-dominant groups. Linking these insights to educational track identification, more chauvinistic students are assumed to have more negative outgroup evaluations, whereas patriotism leads to neutral or positive outgroup evaluations, called the track identification hypothesis going forward. These effects are hypothesised for all three tracks.

2.2. School determinants of prejudice

Social norms are rules that guide/constrain social behaviour, are generally known by members of the same group and shared through social networks, not enforced by law (Cialdini & Trost, 1998). For educational tracking, the societal norm on education and employment enforces thinking and acting in terms of a track status hierarchy, with academic education on top and vocational education at the bottom. Social norms not only guide behavior but can also become internalised values (e.g. Moss, 2003). Therefore, we hypothesise that students will judge tracks in accordance with the societal track hierarchy norm, but whether that is the case for all tracks or just the outgroups might depend on ingroup favouritism.

Educational research on ethnicity has shown that students’ exposure to diversity is associated with lower levels of prejudice towards outgroups (Vervaet et al., 2018). Regarding educational tracking, this diversity is represented by schools that offer multiple tracks (e.g. the United States and United Kingdom) or schools that provide one track (e.g. Japan) (Chmielewski, 2014). Yet in countries like Belgium and Germany tracks are organised both between and within schools. So-called categorical schools provide one track and multilateral schools provide multiple tracks. Belgian education providing both is ideal to test the effects of school composition and mere exposure on tracked intergroup dynamics within the same (societal) status hierarchy.

We study the importance of school composition on track prejudice through the mere contact hypothesis. Exposure to an unknown stimulus, in our case students from another track, can lead to lower prejudice, by gaining familiarity with and a disappearing feeling of threat from the stimulus through neutral, positive, or sometimes even negative interactions (Zajonc, 1968). Bornstein’s review (1993) confirmed that mere exposure is robust and reliable in lowering prejudice. The likelihood of social contact between students from different tracks will increase when both groups co-exist in the same school environment since co-existing fosters exposure (Van Houtte & Stevens, 2015). Therefore, we hypothesise that the mere contact hypothesis takes effect more often in multilateral schools compared to categorical schools, causing lower track prejudices in multilateral schools.

Secondly, students can create between group friendships. A meta-analysis by Pettigrew and Tropp (2008) indicates that cross-group friendships lower prejudice through mutual perspective-taking and empathising. Although it could be argued that it is more likely for students with the most positive outgroup attitudes to search for more cross-group friendships, the positive effects of cross-group friendships on reducing outgroup prejudice are real, even beyond initial outgroup attitudes (Titzmann et al., 2015). Therefore, we hypothesise that social interactions between tracks will lead to more positive sentiment towards each other. The extent to which these positive effects of school composition can be attributed to mere exposure or it leading to more cross-group friendships can currently not be specified.

2.3. Current study

This paper looks at how tracked students evaluate their own tracked ingroup and outgroups, and the social-psychological and school structural features influencing these judgements. In doing so we test the applicability of social-psychological theories (SIT, mere contact hypothesis) and ethnic studies concepts (patriotism, chauvinism). Since there is no prior application of these theories and concepts in educational tracking, the hypothesised directions of effect for the hypotheses will be adopted from their original fields. Gaining insight into track identification and possible track prejudices could help identify the impact ability grouping has on students’ self-image and how they view outgroup students.

The literature led to three hypotheses. Firstly, the track status hierarchy hypothesis assumes that all tracks rate themselves the highest based on ingroup favouritism and other tracks according to the social hierarchy norm. Secondly, the track identification hypothesis expects patriotism to have a positive or neutral and chauvinism a negative influence on outgroup evaluations, while both positively influencing ingroup evaluation. Although most literature on patriotism and chauvinism adopts the dominant group’s view, we expect these effects to be the same for the non-dominant technical and vocational track (Carter & Pérez, 2016). Lastly, the social influence hypothesis adopts the established effects of mere contact and between group friendship from identity research, assuming that any social interaction, through mere exposure and cross-group friendships, will stimulate a more positive outlook on the tracked outgroups.

2.4. Belgian education

Starting secondary education, students must choose either the general A-stream or remedial B-stream. The A-stream is considered the regular programme. Yet within this stream students must choose elective courses in the first year, such as Latin or technology, preparing them for the official track choice after the second year. From the third year onwards, students are divided into tracks they chose based on educational interest and grades. In the third through sixth year of Belgian secondary education (USA: Grades 9–12; UK: Years 10–13), students are offered four tracks: academic, technical, vocational or arts. These are the same for the Flemish and Francophone community (representing two separate educational systems), sampled in this study (Vlaams Ministerie van Onderwijs en Vorming, 2019; Wallonie-Bruxelles, 2019). The academic track prepares students for higher education. The technical track provides both general and technical-theoretical courses. The vocational track focuses on training students for a craft (Vlaams Ministerie van Onderwijs en Vorming, 2019). Very few students are enrolled in the arts track, therefore it is not included in this study. A typical feature of tracking in Belgium is the ‘cascade-effect’ (see introduction), which shows that the academic track is interpreted by students and parents as the educational standard. A clear track hierarchy with rigid boundaries and downward movement pattern shows the emphasis on status differences in this tracking system.

3. Methods

3.1. Sample

This study uses The School, Identity and Society (SIS) survey (Maene et al., 2021). The ethics committee of Political and Social Sciences at Ghent University approved this survey in accordance with the ethical and confidentiality requirements. The dataset relies on a mixed-method research design (QUAL > QUAN). Qualitative explorative research was used to identify inductively which collective identities were important to Belgian secondary school students; the importance of tracking as an identity became apparent. Consequently, the quantitative survey included students’ track identities.

The quantitative SIS-dataset contains 4,540 third year secondary school students. Sixty-four schools participated: seven schools in Wallonia, 29 in Flanders and 28 in the Brussels Capital Region. The schools were selected through multistage sampling: first, cities with a history of migration were randomly selected, and secondly the secondary schools were divided into strata based on their track variety. This led to a division of purely academic schools, technical-vocational and multilateral schools. Schools were randomly invited from each stratum to participate in the survey.

The data collection took place from September 2017 to December 2017. The school principals distributed an information letter to all pupils and their parents, informing them on the research theme, timing and the anonymous and voluntary nature of the study. This letter gave parents the option to permit their child to participate in the survey. In the Brussels Capital Region more parents (8%) withheld permission than in Flanders (6%) or Wallonia (6%). Of the students with parental consent, 80% participated by completing questionnaires which were distributed and filled out in their classroom. There was no pattern detected in the non-consent. This survey moment was monitored by a researcher, answering pupils’ questions, and teachers, supervising pupils while respecting their privacy.

3.2. Variables

3.2.1. Dependent variable

Students’ track evaluations: measuring students’ feelings towards their own and other tracks, formulated as: ‘Express your feelings towards the following groups on a scale from 0 (totally negative) to 100 (totally positive). Fifty means you are neutral towards that group’. For the total sample the academic track is most highly rated with an average score of 76.25 (Table 1, N = 4540; SD = 24.13), followed by the technical track at 71.16 (N = 4540; SD = 21.45), and the vocational track at 56.98 (N = 4540; SD = 28.67). This ‘feeling thermometer’ is widely used to measure group evaluations, and is a good indicator of global in- and outgroup attitudes (e.g. Verkuyten, 2005).

Table 1. Descriptive statistics for student-level variables – frequencies, means, standard deviations, and F-tests or Cramer’s V comparing the academic, technical and vocational track.

Variable (N = 4540)	Total		Academic (N = 2723)		Technical (N = 1034)		Vocational (N = 783)		Difference
	Mean or %	SD	Mean or %	SD	Mean or %	SD	Mean or %	SD	F-test or Cramer’s V
Opinion on academic track (N = 4198)	76.25	24.13	83.34	17.99	67.37	26.71	60.97	29.38	373.26***
Opinion on technical track (N = 4181)	71.16	21.45	69.76	20.49	79.14	19.19	64.36	24.80	109.12***
Opinion on vocational track (N = 4236)	56.98	28.67	55.94	26.91	49.14	30.34	71.55	27.49	135.00***
Patriotism (N = 4405)	11.80	2.60	12.11	2.35	11.67	2.67	10.92	3.09	63.40***
Chauvinism (N = 4458)	7.23	2.83	6.83	2.76	7.65	2.81	8.08	2.83	74.10***
Cross-group friendships (ref.= less than half) (N = 3793)	45.3		59.8		20.4		27.3		0.362***
Sex (ref.= male) (N = 4540)	48.8		44.5		56.3		54.8		0.113***
SES (N = 4540)	48.08	16.65	50.11	17.28	45.51	15.78	44.40	14.42	52.74***

3.2.2. Independent variables, student level

Track membership: Students are part of only one track. There are 2,723 surveyed academic track students, 1,034 technical and 783 vocational track students.

Chauvinism: This 5-item scale is based on exploratory qualitative interviews that preceded the quantitative SIS-survey and aims at measuring students’ sense of track superiority, through the items: (1) ‘students in my track are smarter than those in other tracks’, (2) ‘are more capable’, (3), ‘are more creative’, (4) ‘students in my track are cooler’ and (5) ‘my track is harder’. These were measured on 5-point scales with 1 signifying ‘absolutely disagree’ and 5 signifying ‘totally agree’, recoded so a high score represents high chauvinism. ‘Students in my track are cooler’ was omitted as it theoretically seems too distinct from the other, more academically oriented items. After performing a factor analysis for the total sample and per track, ‘my track is harder’ was omitted due to considerably lower standardised factor loading, below or around 0.5, whereas the other items scored above 0.7 (Hair et al., 2006). Therefore, only the other three items were retained (Table 1, N = 4458; M = 7.235; SD = 2.831; α = 0.839) (analyses available on request). All three tracks differed significantly from each other at the 5% level.

Patriotism: To measure students’ sense of track pride, the MIBI-teen scale (Scottham et al., 2008) was adapted. This contains three items: ‘I am proud of this track’, ‘I am happy in this track’ and ‘I feel good about this track’. These were measured on 5-point scales with 1 signifying ‘absolutely disagree’ and 5 signifying ‘totally agree’ (Table 1, N = 4413; M = 11.80; SD = 2.60; α = 0.89). In contrast with the feelings thermometer, which gives a general evaluation of a group, the patriotism and chauvinism scales allow us to look more specifically at how affective and evaluative ingroup attitudes affect both in- and outgroup evaluations (Sellers et al., 1998). All three tracks differed significantly from each other at the 5% level.

Cross-track friendships: students were asked what proportion of their friendships are with students from other tracks. These were rescaled to a dichotomous variable with 0 being a minority to half of their friendships, and 1 being more than half their friendships. Most students have mostly cross-track friendships (N = 2073; 54.7%), with a minority having less than half of their friendships across tracks (N = 1720; 45.3%) (Table 1). All three tracks differed significantly from each other at the 5% level.

3.2.3. School level

School structure: this is a school level variable dividing schools into two categories: schools that do not organise the track studied in the outcome variable (scored as 0, reference category) and schools that do (scored as 1). The reference group is called ‘separate schools’, since they are separated from the studied outcome track, and the dummy is named ‘same schools’ (Table 2). The tracks organised in each school were established through official school websites or when school data were missing, self-reported student data.

Table 2. Descriptive statistics for school level variable: school structure.

Variable	School structure relative to academic track		School structure relative to technical track		School structure relative to vocational track
	No academic track	Academic track present	No technical track	Technical track present	No vocational track	Vocational track present
Academic track students	/	2723 100%	1648 60.52%	1075 39.48%	2425 89.06%	298 10.94%
Technical track students	496 47.96%	538 52.03%	/	1034 100%	308 29.79%	726 70.21%
Vocational track students	555 70.88%	228 29.12%	90 11.49%	693 88.51%	/	783 100%
No. of Schools	20	44	27	37	36	28

3.2.4. Control variables

Sex: Sex is almost equally distributed in our sample: 48.8% of the sample were male. The academic track had mostly girls (55.5%); in the technical and vocational track the majority were male (56.3% and 54.8% respectively). Previous research showed that male students are usually more prejudiced than women (e.g. Vervaet et al., 2018).

Socio-economic status: Students’ SES is based on the profession of the parent with the highest occupational status, which was matched to the International Socio-Economic Index of Occupational Status (ISEI) (Ganzeboom et al., 1992). The higher the score on the ISEI, the higher the SES. The scale theoretically ranges from 10 to 90. For our sample the minimum score was 15, the maximum 90, with an average of 48.08 (N = 4540, SD = 16.67) (Table 1).

3.3. Design

First, to study the track status hierarchy hypothesis, we ran a repeated measures ANOVA on the track evaluation measures. This allows to test for both ingroup favouritism and whether students evaluate tracks hierarchically. Through an ANOVA with a post-hoc Bonferroni test we checked if track evaluations differ between tracks.

To test the track identification hypothesis and the social influence hypothesis we ran stepwise multilevel regression models using HLM6 software (Raudenbush et al., 2013). The stepwise regression started with an unconditional model (Model 0); next we added students’ track as dummies (Model 1). Students who are members of the track that is considered the dependent variable were used as the reference category for track membership. Subsequently, patriotism (Model 2), chauvinism (Model 3), school structure and the control variables (Model 4) are added. The continuous patriotism, chauvinism and SES variables were grand mean centred.

In Model 5 we added the cross-group friendship variable, with Models 4 and 5 testing the social influence hypothesis. By adding the interaction term between track membership and patriotism in Model 6 and the interaction between track membership and chauvinism in Model 7, we can determine the direction and significance of patriotism and chauvinism on track judgement for all tracks. To do this we perform a simple slope analysis of the interaction effects in Models 6 and 7 (Aiken & West, 1991). The models with the evaluated track as reference category are given in Tables 3 to 5. For the simple slope analysis only the relevant interaction coefficients are given in the results section. The full tables of these simple slope analyses are available on request. This regression was repeated three times, once for each track evaluation as the dependent variable, being the academic (Table 3), technical (Table 4) and vocational track evaluation (Table 5). All results in the tables are unstandardised, but standardised coefficients (y*) were also calculated to obtain comparable effect sizes and are presented in text. These coefficients were obtained by multiplying the regression coefficient by the standard deviation of the independent variable and dividing the multiplication by the standard deviation of the dependent variable.

Table 3. The Association between Track Membership, Chauvinism, Patriotism, Cross-Group Friendships, School Structure and Evaluation of Academic Track: Stepwise Two-level Multiple Regression (HLM 6) with evaluation of Academic Track as Outcome.

Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7
Intercept	83.576***	83.138***	83.221***	85.371***	81.913***	81.095***	81.905***
School level
School structure (ref.= separate school)				−3.080 (1.871)	−0.959 (1.804)	−0.352 (1.795)	−0.860 (1.864)
Student level
Technical track (ref.= academic track)	−14.965*** (1.246)	−14.067*** (1.321)	−14.150*** (1.277)	−15.168*** (1.610)	−15.020*** (1.588)	−14.902*** (1.509)	−14.990*** (1.629)
Vocational track (ref.= academic track)	−23.44*** (1.464)	−21.051*** (1.639)	−21.215*** (1.542)	−22.903*** (2.121)	−22.476*** (2.083)	−24.116*** (2.268)	−22.246*** (2.130)
Patriotism		1.400* (0.237)	1.107*** (0.230)	1.079*** (0.231)	1.114*** (0.246)	2.302*** (0.233)	1.128*** (0.248)
Chauvinism			0.212* (0.141)	0.281 (0.141)	0.328* (0.163)	0.356* (0.164)	0.584*** (0.140)
Cross-track friendships (ref.= less than half)					1.920* (0.813)	1.774* (0.829)	1.873* (0.807)
Sex (ref. = Male)				1.582* (0.626)	2.143** (0.695)	2.123** (0.656)	2.218** (0.694)
SES				−0.011 (0.022)	−0.017 (0.022)	−0.016 (0.023)	−0.017 (0.022)
Technical Track*Patriotism						−2.555*** (0.514)
Vocational Track*Patriotism						−2.877*** (0.556)
Technical Track*Chauvinism							−0.505 (0.334)
Vocational Track*Chauvinism							−0.905 (0.722)
Variance components
Intercept	11.508	7.149	6.492	11.112	10.948	8.112	9.939
Technical track	15.467	22.400*	20.529*	27.366*	26.574*	19.144*	29.002*
Vocational track	11.673	28.316	19.476	17.903	24.635	31.596	28.054
Patriotism		1.961*	1.772	1.775*	1.940*	0.781*	1.954*
Chauvinism			0.233*	0.232*	0.397**	0.449**	0.250**
Cross-track Friendships					9.221	9.402	8.578
Sex				2.993	2.816	1.834	2.836
SES				0.005	0.004	0.005	0.004

Note. All presented coefficients are unstandardised. *p < 0.05.; **p < 0.01; ***p < 0.001.

Table 4. The Association between Track Membership, Chauvinism, Patriotism, Cross-Group Friendships, School Structure and Evaluation of Technical Track: Stepwise Two-level Multiple Regression (HLM 6) with evaluation of Technical Track as Outcome.

Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7
Intercept	79.093***	79.609***	79.634***	77.675***	76.407***	76.751***	75.892***
School level
School structure (ref.= separate school)				1.494 (1.691)	1.560 (1.649)	1.353 (1.626)	1.466 (1.655)
Student level
Academic track (ref.= technical track)	−9.652*** (1.272)	−10.837*** (1.185)	−11.282*** (1.079)	−10.527*** (1.418)	−9.203*** (1.501)	−9.239*** (1.540)	−8.414*** (1.569)
Vocational track (ref. = technical track)	−14.996*** (1.675)	−14.101*** (1.686)	−14.490*** (1.634)	−13.926*** (1.581)	−14.615*** (1.475)	−15.667*** (1.479)	−14.221*** (1.470)
Patriotism		0.942*** (0.167)	1.033*** (0.154)	1.011*** (0.157)	1.029*** (0.161)	2.269*** (0.227)	1.005*** (0.160)
Chauvinism			−0.474** (0.146)	−0.439** (0.147)	−0.457** (0.165)	−0.442* (0.166)	0.541* (0.238)
Cross-track friendships (ref.= less than half)					1.589 (0.808)	1.413 (0.812)	1.715* (0.810)
Sex (ref.= Male)				0.877 (0.752)	1.170 (0.774)	1.162 (0.791)	1.139 (0.775)
SES				−0.021 (0.019)	−0.027 (0.022)	−0.027 (0.022)	−0.025 (0.022)
Academic Track*Patriotism						−1.479*** (0.337)
Vocational Track*Patriotism						−2.305*** (0.517)
Academic Track*Chauvinism							−1.600*** (0.337)
Vocational Track*Chauvinism							−0.461 (0.432)
Variance components
Intercept	4.458	3.744	4.286	10.846*	22.263*	26.861*	22.012*
Academic track	33.878*	24.281*	18.220*	19.599*	26.957**	26.518**	29.763**
Vocational track	37.904	44.171	37.288	36.743*	24.888	18.323	21.950
Patriotism		0.556	0.493	0.532	0.471	0.369	0.441
Chauvinism			0.428	0.467	0.557	0.590	0.166
Cross-track Friendships					7.875*	8.172*	8.084*
Sex				9.440	9.000	11.166	9.423
SES				0.002	0.006	0.006	0.006

Note. All presented coefficients are unstandardised. *p < 0.05.; **p < 0.01; ***p < 0.001.

Table 5. The Association between Track Membership Chauvinism, Patriotism, Cross-Group Friendships, School Structure and Evaluation of Vocational Track: Stepwise Two-level Multiple Regression (HLM 6) with evaluation of Vocational Track as Outcome.

Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7
Intercept	71.490***	72.573***	72.770***	69.924***	70.940***	73.745***	70.026***
School level
School structure (ref.= separate school)				1.205 (1.738)	0.268 (1.851)	−0.670 (1.844)	0.561 (1.817)
Student level
Academic track (ref.= vocational track)	−16.708*** (1.704)	−18.626*** (1.588)	−18.644*** (1.600)	−17.335*** (2.080)	−16.233*** (2.144)	−18.161*** (2.080)	−15.618*** (2.188)
Technical track (ref. = vocational track)	−21.489*** (2.156)	−22.802*** (2.094)	−22.125*** (2.225)	−21.703*** (2.158)	−21.231*** (2.177)	−23.356*** (2.119)	−20.391*** (2.201)
Patriotism		0.952*** (0.234)	1.083*** (0.222)	0.994*** (0.221)	0.736** (0.247)	2.116*** (0.496)	0.695** (0.245)
Chauvinism			−0.934*** (0.214)	−0.868*** (0.201)	−0.928*** (0.227)	−0.948*** (0.227)	1.259* (0.551)
Cross-track friendships (ref.= less than half)					−1.540 (1.091)	−1.575 (1.094)	−1.484 (1.113)
Sex (ref.= Male)				1.917 (1.048)	1.691 (1.083)	1.570 (1.094)	1.741 (1.085)
SES				−0.025 (0.027)	−0.038 (0.031)	−0.036 (0.031)	−0.037 (0.031)
Academic Track*Patriotism						−1.982*** (0.542)
Technical Track*Patriotism						−1.302 (0.671)
Academic Track*Chauvinism							−2.914*** (0.615)
Technical Track*Chauvinism							−1.970** (0.720)
Variance components
Intercept	20.405***	9.666**	18.175**	30.769*	82.143***	71.257***	71.172***
Academic track	49.017**	55.519**	44.618**	48.510**	59.803**	55.203**	60.990**
Technical track	85.646**	78.419*	98.570**	88.762**	93.814**	88.246**	89.127**
Patriotism		1.351**	1.189*	1.106*	1.344*	1.243*	1.312
Chauvinism			1.137	1.011	1.225**	1.219**	0.803
Cross-track Friendships					17.051*	17.886*	17.875*
Sex				20.234**	19.450**	22.062**	18.438**
SES				0.009	0.013	0.013	0.014

Note. All presented coefficients are unstandardised. *p < 0.05.; **p < 0.01; ***p < 0.001.

4. Results

Before interpreting the repeated measures ANOVA, we control and, if needed, adjust for sphericity. Mauchly's Test showed that all three tracks violate the sphericity condition. To correct for this and since the Greenhouse-Geisser Epsilon was bigger than 0.75 in all tracks, the Huynh-Feldt results are given, following the sphericity corrections by Howell (2002) and Field (2013). The within-group difference is statistically significant for the academic F(2, 4361) = 1405.69, p < 0.001; the technical F(2, 1708) = 382.29, p < 0.001; and the vocational track students F(2, 932) = 32.165, p < 0.001.

The repeated measures ANOVA showed that all tracks rate their ingroup the highest. The academic and technical track follow the societal hierarchy, from most to least academic, in their judgement of the other tracks. Yet the vocational track goes against the societal hierarchy, with their own track rated highest, then the technical track and lastly the academic track (Table 1).

The differences in judgement by students between their highest and lowest evaluated track is similar for academic and technical track students (27.4 and 30 points respectively). This difference is considerably smaller within vocational track students, only 10.58 points. It is noteworthy that in their self-judgement, the societally highest perceived academic track rates itself the highest (83.34), 4.2 points more than the technical track’s self-rating (79.14) and even 11.79 points more than the vocational students’ self-rating (71.55).

When studying the ANOVAs on the track evaluations, with track membership as factor, the ANOVA on the evaluation of the academic [F(2,4169) = 373.260, p < 0.001], technical [F(2,4152) = 109.126, p < 0.001] and vocational [F(2,4207) = 134.996, p < 0.001] tracks all showed significant differences, meaning that between the tracks there are significant differences in how each track is judged. All mean differences were significant at the 0.1% level, tested through the Bonferroni post-hoc test.

The intraclass correlation coefficients of the null-models of the academic, technical and vocational track evaluation indicated that there are significant differences at the school level in these judgements, with respectively 14.71 (p < 0.001), 5.94 (p < 0.001) and 5.85 (p < 0.001) per cent of variance at the school level. Multilevel analysis is therefore advised to account for possible school-level effects.

Table 3 shows the judgement of the academic track by all three tracks. Model 1 indicates that the academic track’s self-evaluation is significantly more positive than the evaluation of the academic track by their technical (y=-14.965; p > 0.001; y*=-0.260) and vocational counterparts (y=-23.436; p < 0.001; y*=-0.367). Patriotism has a significant positive association with track judgement (Model 2) (y = 1.110; p < 0.001; y* = 0.120). Chauvinism (Model 3) does not show a significant effect on academic track judgement (y = 0.212; p = 0.136; y* = 0.025). Adding school structure, sex and SES (Model 4) yielded one significant association, namely sex, with girls rating the evaluated group higher than boys (y = 1.582; p = 0.014; y* = 0.033). Cross-track friendships (Model 5) were significant: students with mostly cross-group friendships show a more positive academic track judgement (y = 1.920; p = 0.021; y* = 0.040). Adding cross-group friendships did not alter the effect of school structure, yet chauvinism did become significant (y = 0.328; p = 0.049; y* = 0.038). Model 6 studied the interaction of patriotism with track membership. Patriotism has a significantly positive effect for academic track students on their ingroup evaluation (y = 2.302; p < 0.001; y* = 0.248). According to the simple slope analysis, there is no effect of patriotism on evaluating the academic track for the technical (y=-0.255; p = 0.571; y*=-0.028) and vocational students (y=-0.573; p = 0.280; y*=-0.062). Model 7 tested the interaction of chauvinism with track membership. The academic students rate themselves significantly higher when feeling more chauvinistic (y = 0.584; p < 0.001; y* = 0.069). The interactions show no significant difference between the academic and the other tracks for the effect of chauvinism. For the technical (y = 0.079; p = 0.807; y* = 0.009) and vocational students (y=-0.321; p = 0.658; y*=-0.038) chauvinism does not affect how they rate the academic track.

For the evaluation of the technical track by all tracks (Table 4), Model 1 indicated that the technical track’s ingroup evaluation is significantly higher than the evaluation of the technical track by their academic (y=-9.652; p < 0.001; y*=-0.220) and vocational counterparts (y=-14.996; p < 0.001; y*=-0.264). Patriotism (Model 2) has a significant positive effect on the technical track judgement (y = 0.942; p < 0.001; y* = 0.114). Chauvinism (Model 3) initially showed a negative significant effect (y=-0.474; p = 0.002; y*=-0.063). The addition of sex, SES and school structure (Model 4) added no significant associations, nor did it alter any associations from Model 3. Model 5 showed a borderline insignificant effect of cross-group friendships on students’ opinion of the technical track (y = 1.589; p = 0.054; y* = 0.037). Model 6 shows a significant positive influence of patriotism on technical students’ ingroup judgement (y = 2.269; p < 0.001; y* = 0.275), which is significantly smaller for the other tracks. Based on the simple slope analysis, academic track students’ patriotism has a significantly positive effect in evaluating the technical track (y = 0.790; p < 0.001; y* = 0.096), yet vocational students’ patriotism has no effect (y=-0.035; p = 0.937; y*=-0.004). Model 7 shows that chauvinism has a significantly positive effect on the technical track’s ingroup judgement (y = 0.541; p = 0.027; y* = 0.071); the effect of chauvinism is significantly different in the other tracks. For the academic track (y=-1.051; p < 0.001; y*=-0.139) there is a significantly negative effect of chauvinism on the evaluation of the technical track. No discernable effect is seen with the vocational students (y = 0.077; p = 0.857; y* = 0.010).

Lastly, we examine the evaluation of the vocational track by all tracks (Table 5). Vocational students show a significantly higher opinion of the vocational track than the academic (y=-16.708; p < 0.001; y*=-0.286) and technical track (y=-21.489; p < 0.001; y*=-0.314), remaining significant throughout all models. Patriotism (Model 2) relates significantly positively with vocational track judgement in all models (y = 0.952; p < 0.001; y* = 0.086). Chauvinism (Model 3) initially shows a significantly negative effect (y=-0.934; p < 0.001; y*=-0.092). Neither sex, SES and school structure (Model 4), nor cross-track friendships (Model 5) show significant associations with vocational track judgement. Vocational students have a positive effect of ingroup patriotism on their track judgement (Model 6) (y = 2.116; p < 0.001; y* = 0.192). There is no significant effect of ingroup patriotism for the academic (y = 0.134; p = 0.606; y* = 0.012) and technical tracks (y = 0.817; p = 0.113; y* = 0.074) towards their vocational counterparts. In Model 7, the effect of chauvinism shifted from negative to positive (y = 1.259; p = 0.026; y* = 0.124). Chauvinism therefore has a significantly positive effect on vocational students’ ingroup judgement. This is significantly different for the other tracks. There is a significantly negative influence of chauvinism on the academic students’ judgement (y=-1.649; p < 0.001; y*=-0.163) of the vocational track. The technical students (y=-0.713; p = 0.130; y*=-0.070) show no significant effect of chauvinism.

5. Discussion

This paper looked at how tracked students evaluate their own ingroup and outgroups, and the social-psychological and school structural features influencing these judgements. The track status hierarchy hypothesis shows two clear tendencies. Firstly, every track prefers its own ingroup, in line with SIT (Tajfel et al., 1979). Additionally, social norms seemingly influence the judgement of academic and technical students, since they follow the societal hierarchy in judging the other tracks. Vocational students are the exception; they rate their academic counterparts the lowest. A possible explanation is that vocational students feel looked down upon by academic track students, leading to oppositional attitudes. Secondly, vocational students show less hierarchical division in track judgements, which can be linked to social norm theory. The societal norm enforces thinking in terms of hierarchy. Yet vocational students think less hierarchically, probably because they are ‘the victim’ of this norm. The track status hierarchy hypothesis is confirmed for academic and technical students, but not for vocational students.

In the academic track, rated highest by society, chauvinism and patriotism show the effects expected by the track identification hypothesis (e.g. Carter & Pérez, 2016; Huddy & Del Ponte, 2019; Stevens et al., 2014). Chauvinism causes more positive ingroup and significantly lower outgroup evaluations. Patriotism causes more ingroup positivity and is significantly related with a positive judgement of the technical track. For the technical and vocational tracks patriotism and chauvinism both significantly positively influence their ingroup judgement, as expected. Towards the other tracks, their ingroup patriotism and chauvinism hasve no significant effects. Based on the standardised coefficients, the effects of patriotism on ingroup judgement are generally stronger than those of chauvinism, whereas in the academic track, the effect of chauvinism on outgroup judgement is stronger than that of ingroup patriotism. These effects are smaller than the track differences, but still sizeable. The effects of the control variables are considerably weaker.

Interestingly, the negative effect of ingroup chauvinism on outgroup evaluation only occurs with the academic students, despite the vocational students being the most chauvinistic (Table 1). Vocational students being most chauvinistic is unexpected, as academic students are considered ‘superior’ in tracked education. Vocational students are placed at the bottom of the status hierarchy, making them more sensitive to their status position (e.g. Brown et al., 2011). This sensitivity can cause a defensive reaction by heightening chauvinism (Huddy & Del Ponte, 2019). To justify these feelings of superiority and feeling smarter, vocational students might attach more value to skills and labour experience than to ‘being book smart’.

The negative effect of ingroup chauvinism on outgroup evaluations might indicate that academic students internalise the idea that other students failed to reach the standards of the academic track (Stevens & Vermeersch, 2010). For the technical and vocational track their higher chauvinism might only be employed to defend themselves against their lower societal valuation by strengthening their ingroup identification (Huddy & Del Ponte, 2019).

The technical track’s position is ambiguous, since this track is societally placed above the vocational track and they do rate vocational students lowly (49.14), but this does not seem connected to feelings of superiority towards vocational students. Attributing this to bad interpersonal relationships does not seem likely since vocational students rate technical students more positively (64.36) than academic students (60.97). Future research should look at how chauvinism and patriotism are interpreted and employed by different groups within a status hierarchy and the technical-vocational track dynamic.

The social influence hypothesis has two components. Positive effects of cross-track interacting were expected, but whether these would occur through mere exposure based on school structure, cross-track friendships, or both, was not specified. Having mostly cross-track friendships has a significantly positive influence on the judgement towards the academic track, and only borderline insignificant towards the technical track (Pettigrew & Tropp, 2008; Titzmann et al., 2015). These effects are not replicated when evaluating the vocational track. There is no association between school structure and the evaluation of any track, nor does the addition of school structure alter any cross-track friendship effects. These results go against the hypothesised effect of exposure on outgroup evaluation. Future research could study if there are differences between tracks in how they experience school composition and how school structure shapes their opinion of other tracks.

The first limitation of this paper is that friendships were surveyed in a general sense, being within track or between track, but the respondents could not specify with which tracks they have more friendships. Future research could look deeper into these friendships and clarify whether the lack of cross-group friendship effect on opinions about the vocational track is due to students having fewer vocational friends or due to other reasons. Cross-track friendships seemingly benefit the perception of academic and technical track students. Stimulating cross-track friendships, which can be done both in school and through leisure activities, could help reduce prejudiced thinking in Belgian education.

The second limitation lies in the school structure variable. It asked whether students are part of a school that organises both their own track and the evaluated track. This does however leave variation in how school life is organised in the ‘same schools’ category. Some schools might facilitate close relationships between the offered tracks or stimulate a communal school culture, while others might separate the school life between tracks. So, although our results provide no proof for the mere contact hypothesis (Zajonc, 1968), they cannot deny its effect either.

Our findings suggest that track membership is meaningful for Belgian secondary school students and that status differences between tracks can structure how students look at their own and other tracks. While developing cross-track friendships can reduce track prejudice, this does not seem the case towards the vocational track.

Educators should devote extra attention to the connection students create with their track identity. Patriotism can for some lead to more positive outgroup attitudes. Chauvinism, to the contrary, causes feelings of superiority and a more negative perception from the highest regarded track towards ‘lower tracks’. Postponing track choice and valuing other competences in a more comprehensive system could lower chauvinistic attitudes. Prior research has already shown the benefits of more comprehensive systems (e.g. Chmielewski, 2014), but we would like to stress this again based on the results of this study, as tracking seems to stimulate feelings of superiority and negative between-group attitudes. Additionally, our results show that mere contact in school is not sufficient to change intergroup evaluations, yet friendships can do this. A difference between these two is that friendships are based on equal status, whereas for mere contact this is not always the case. Based on these findings, we propose that schools/policy makers should create more between-track contact based on equal-status, to lower prejudice and chauvinism (Pettigrew et al., 2011). Additionally, policy makers should advertise more that vocational jobs are well paid, in demand, enriching and require cognitive and technical skills to master, to counteract the negative image vocational education has. Educators should also pay extra attention to (vocational) students who receive outgroup derogation and to students internalising outside track judgement.

This study shows that students rank tracks in terms of status and sense prejudice towards other tracks, partly through how they see their own track. Given the need for young people to develop positive identities and the labour market shortage for highly motivated, technically skilled employees, more research is required to understand how young people, particularly in lower status tracks, can develop positive identities in school contexts. Future research can build on this study by developing more insight into the determinants of educational group prejudice in different (nationally specific) educational grouping systems, into how students cope with outgroup derogation experiences, based on the negative effects from chauvinism towards them, into the consequences of such experiences and the role played by teachers and school policy in this.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

We would like to thank both the Research Foundation Flanders (FWO) grant number [G024516N] and the BOF Special Research Fund (Ghent University) grant number [BOF.24Y.2021.0006.01] for providing funding on which this article is based.

Notes on contributors

Lorenz Dekeyser

Lorenz Dekeyser is a doctoral researcher in Sociology, funded by the FWO (Scientific Research Foundation Flanders) and BOF Ghent University (Special Research Fund). He is a part of the CuDOS research team at the Department of Sociology at Ghent University (Belgium). His main field of research interest is educational tracking in secondary education and how this affects students and teachers in terms of educational and identity attitudes.

Mieke Van Houtte

Mieke Van Houtte, PhD sociology, is full professor and head of the CuDOS research team at the Department of Sociology at Ghent University (Belgium). Her research interests cover diverse topics within the sociology of education, particularly the effects of structural and compositional school features on several outcomes for students and teachers. She is a member of the Royal Flemish Academy of Belgium for Science and the Arts.

Charlotte Maene

Charlotte Maene is a doctoral researcher in Sociology and teaching assistant for the Department of Sociology at Ghent University. She is part of the CuDOS research team. Her main field of research interest is ethnic inequality in secondary education and how this affects students and teachers in terms of educational and identity attitudes.

Peter Stevens

Peter A. J. Stevens is Associate Professor at the Department of Sociology, Ghent University and member of the research groups CuDOS and CESSMIR. His main fields of interest are race and ethnic relations, sociology of education and qualitative research.