The Role of Teacher–Pupil Relations in Stereotype Threat Effects in Flemish Secondary Education

Abstract

This study aims to test stereotype threat theory hypotheses using a pupil survey database from Flemish urban secondary education characterized by a stratified tracking system. We relate these systemic features to stereotype threat effects by adding teacher–pupil relations to our analyses. Our results show that stigmatized groups—ethnic minority pupils in vocational education—experience the most negative teacher–pupil relations. To protect their academic self-concept from stereotype threat, they are also most vulnerable to psychological disengagement, discounting negative teacher feedback, and to disidentification from education. Moreover, teacher–pupil relations play an important role in explaining stereotype threat effects.

Keywords

stereotype threat teacher-pupil relations urban education academic achievement gap student self-esteem

Individual members of socially disadvantaged ethnic minority groups are often stigmatized for having low levels of academic motivation and performance levels as characteristic for their ethnic group (Gibson, Carrasco, Pàmies, Ponferrada, & Rios, 2013; Valencia, 2010; Valencia & Black, 2002). Moreover, in noncomprehensive secondary education systems characterized by high socioethnic stratification between educational tracks, pupils with a socially disadvantaged ethnic minority background can suffer a double stigmatizing because ethnic minority pupils are often overrepresented in lower status educational tracks (Baysu, Phalet, & Brown, 2011). In many school systems, characterized by early tracking in secondary education, this often also generates stigmatization processes related to the (implicit) status hierarchy and the different societal appreciation of the various educational tracks (Kelly, 2009; Stevens & Vermeersch, 2010; Van Houtte, Demanet, & Stevens, 2012). What we aim to do in this article is to relate these processes of double stigmatization related to these structural inequalities in the educational domain to effects on teacher–pupil relations and pupil’s responses to the stereotype threat.

The underlying theoretical assumptions for formulating our hypotheses are based on the stereotype threat theory that argues that in response to these negative stereotypes, members of stigmatized groups can develop response strategies to protect their self-image from the threat these negative stereotypes pose. In the context of education, pupils can psychologically disengage their academic self-concept from their actual performances, which means that pupils’ self-perception about their academic competences does not (longer) reflect their actual performance levels. Other response strategies described in stereotype threat literature are pupils that discount negative feedback from teachers for being unfairly biased and/or pupils from stigmatized groups that disidentify from education (Steele, Spencer, & Aronson, 2002). What is crucial for these stereotype threat effects to occur is that a negative stereotype is present in a specific domain for a specific group and that the members of this group are conscious about these negative stereotypes (Steele & Aronson, 1995). Fundamental to our study of the impact of teacher–pupil relations on stereotype threat effects is that we can argue that Flemish secondary education constitutes such a “stigmatized” domain for particular groups of pupils, that is, pupils with a Turkish or Moroccan immigration background and/or those in the lower status vocational track. We will therefore first present existing empirical evidence for the stigmatization of these groups in Flemish education before explaining more about the response strategies described in stereotype threat theory.

In general, this article aims to complement previous findings on the effects of stratification processes due to educational tracking by focusing primarily on the consequences of negative stereotyping within these educational structures, taking into account the ethnic background of pupils and focusing on urban educational contexts. The theory of stereotype threat effects is applied to disentangle whether and how pupils in stigmatized groups respond to negative stereotypes in education and to study to what extent stereotype threat effects related to particular social identities—that is, pupils with an ethnic minority background and/or in a lower status track—are mediated by teacher–pupil relations in Flemish urban secondary schools.

More concretely, we rely on survey data collected from a large sample of native and ethnic minority pupils in Flemish urban secondary schools to try to answer the following research questions:

Research Question 1: Do youngsters in negatively stereotyped groups in Flemish education—that is, youngsters with a Moroccan/Turkish immigration background and/or in the lower status vocational tracks—experience more negative teacher–pupil relations than other youngsters?

Research Question 2: Do youngsters in these stigmatized groups respond to negative stereotypes by (a) psychologically disengaging their academic self-concept more often from their actual educational performances, (b) discounting negative feedback from teachers more easily as being unfair, and (c) disidentifying themselves more often from educational goals?

Research Question 3: Do the teacher–pupil relationships as experienced by pupils influence these stereotype threat effects?

We thus aim to contribute to educational research in urban contexts by relating the structural and institutional educational context to concrete teacher–pupil relations in urban schools characterized by a segregation of vulnerable ethnic minority groups in lower status vocational schools, and study whether and how stigmatization processes in these urban educational contexts lead pupils in these stigmatized groups to develop response strategies toward stereotype threat.

Structural and Institutional Features of Flemish Urban Secondary Schools as Educational Contexts for Studying Stereotype Threat Effects

Educational research has repeatedly shown that structural and institutional contexts remain important factors when explaining the position of (groups of) individuals in education (Crul, Schneider, & Lelie, 2012). Many scholars have shown that broader societal inequalities between social groups are also reflected in educational inequalities (Bernstein, 1971; Giroux, 1980). As also Anyon has argued, educational systems are to a large extent set up to maintain some sort of social order by channeling most of a society’s children through a more or less homogenized educational system. Reproduction with the aim to uphold social cohesion is thus central to educational systems which are then also by definition biased for being developed by the dominant group in society. As Anyon (1980, 2006) argues, a hidden curriculum underlies these educational systems and leads to reproduction of class and other inequalities in society.

Whereas in the 1960s and 1970s, research deconstructed the implicit power relations in the educational system impeding working class pupils to be successful in education (Bourdieu & Passeron, 1977), recent research focuses more on the educational position of ethnic minority pupils (Crul et al., 2012; Gibson et al., 2013). Various scholars have pointed out that the weak position of minority pupils is often explained in terms of a problematic home environment, while structural features within the educational system itself are seldom problematized or discussed (Gibson et al., 2013; Valencia, 2010). Although these findings often show a strong overlap between the explanatory power of having an “ethnic minority origin” and that of having a “low SES background,” culturalization discourses often ignore the importance of a socially disadvantaged background in the educational position of some ethnic minorities (Roosens, 1995; Valenzuela, 1999). Educational achievement gaps are rather explained by discourses in which members of ethnic minority groups are often accused of holding on too much to their mother tongue and of being unwilling to participate in society and its institutions such as education (Valencia & Black, 2002).

More specifically with respect to social reproduction in Flemish secondary education, structural features such as early tracking, socioethnic segregation, and the (implicit) status hierarchy between educational tracks are crucial systemic features to discuss. In theory, tracking only starts from the third year of secondary education.¹ Yet, in reality, also the supposedly comprehensive first 2 years of secondary education already initiate an early form of tracking by sorting pupils in academically or vocationally oriented schools. Furthermore, in the first year pupils are already tracked into an A-stream that leaves all track options open and a B-stream that de facto almost exclusively prepares pupils for vocational education. Starting from the third year, a full-fledged tracking system is installed where the general or academic track steers pupils toward higher education, while the technical track prepares pupils for either higher education or directly for skilled technical professions. The vocational track directs pupils to lower skilled professions after secondary education.² This hierarchical tracking structure almost exclusively allows for “downward mobility” between tracks, which is often referred to as streaming down the “waterfall” from the general to the technical, to the vocational track (Baysu et al., 2011). Studies have shown that the orientation of pupils into different educational tracks in Flemish secondary education is socially biased and pupils with a lower socioeconomic status (SES) and an ethnic minority background are more often oriented to lower status tracks, even when controlled for their academic performances (Boone & Van Houtte, 2012; Spruyt, Laurijssen, & Van Dorsselaer, 2009).

Educational systems characterized by (early) tracking are also shown to be vulnerable for processes of stigmatization, as these systems generally contain some form of explicit or implicit status hierarchy between tracks (Kelly, 2009; Stevens & Vermeersch, 2010; Trautwein, Lüdtke, Marsh, Köller, & Baumert, 2006; Van Houtte & Stevens, 2009). As a consequence, the vocational track and its pupil population suffers the negative stereotype of being a reservoir of pupils with all kinds of learning disabilities (Clycq, Nouwen & Vandenbroucke, 2014; Van Houtte et al., 2012). Vocational tracks are often perceived as the “waste basket” of hierarchical educational system comprised of pupils who are not capable or insufficiently motivated for more academically oriented tracks (Stevens & Vermeersch, 2010). Pupils in lower status tracks can therefore be considered to belong to a specific stigmatized social category (Van Houtte et al., 2012), and this tracking has shown to affect future aspirations and performances (Chun, 1988; Giroux, 1980).

Moreover, research on the educational context of urban education in Flanders clearly shows that particularly Turkish and Moroccan pupils are overrepresented in the vocational track of urban schools (Baysu et al., 2011). Turkish and Moroccan pupils and parents are often perceived by teaching staff as having a deficient proficiency in Dutch and lacking the competences and attitudes to support a successful school career. Self-reports on experiences of discrimination in the educational context indicate that Turkish and Moroccan pupils and parents are often aware of these stigma’s (Baysu et al., 2011; Phalet, Deboosere, & Bastiaenssen, 2007). A recent study of educational trajectories of minority pupils in secondary schools in the same Flemish urban areas has shown the problematization of ethnic minority habitus as not being compatible with the educational domain and the culturalization of “educational failure” by teachers (Clycq et al., 2014).

As one can expect that these structural and institutional features affect relationships in these urban schools, we will now explore some relevant findings regarding the importance of stigmatization in teacher–pupil relations.

The Role of Teacher–Pupil Relationships in Stigmatization

As Klem and Connell (2004) argue, in a school context, the teacher–pupil relationship is assumed to be the most important as it can strongly influence pupils’ school engagement and performance. Teachers are often key actors regarding pupils’ social integration and experiences in school, and feeling supported by teachers is shown to be crucial in pupils’ educational trajectories (Goodenow, 1993a; Van Maele & Van Houtte, 2011). Given the importance of these teacher–pupil relations, it is necessary to study how these relationships unfold in specific educational contexts. Flemish urban education is particularly relevant to study stratification effects due to tracking on teacher–pupil relationships as also teachers’ expectations and motivation are found to be influenced by the tracks and composition of the schools teachers are assigned to (Finley, 1984; Van Houtte, 2004).

Another important finding from previous research in Flanders is that teachers often hold negative stereotypes of ethnic minority pupils and pupils in the lower status vocational track (Clycq et al., 2014). Furthermore, teacher perceptions of the teachability and capabilities of their pupils proved to be an important intervening variable in explaining educational outcomes (Agirdag, Van Houtte, & Van Avermaet, 2012; Van Houtte, 2004, 2006). As shown above, pupils’ social and ethnic backgrounds are strong predictors of pupil’s orientation into a specific track (Boone & Van Houtte, 2012). Therefore, mainly pupils with a lower than average SES and ethnic minority background are overrepresented in urban vocational schools (De Meyer & Warlop, 2010). As a consequence, negative stereotyping of the lower status vocational track intersects with negative stereotyping of social groups such as the socially disadvantaged ethnic minorities, which make socially disadvantaged ethnic minority pupils in vocational education particularly vulnerable for stigmatization in education (Kelly, 2009; Stevens & Vermeersch, 2010).

The literature review outlined above clearly showed that negative stereotypes about ethnic minorities and lower status tracks are present in Flemish secondary education and that these negative stereotypes can influence teacher–pupil relationships. Urban secondary schools in Flanders can therefore be considered a relevant context to study stereotype threat effects.

Pupils Responding to Stigmatization in Education

In this section, we build on the theoretical framework of stereotype threat to discuss how pupils in stigmatized groups can respond to the negative stereotypes in education that threaten their (academic) self-concept. A prerequisite for feeling threatened by a certain stereotype is that the stigma is related to a relevant group identity that exists in a domain that is valued by the individual (Aronson et al., 1999; Steele et al., 2002). Nonetheless, it is not necessary that an individual is convinced that the stereotype will impede success in this domain, it is sufficient that the individual is conscious of the stereotype and values the domain enough to invest in negating the stereotype’s hampering effects (Aronson et al., 1999). The valuing of the educational domain seems clear, as it is universally held in high regard in society and success in education is seen as being crucial for future live opportunities (Clycq et al., 2014; Steele et al., 2002).

When pupils strongly identify with education and are conscious of specific stigmas related to their group’s performance in this domain, a stereotyped pupil is likely to reduce the negative effects of this stereotype threat on their personal well-being (Aronson et al., 1999; Steele & Aronson, 1995; Steele et al., 2002). Eventually, disidentification from the educational domain is possible as a more chronic response strategy, while various acute reactions can also be displayed, possibly leading up to the more chronic disidentification response strategy. A basic response to stereotype threat is the psychological disengagement of one’s academic self-concept and broader self-esteem from one’s educational performance. Psychological disengagement is therefore defined as a disconnection between one’s academic self-esteem from their actual performance in education (Major, Spencer, Schmader, Wolfe, & Crocker, 1998; Régner & Loose, 2006; Steele et al., 2002). This strategy serves the protection of a pupils’ academic self-concept in case the pupil feels threatened to confirm negative stereotypes about one’s social group in education, that is, performing poorly as an ethnic minority pupil (Régner & Loose, 2006; Steele et al., 2002). This basic principle of stereotype threat effects leads us to our first hypothesis: “Ethnic minority pupils and pupils in lower status tracks are more likely to disengage their academic self-concept from their actual achievements in education.”

Pupils’ protection of their academic self-concept from stereotype threat is also found to act as a mediator in the relationship between lower educational performance levels and both discounting of negative feedback and disidentification from education (Régner & Loose, 2006). Pupils’ psychological disengagement is therefore often accompanied by discounting the validity of negative feedback on their performances (Régner & Loose, 2006; Schmader, Major, & Gramzow, 2001). Moreover, the discounting of teachers’ fairness in evaluations strengthens the psychological disengagement response strategy because pupils’ academic self-concept is not threatened when the focus is shifted toward external factors: the test and the tester, but not the testee (Régner & Loose, 2006; Y. Stephan, Caudroit, Boiché, & Sarrazin, 2011). Furthermore, the pupil’s academic self-concept can remain “intact” and therefore discounting can go hand in hand with a high valuing of the educational domain, making it a less radical response strategy than a general disidentification from education (Nussbaum & Steele, 2007). This leads us to our second hypothesis: “Pupils belonging to stigmatized groups in education will discount the validity of teachers’ assessments more often when receiving “bad grades” on their academic performances.”

While discounting negative feedback on academic performances is a rather acute reaction to stereotype threat in education, a more chronic reaction is to personally disidentify from the domain of education (Steele et al., 2002; Y. Stephan et al., 2011). Here, one can still value the role of education for society but to protect one’s own academic self-concept, the individual devalues the goals set in the educational domain in relation to their own personal goal-setting (Major et al., 1998; Nussbaum & Steele, 2007). Disidentification from education is likely to have detrimental effects on a pupils’ motivation to perform well in the educational domain and therefore in a higher risk of dropping out before reaching graduation (Schmader et al., 2001; Steele, 1997). The stereotype threat effect of disidentification from education leads us to our third hypothesis: “Pupils belonging to stigmatized groups in education are more likely to personally disidentify from the goals set in the educational domain.”

Since negative stereotypes about ethnic minority pupils and pupils in the vocational track are found to influence teachers’ perceptions of pupils in these stigmatized groups, we expect that these pupils also experience more negative teacher–pupil relations. A final overall hypothesis for this study is that we can find a reinforcing effect on stereotype threat effects for pupils in stigmatized groups in Flemish urban education who experience stigmatization from teachers, measured through the lack of respect and acceptance pupils experience from their teachers. In other words, our final hypothesis states that “the occurrence of responses to protect one’s academic self-concept from negative stereotypes are influenced by the respect and acceptance pupils—and particularly pupils in stigmatized groups—experience from their teachers.” The role of feeling respected and accepted by teachers—measured by a subsample of items from Goodenow’s (1993b) measurement of school belonging—is studied as a proxy for the teacher–pupil relationships within the analyses each of the three hypotheses described above. We can therefore study how teacher–pupil relations influence stereotype threat effects.

Method

To test the hypotheses presented above, we rely on a database that contains data collected in 2010 from 6,244 pupils in the third or fourth year of Flemish secondary education in the cities Antwerp, Genk, and Ghent.³ The initial database comprises 11,015 unique records, representing almost 50% of the total pupil population in the aforementioned year cohorts. A few methodological considerations forced us to limit the scope of the data to 6,244 records. First, we chose not to take into account special education, part-time vocational or arts education due to the particular character of these tracks and therefore relatively small proportions in the sample. Second, we restricted our analyses to pupils without an immigration background and to those with an immigration background that is exclusively linked with Turkey or Morocco.⁴ This restriction is based on two different reflections: (a) There is extensive empirical evidence for the stigmatization of these ethnic groups in Flemish education, which is not available for the wide variation of pupils with other ethnic backgrounds in our initial sample; (b) the hyper-diversity of other ethnic backgrounds in the sample resulted in insufficient group sizes for the purposes of our analyses.

By taking together the variables educational track and ethnic background, we were able to define six groups. Figure 1 shows the relative group sizes of pupils in different educational tracks by ethnic background, which visualizes the socioethnic stratification described above through the strong overrepresentation of the native group in general education and pupils with Moroccan or Turkish backgrounds in vocational education.

Background Variables

A first background variable used in most of the analyses presented in this article is the gender of the pupil. Although there exists a large body of literature concerning gender-induced stereotype threat in education, the pupil’s gender is in this study only used as a control variable to test stereotype threat related to the pupil’s educational track and ethnicity. Gender is entered as a dummy variable wherein male pupils serve as reference category.

SES is also considered a control variable in our analyses. Notwithstanding the strong correlations of SES with both the pupil’s educational track as well as their ethnic background, multicollinearity did not cause problems in our analyses.⁵ The SES variable was constructed by running a factor analysis on a series of variables connected to the pupils SES. The final factor was determined by the labor market situation of both parents, their educational attainment, as well as the availability of a quiet room to study, a computer, and Internet at home.

To determine the pupil’s ethnic background, we made use of the available data on the pupil’s ethnic background, that is, the pupil’s place of birth, as well as the country of birth of their father, mother, and both of the pupil’s grandmothers. As we addressed in the description of our sample, the ethnic background is applied as a dichotomous variable distinguishing pupils with no immigration background and those with nonmixed Turkish or Moroccan immigration backgrounds. For pupils in the latter category, at least both grandmothers are born in Turkey or Morocco. It is important to emphasize that we do not argue that there are no ethnic group related differences within the latter category. However, as argued above, in terms of their position and stereotypes within Flemish education, these ethnic groups find themselves in a similar position.

A final background variable is the pupils’ educational track. We therefore distinguish pupils in general, technical, and vocational education. For some analyses, we chose to analyze relations by splitting the dataset in groups that combine ethnic background and educational track. In other tests, we used dummy variables for vocational and technical education, wherein pupils in general education served as reference category. We will further elaborate on this in the results section.

Academic Performance Indicators

Another main variable in our analyses is the pupil’s academic performance level, measured through their grade point average (GPA) at the end of the second year of secondary education. This GPA value is based on a self-reported scale and categorizes the pupil’s GPA as follows: less than 50%, 50% to 60%, 60% to 70%, 70% to 80%, and more than 80%. Yet, the 5-point scale in the survey data did not limit us to working with a categorical variable. Normality testing encouraged us to use the self-reported GPA as a proxy for a continuous variable in the analyses.

School-Related Beliefs and Attitudes

Because stereotype threat effects literature mainly concerns effects on pupil’s beliefs and attitudes toward education, the measurement of these psychological traits plays a central role in our methodology. We now discuss our measurement teacher–pupil relations through the respect and acceptance experienced from teachers, the academic self-concept, discounting of teacher’s assessments and disidentification from educational goals.

To address teacher–pupil relations or more specifically the respect and acceptance pupils experience from teachers, we made use of a subscale extracted from Goodenow’s (1993b) Sense of Belonging in School measurement. We used factor analysis that showed that the initial latent construct could better be divided in one factor addressing a broader concept “school belonging” and another factor that can be best described as “feeling respected and accepted by teachers.” The latent construct is made up by six items and have a strong internal consistency, reflected in a Cronbach’s alpha value of .75. Some of sample items are “Teachers in this school are not interested in people like me”; “I am treated with as much respect as other pupils”; “Teachers at my school respect me.” For the purpose of our analyses, the measurement scale was constructed so that a higher value means feeling less respected/accepted by teachers.

To measure pupil’s academic self-concept, we made use of Shavelson and Marsh’s (1986) measurement of a general academic self-concept. Because previous research has shown that academic achievement is more correlated with academic self-concept than with general self-esteem (Marsh, 1987; Van Houtte et al., 2012), we argue to test the psychological disengagement hypotheses using this scale rather than a global self-esteem scale (Régner & Loose, 2006). Our measurement of academic self-concept is based on nine items that show an internal consistency of Cronbach’s alpha of .77. Some of the sample items are “I think I’m good at learning”; “My classmates can learn better than me”; “When I make a test, I mostly feel that I can do good.” The academic self-concept measurement scale was constructed so that a higher value shows a higher academic self-concept.

To test whether and how psychological disengagement manifests itself through discounting and/or disidentification, we use proxies available in our survey data. Discounting could only be operationalized by a single item where pupils could attribute “bad results” to unfair scoring by teachers. Although we only had access to a single item, this item captures the idea of discounting very well as ascribing low scores to the biased evaluation by teachers is a very close operationalization of the discounting concept for protecting one’s academic self-concept from stereotype threat (Schmader et al., 2001). Last but not least, disidentification from education was measured through the mirrored goal valuation scale from the revised School Attitude Assessment Survey (McCoach, 2002), containing six items such as “I want to do good in school” and “it’s important that I’m doing good in school.” The strong internal consistency of these items is shown in the Cronbach’s alpha value of .90.

In Table 1, we present mean scores and standard deviations for a lack of respect and acceptance experienced from teachers as a measurement of teacher–pupil relations, academic self-concept, and disidentification of educational goals per subgroup delimited by both educational track and ethnic background. The table also shows absolute frequencies and percentages of pupils per subgroup that discount their teacher’s scoring as being unfairly biased in case of “bad results” in italics.

Table 1.

Descriptive Statistics of School-Related Beliefs and Attitudes Per Subgroup.

Subgroups	n	Label	N f	M (centered) percentage	SD
Native pupils general education	2,888	Lack of respect from teachers	2,695	−0.233	0.841
		Academic self-concept	2,802	0.022	0.880
		Discounting	510	17.95%
		Disidentification	2,820	0.023	0.938
Native pupils vocational education	699	Lack of respect from teachers	628	0.180	0.866
		Academic self-concept	652	0.011	0.902
		Discounting	152	22.39%
		Disidentification	672	0.244	1.027
Native pupils technical education	1,340	Lack of respect from teachers	1,234	0.139	0.844
		Academic self-concept	1,278	−0.146	0.889
		Discounting	238	18.02%
		Disidentification	1,313	0.201	0.961
Minority pupils general education	353	Lack of respect from teachers	320	−0.006	0.940
		Academic self-concept	333	0.091	0.859
		Discounting	74	21.26%
		Disidentification	344	−0.477	0.852
Minority pupils vocational education	611	Lack of respect from teachers	502	0.264	1.053
		Academic self-concept	532	0.225	0.924
		Discounting	171	30.11%
		Disidentification	563	−0.166	1.119
Minority pupils technical education	353	Lack of respect from teachers	306	0.182	0.901
		Academic self-concept	325	0.161	0.866
		Discounting	83	23.92%
		Disidentification	338	−0.298	1.026

Note. For discounting, we presented the number and percentage of pupils that reported to discount their teachers’ assessment when receiving negative feedback in italics.

Results

As described in the literature section above, a precondition for stereotype threat effects to manifest itself is that the members of a stereotyped group in a certain domain are aware of the stereotypes about their group. Table 1 shows that pupils in stigmatized groups—that is, pupils in lower status tracks and pupils with an ethnic minority background—feel the least respected and accepted by teachers.⁶ This finding strengthens previous findings on the fact that pupils in stigmatized groups are conscious about stigmas about their social group in education.

In this section, we test the first three hypotheses concerning stereotype threat effects drawn from theory and previous empirical research. The testing of the final overall hypothesis will be discussed throughout the sections that present the testing of the hypotheses based on stereotype threat literature. Each of the analyses are adapted to the data available in our research project on educational trajectories in Flemish urban education. The methodological choices made in the process are explicated throughout the results section.

Hypothesis 1: Pupils in stigmatized groups are more likely to disengage their academic self-concept from their actual achievements in the educational domain.

Since the existing literature clearly shows that pupils in vocational education and those with Turkish or Moroccan backgrounds can be considered to belong to a stigmatized group in education, one could expect that these negative stereotypes lead stigmatized pupils to show response strategies for protecting their self-concept.

An underlying theoretical assumption of stereotype threat literature is that pupils in stigmatized groups respond to the threat of negative stereotypes in education by disconnecting their academic self-concept from their achievements in education. As we hypothesized that a double stigmatization occurs for those pupils with a minority background in vocational education, we tested the effect of pupils’ GPA on their academic self-concept separately per subgroup, defined by both their educational track and ethnic background.

In table 2, per subgroup, we engaged in a two-step regression analyses of the effect of pupils’ prior achievements on their current academic self-concept, controlling for the effects of gender and SES in a first step and we added the effect of the reversed measurement scale for respect and acceptance experienced from teachers—as a proxy for teacher–pupil relations—on the academic self-concept in the second step. By inserting the teacher–pupil relations measurement in the second step, we can test our overall hypotheses by studying how teacher–pupil relations influence the initial effect of pupils’ GPA on their academic self-concept by comparing the regression correlations and the total of variance explained (adjusted R²) in each model. It is important to acknowledge that no causal relations can be tested since the direction of the relation between pupils’ GPA and academic self-concept is not clear. Nonetheless, the regression analyses allow us to control for the pupils’ background variables and the additional influence of teacher–pupil relations in the correlation between pupils’ GPA and academic self-concept.⁷

Table 2.

Standardized Regression Parameter Estimates (β) and Standard Errors (in Parentheses) for a Two-Step Regression Analyses Predicting Academic Self-Concept per Subgroup.

	Native pupils in general education		Native pupils in technical education		Native pupils in vocational education
	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2
Gender (I = girls)	−0.171* (0.030)	−0.200* (0.029)	−0.078* (0.049)	−0.084* (0.049)	−0.079* (0.071)	−0.093 * (0.070)
SES	0.028 (0.032)	0.013 (0.031)	0.064* (0.049)	0.016 (0.049)	−0.003 (0.068)	−0.009 (0.067)
GPA	0.441* (0.019)	0.377* (0.019)	0.236* (0.029)	0.191* (0.029)	0.205* (0.033)	0.152* (0.033)
Respect	—	−0.296* (0.018)	—	−0.327* (0.029)	—	−0.343* (0.040)
Adjusted R²	.21	.30	.07	.16	.04	.15
n	2,778	2,606	1,260	1,168	643	590
	Minority pupils in general education		Minority pupils in technical education		Minority pupils in vocational education
	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2
Gender (I = girls)	−0.280* (0.090)	−0.275* (0.090)	−0.104 (0.097)	−0.083 (0.099)	−0.054 (0.081)	−0.076 (0.082)
SES	−0.059 (0.070)	−0.099 (0.087)	0.016 (0.093)	0.020 (0.095)	0.067 (0.074)	0.096* (0.074)
GPA	0.276* (0.055)	0.241* (0.057)	0.146* (0.054)	0.126* (0.055)	0.151* (0.038)	0.074 (0.039)
Respect	—	−0.261* (0.048)	—	−0.282* (0.054)	—	−0.299* (0.039)
Adjusted R²	.14	.20	.02	.09	.02	.11
n	328	301	314	279	519	447

Note. SES = socioeconomic status; GPA = grade point average.

p < .05.

We start by reporting the influence of the control variables gender and SES on pupils’ academic self-concept for the different subgroups in both steps of the analyses. Because the control variables gender and SES can be defined as exogenous variables the direction of the effects are clear. The pupil’s SES only shows significant effects on the academic self-concept of native pupils in technical education in Step 1 (β = .06363; p < .05). This effect, however, disappears when also taking into account the effect of respect and acceptance from teachers in Step 2. In Step 2, minority pupils in vocational education show a higher positive effect of their SES on their academic self-concept (β = .09579; p < .05). Gender mainly affects pupils’ academic self-concept for native pupils, for the pupils with a minority background the results only show a (although rather strong) negative effect in general education. For all significant effects of gender, we can report that female pupils report a lower academic self-concept than male pupils, again controlled for the other variables in the first step.

Next, we report on the effect of pupils’ GPA on their academic self-concept in Step 1, that is, without taking into account effects of teacher–pupils relations. When exploring the results from the first step, we can confirm our psychological disengagement hypothesis: Pupils in the least stigmatized group—that is, native pupils in general education—show a stronger effect of their GPA on their academic self-concept than minority pupils in general education (native: β = .44057, p < .05; minority: β = .27641, p < .05). This strong positive effect of pupils’ GPA is significantly lower when considering the parameters for pupils in technical and vocational education. The lowest correlations are found for pupils in the most stigmatized group, that is, pupils with an immigrant background in technical and vocational education.

When we include our measurement of teacher–pupil relationships in Step 2, the results show an overall strong negative effect of not feeling respected and accepted by teachers on pupils’ academic self-concept. The strength of these regression parameters is rather similar for the different subgroups and ranges from β = −.26095 for minority pupils in general education to β = −.34298 for native pupils in vocational education. More importantly, we can report that all initial regression parameters of pupils’ GPA on their academic self-concept decrease and even lose statistical significance for pupils in the most stigmatized groups, that is, pupils with a Turkish or Moroccan origin in vocational education. Another finding that supports our overall hypothesis is the growing explained variance in Step 2 in comparison to Step 1 for all subgroups but more particularly for pupils in vocationally oriented education. The explained variance in pupils’ academic self-concept is very low for the analyses in Step 1 and increases up to five times the initial amount for minority pupils in vocational education when inserting the teacher–pupil relations are inserted in the second step of the analyses.

We can therefore conclude that the strong positive relation between the GPA and the academic self-concept of pupils in the least negatively stereotyped groups disengages for pupils in the most stigmatized groups. Furthermore, teacher–pupil relationships do in fact also play a stronger role in relation to pupils’ academic self-concept for pupils in the most stigmatized groups, that is, pupils in vocationally oriented education, especially those with an immigrant background.

Hypotheses 2: Pupils in stigmatized groups will discount the validity of teachers’ assessments more often in case of negative feedback on their academic performances.

Our next hypothesis concerns an acute response to stereotype threat. In the short run, pupils who suffer stigmatization in education can discount teachers’ negative feedback on their educational performances to protect their academic self-concept. By discounting negative feedback from teachers for being unfairly biased, a pupil—who feels threatened by stereotypes about their social group not performing well in education—can protect his or her academic self-concept. We therefore expect pupils in the most stigmatized groups to discount negative feedback on their academic performances more often.

Figure 2 shows the results of a logistic regression estimating probabilities to discount teachers’ negative feedback on academic performances shows significant effects for gender, SES, ethnic background, pupils’ educational track and respect and acceptance from teachers (χ² = 265.12; p < .001). Female pupils discount their teachers’ scoring in case of “bad results” significantly less than male pupils. The pupils’ SES has an additional positive significant effect on discounting.⁸ Pupils in vocational, but not pupils in technical education, differ significantly from pupils in general education in discounting negative feedback. Our second hypothesis is therefore confirmed as pupils in vocational education discount teachers’ scoring when receiving “bad results” 24% more often than in general education. Furthermore, pupils with Turkish or Moroccan backgrounds have 49% more probability to discount negative feedback than native pupils.

The strongest predictor for discounting negative feedback from teachers for being unfairly biased, however, is experiencing negative teacher–pupil relations, a one unit increase of in the measurement for a lack of respect/acceptance from teachers predicts a 71% increase in the probability to discount negative feedback. This further strengthens our overall hypothesis stating that experiencing low levels of respect and acceptance from teachers—as a proxy for negative teacher–pupil relations—strongly increases the occurrence of stereotype threat effects.

Hypothesis 3: Pupils in stigmatized groups in education are more likely to personally disidentify from the goals set in the educational domain.

Our third hypothesis concerns psychological disengagement by disidentifying from the goals set in education. In figure 3, similar to the stereotype threat effect of discounting, we estimated pupils’ gender, SES, ethnic background, educational track and the amount of respect and acceptance experienced from teachers—as a proxy for teacher–pupil relations—as predictors for disidentification.

Female pupils again reported less stereotype threat effects than males (β = −.07; p < .05). Pupil’s SES does not have a significant predictive value toward disidentificaion.⁹ Pupils in vocational education have significantly more chance to disidentify from education than pupils in general education (β = .05; p < .05), while those in technical education do not differ significantly from pupils in the general track. The predictive value of pupils’ gender and educational track, however, is rather small compared to the pupils’ ethnic background and the respect/acceptance experienced from teachers. Remarkably, pupils with Turkish or Moroccan backgrounds value the domain of education more than native pupils (β = −.2; p < .05). Nonetheless, feeling less respected and accepted by teachers does predict more disidentification from education (β = .32; p < .05). Even though a Turkish or Moroccan background protects pupils from disidentification from education—which does not concur with the stereotype threat effects inspired hypothesis—the predictive value of not experiencing positive teacher–pupil relationships is more pronounced. Keeping in mind that pupils in the most stigmatized groups feel least respected and accepted by teachers (see Table 1), one can expect more disidentification for pupils in vocational education and pupils with Turkish or Moroccan backgrounds in case stigmatized pupils experience more negative teacher–pupil relations.

In accordance with Régner and Loose (2006), we tested the mediational role of stereotype threat on pupils’ academic self-concept that is implicated by the discounting and disidentifcation hypotheses. The hypotheses imply that poor grades threaten to confirm negative stereotypes and could thereby negatively affect the academic self-concept. To protect their academic self-concept, stigmatized pupils will discount the validity of negative feedback or disidentify from educational goals more often. The four steps procedures to test the mediational role of the threat on pupils’ academic self-concept toward predicting discounting and disidentification by pupils are shown below.

The first step consists of showing that lower grades predict discounting. In the second step, the GPA needs to predict academic self-concept. Moreover, academic self-concept needs to predict discounting, while controlling for grades. In addition, the direct effect of grades on discounting needs to be significantly reduced when academic self-concept is included in the analysis. Figure 4 shows that all four conditions for the mediation hypothesis for discounting are met. The initial significant negative relation between GPA and discounting (β = −.07; p < .05) shows that lower grades do heighten the chance to discount teachers’ scoring for being unfairly biased. The pupils’ GPA does overall have a strong relation with pupils’ academic self-concept (β = .31; p < .05) and a higher academic self-concept negatively relates to the probability to discount negative feedback when controlled for pupils’ GPA (β = −.12; p < .05). Last but not least, the initial significant relation between GPA and discounting diminishes in strength and loses its statistical significance. It is important to notice that the analyses for all four conditions also included the pupils’ gender, SES, ethnic background, and educational track as control variables.¹⁰

As we did for the discounting hypothesis testing, we now present a test for the mediational role of the protecting one’s academic self-concept in the relation between pupils’ GPA and disidentification from education. Figure 5 demonstrates that all four conditions for a mediation effect are indeed fulfilled. The negative relation between pupils’ GPA and disidentification (β = −.15; p < .05), that is, before controlling for academic self-concept, confirms the direct effect of GPA. Moreover, after controlling for academic self-concept, the standardized parameter for GPA toward disidentification almost halves (β = −.08; p < .05). Taking into account that the pupils’ GPA does positively relate to academic self-concept (β = .31; p < .05) and academic self-concept tempers the relation between the GPA and disidentification (β = −.22; p < .05), we can conclude that the mediation hypothesis for devaluation is confirmed. Similar to the testing of the mediation effect for discounting, we controlled for the pupils’ gender, SES, ethnic background, and educational track.¹¹

Discussion

As the literature review shows, the systemic and institutional context of Flemish urban education is a relevant context to study pupils’ responses to stereotype threat (Baysu et al., 2011; Van Houtte et al., 2012). Flemish urban secondary schools are highly stratified and segregated, both on the socioeconomic and the ethnic dimension (Danhier, Jacobs, Devleeshouwer, Martin, & Alarcon, 2014; De Meyer & Warlop, 2010). The literature review also showed that clear stigmatization processes are present in this educational context (Stevens & Vermeersch, 2010). These stratification and stigmatization processes influence the way pupils engage in and identify with education. As argued, in particular, the overrepresentation of minority pupils in lower status tracks makes stereotype threat effects related to both ethnic background and educational track particularly relevant for Flemish urban secondary education. Our contribution to urban education research is to study how teacher–pupil relations can be affected by these systemic features and how this affects pupils’ responses toward negative stereotypes related to their educational track and/or ethnic minority status.

More in particular, this article tried to disentangle whether and how pupils who are negatively stereotyped for having low motivation and performance levels try to prevent this stereotype threat from harming their academic self-concept by (a) disconnecting their academic self-concept from their actual performances in education, (b) discounting teachers’ negative feedback for being unfairly biased, and (c) disidentifying from the goals set in education. Our results to a large extent support and illustrate the workings of these stereotype threat effects in education as theorized and empirically supported by previous studies (e.g., Aronson et al., 1999; Steele & Aronson, 1995; Steele et al., 2002). As argued, an important contribution of our study is that we focused on the impact of teacher–pupil relations on these stereotype threat effects. This is relevant as the degree to which teacher–pupil relations are often discussed as one of the most important influences effecting educational outcomes (Goodenow, 1993a; Klem & Connell, 2004; Van Maele & Van Houtte, 2011). Another contribution of this study is to widen the scope of stereotype threat related studies to a stigmatized group identity that relates to a lower status educational track in which pupils are enrolled.

Limitations of the Study

Before discussing our main findings in relation to our specific research questions and hypotheses, we elaborate on two important limitations of our study. A first caveat concerns the causal direction of the relationships we focused on. Because the study was not longitudinal in design, the direction of relationships implicated in testing hypotheses related to stereotype threat effects should be taken with some caution. Our cross-sectional study design only captures the theorized relationships implied by the psychological disengagement, discounting, and disidentification hypotheses at a particular moment in time. Well-designed longitudinal studies can bring more clarity to study long-term stereotype threat effects, while experimental settings can further strengthen explanatory analyses on situational reactions of pupils to specific stereotype threat impulses through the use of manipulations.

A second important limitation concerns the fact that all measurements are self-reported by pupils, which make them vulnerable for a social desirability bias. Nonetheless, self-reports do have the strength to address the perceptions of the pupils and therefore measure what is real to them, not what is understood through indirect assessment of teachers or other professionals. Further qualitative and school ethnographic research can, however, address how the hypothesized protection of one’s academic self-esteem against negative stereotypes looks like in every everyday class practices and can provide a deeper understanding of the meaning both pupils and teachers attach to the social imaginaries about these stigmatized groups of pupils in the stratified and segregated context such as Flemish urban secondary education.

Main Research Findings

Despite these limitations, we do have found some interesting results that complement current insights on the processes of stereotype threat effects, and more directly on the impact of teacher–pupil relations on these processes. We will summarize our main findings before discussing some implications of this study. In doing so, we first discuss the testing of our three singular hypotheses based on stereotype threat theory before engaging with a more in-depth discussion of our overarching hypothesis regarding the influence of negative teacher–pupil relationships on these stereotype threat effects.

With regard to our first hypothesis on pupils reacting to stereotype threat by disconnecting their academic self-concept from their performances in education, we can confirm that minority pupils in all educational tracks relate their academic self-concept less with their GPA than their native counterparts (Aronson et al., 1999; Steele & Aronson, 1995; Steele et al., 2002). Furthermore, our results show that pupils in general education hold the strongest connection between their achievements in education and their academic self-concept, suggesting that stigmatization based on pupils’ educational track also affects psychological disengagement.

In the analyses for testing the discounting hypothesis, the results show that pupils with Turkish or Moroccan backgrounds and pupils in vocational education show the highest occurrence of discounting negative feedback by teachers (Steele et al., 2002; Y. Stephan et al., 2011). With regard to the disidentification hypotheses, our findings show that being in vocational education is a risk factor for disidentification from education. Given our findings concerning Hypotheses 1 and 2, and the assumption that stigmatized pupils in general—be it based on track enrolment and/or ethnicity—we could expect that Turkish and Moroccan pupils disidentify more strongly from education. Yet, it is precisely these pupils that in fact value educational goals more than their native counterparts. This finding matches with previous research that showed that stigmatized ethnic minority pupils show higher levels of identification with education (e.g., African American pupils in comparison with European American pupils in Wang, Willett, & Eccles, 2011). Moreover, other studies found that ethnic minority pupils who initially show a higher valuing of education are most likely to withdraw from school due to being more vulnerable for stereotype threat effects (Osborne & Walker, 2006). This further strengthens the idea that teachers’ perceptions and expectations play an important role in the engagement and identification of pupils, particularly those with a stigmatized group identity in education. To understand this finding better, it is important to take a closer look at the impact of teacher–pupil relations on these processes.

With respect to our overarching hypothesis regarding the influence of negative teacher–pupil relationships on the occurrence of stereotype threat effects, we found that the teacher–pupil relations play an important role in the occurrence of psychological disengagement, discounting of negative feedback and disidentification from education. In testing the psychological disengagement hypothesis by a two-step regression analyses predicting pupils’ academic self-concept, the introduction of the level to which pupils experience respect and acceptance from teachers significantly increased the explained variance. The strengthening explanatory power of the regression models was especially striking for pupils in stigmatized groups, where the GPA and control variables only mounted to a very weak explanation of the variance in pupils’ academic self-concept. Furthermore, the lack of respect and acceptance pupils experience from their teachers—as a proxy for negative teacher–pupil relations—was also the strongest predictor for both discounting negative feedback from teachers and disidentification from education.

This study therefore also further disentangles high levels of academic self-esteem of Turkish and Moroccan pupils in vocational education—while being stigmatized for having low overall academic motivation and achievement levels—by introducing the role of negative teacher–pupil relationships in relation to stereotype threat effects. These findings thus support the hypothesis that stigmatized pupils protect their academic self-concept from confirming stereotypes and adds the role of teacher–pupil relations in the understanding of this process.

Implications of This Study

Although our research mainly focused on the micro-level effects of educational stratification on pupils’ academic self-concept, identification with the educational domain and their relationship with teachers, we embedded these interactions within a broader educational structure that was argued to be particularly harmful for socioeconomically vulnerable ethnic minority pupils. When addressing the findings and proposing solutions to alter these processes, one therefore also needs to focus on the macro level of systemic inequalities present in the educational system, on the institutional level of urban secondary schools and the reproduction of inequality in these meso-level contexts. When only aiming to change barriers on the micro level, one ignores the importance of the structural and institutional levels. For example, pupil–teacher relations are crucial for successful educational trajectories; however, these relations are influenced by processes and inequalities surpassing the context of a specific school or classroom. With this study, we tried to take into consideration the effects of stereotype threat and uncover underlying mechanisms explaining pupils’ disidentification from education. We therefore believe these insights can also inspire the development of strategies to tackle issues such as academic disengagement and school dropout.

On the structural as well as on the institutional level, we discussed the negative impact of the current educational structures in Flanders reproducing rather than transforming existing socioethnic inequalities. In particular, the early selection of pupils into a hierarchical tracking structure with little room for “upward” social mobility is detrimental to the educational opportunities of the most vulnerable pupils (e.g., pupils with a Turkish or Moroccan background). It is hopeful that since 2009, a public and political discussion was triggered to restructure Flemish secondary education. These ideas were, however, met with a lot of criticism and up until now little change has occurred (Clycq, Nouwen, Van Caudenberg & Timmerman, 2015). Nonetheless, even though some aspects of the restructuring plan for Flemish secondary education has met political opposition (e.g., postponing early tracking and releasing the structural barriers for upward mobility between tracks), other structural reforms are being designed and will be implemented in the near future. Of particular importance is, on the one hand, the reappraisal of the formal teacher training courses that requires a broader and more intensive training of future teachers, and on the other, policy intentions for a stronger inclusion of teachers with an ethnic minority background. This is a hopeful transformation as research has repeatedly shown how important teacher–pupil relations are to counter stigmatization and to heighten vulnerable groups of pupils’ sense of belonging in schools and identification with education (Elffers, Oort, & Karsten, 2012; Wang & Eccles, 2012).

What our study shows is that the teacher–pupil relationships, and in particular, the respect and acceptance pupils experience from their teachers is a crucial aspect for their identification with education. As a consequence different researchers designed concrete actions and point, for example, to a better matching of teacher profiles, curriculum, and didactics (Inzlicht, Aronson, & Mendoza-Denton, 2009). All these strategies in some way aim at creating the school as an “identity safe” school that reduces stereotype threat. The recruiting of teachers with a minority background has, for instance, has proven to reduce the effects of negative stereotypes by providing positive role models and deemphasizing threatened social identities (Marx & Goff, 2005; Shih, Bonam, Sanchez, & Peck, 2007).¹² This is of particular interest for the Flemish educational system where only very few minority teachers are present. Research also showed that cooperative learning methods between pupils with and without stigmatized identities also reduces negative stereotype threat effects by encouraging the creation of stereotype-inconsistent information due to pupil’s interdependence (W. G. Stephan & Stephan, 2004). Teacher training should therefore also focus more on teaching in diverse classrooms and in being sensitive to issues related to stereotype threat (Baysu et al., 2011).

These findings illustrate that stereotype threat effects are certainly malleable. Even small everyday didactical adjustments can generate positive outcomes and teacher–pupil relations often play a key role in these processes. Supportive teacher–pupil relations in current diversifying urban educational contexts—that are simultaneously characterized by vast socioethnic inequalities—can provide necessary emotional support, trust, and respect in these schools. Therefore, teaching as merely traditional knowledge transfer has to be complemented with competencies to deal with stigmatization and identification processes in the classroom. Nonetheless, as was shown specific structural and institutional features of educational systems can facilitate or hamper stratification and stigmatization processes. Therefore, also structural adjustments in the Flemish educational system are crucial to tackle the strong social and ethnic stratification between educational tracks, which is fundamental for countering the double stigmatization of ethnic minority pupils in lower status tracks.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Agency for Innovation by Science and Technology Flanders (SBO080051).

Notes

Author Biographies

Ward Nouwen is a sociologist of education at the University of Antwerp (Belgium). He is currently working as a doctoral researcher in the EU funded research ‘Reducing Early School Leaving in Europe’. His main research interests are social and ethnic inequalities in education, educational tracking, and school segregation.

Noel Clycq is professor of Chair in European values: discourses and propects; and postdoctoral research coordinator in the EU funded research ‘Reducing Early School Leaving in Europe’, at the University of Antwerp, Belgium.

References

Agirdag

Van Houtte

Van Avermaet

(2012). Why does the ethnic and socioeconomic composition of schools influence math achievement? The role of sense of futility and futility culture. European Sociological Review, 28, 366-378. doi:10.1093/esr/jcq070

Anyon

(1980). Social class and the hidden curriculum of work. Journal of Education, 162(1), 7-92.

Anyon

(2006). Social class, school knowledge, and the hidden curriculum revisited. In Weis

Dimidiates

(Eds.), Ideology, curriculum, and the new sociology of knowledge (pp. 37-46). New York: Routledge.

Aronson

Lustina

M. J.

Good

Keough

Steele

C. M.

Brown

(1999). When white men can’t do math: Necessary and sufficient factors in stereotype threat. Journal of Experimental Social Psychology, 35, 29-46. doi:10.1006/jesp.1998.1371

Clycq

Nouwen

Vandenbroucke

(2014). Meritocracy, deficit thinking and the invisibility of the system: Discourses on educational success and failure. British Educational Research Journal, 40, 796–819. doi:10.1002/berj.3109

Clycq

Nouwen

Ward

Rut

Van Caudenberg

Chris

Timmerman

(2015). Education in Flanders: balancing social and economic rationales while tackling early school leaving. Educação, Sociedade & Culturas, 45, 13-32.

Baysu

Phalet

Brown

(2011). Dual identity as a two-edged sword: Identity threat and minority school performance. Social Psychology Quarterly, 74, 121-143. doi:10.1177/0190272511407619

Bernstein

(1971). Class, codes, and control: Theoretical studies towards a sociology of language. London, England: Routledge.

Boone

Van Houtte

(2012). Social inequalities in educational choice at the transition from primary to secondary education. Further evidence for a rational choice? Culture and Education, 91, 189-214.

10.

Bourdieu

Passeron

J. C.

(1977). Reproduction in education, society and culture. London, England: Sage.

11.

Chun

E. W.

(1988). Sorting Black pupils for success and failure: The inequality of ability grouping and tracking. Urban League Review, 11, 93-106.

12.

Crul

M. R. J.

Schneider

Lelie

(Eds.). (2012). The European second generation compared: Does the integration context matter? (IMISCOE Research). Amsterdam, The Netherlands: Amsterdam University Press.

13.

Danhier

Jacobs

Devleeshouwer

Martin

Alarcon

(2014). Naar kwaliteitsscholen voor iedereen? Analyse van de resultaten van het PISA 2012 onderzoek in Vlaanderen en in de Federatie Wallonië-Brussel [To quality schools for everybody? Analysis of the results of the PISA 2012 research in Flanders and in the Federation Walloons-Brussels]. Brussels, Belgium: Koning Boudewijnstichting.

14.

De Meyer

Warlop

(2010). Leesvaardigheid van 15–jarigen in Vlaanderen. De eerste resultaten van PISA 2009 [Literacy of 15-year olds in Flanders. The first results of the PISA 2009]. Ghent, Belgium: Vakgroep Onderwijskunde.

15.

Elffers

Oort

F. J.

Karsten

(2012). Making the connection: The role of social and academic school experiences in students’ emotional engagement with school in post-secondary vocational education. Learning and Individual Differences, 22, 242-250.

16.

Finley

M. K.

(1984). Teachers and tracking in a comprehensive high school. Sociology of Education, 57, 233-243.

17.

Gibson

Carrasco

Pàmies

Ponferrada

Rios

(2013). Different systems, similar results: Youth of immigrant origin at school in California and Catalonia. In Alba

Holdaway

(Eds.), The children of immigrants at school: A comparative look at integration in the United States and Western Europe (pp. 84-119). New York: New York University Press.

18.

Giroux

H. A.

(1980). Beyond the correspondence theory: Notes on the dynamics of educational reproduction and transformation. Curriculum Inquiry, 10, 225-247.

19.

Goodenow

(1993a). Classroom belonging among early adolescent pupils: Relationships to motivation and achievement. Journal of Early Adolescence, 13, 21-43.

20.

Goodenow

(1993b). The psychological sense of school membership among adolescents: Scale development and educational correlates. Psychology in the Schools, 30, 79-90. doi:10.1002/1520-6807(199301)30:1

21.

Inzlicht

Aronson

Mendoza-Denton

(2009). On being the target of prejudice: Educational implications. In Butera

Levine

(Eds.), Coping with minority status: Responses to exclusion and inclusion (pp. 13-36). Cambridge, UK: Cambridge University Press.

22.

Kelly

(2009). Social identity theories and educational engagement. British Journal of Sociology of Education, 30, 449-462. doi:10.1080/01425690902954620

23.

Klem

A. M.

Connell

J. P.

(2004). Relationships matter: Linking teacher support to student engagement and achievement. Journal of School Health, 74, 262–273. doi:10.1111/j.1746-1561.2004.tb08283.x

24.

Major

Spencer

S. J.

Schmader

Wolfe

C. T.

Crocker

(1998). Coping with negative stereotypes about intellectual performance: The role of psychological disengagement. Personality and Social Psychology Bulletin, 24, 34-50. doi:10.1177/0146167298241003

25.

Marsh

H. W.

(1987). The big-fish-little-pond effect on academic self-concept. Journal of Educational Psychology, 79, 280-295. doi:10.1037//0022-0663.79.3.280

26.

Marx

D. M.

Goff

P. A.

(2005). Clearing the air: The effect of experimenter race on targets’ test performance and subjective experience. British Journal of Social Psychology, 44, 645-657. doi:10.1348/014466604X17948

27.

McCoach

D. B.

(2002). A validity study of the School Attitude Assessment Survey (SAAS). Measurement and Evaluation in Counseling and Development, 35, 66-77. doi:10.1177/0013164403063003005

28.

Nussbaum

A. D.

Steele

C. M.

(2007). Situational disengagement and persistence in the face of adversity. Journal of Experimental Social Psychology, 43, 127-134. doi:10.1016/j.jesp.2005.12.007

29.

Osborne

J. W.

Walker

(2006). Stereotype threat, identification with academics, and withdrawal from school. Educational Psychology, 26, 563-577. doi:10.1080/01443410500342518

30.

Phalet

Deboosere

Bastiaenssen

(2007). Old and new inequalities in educational attainment: Ethnic minorities in the Belgian census 1991-2001. Ethnicities, 7, 390-415. doi:10.1177/1468796807080235

31.

Régner

Loose

(2006). Relationship of sociocultural factors and academic self-esteem to school grades and school disengagement in North African French adolescents. British Journal of Social Psychology, 45, 777-797. doi:10.1348/014466605X83610

32.

Roosens

(1995). Rethinking culture, “multicultural society” and the school. Oxford, UK: Pergamon.

33.

Schmader

Major

Gramzow

R. H.

(2001). Coping with ethnic stereotypes in the academic domain. Journal of Social Issues, 57, 93-111. doi:10.1111/0022-4537.00203

34.

Shavelson

R. J.

Marsh

H. W.

(1986). On the structure of self-concept. In Schwarzer

(Ed.), Anxiety and cognitions (pp. 305-330). Hillsdale, NJ: Lawrence Erlbaum.

35.

Shih

Bonam

Sanchez

Peck

(2007). The social construction of race: Biracial identity and vulnerability to stereotypes. Cultural Diversity & Ethnic Minority Psychology, 13, 125-133. doi:10.1037/1099-9809.13.2.125

36.

Spruyt

Laurijssen

Van Dorsselaer

(2009). Kiezen en verliezen. Een analyse van de keuze na het krijgen van een B-attest in het Vlaams secundair onderwijs als een replicatie van Kloosterman en De Graaf [Choosing and losing. An analysis of the choice after receiving a B-certificate in Flemish secondary education as a replication of Kloosterman and De Graaf]. Mens & Maatschappij, 84, 279-299.

37.

Steele

C. M.

(1997). A threat in the air: How stereotypes shape intellectual identity and performance. American Psychologist, 52, 613-629. doi:10.1037/0003-066X.52.6.613

38.

Steele

C. M.

Aronson

(1995). Stereotype threat and the intellectual test performance of African Americans. Journal of Personality and Social Psychology 69, 797-811. doi:10.1037/0022-3514.69.5.797

39.

Steele

C. M.

Spencer

S. J.

Aronson

(2002). Contending with group image: The psychology of stereotype and social identity threat. In Zanna

M. P.

(Ed.), Advances in experimental social psychology (pp. 379-440). San Diego, CA: Academic Press.

40.

Stephan

W. G.

Stephan

C. W.

(2004). Intergroup relations in multicultural education programs. In Banks

J. A.

Banks

C. A. M.

(Eds.), Handbook of research on multicultural education (2nd ed., pp. 782-798). San Francisco, CA: John Wiley.

41.

Stephan

Caudroit

Boiché

Sarrazin

(2011). Predictors of situational disengagement in the academic setting. British Journal of Educational Psychology, 81, 441-455. doi:10.1348/000709910X522285

42.

Stevens

P. A. J.

Vermeersch

(2010). Streaming in Flemish secondary schools: Exploring teachers’ perceptions of and adaptations to pupils in different streams. Oxford Review of Education, 36, 267-284. doi:10.1080/03054981003629862

43.

Trautwein

Lüdtke

Marsh

H. W.

Köller

Baumert

(2006). Tracking, grading, and pupil motivation: Using group composition and status to predict self-concept and interest in ninth-grade mathematics. Journal of Educational Psychology, 98, 788-806. doi:10.1037/0022-0663.98.4.788

44.

Valencia

R. R.

(2010). Dismantling contemporary deficit thinking: Educational thought and practice. New York, NY: Routledge.

45.

Valencia

R. R.

Black

(2002). “Mexican Americans don’t value education!” On the basis of myth, mythmaking and debunking. Journal of Latinos and Education, 1, 81-103. doi:10.1207/S1532771XJLE0102_2

46.

Valenzuela

(1999). Subtractive schooling: U.S. Mexican youth and the politics of caring. Albany: State University of New York Press.

47.

Van Houtte

(2004). Tracking effects on school achievement: A quantitative explanation in terms of the academic culture of school staff. American Journal of Education, 110, 354-388. doi:10.1086/422790

48.

Van Houtte

(2006). School type and academic culture: Quantitative evidence for the differentiation–polarisation theory. Journal of Curriculum Studies, 38, 273-292. doi:10.1080/00220270500363661

49.

Van Houtte

Demanet

Stevens

P. A. J.

(2012). Self-esteem of academic and vocational pupils: Does within-school tracking sharpen the difference? Acta Sociologica, 55, 73-89. doi:10.1177/0001699311431595

50.

Van Houtte

Stevens

P. A. J.

(2009). Study involvement of academic and vocational pupils: Does between-school tracking sharpen the difference? American Educational Research Journal, 46, 943-973. doi:10.3102/0002831209348789

51.

Van Maele

Van Houtte

(2011). The quality of school life: Teacher-student trust relationships and the organizational school context. Social Indicators Research, 100, 85-100.

52.

Wang

M.-T.

Eccles

J. S.

(2012). Social support matters: Longitudinal effects of social support on three dimensions of school engagement from middle to high school. Child Development, 83, 877-895. doi:10.1111/j.1467-8624.2012.01745

53.

Wang

M.-T.

Willett

J. B.

Eccles

J. S.

(2011). The assessment of school engagement: Examining dimensionality and measurement invariance across gender and race/ethnicity. Journal of School Psychology, 49, 465-480.