The Effects of Neighborhood and Parental Educational Expectations on Teacher Judgments of Children in the Fifth and Sixth Grades of Primary School

Abstract

Research has studied less the embeddedness of the relationship between parental educational expectations and teacher judgments in the neighborhood. We use data that were collected in a longitudinal study conducted in Belgium, which focused on children in the fifth and sixth grades of primary school. We found that the effect of neighborhood on teacher judgments is mediated by the nonnative background of students. The effect of neighborhood seems to be mainly one of selecting pupils in the school. Teachers have—under control of deductive ability and previous teacher judgments—a lower judgment of students with an immigrant background.

Keywords

neighborhood academic achievement urban education teachers

Introduction

Parental involvement is a key factor in scholastic success: Previous research shows a positive association between educational outcomes and parental involvement (e.g., Hattie, 2009; Hornby & Lafaele, 2011). A strong predictor of educational outcomes is parental educational expectations or what parents expect that their children will achieve in their school career (Jeynes, 2005, 2007). These expectations are a specific form of parental involvement because they imply a “dedication of resources by the parent to the child within a given domain” (Grolnick et al., 1997, p. 538; Grolnick & Slowiaczek, 1994, p. 238). Parents exhibiting high educational expectations transmit the value of education and study to their children through their interactions (Kim et al., 2013).

Parental educational expectations do not occur in a social vacuum. They are situated in a process of socialization in which children learn different ideas, skills, and develop an identity during different stages of their life. This socialization process is initiated during early childhood in the family environment (i.e., primary socialization) and continues in the school environment (i.e., secondary socialization; Corsaro, 2010).

Research that has focused on the socialization in school has emphasized the role of teachers. In the school environment, teachers form attitudinal, ability, and behavioral judgments about their pupils (Bourdieu, 1998) that in some cases may influence later academic achievement (Hattie, 2009; McKown & Weinstein, 2002; see also Agirdag et al., 2012). When children start school, these teacher judgments begin to coexist and possibly interrelate with parental educational expectations (see also Bourdieu, 1984, 1990; Durkheim, 1925/1961; Epstein, 1987; Feinstein et al., 2004). Parental educational expectations can affect educational success through the teachers’ judgments of their pupils. When teachers perceive that children are exposed to high parental educational expectations, the likelihood increases that they perceive a positive congruence between school goals (classroom efforts) and home values (Yamamoto & Holloway, 2010). They might consequently also raise their own judgments of these pupils with perceived higher parental involvement (see also Lareau, 1989).

Research on this process, however, has studied less the embeddedness of the relationship between parental educational expectations and teacher judgments in an additional layer of social life: the neighborhood. This can be a relatively small or large place in which people are situated. In this contribution, we argue that a study on this mechanism needs to be contextualized with an understanding that children are also embedded in the neighborhood (Boardman & Saint Onge, 2005; Bronfenbrenner, 1979, 1986; Brooks-Gunn et al., 1993; Sadler & Lafreniere, 2017; Stewart et al., 2007). Schools, after all, tend to reflect the neighborhood in which they are situated. For example, in Belgium, there is a parental freedom of school choice and they tend to select primary schools in their neighborhood (Creten et al., 2000). Schools, thus, tend to enroll pupils that are living in their vicinity and/or immediate surroundings. As a result, the socioeconomic and ethnic composition of the school and also the school processes (including teacher judgments of pupils) are influenced by its geographical location (Kauppinen, 2008; Owens, 2010; Sykes & Musterd, 2011).

Besides the school and family environment, pupils also spend time in the neighborhood in which they are currently living. The characteristics of that neighborhood as a context of social interaction surround the family and also contextualize the internal events of the pupils’ schools (Brännström, 2008). We therefore argue that pupils’ differential geographical allocation to schools is an important element to consider in the study of parental educational expectations and teacher judgments. This article aims to investigate this embeddedness in the neighborhood. More specifically, the article is designed to contribute to the study of how the embeddedness of specific social relationships (i.e., parental expectations and teacher judgments) in the neighborhood is beneficial or detrimental for educational processes in urban schools (Milner & Lomotey, 2014). We focus on Belgian schools in Antwerp and Ghent. These cities were selected with the aim of studying the educational choices of students with an immigrant background. Moreover, the secondary school dropout rate is high in both these cities. In short, they face similar school issues to other urban areas elsewhere in the world (see also Milner, 2012). The current contribution also has a policy relevance. In certain school systems, financial benefits are given because children live in more deprived neighborhoods. It is therefore interesting for policy to scrutinize the role of neighborhood on educational outcomes.

Theoretical Framework

Parental Educational Expectations and Teacher Judgments

Since the early studies of Sewell and colleagues on status attainment and educational plans (Sewell et al., 1969; Sewell et al., 1970), a body of scholarly work has emerged on educational expectations, or what students expect to achieve in education and what parents expect that their children will achieve (see, for example, Alexander et al., 1994; Beutel & Anderson, 2008; Buchmann & Dalton, 2002; Kim et al., 2013).

In contrast with similar concepts, such as educational aspirations, educational expectations appear as a more realistic/concrete assessment of pupils’ future educational attainment (Boxer et al., 2011; Hanson, 1994; Mickelson, 1990). Researchers also claim that educational expectations have a stronger positive relation with later educational success than educational aspirations (Feliciano, 2006; Jerrim, 2014; Khattab, 2014, 2015). In what follows, we will therefore focus on educational expectations and more specifically on parental educational expectations. As a form of parental involvement, parental educational expectations are particularly strong predictors of later educational success. Previous research has found support for the notion that high parental educational expectations are influential and an important element of children’s success at school (Jeynes, 2005, 2007).

One mechanism that explains how parental educational expectations might affect educational outcomes is through the teachers’ judgments of their pupils. When teachers perceive that children are exposed to high parental educational expectations, they might believe that “their efforts in the classroom are being reinforced at home” (Yamamoto & Holloway, 2010, p. 205). They might consequently also raise their own judgments of these pupils with higher perceived parental involvement (see also Lareau, 1989). Conversely, when teachers have the perception of a low parental involvement, they would tend to hold more negative judgments of those students (e.g., for Belgium: Hermans, 1995). We hypothesize the following:

Hypothesis 1: Higher (lower) parental educational expectations lead to more positive (negative) teacher judgments.

Embeddedness in the Neighborhood

Previous research has neglected the embeddedness of the relationship of parental educational expectations and teacher judgments in the neighborhood. We argue that the characteristics of individuals, families, and schools should be contextualized by studying all these environments, including the neighborhood in which they are situated (Brännström, 2008; Carlson & Cowen, 2014; Goldsmith, 2009; Sanbonmatsu et al., 2006). This is because besides spending time in the family and the school, children also spend time in the neighborhood in which they are currently living. Past research in the United States and Europe has found that the composition of the neighborhood or the spatial segregation of resources can have an additional positive or negative effect on different outcomes, including educational success of youth (e.g., Andersson & Subramanian, 2006; Brattbakk & Wessel, 2013; Ceballo et al., 2004; Dupere et al., 2010; Garner & Raudenbush, 1991; Leventhal & Brooks-Gunn, 2004; Murry et al., 2011; Sharkey & Elwert, 2011; Wilson, 1987; Wodtke et al., 2011).

There are three main responses to the question why there should be a direct effect of neighborhood composition on parental educational expectations: the contagion, collective socialization, and institutional explanations (Jencks & Mayer, 1990; Kauppinen, 2006; see also Ainsworth, 2010). The contagion explanation describes the effect of neighborhood composition with the use of the metaphor of an epidemic. Neighborhoods are characterized by different levels of economic and social resources and people residing in them can have heterogeneous or homogeneous socioeconomic and ethnic backgrounds. Neighborhoods can therefore show segregation or not of resources and backgrounds with a specific, underlying spatial logic (Roscigno et al., 2006). When parents and students live in a neighborhood in which many people with possible problematic behaviors and attitudes are residing (Kauppinen, 2006), parents and students are exposed to them. These behaviors and attitudes might include grade retention, low attitudes toward schooling, and low parental educational expectations, and are associated with lower socioeconomic backgrounds (on the relation between low socioeconomic background and grade retention, see González-Betancour & López-Puig, 2016; Hauser et al., 2007; parental attitudes toward schooling, see Gorman, 1998; and parental educational expectations, see Crosnoe et al., 2002; Davis-Kean, 2005).

The collective socialization explanation claims that other adults in the neighborhood can act as role models and/or norm enforcers (Bucher, 1998; Pong & Hao, 2007). If many adults in the neighborhood have, for example, a high socioeconomic status, they can act as role models for students and/or parents. The presence of affluent people in a neighborhood might prove to them that (educational) success is possible if they follow their example (Jencks & Mayer, 1990). The reverse is also true if there are a lot of poor people in the neighborhood (Ainsworth, 2002), insofar as the pupils might follow those examples due to lower or a lack of exposure to alternative role models.

The institutional explanation hypothesizes that internal processes in neighborhood institutions (such as schools) are affected by the neighborhood composition. They receive, for example, less (or more) financial means to organize their activities based on the environment in which they are embedded (i.e., a relatively poor or rich neighborhood). This may affect parents’ willingness to invest in education or their educational expectations (see Roscigno et al., 2006). Based on these explanations, we hypothesize the following:

Hypothesis 2: There is a direct effect of the neighborhood on parental educational expectations: A higher (lower) socioeconomic status of the neighborhood increases (decreases) parental educational expectations.

Mediation of Neighborhood Effects

The three possible explanations also allow us to hypothesize that the effects of neighborhood composition are operating indirectly or—in other words—are mediated (Kauppinen, 2008; Sykes & Musterd, 2011).

Schools tend to be closely connected to the neighborhood in which they are providing and organizing education for children (Sykes & Kuyper, 2009). More specifically, schools tend to enlist and teach children who are living in the (immediate) vicinity of the school. For example, in Belgium, there is a parental freedom of school choice and they tend to select primary schools in their neighborhood. The proportion of people with an immigrant or low socioeconomic background living in a particular neighborhood is reflected in the background profiles of the pupils in the schools located in those neighborhoods (i.e., in Belgium, the so-called “concentration schools” or segregated schools that are located predominantly in immigrant and/or less affluent neighborhoods, see De Rycke & Swyngedouw, 1999). This results in a close relationship between the neighborhood and the school (Brännström, 2008; Roscigno & Crowley, 2001). It is through this embeddedness of the school in their geographical surroundings that internal school processes are influenced by the neighborhood. The influence of the neighborhood on internal school processes (including, for example, on the teacher judgments of pupils) occurs by influencing the school (see also Martinez, 2018). Among other variables such as the cognitive ability of pupils and academic achievement, this has an effect on the opinions that teachers hold about their students. These ideas might also influence the teaching practices of the school personnel (see Agirdag et al., 2013; Owens, 2010; Van Maele & Van Houtte, 2009). The opinions and teaching practices of school personnel matter because these personnel are significant others and influential (Martinez, 2018). We hypothesize the following:

Hypothesis 3: The effect of neighborhood on teacher judgments is mediated by the socioeconomic/ethnic characteristics of the students. We expect that a neighborhood with a lower (higher) socioeconomic status has a negative (positive) indirect effect on teacher judgments.

Furthermore, we can also expect that the effect of neighborhood is mediated by the level of parental educational expectations. We already hypothesized that parental educational expectations are affected by the neighborhood and that parental educational expectations affect teacher judgments.

Hypothesis 4: The effect of neighborhood on teacher judgments is mediated by parental educational expectations. We expect that a neighborhood with a lower (higher) socioeconomic status has a negative (positive) indirect effect on teacher judgments.

Method

Introduction

In this article, we use data that were collected in the “Transbaso” study (Boone, 2016; Thys & Van Houtte, 2016). This project focuses on the determinants of and differences in the transition of pupils in primary to secondary education in the Flemish cities of Antwerp and Ghent (Belgium). These cities were selected with the aim of studying the educational choices of students with an immigrant background. Moreover, there is a substantial dropout at the secondary school level in both contexts. In this article, we focus on children in the fifth and sixth grades of primary school in Flanders. For these students, we have measurements of their parental educational expectations and teacher judgments at the end of the 2013–2014 school year (Wave 1, fifth grade) and at the end of the 2014–2015 school year (Wave 2, sixth grade). For an overview of the descriptive statistics of this sample, see Table 1.

Table 1.

Descriptive Statistics of the Variables: Frequency, Range, Percentages, Mean, Standard Deviations, and % Imputed.

Variables	Respondents	Range	M or % (SD)	% imputed
Level 1: Student
1. Nonnative background	757	0–1	39.70%	0
2. Male student	757	0–1	50%	1.45
3. Low parental educational background	757	0–1	41.23%	4.62
4. Teacher judgment (Wave 1)	757	−3.23 to 1.89	0 (1)	1.85
5. Teacher judgment (Wave 2)	757	−3.26 to 1.72	0 (1)	7.93
6. High parental educational expectations	757	0–1	68.20%	13.34
7. Learning disability	757	0–1	18.65%	0.40
8. Deductive ability (Raven score)	757	13–58	42.12 (6.51)	1.72
9. Individual neighborhood score	757	−1.95 to 3.11	0 (1)	0
10. Single-parent family	757	0–1	20.15%	1.45
11. Number of children in the family	757	1–10	2.69 (1.22)	0.40
Level 2: School
12. Nonnative composition (%)	35	0–100	41.50 (29.14)	0
13. Low parental education composition (%)	35	0–88.89	42.19 (25.39)	0
14. Antwerp (school location)	35	0–1	48.57%	0

In Belgium, education is compulsory from the age of 6 until the age of 18 (Baysu & de Valk, 2012; De Groof & Franck, 2013). Primary education normally lasts 6 years with a transition to secondary education at the theoretical age of 12. Enrollment is driven by the principle of parental freedom of school choice.

In the Transbaso project, 36 primary schools were selected with a stratified random sample based on three selection criteria. The selection criteria were city (Ghent or Antwerp), school sector (state or private schools), and percentage of students with a low socioeconomic background (low, medium, or high). To select those schools, the project used the data from the Flemish department of education. The project selected 36 schools (18 in Antwerp and 18 in Ghent). Within each city, nine state and nine private schools were selected. Within each school sector (state or private school) in each city, there are three schools with a low, medium, and high percentage of students with a low socioeconomic background. In Figure 1, we situate our sample of 36 schools in the population of Flemish primary schools in the province of Antwerp and East Flanders. The red squares are the selected schools. The black circles are nonselected schools.

Figure 1.

Scatterplot of the population and sample of Flemish primary schools in the province of East Flanders and Antwerp.

The resulting samples of Ghent and Antwerp are representative of the primary school system in both cities with regard to two composition variables: the percentage of children with a mother without a higher secondary degree and the percentage of children of which no or only one family member in a family consisting of three people (excluding the child) speaks the language of instruction (Dutch) at home (population data provided by the Flemish department of education).¹

We found with goodness-of-fit chi-square tests that the distributions of the composition variables (education and language) in the population of nonselected schools in Ghent and Antwerp are equal to the distributions of the selected schools in Ghent and Antwerp. We first categorized the nonselected schools into four groups of approximately 25%. We then imposed the same categorization on the selected group of schools. We then compared the distribution of the selected and nonselected schools in the categories of, respectively, the education and language composition variables (including χ² tests)—for Ghent: χ² = 4.15, df = 3, p = .24; χ² = 3.74, df = 3, p = .29; for Antwerp: χ² = 1.92, df = 3, p =.59; χ² = 1.77, df = 3, p = .62.

The focus of this article lies on the explanation of outcomes of children who were enlisted in the school in Wave 2 that were also enlisted in Wave 1 based on the registration of the school personnel and with data on a parental questionnaire. We also linked students to the available neighborhood data and have a sample of 757 students in 35 schools.² We used multiple imputation to deal with missing values. The advantage of multiple imputation is that the available information of cases in both waves can be used to impute the missing data: No information is neglected. It produces several data sets with the use of Bayesian statistics and analyses with these data are pooled to produce estimates (Enders, 2010). We use 10 imputed data sets. In the imputation phase, we use variables that will be overviewed in the following section (gender, educational level of parents, nonnative background, teacher judgments [Waves 1 and 2], parental educational expectations, learning disability, family structure, number of children in the family, and deductive ability).

Variables

Outcome variables

We use two major outcome variables: teacher judgments and parental educational expectations (see Table 1 for descriptive statistics). In a parental questionnaire administered in Wave 1, parents were asked which degree they expect their children will achieve in the future. We distinguish the highest expectations (i.e., higher education) versus the lowest (i.e., no higher education) and no expectations (these are not missing responses). To clearly contrast people with the highest and relatively lower expectations, we dichotomized this variable.

In teacher questionnaires during Waves 1 and 2, teachers of the fifth- and sixth-grade class were asked to indicate the language, mathematics, and technical skills of each student in their class. These three questions were scored on a scale of 1 (very weak) to 5 (very strong). The three questions of each wave were transformed in regression factor scores with the use of principal components analysis (with varimax rotations). One principal component explained around 65% of the variance of the Wave 2 teacher judgments. We extracted this component. The other components had eigenvalues below 1. The resulting scale has a mean of 0 and a standard deviation of 1 and indicates very low to very high teacher judgments. The Cronbach’s alpha is .74 (listwise deleted data = 697 students).

Similarly, one principal component explained around 61% of the variance of the Wave 1 teacher judgments. We also extracted this component. The other components had eigenvalues below 1. The resulting scale has a mean of 0 and a standard deviation of 1 and indicates very low to very high teacher judgments. The Cronbach’s alpha is .69 (listwise deleted data = 743 students).

Student-level variables

At the student level, we include background information on gender, highest level of parental education, learning disability, number of children in the family, deductive ability, family structure, and nonnative background. Gender was coded as female (0) and male (1). The information on the educational levels of parents was collected in a parental questionnaire during Wave 1. The highest level of parental education is coded from higher secondary education or lower (1) and above higher secondary education (0). In this way, we can clearly distinguish between higher and lower educated parents.

We have also measurements on the number of children in the family (standardized at the student level with a mean of 0 and a standard deviation of 1; parental questionnaire, Wave 1), family structure (single-parent family or not; student questionnaire, Wave 1), the presence (1) or absence (0) of a learning disability of the child (dyslexia, dyscalculia, attention deficit hyperactivity disorder [ADHD], autism or other mental disorders; parental questionnaire, Wave 1), and nonnative background (“native” defined here as Belgium or Northern, Western Europe [0] and “nonnative” as Eastern Europe, Maghreb, Turkey, Asia, sub-Saharan Africa, Latin America, the Middle East or Southern Europe [1], parental questionnaire, Wave 1). The last question was based on the country of birth of the grandmother on the mother side.

We also include scores on the Standard Progressive Matrices of Raven that measures the deductive skills of the children (Raven, Raven, & Court, 2003).³ This test was administered in Wave 1. We standardized this variable at the student level and have a variable with a mean of 0 and a standard deviation of 1.

We obtained census data from 2011 on a low geographical level in Belgium: the so-called statistical sector. A sector is a statistical district roughly corresponding with a neighborhood consisting of a small number of streets within municipalities.⁴ Each statistical sector gives aggregated geographical information on socioeconomic and demographic indicators. We linked each pupil in our sample to the statistical sector in which they were residing at the moment of Wave 1. We were then able to link them to these different socioeconomic and demographic realities.

We obtained the following neighborhood variables at the statistical sector level in the Belgian census of 2011: the proportion of single-parent families (single father or single mother), the proportion of people living in a house that they own themselves, people with a low educational level (lower secondary school or lower), unemployed people, and people with a non-Belgian nationality (including those without a state and an unknown nationality). With a principal components analysis (with varimax rotations), we extracted one regression factor score that explains around 66% of the variance of the neighborhood variables. The resulting indicator has a mean of 0 and a standard deviation of 1. The Cronbach’s alpha of these variables is .70. This indicator provides a generalized insight on the affluency versus the social deprivation of a neighborhood (a higher score indicates more social deprivation). Every student received a value that represents the status of their immediate neighborhood. Aggregating this to a higher neighborhood level does not result in many cluster units because we only have two major cities in the data set and students tend to be alone in a specific neighborhood unit. “Neighborhood” should therefore not be used as an additional level in the analysis. We included this variable as a student-level variable.

When we look at the addresses of the children, we notice that they always live in the city of Antwerp or Ghent or close to one of them. We also calculated that in a randomly drawn school in our sample, the correlation of the status of the neighborhood between two randomly chosen individuals is .41.

School-level variables

At the school level, we use two composition variables that are based on the characteristics of students in the schools: the parental educational level and nonnative composition. The parental educational level composition variable is the school percentage of the students in our sample with a mother who possesses a higher secondary education or lower. A higher score on this variable therefore indicates a lower parental educational level composition. The nonnative composition variable is the school percentage of the students in our sample who have a nonnative grandmother on the mother’s side. Furthermore, we included the region in which the school is situated (Antwerp or Ghent).

Research Procedures

We use a two-wave data set on children in the fifth and sixth grades of primary school in Flanders (collected at the end of the 2013–2014 and 2014–2015 school years). The data set includes respondents (i) (Level 1) clustered in schools (j) (Level 2) with k covariates at level 1 and z covariates at level 2. We take this nesting of respondents into account with multilevel regression models (Gaskins et al., 2013; Geiser, 2013; Gelman & Hill, 2007; Hox, 2010).

More specifically, to test our hypotheses, we use random intercept regression models (Heck & Thomas, 2015) in Tables 2 and 3. We use a school-level variance ( $u_{0 j}$ ) and/or an error term ( $e_{i j}$ ). The format of the equation is as follows:

Y_{i j} = γ_{00} + β_{1} X_{1 i j} + \dots + β_{k} X_{k i j} + γ_{01} W_{z j} + \dots + γ_{0 z} W_{z j} + e_{i j} + u_{0 j} .

Table 2.

Multilevel Linear Regression of Teacher Judgments: Parameters With Standard Errors in Parentheses.

Variables	Model 0	Model 1	Model 2	Model 3
Level 1: Student
1. Individual neighborhood score		−0.151 (0.043)***	−0.048 (0.038)	−0.030 (0.035)
2. Nonnative background			−0.550 (0.113)***	−0.260 (0.109)*
3. Male student			0.100 (0.070)	0.064 (0.058)
4. Low parental educational background			−0.252 (0.073)**	−0.094 (0.059)
5. Learning disability			−0.481 (0.072)***	−0.229 (0.060)***
6. Deductive ability (Raven score)			0.352 (0.035)***	0.166 (0.031)***
7. Single-parent family			−0.025 (0.074)	−0.044 (0.073)
8. Number of children in the family			−0.044 (0.045)	−0.022 (0.040)
9. High parental educational expectations			0.193 (0.071)**	0.080 (0.059)
10. Teacher judgments (Wave 1)				0.554 (0.033)***
Level 2: School
11. Nonnative composition (%)			0.005 (0.003)^†	0.006 (0.002)**
12. Low parental education composition (%)			0.001 (0.003)	−0.003 (0.003)
13. Antwerp (school location)			0.123 (0.132)	0.104 (0.115)
Intercept	0.008 (0.050)	0.004 (0.051)	−0.085 (0.126)	−0.072 (0.113)
Variance at the student level	0.97 (0.069)***	0.947 (0.067)***	0.654 (0.033)***	0.454 (0.027)***
Variance at the school level	0.03 (0.025)	0.037 (0.030)	0.052 (0.027)^†	0.038 (0.017)*
Akaike information criterion	2,148.293	2,135.726	1,889.847	1,617.010
Bayesian information criterion	2,162.181	2,154.243	1,959.287	1,691.080

†

p < .10. *p < .05. **p < .01. ***p < .001, two-tailed significance test.

Table 3.

Multilevel Logistic Regression of Parental Educational Expectations (Higher Education): Log Odds With Standard Errors in Parentheses.

Variables	Model 0	Model 1	Model 2	Model 3
Level 1: Student
1. Individual neighborhood score		0.038 (0.094)	−0.015 (0.116)	−0.001 (0.115)
2. Nonnative background			0.993 (0.288)**	1.141 (0.317)***
3. Male student			0.014 (0.170)	−0.009 (0.178)
4. Low parental educational background			−1.007 (0.279)***	−0.924 (0.278)**
5. Learning disability			−0.698 (0.175)***	−0.580 (0.180)**
6. Deductive ability (Raven score)			0.266 (0.096)**	0.174 (0.096)^†
7. Single-parent family			0.042 (0.229)	0.031 (0.235)
8. Number of children in the family			−0.066 (0.104)	−0.051 (0.105)
9. Teacher judgments (Wave 1)				0.286 (0.110)**
Level 2: School
10. Nonnative composition (%)			−0.003 (0.006)	−0.003 (0.006)
11. Low parental education composition (%)			0.001 (0.006)	−0.001 (0.006)
12. Antwerp (school location)			0.192 (0.164)	0.176 (0.185)
Intercept	0.763 (0.080)***	0.764 (0.078)***	0.930 (0.217)***	0.932 (0.214)***
Variance at the school level	0.004 (0.024)	0.003 (0.017)	0.000 (0.000)	0.000 (0.000)
Akaike information criterion	950.665	952.329	903.745	897.000
Bayesian information criterion	959.924	966.217	963.927	961.811

†

indicates p < .10. *indicates p < .05. **indicates p < .01. *** indicates p < .001, two tailed significance test.

We estimate the models in Tables 2 and 3 with regression techniques that are appropriate to the measurement level of the several outcome variables. We use multilevel logistic regression for a dichotomous outcome variable with a logit link function and linear regression for the continuous outcome variables.⁵ Because there was not much variance at the school level, we also estimated models with fixed school effects instead of random school effects (in Table 4). In these models, there are 34 school dummies.

Table 4.

Multivariate Regression (Multiple Outcomes) With Fixed School Effects and Counterfactually Defined Effects: Parameters With Standard Errors in Parentheses.

Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7
Teacher judgments
Individual neighborhood score	−0.062 (0.040)	−0.062 (0.040)	−0.062 (0.040)	−0.062 (0.040)	−0.062 (0.040)	−0.049 (0.034)	−0.049 (0.034)
Nonnative background	−0.547 (0.083)***	−0.547 (0.083)***	−0.547 (0.083)***	−0.547 (0.083)***	−0.547 (0.083)***	−0.251 (0.069)***	−0.251 (0.069)***
Male student	0.095 (0.060)	0.095 (0.060)	0.095 (0.060)	0.095 (0.060)	0.095 (0.060)	0.060 (0.051)	0.060 (0.051)
Low parental educational background	−0.246 (0.078)**	−0.246 (0.078)**	−0.246 (0.078)**	−0.246 (0.078)**	−0.246 (0.078)**	−0.087 (0.063)	−0.087 (0.063)
Learning disability	−0.487 (0.089)***	−0.487 (0.089)***	−0.487 (0.089)***	−0.487 (0.089)***	−0.487 (0.089)***	−0.222 (0.074)**	−0.222 (0.074)**
Deductive ability (Raven score)	0.357 (0.032)***	0.357 (0.032)***	0.357 (0.032)***	0.357 (0.032)***	0.357 (0.032)***	0.168 (0.029)***	0.168 (0.029)***
Single-parent family	−0.024 (0.074)	−0.024 (0.074)	−0.024 (0.074)	−0.024 (0.074)	−0.024 (0.074)	−0.043 (0.066)	−0.043 (0.066)
No. of children in the family	−0.043 (0.037)	−0.043 (0.037)	−0.043 (0.037)	−0.043 (0.037)	−0.043 (0.037)	−0.022 (0.032)	−0.022 (0.032)
High parental educational expectations	0.200 (0.072)**	0.200 (0.072)**	0.200 (0.072)**	0.200 (0.072)**	0.200 (0.072)**	0.088 (0.060)	0.088 (0.060)
Teacher judgments (Wave 1)						0.556 (0.032)***	0.556 (0.032)***
High parental educational expectations
Individual neighborhood score	0.039 (0.083)
Low parental educational background
Individual neighborhood score		0.559 (0.081)***					0.559 (0.081)***
Nonnative background
Individual neighborhood score			1.011 (0.098)***			1.011 (0.098)***
Individual neighborhood score
Low parental educational background				0.530 (0.071)***
Nonnative background					0.861 (0.067)***
Intercepts
Teacher judgments (Wave 2)	1.070 (0.170)***	1.070 (0.170)***	1.070 (0.170)***	1.070 (0.170)***	1.070 (0.170)***	0.653 (0.170)***	0.653 (0.170)***
High parental educational expectations	0.764 (0.080)***
Low parental educational background		−0.380 (0.078)***					−0.380 (0.078)***
Nonnative background			−0.512 (0.083)***			−0.512 (0.083)***
Individual neighborhood score				−0.218 (0.047)***	−0.342 (0.044)***
Residuals
Teacher judgments (Wave 2)	0.624 (0.033)***	0.624 (0.033)***	0.624 (0.033)***	0.624 (0.033)***	0.624 (0.033)***	0.431 (0.022)***	0.431 (0.022)***
Individual neighborhood score				0.931 (0.040)***	0.821 (0.038)***
Akaike information criterion	2,832.047	2,858.822	2,757.596	3,981.616	3,886.752	2,479.415	2,580.641
Bayesian information criterion	3,049.627	3,076.402	2,975.176	4,203.826	4,108.962	2,701.624	2,802.850
Indirect effect: low educational background → neighborhood → teacher judgments (Wave 2)				−0.033 (0.022)
Indirect effect: nonnative background → neighborhood → teacher judgments (Wave 2)					−0.053 (0.035)
Counterfactually defined effects
Total natural indirect effect	0.004 (0.007)	−0.065 (0.022)**	−0.242 (0.043)***			−0.111 (0.032)**	−0.023 (0.017)

†

p < .10. *p < .05. **p < .01. ***p < .001, two-tailed significance test.

We performed mediation analysis to test Hypotheses 3 and 4 (see Table 4). These stated that the effect of neighborhood is mediated by socioeconomic status, ethnic background, or parental educational expectations. This amounts to a continuous variable (neighborhood) that is mediated by binary variables (socioeconomic status, ethnic background, or parental educational expectations) with a continuous outcome variable (teacher judgments at Wave 2).

The fact that there is a binary mediator calls for specific approaches. We performed mediation analyses with binary mediators with the potential outcomes and counterfactual reasoning behind Muthén et al. (2016). We use this to analyze mediation in which we want to know the indirect effect of neighborhood on teacher judgments.⁶ The value of an outcome that is potential is “the value . . . that would have been observed under a certain condition. It is potentially observed when that condition is observed and is otherwise unobserved. The unobserved value is called a counterfactual” (Muthén et al., 2016, p. 189). Imagine, for example, that teachers have a negative judgment about students because the students have a low socioeconomic background. If a researcher wanted to know whether the teachers would still have those perceptions if those students had a high socioeconomic background, he would have to make a counterfactual statement. This is because the researcher cannot observe the latter situation but has observed the former situation. To find the indirect effects, we calculate a so-called total natural indirect effect (TNIE). Research using this indirect effect compares two values of a continuous variable that is hypothesized to be mediated (i.e., neighborhood) by binary mediators. $TNIE$ shows the indirect effect of neighborhood on teacher judgments through its effect on the mediator: $TNIE = E [Y (1, M (1)) - Y (1, M (0))]$ . The first term in the equation can be defined as the mean of teacher judgments when students live in a specific neighborhood while the mediator (i.e., socioeconomic status, ethnic background, or parental educational expectations) varies as it would be for that neighborhood. The second term in the equation can be defined as the mean of teacher judgments when students live in a specific neighborhood while the mediator (i.e., socioeconomic status, ethnic background, or parental educational expectations) varies as it would be if students would have lived in another neighborhood. The latter counterfactual mean shows how teacher judgments are influenced by a specific neighborhood without socioeconomic status, ethnic background, or parental educational expectations being influenced by the characteristics of that neighborhood. The difference between both terms gives the $TNIE$ .

The models are estimated with robust maximum likelihood. Data analysis was done with Mplus (version 8) for the regression modeling and SPSS (version 21), and Excel for descriptive statistics and data management.⁷

Results

In this section, we build toward multivariate multilevel regression models. In Tables 2 and 3, we report multilevel logistic and linear regressions of parental educational expectations and Wave 2 teacher judgments.

In Model 0 of both tables, we report the variances at the school and student levels in an empty model. We see that for both outcomes, there is not much school variance. In Table 2, we examine effects on teacher judgments. In Model 1, we include the effect of neighborhood in a separate model (without the other variables). The effect is significant but loses its significance after controlling for student and school variables (Model 2). There is therefore no direct significant effect of neighborhood on teacher judgments after including these variables.

Parental educational expectations (in Wave 1) are significantly related to teacher judgments (in Wave 2; in Model 2, controlling for background variables; Hypothesis 1). For children with high parental educational expectations in Wave 1, teacher judgments increase (i.e., are more positive) on average by 0.19 standard deviation in Wave 2 compared with children with low parental educational expectations in Wave 1. This effect lowers and becomes insignificant when we include Wave 1 teacher judgments (Model 3).

In Table 3, Model 1, we entered the neighborhood variable to test whether pupils living in socioeconomically deprived neighborhoods are more likely to experience lower parental educational expectations than students living in a more socioeconomically affluent neighborhood (Hypothesis 2). Results show that the socioeconomic status of a neighborhood is not directly related in a significant way to parental educational expectations. This result holds after controlling for all student- and school-level variables (Table 3, Models 2 and 3).

After testing the direct effects, we performed mediation analysis to test Hypotheses 3 and 4. The neighborhood data were collected in 2011 (census data) while the ethnic and socioeconomic background of students were collected after 2011. The teacher judgment data of Wave 2 were collected at the end of the 2014–2015 school year. We use specifications with fixed school effects because there is not much school variance of teacher judgments to explain (see Table 2).

We used counterfactual and potential outcome effect modeling as reported in Muthén et al. (2016). In Table 4, we report multivariate results with indirect effects. We see that there are significant indirect effects from neighborhood to ethnic background/socioeconomic background to teacher judgments. The mediation from neighborhood to high parental expectations to teacher judgments is insignificant. In Columns 4 and 5, we also see that there are no significant mediations from socioeconomic/ethnic background to neighborhood to teacher judgments.

As can be seen from Table 4, Column 2, there is a −0.065 standard deviation decrease of teacher judgments when neighborhood goes from −1 to 1 standard deviation (neighborhood → socioeconomic background → teacher judgments). This is an indirect effect that is mediated by low parental educational background. This effect, however, becomes insignificant after the inclusion of Wave 1 teacher judgments (Column 7).

Column 3 shows that there is a −0.24 standard deviation decrease of teacher judgments when neighborhood goes from −1 to 1 standard deviation. This is an indirect effect mediated by nonnative background (neighborhood → nonnative background → teacher judgments).

With the inclusion of Wave 1 teacher judgments (Columns 6 and 7), only the indirect effect with nonnative background stays significant but with a smaller coefficient.

Discussion and Conclusion

The aim of this article was to study the effects of neighborhood on teacher judgments and parental educational expectations. We expected direct and indirect relationships between these concepts. We formulated four hypotheses. Hypotheses 1 and 2 stated that there is a direct effect of parental educational expectations on teacher judgments and of the neighborhood on parental educational expectations. Hypotheses 3 and 4 stated the effect of neighborhood on teacher judgments is mediated by the socioeconomic/ethnic characteristics and parental educational expectations of the students. We use data (a relatively large sample) that were collected in the “Transbaso” study conducted in the cities of Antwerp and Ghent, which focused on children in the fifth and sixth grades of primary school in Flanders.

We found that the effect of neighborhood on teacher judgments is mediated by the socioeconomic and ethnic background of students. We also found that there is a direct effect of parental educational expectations on teacher judgments and no direct effect of neighborhood on parental educational expectations. The direct effect of parental educational expectations on teacher judgments becomes insignificant after the inclusion of Wave 1 teacher judgments. We then showed that the effect of neighborhood on teacher judgments is not mediated by parental educational expectations. In the sample of Antwerp and Ghent, it seems that the effect of neighborhood exists because of the selection of students with a particular background in schools.

The mediation of the effect of neighborhood on teacher judgments is stronger for nonnative than for socioeconomic background. The latter also disappears after including Wave 1 teacher judgments while the former stays (even after controlling for deductive ability). The effect of neighborhood seems to be mainly one of selecting pupils in the school. Teachers have—under control of deductive ability and previous teacher judgments—a lower judgment of students with a nonnative background. The fact that the mediation of the effect of the neighborhood on teacher judgments stays for nonnative students might also be related to the deficit perspective (Blake et al., 2016; Garcia & Guerra, 2004; Milner, 2008, 2013). According to this perspective, teachers can focus on certain characteristics of students that are perceived as problematic or deficit (Walker, 2011). Consequences of adopting a deficit perspective in education-related matters can be more negative teacher judgments about certain groups of students. Teachers may make implicit assumptions about (the performance of) students that are partly or wholly independent of the achievement of their students (see also the Kirwan Institute for the Study of Race and Ethnicity, 2017). The teachers in our sample may adopt a deficit approach. This is a topic for further research.

Another theoretical framework that is relevant here is critical race theory (CRT), in which the concept of “race” is used to interpret and study educational processes in the United States (Dixson & Rousseau, 2005; Ladson-Billings, 1999; Ladson-Billings & Tate, 1995). According to this framework, inequality at school is related to racism that is institutionally embedded in society. This racism may also affect educational processes. People of color are discriminated against through, for example, lower teacher judgments. Whether the lower teacher judgments in the current findings of this article are due to similar racism/discrimination processes is also a topic that needs further investigation.

This contribution showed that it is wise to combat residential segregation (see also Fossett, 2006; Van der Bracht et al., 2015; Verhaeghe et al., 2012). The continuing immigration in various Western European countries has triggered social and political concerns. Immigration creates new challenges for the education of coming generations. For instance, many policymakers worry about the growing school segregation (i.e., the concentration of pupils with a low socioeconomic and/or an immigrant background) happening in many European cities. The concentration of pupils with such background in specific schools is perceived as unfavorable for educational performance, but also as an obstruction to social integration and cohesion (Jenkins et al., 2008). A contribution of this text is to show that policymakers should try to prevent residential segregation. Combating segregation in primary schools means preventing and countering segregation in neighborhoods. Formulating policy on school segregation should necessarily imply policy on residential segregation. The two cannot be easily separated from each other.

Based on the findings of this article, there are other recommendations that we can give. In Flanders, there are not many teachers with a nonnative background (see also Agirdag et al., 2016). One policy option might be to expand the teaching profession to people with a nonnative background. It might also be relevant to actively recruit more students with a nonnative background to teacher training courses. Research has shown that a teacher body that is more similar to the student composition of the school can be beneficial for student outcomes (see also Blake et al., 2016).

When schools have many students with diverse cultural backgrounds, it might also be wise to train and immerse teachers in (cultural) contexts other than their own (see Blake et al., 2011; Ladson-Billings, 2009; Milner, 2008; Walker, 2011). In this way, they can also increase their knowledge of the background of their students. These backgrounds can also be used as a way to apply the curriculum that they have to teach.

As a possible point of criticism, it might be said that the effects of neighborhood on the outcomes of pupils are not due to the neighborhood itself but rather to how (the families of) pupils were assigned to neighborhoods (Sampson et al., 2002). This possible criticism is overstated for the current text because it is focused on the effect of neighborhood through its effect on the selection and sorting of pupils in school. What lies behind the selection of families (and pupils) in neighborhoods is not the focus of the current contribution. Further research should also include more information on the teachers (e.g., their working experience and native/nonnative background). A qualitative study might also be devised to study parental educational expectations and teacher judgments.

A comparison between research on neighborhood effects in the United States and European settings should only be done with great attention to differences in settings. The residential situation in Antwerp and Ghent is different compared with the multivarious situations in the United States, especially insofar as Antwerp and Ghent, like other cities in Belgium, do not have any exceptionally highly socioeconomically segregated neighborhoods. This, among other differences, should be taken into account when comparing research on European (i.e., Belgian) and cities in the United States—where the bulk of research on neighborhoods is conducted (see also Kauppinen, 2006). This might account for the fact that we did not find direct effects of neighborhood on teacher judgment and parental educational expectations.

Footnotes

Authors’ Note

Simon Boone is now affiliated with Université Libre de Bruxelles, Brussels, Belgium.

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 is a Flanders Innovation and Entrepreneurship (IWT/SBO) funded study (project no. 130074).

Notes

Author Biographies

Kenneth Hemmerechts is a post-doctoral researcher at the Free University of Brussels in Etterbeek (Brussels; VUB) where he is currently working in the field of the sociology of education. Previously, he has worked at different universities in Belgium on a variety of topics, including migration, criminal recidivism, fraud, police capacity, employment and trade unionism, genocide, and social theory.

Dimokritos Kavadias is professor at the political science department of the VUB. He currently teaches methodology courses and political psychology. His current research activities focus on political socialization, political psychology, civic education, and educational policy.

Simon Boone (PhD in sociology, Ghent University) is currently director of the inter-university research platform Brussels Studies Institute. He obtained his PhD in sociology from Ghent University (2013), with a dissertation focusing on social disparities in educational choice at the transition from primary to secondary education in Flanders. His research revolved around the determining factors at the family and school level. His research interests include educational inequality, education systems, school effects and mixed-methods research.

References

Agirdag

Merry

M. S.

Van Houtte

(2016). Teachers’ understanding of multicultural education and the correlates of multicultural content integration in Flanders. Education and Urban Society, 48(6), 556–582.

Agirdag

Van Avermaet

Van Houtte

(2013). School segregation and math achievement: A mixed-method study on the role of self-fulfilling prophecies. Teachers College Record, 115(3), 1–50.

Agirdag

Van Houtte

Van Avermaet

(2012). Ethnic school segregation and self-esteem: The role of teacher-pupil relationships. Urban Education, 47(6), 1135–1159.

Ainsworth

(2002). Why does it take a village? The mediation of neighborhood effects on educational achievement. Social Forces, 81(1), 117–152.

Ainsworth

(2010). Does the race of neighborhood role models matter? Collective socialization effects on educational achievement. Urban Education, 45(4), 401–423.

Alexander

K. L.

Entwisle

D. R.

Bedinger

S. D.

(1994). When expectations work: Race and socioeconomic differences in school performance. Social Psychology Quarterly, 57(4), 283–299.

Andersson

Subramanian

S. V.

(2006). Explorations of neighbourhood and educational outcomes for young Swedes. Urban Studies, 43(11), 2013–2025.

Baysu

de Valk

(2012). Navigating the school system in Sweden, Belgium, Austria and Germany: School segregation and second-generation school trajectories. Ethnicities, 12(6), 776–799.

Beutel

Anderson

(2008). Race and the educational expectations of parents and children: The case of South Africa. Sociological Quarterly, 49(2), 335–361.

10.

Blake

J. J.

Butler

B. R.

Lewis

C. W.

Darensbourg

(2011). Unmasking the inequitable discipline experiences of urban black girls: Implications for urban educational stakeholders. The Urban Review, 43(1), 90–106.

11.

Blake

J. J.

Smith

D. M.

Marchbanks

M. P.

Seibert

A. L.

Wood

S. M.

Kim

E. S.

(2016). Does student–teacher racial/ethnic match impact black students’ discipline risk? A test of the cultural synchrony hypothesis. In Skiba

R. J.

Mediratta

Rausch

M. K.

(Eds.), Inequality in school discipline. Research and practice to reduce disparities (pp. 79–98). Palgrave Macmillan.

12.

Boardman

J. D.

Saint Onge

J. M.

(2005). Neighborhoods and adolescent development. Child Youth Environment, 15(1), 138–164.

13.

Boone

(2016). Intern rapport Transbaso. Vrije Universiteit Brussel.

14.

Bourdieu

(1984). Distinction. Routledge.

15.

Bourdieu

(1990). The logic of practice. Polity Press.

16.

Bourdieu

(1998). The state nobility. Stanford University Press.

17.

Boxer

Goldstein

DeLorenzo

Savoy

Mercado

(2011). Educational aspiration-expectation discrepancies: Relation to socioeconomic and academic risk-related factors. Journal of Adolescence, 34(4), 609–617.

18.

Brännström

(2008). Making their mark: The effects of neighbourhood and upper secondary school on educational achievement. European Sociological Review, 24(4), 463–478.

19.

Brattbakk

Wessel

(2013). Long-term neighbourhood effects on education, income and employment among adolescents in Oslo. Urban Studies, 50(2), 391–406.

20.

Bronfenbrenner

(1979). The ecology of human development: Experiments by nature and design. Harvard University Press.

21.

Bronfenbrenner

(1986). Ecology of the family as a context for human development: Research perspectives. Developmental Psychology, 22(6), 723–742.

22.

Brooks-Gunn

Duncan

G. J.

Klebanov

P. K.

Sealand

(1993). Do neighborhoods influence child and adolescent development? American Journal of Sociology, 99(2), 353–395.

23.

Bucher

(1998). The influence of models in forming moral identity. International Journal of Educational Research, 27(7), 619–627.

24.

Buchmann

Dalton

(2002). Interpersonal influences and educational aspirations in 12 countries: The importance of institutional context. Sociology of Education, 75(2), 99–122.

25.

Carlson

Cowen

J. M.

(2014). Student neighborhoods, schools, and test score growth: Evidence from Milwaukee, Wisconsin. Sociology of Education, 88(1), 1–17.

26.

Ceballo

McLoyd

V. C.

Toyokawa

(2004). The influence of neighborhood quality on adolescents’ educational values and school effort. Journal of Adolescent Research, 19(6), 716–739.

27.

Corsaro

(2010). The sociology of childhood. SAGE.

28.

Creten

Douterlungne

Verhaeghe

J.-P.

De Vos

(2000). Voor elk wat wils: schoolkeuze in het basis- en secundair onderwijs [Something for everyone: choice of school in primary and secondary education]. Hoger Instituut voor de Arbeid.

29.

Crosnoe

Mistry

R. S.

Elder

G. H.

(2002). Economic disadvantage, family dynamics, and adolescent enrollment in higher education. Journal of Marriage and Family, 64(3), 690–702.

30.

Davis-Kean

P. E.

(2005). The influence of parent education and family income on child achievement: The indirect role of parental expectations and the home environment. Journal of Family Psychology, 19(2), 294–304.

31.

De Groof

Franck

. (2013). Belgium (Flemish community). In Ainley

Schulz

Friedman

(Eds.), ICCS 2009 encyclopedia: Approaches to civic and citizenship education around the world (pp. 39–49). International Association for the Evaluation of Educational Achievement.

32.

De Rycke

Swyngedouw

. (1999). The value of concentration schools as appreciated by Moroccan, Turks and unskilled Belgians in Brussels. International Journal of Educational Research, 31(4), 267–281.

33.

Dixson

A. D.

Rousseau

C. K.

(2005). And we are still not saved: Critical race theory in education ten years later. Race Ethnicity and Education, 8(1), 7–27.

34.

Dupere

Leventhal

Crosnoe

Dion

(2010). Understanding the positive role of neighborhood socioeconomic advantage in achievement: The contribution of the home, child care, and school environments. Developmental Psychology, 46(5), 1227–1244.

35.

Durkheim

(1961). Moral education. The Free Press. (Original work published 1925)

36.

Enders

(2010). Applied missing data analyses. Guilford Press.

37.

Epstein

(1987). Parent involvement: What the research says to administrators. Education and Urban Society, 19(2), 19–36.

38.

Feinstein

Duckworth

Sabates

(2004). A model of the inter-generational transmission of educational success (Wider Benefits of Learning, Research Report No. 10). Institute of Education.

39.

Feliciano

(2006). Beyond the family: The influence of pre-migration group status on the educational expectations of immigrants’ children. Sociology of Education, 79(4), 281–303.

40.

Fossett

(2006). Ethnic preferences, social distance dynamics, and residential segregation: Theoretical explorations using simulation analysis. The Journal of Mathematical Sociology, 30(3–4), 185–273.

41.

Garcia

S. B.

Guerra

P. L.

(2004). Deconstructing deficit thinking: Working with educators to create more equitable learning environments. Education and Urban Society, 36(2), 150–168.

42.

Garner

C. L.

Raudenbush

S. W.

(1991). Neighbourhood effects on educational attainment: A multilevel analysis. Sociology of Education, 64(4), 251–262.

43.

Gaskins

Golder

Siegel

D. A.

(2013). Religious participation and economic conservatism. American Journal of Political Science, 57(4), 823–840.

44.

Geiser

(2013). Data analysis with Mplus. Guilford Press.

45.

Gelman

Hill

(2007). Data analysis using regression and multilevel/hierarchical models. Cambridge University Press.

46.

Goldsmith

P. R.

(2009). Schools or neighborhoods, or both? Race and ethnic segregation and educational attainment. Social Forces, 87(4), 1913–1942.

47.

González-Betancour

López-Puig

(2016). Grade retention in primary education is associated with quarter of birth and socioeconomic status. PLOS ONE, 11(11), Article e0166431.

48.

Gorman

(1998). Social class and parental attitudes toward education. Resistance and conformity to schooling in the family. Journal of Contemporary Ethnography, 27(1), 10–44.

49.

Grolnick

W. S.

Benjet

Kurowski

C. O.

Apostoleris

(1997). Predictors of parent involvement in children’s schooling. Journal of Educational Psychology, 89(3), 538–548.

50.

Grolnick

W. S.

Slowiaczek

M. L.

(1994). Parent involvement in children’s schooling: A multidimensional conceptualization and motivational model. Child Development, 65(1), 237–252.

51.

Hanson

(1994). Lost talent: Unrealized educational aspirations and expectations among U.S. youths. Sociology of Education, 67(3), 159–183.

52.

Hattie

(2009). Visible learning: A synthesis of over 800 meta-analyses relating to achievement. Routledge.

53.

Hauser

R. M.

Frederick

C. B.

Andrew

(2007). Grade retention in the age of accountability. In Gamoran

(Ed.), No Child Left Behind and the poverty gap (pp. 120–153). Brookings Institution.

54.

Heck

R. H.

Thomas

S. L.

(2015). An introduction to multilevel modeling techniques: MLM and SEM approaches using Mplus. Routledge.

55.

Hermans

(1995). Moroccan immigrants and school success. International Journal of Educational Research, 23(1), 33–44.

56.

Hornby

Lafaele

(2011). Barriers to parental involvement in education: An explanatory model. Educational Review, 63(1), 37–52.

57.

Hox

(2010). Multilevel analysis: Techniques and applications. Routledge.

58.

Jencks

Mayer

(1990). The social consequences of growing up in a poor neighborhood. In Lynn

L. E.

McGeary

M. F. H.

(Eds.), Inner-city poverty in the United States (pp. 111–186). National Academies Press.

59.

Jerrim

(2014). The unrealistic educational expectations of high school pupils: Is America exceptional? The Sociological Quarterly, 55(1), 196–231.

60.

Jeynes

W. H.

(2005). A meta-analysis of the relation of parental involvement to urban elementary school student academic achievement. Urban Education, 40(3), 237–269.

61.

Jeynes

W. H.

(2007). The relationship between parental involvement and urban secondary school student academic achievement: A meta-analysis. Urban Education, 42(1), 82–110.

62.

Jenkins

S. P.

Micklewright

Schnepf

(2008). Social segregation in secondary schools: How does England compare with other countries. Oxford Review of Education, 34(1), 21–37.

63.

Kauppinen

T. M.

(2006). Neighborhood effects in a European city: Secondary education of young people in Helsinki. Social Science Research, 36(1), 421–444.

64.

Kauppinen

T. M.

(2008). Schools as mediators of neighbourhood effects on choice between vocational and academic tracks of secondary education in Helsinki. European Sociological Review, 24(3), 379–391.

65.

Khattab

(2014). How and when do educational aspirations, expectations and achievement align? Sociological Research Online, 17(4), 1–7.

66.

Khattab

(2015). Students’ Aspirations, Expectations and School Achievement: What really matters? British Educational Research Journal, 41(5), 731–748.

67.

Kim

Sherraden

Clancy

(2013). Do mother’s educational expectations differ by race and ethnicity, or socioeconomic status? Economics of Education Review, 33, 82–94.

68.

Kirwan Institute for the Study of Race and Ethnicity. (2017). Implicit bias review. Race and ethnicity: Views from inside the unconscious mind. The Ohio State University.

69.

Ladson-Billings

(1999). Preparing teachers for diverse student populations: A critical race theory perspective. Review of Research in Education, 24, 211–247.

70.

Ladson-Billings

(2009). The dreamkeepers: Successful teachers of African American children. John Wiley.

71.

Ladson-Billings

Tate

W. F.

(1995). Toward a critical race theory of education. Teachers College Record, 97(1), 47–68.

72.

Lareau

(1989). Home advantage: Social class and parental intervention in elementary education. Falmer Press.

73.

Leventhal

Brooks-Gunn

(2004). A randomized study of neighborhood effects on low income children’s educational outcomes. Developmental Psychology, 40(4), 488–507.

74.

Martinez

M. J.

(2018). Does school racial composition matter to teachers: Examining racial differences in teachers’ perceptions of student problems. Urban Education. Advance online publication. https://doi.org/10.1177/0042085918770709

75.

McKown

Weinstein

R. S.

(2002). Modeling the role of child ethnicity and gender in children’s differential response to teacher expectations. Journal of Applied Social Psychology, 32(1), 159–184.

76.

Mickelson

(1990). The attitude-achievement paradox among black adolescents. Sociology of Education, 63, 44–61.

77.

Milner

(2008). Disrupting deficit notions of difference: Counter-narratives of teachers and community in urban education. Teaching and Teacher Education, 24(6), 1573–1598.

78.

Milner

(2012). But what is urban education. Urban Education, 47(3), 556–561.

79.

Milner

(2013). Rethinking achievement gap talk in urban education. Urban Education, 48(1), 3–8.

80.

Milner

Lomotey

(2014). Handbook of urban education. Routledge.

81.

Murry

Berkel

Gaylord-Harden

N. K.

Copeland-Linder

Nation

(2011). Neighborhood poverty and adolescent development. Journal of Research on Adolescence, 21(1), 114–128.

82.

Muthén

L. K.

Asparouhov

(2016). Regression and mediation analysis using Mplus. Muthén & Muthén.

83.

Owens

(2010). Neighborhoods and schools as competing and reinforcing contexts for educational attainment. Sociology of Education, 83(4), 287–311.

84.

Pong

S. L.

Hao

L. X.

(2007). Neighborhood and school factors in the school performance of immigrants’ children. International Migration Review, 41(1), 206–241.

85.

Raven

J.C.

Court

J.H.

(2003). Manual for Raven’s Progressive Matrices and Vocabulary Scales. Harcourt Assessment.

86.

Roscigno

V. J.

Crowley

M. L.

(2001). Rurality, institutional disadvantage, and achievement/attainment. Rural Sociology, 66(2), 268–292.

87.

Roscigno

V. J.

Tomaskovic-Devey

Crowley

(2006). Education and the inequalities of place. Social Forces, 84(4), 2121–2145.

88.

Ruiter

Van Tubergen

(2009). Religious attendance in cross-national perspective: A multilevel analysis of 60 countries. American Journal of Sociology, 115(3), 863–895.

89.

Sadler

R. C.

Lafreniere

D. J.

(2017). You are where you live: Methodological challenges to measuring children’s exposure to hazards. Journal of Children and Poverty, 23, 189–198.

90.

Sampson

R. J.

Morenoff

J. D.

Gannon-Rowley

(2002). Assessing “neighborhood effects”: Social processes and new directions in research. Annual Review of Sociology, 28, 443–478.

91.

Sanbonmatsu

Kling

J. R.

Duncan

G. J.

Brooks-Gunn

(2006). Neighborhoods and academic achievement: Results from the moving to opportunity experiment. The Journal of Human Resources, 41(4), 649–691.

92.

Sewell

W. H.

Haller

Ohlendorf

G. W.

(1970). The educational and early occupational attainment process: Replication and revision. American Sociological Review, 35(6), 1014–1027.

93.

Sewell

W. H.

Haller

Portes

(1969). The educational and early occupational attainment process. American Sociological Review, 34(1), 82–92.

94.

Sharkey

Elwert

(2011). The legacy of disadvantage: Multigenerational neighborhood effects on cognitive ability. American Journal of Sociology, 116(6), 1934–1981.

95.

Stewart

Simons

(2007). The effect of neighborhood context on the college aspirations of African American adolescents. American Educational Research Journal, 44(4), 896–919.

96.

Sykes

Kuyper

(2009). Neighbourhood effects on youth educational achievement in the Netherlands: Can effects be identified and do they vary by student background characteristics? Environment and Planning A, 41(10), 2417–2436.

97.

Sykes

Musterd

(2011). Examining neighbourhood and school effects simultaneously: What does the Dutch evidence show? Urban Studies, 48(7), 1307–1331.

98.

Thys

Van Houtte

(2016). Ethnic composition of the primary school and educational choice: Does the culture of teacher expectations matter? Teaching and Teacher Education, 59, 383–391.

99.

Van der Bracht

Coenen

Van de Putte

. (2015). The not-in my property syndrome: The occurrence of ethnic discrimination in the rental housing market in Belgium. Journal of Ethnic and Migration Studies, 41(1), 158–175.

100.

Van Maele

Van Houtte

. (2009). Faculty trust and organizational school characteristics: An exploration across secondary schools in Flanders. Educational Administration Quarterly, 45(4), 556–589.

101.

Verhaeghe

P. P.

Van der Bracht

Van de Putte

(2012). Migrant zoekt toekomst: Gent op het keerpunt tussen oude en nieuwe migratie [Migrant is looking for a future: Ghent at the turning point between old and new migration]. Garant.

102.

Walker

K. L.

(2011). Deficit thinking and the effective teacher. Education and Urban Society, 43(5), 576–597.

103.

Wilson

W. J.

(1987). The truly disadvantaged: The inner city, the underclass, and public policy. University of Chicago Press.

104.

Wodtke

Harding

Elwert

(2011). Neighborhood effects in temporal perspective: The impact of long-term exposure to concentrated disadvantage on high school graduation. American Sociological Review, 76(5), 713–736.

105.

Yamamoto

Holloway

S. D.

(2010). Parental expectations and children’s academic performance in sociocultural context. Educational Psychology Review, 22(3), 189–214.