Mediational Effect of Teacher-Based Discrimination on Academic Performance: An Intersectional Analysis of Race, Gender, and Income/Class

Abstract

Drawing on prior research, this study applies an intersectional framework to investigate discrimination in the context of teacher–student relationships and its influence on students’ academic outcomes. Outcomes assessed were inclusive of self-efficacy, school attendance, and grade point average (GPA). For this analysis, structural equation modeling was used with a cross-sectional sample of the Maryland and Adolescent Development in Context Study (MADICS) and the youth self-administered (YSA) questionnaires administered when the youth were in 8th grade (Wave 3). A total of 1182 students completed the survey, of whom 704 were selected for this study. Findings show teacher discrimination as a mechanism to uncover some of the ways race, gender, and income simultaneously intersect to affect students’ academic outcomes. The current study confirms and extends prior work establishing associations among race, gender, income, and teacher discrimination and academic outcomes among African American youth. African American students, especially males, regardless of income levels, may benefit directly—evidenced in visible academic performance—from more positive and race-conscious interactions with teachers. Future implications for practice are shared.

1. Introduction

Conceptually distinct in and of themselves, race, gender, and income are inextricably linked to discriminatory student experiences, particularly for those students of color [1,2,3,4]. The extant literature paints the US educational canvas as racialized, gendered, and stratified on income, particularly for those classified as being from minority backgrounds [1,3,4,5,6,7,8,9,10]. Regardless of demonstrated academic ability, African American students in particular are shown to experience significant bouts of discrimination through teacher–student interactions [4,11,12,13].

Research examining the independent relations of race, gender, income, and discrimination has very minimally examined the intersectionality of these variables on student experiences through interactions with teachers. Studies where differences in race, gender, and income/social class have been examined suggest associations [1] that provide insights, though limited, in relaying the nuances which operate to influence student outcomes. Therefore, the current study attempts to move beyond independent associations to an intersectional analysis. Such an analysis examines the mechanisms by which race, gender, and income simultaneously and intersectionally relate to reported incidents of discrimination—through teacher–student relationships—to influence students’ academic outcomes (e.g., academic self-efficacy, attendance: see Figure 1).

2. Theoretical Framework

Developed by Crenshaw [14]) the intersectional analysis uses an analytical framework [15]. Crenshaw first used an intersectional perspective to capture and explain unique ecological forces (e.g., racism, and sexism) which simultaneously characterized black women in pursuit of anti-discrimination claims in the U.S. Courts. Crenshaw [14] argued that it was difficult to separate gendered experiences from racial experiences. This was the birth of a novel framing in the social sciences. Its growth deepened our understanding of relationships between and among social stratification and identity markers. It spotlighted how racism and sexism may operate simultaneously to jointly produce inequities contributing to disparities in group experiences and outcomes [5,15]. Research analysis, as it relates to the “what and how” of the intersectionality perspective, remains debatable, and the body of work continues to grow [16]. However, to date most research using quantitative approaches looks for variables with prominent effects instead of their intersectionality. Intersectionality as an analytical strategy contradicts the notion of indicators intertwining to shape experiences and outcomes [17]. Using a structural equation model, the current study examines the intersectional effects of race, gender, and income as markers for discrimination observed through teacher–student interactions.

We accept the premise that herein, race, gender, and income are structural constructions used to organize social relations. Each variable provides “an initial basis for deciding who the other is, who we are in comparison with, and therefore how each of us is likely to behave” [16] (p. 298). Although these stratification markers are cultural constructions, because of their role in organizing social relations, they are consequential to resources—material, informational, and symbolic accessibility. In social contexts, these constructs are simultaneously present. They associate with cultural meanings (e.g., stereotypes) to shape the resources available within such contexts [16]. This implies that cultural and stereotypical societal assumptions constructed around race, gender, and income jointly affect the activation and distribution of resources in schools. These variables may influence differential access to educational resources and subsequently, differential academic outcomes (Stanton–Salazar, 2011). Unlike income or class, race and gender appear to be essential social and cultural constructs. For instance, income and class are deeply embedded in race and gender, and when salient, can powerfully influence social behaviors to produce inequalities, making income/class an additive factor to its intersection with race and gender [1,16].

For example, Bécares–Priest [17], in a study across socio-economic status, revealed race and gender as prominent in explaining behavioral problems and academic outcomes among White, Black, and Latino adolescents. They concluded that “Growing up Black, Latino, or White in the U.S is not the same for boys and girls and growing up as a boy or a girl in America does not lead to the same outcomes and opportunities for Black, Latino and White children as they become adults” (p. 12). Two other studies of students which used intersectional analyses, conducted by Diamond and Lewis [1] and Gillborn et al. [18], offered qualitative findings to show students’ race, gender, and income/class were important social position variables influencing how school personnel interacted with students. In an adult study, using a nationally representative sample, Penner and Saperstein [15] concluded that intersectional—not independent—social stratification forces shaped individuals’ everyday lives. Furthermore, the authors reported, “race and class are gendered, and gender and class are racialized” (p. 320). These outcomes point to the necessity of further investigations which leverage an intersectional framework, investigations which examine the role of race, gender, and income through teacher–students’ relationships.

In sum, scholars who have used an intersectional perspective continue to underscore the claim that race, gender, and income/class are not independent. The school context evidences these social stratification markers which operate jointly to organize students’ differential access to educational resources and differences in academic outcomes [19]. Discrimination through the teacher–student interaction context may be one vehicle through which to learn of racialized and gendered intersections [1,20]. Teacher–student interactions refer to consistent engagements between teachers and students in the immediate classroom and through various instruction, pedagogical, and learning practices. These interactions are the central mechanism through which students access and use social, material, cultural, and psychological school resources [19,21,22].

3. Teacher-Based Discrimination

Given the current knowledge of the authors to date, intersectional forces with students’ educational outcomes have received very little attention in the literature but have reinforced that the quality of teacher–student relationships remain important to learning, school connectedness, and achievement [22,23,24,25]. Other researchers have emphasized that these variables, when combined, promote and inhibit students’ psychosocial change and academic excellence, particularly for racialized and working-class students [19,26]. The racialized and gendered nuances of schools [1,27], regarding various identity markers, may trigger teacher-based discrimination, undermining the support students receive [4,10,11,12,13,27]. The educational landscape is a mezzo-level actor for the activation of the cultural stereotypes and assumptions around structural constructs which shape teacher–student relationships to produce differential access to educational resources [28,29,30,31,32]. Of note, as institutional agents, teachers directly and indirectly affect the transmission of resources and students’ school engagement [19,26,33].

These discriminations are often expressed as general societal stereotypes about the attributes of a social group [34]. For instance, across primary and secondary education contexts, there is a belief that boys are better in science, technology, engineering, and math (STEM) than girls [8,35,36]. Similarly, there is a racial stereotype that African American students are intellectually inferior, violent, and emotionally disordered [2,4,37,38,39]. Additionally, students from lower-income backgrounds and inner cities—the majority of whom are racial–ethnic minorities—are stereotyped as rude, disruptive, and having educationally disengaged parents [40,41,42]. These stereotypes prime adolescents for discrimination in school through their interaction with teachers. These biases negatively interfere with learning and have far-reaching effects on students’ long-term academic persistence, performance, college attainment, and career opportunities [43].

Admittedly, in schools, the social positions of students do not operate in isolation. They tend to co-mingle to shape gatekeeping access to opportunities for children and their families. Diamond and Lewis [1] assert the following:

The educational marketplace is characterized by economic capital (money and material resources) as well as the exchange of other forms of capital such as social relationships, networks, and symbols of competence, innocence, and legitimacy (symbolic capital). These forms of capital are used to access educational resources, and those that possess more valued forms of capital maintain advantaged positions (p. 843).

These forms of capital and their activation and distribution in schools intersect with race and income/class in a gendered way [3,7,17,27,35,40,44].

4. Teacher-Based Discrimination and Academic Outcomes (Self-Efficacy, Attendance, and GPA)

Academic self-efficacy beliefs—judgments about one’s capabilities to perform an academic-related task despite adversities to achieve successful academic goals—is an integral part of a student’s overall academic performance [45,46,47,48]. Research suggests there are differential associations between students’ academic self-efficacy and their race, gender, and income/class and performance [17,47]. Findings from Bécares–Priest [17] show that across income/class, White students and females were associated with higher reading efficacy as compared with Blacks and Hispanics. The same was true for math, where Whites and boys were associated with higher math academic efficacy as compared with Blacks and Hispanics. Hispanic girls were associated with lower math efficacy compared with boys. Relatedly, Britner–Pajares [17] found that African American girls reported higher science academic efficacy and science achievement than African American males. Across races, White students reported higher science academic efficacy and science achievement compared to Blacks.

Qualitative research studies have also reported evidence of discrimination in schools whereby race, gender, and income/class background of students are linked to differential treatment through the teacher–student interaction context [1,18,22,26,49,50]. Using the Howard [51] study as an example, one participant shared how teachers and school personnel’s interactions activated discriminatory practices, which in turn influenced their educational experiences, resulting in academic outcome disparities.

“A lot of teachers, principals, and counselors are just straight out prejudice. They have all these stereotypes about us because of our culture and race…. It’s like a big tug of war for our minds. We’re pulling on one side telling ourselves that we are smart, bright, and talented. They are pulling on the other side saying that we are dumb, lazy, and will never amount to anything”.

(p. 12)

Additionally, Diamond and Lewis [1] and Gillborn and colleagues [18] corroborate the above quote. They reiterate that the teacher–student interaction intersects with academic self-efficacy. Diamond and Lewis [1] in particular remind us that teachers have good intentions, but stereotypical narratives (e.g., Blacks are inherently criminals and intellectually inferior; poor kids are rude, disruptive; and Whites are inherently innocent and intelligent) interfere to undermine these intentions [16]. All in all, the educators’ relationship, the teacher–student interpersonal interactions via proximal learning ecology (e.g., classroom arrangement, teacher expectations, how teachers affirm or invalidate the cultural asset and identity of students), and self-efficacy as contextual specificity [24,46,52,53,54] directly enhance or deter a student’s academic performance.

While these example studies have linked differential associations in academic outcomes to race, gender, income/class, and teacher discriminations, they were only associational. They failed in their inability to tell us about the mechanism by which these factors operate. Herein, we investigate our contention that discrimination through the teacher–student relationship is one pathway to learn how race, gender, and income intersect to generate unequal academic outcomes.

5. Current Study

We draw on previous studies which associated race, gender, and income/class with academic outcomes (e.g., [17,47]), teacher-based discrimination [11,18,20,55], and intersectional framework [14,15,16]. We apply structural equation modeling to establish a cascading mediation. We hypothesized the following: (a) race, gender, and income/class have both independent and collective direct associations with discrimination through teacher–student interaction and academic outcomes (academic self-efficacy, school attendance, and performance); (b) race, gender, and income/class have indirect associations with attendance and performance through teacher-based discrimination and academic self-efficacy (see Figure 1).

6. Method

6.1. Participant

Participants were chosen from the Maryland Adolescents Development in Context Study (MADICS), a longitudinal study of secondary school students. Their primary caregivers were from an ethnically diverse urban county on the East Coast of the United States. Data were collected between 1991 and 2000, using both face-to-face survey interviews and self-administered questionnaires. The current study used data from Wave 3, which were collected in the summer following 8th grade when youth were transitioning into 9th grade. Of the 1182 students who completed the survey, 704 were selected for this study. The mean age was 14.30 (SD = 0.42). Students were indicated as either African American or White. Sixty-five percent of the respondents were African American (n = 458) and the remaining 35% White (n = 246). Forty-seven percent were male (n = 333). Most of the families’ income levels were less than $49,999 (49%), followed by income levels between $50,000 and $74,999 (33.9%), and more than $75,000 (17%).

6.2. Measures

The MADICS youth self-administered (YSA) questionnaires at Wave 3 were modified to suit the purposes of this study, and latent classes were examined. The modified YSA questionnaire had questions about participants’ demographics, self-efficacy, teacher discrimination, number of absences, and GPA. The teacher discrimination scale developed by the MADICS study team (see Eccles, Wong–Peck [56]; Wong et al., [57]) was used. The scale consisted of five items (e.g., how often students felt teachers graded them more harshly, disciplined them more harshly, called on them less, thought they were not smart, and discouraged them from taking certain courses) because of their race or gender. The items were coded on a five point-Likert scale ranging from 1 = never to 5 = every day. The items were coded as follows: (1) teachers call on you because of your race or gender (q195); (2) teachers grade you harder because of your race or gender (q196); (3) you were disciplined because of your race or gender (197); (4) teachers think you are less smart because of your race or gender (198); and (5) teachers discourage you from taking certain classes because of your race or gender (q200), with internal consistency = 0.88 in this study.

Eccles and colleagues’ [56] scale assessed academic self-efficacy belief. The scale measured students’ own general assessment of their mathematics and other academic capabilities in comparison to others their age. Responses were coded on a seven point-Likert scale ranging from not at all good to very good, and much worse than others to much better than others. The scale was replicated by using exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) in SPSS 26 and AMOS 26. In EFA and CFA, five items and five items were selected as the best measure of self-efficacy, respectively. However, the structure obtained did not support the scale structure, namely YSA Self-Efficacy exhibited by the student, as described in Eccles et al. [56]. The results of the two-factor analysis suggested five items (Chi-square = 108.018, p = 0.000; RMSEA = 0.058; CFI = 0.977). As a result, the two-factor model adopted for this study consisted of self-efficacy which included the following five items: How good are you in math? (q170). How good are you in other subjects? (q171). Compared to others, how do you do in math? (q173). Compared to others, how do you do in other subjects? (q174). How important are other school activities to you? (q178). Their internal consistency was 0.83.

Items for school attendance and GPA were used as endogenous variables and were taken from school records. GPA was measured on a five-point scale (1 = F, to 5 = A: Eccles et al., 2006). In addition, the questionnaire included demographic information about students such as race/ethnicity, gender, and parents’ income. Both race/ethnicity and gender were dichotomously coded: 0 for White and 1 for African American; 0 for female and 1 for male. Total family income was assessed by the primary caregivers’ self-reported income taken in Wave 1 question that asked, “From all sources of income, tell me your total family income before taxes”. Parents’ income was ordinally coded: 1 for less than $5000, 2 for between $5000–9999, 3 for between $10,000–14,999, 4 for between $15,000–19,999, 5 for between $20,000 and 24,999, 6 for between $25,000–29,999, 7 for between $30,000 and 34,999, 8 for between $35,000 and 39,999, 9 for between $40,000 and 44,999, 10 for between $45,000 and 49,999, 11 for between $50,000 and 54,999, 12 for between $55,000 and 59,999, 13 for between $60,000 and 64,999, 14 for between $65,000 and 69,999, 15 for between $70,000–74,999, 16 for more than $75,000. These variables were used as observed variables in the analysis.

6.3. Data Analysis

Multiple regression models used to answer the research question were limited to provide explanation of dynamic relationships from demographic information to school attendance and GPA via teacher-based discrimination and academic self-efficacy. To examine the dynamic relationship, we employed a structural equation modeling to test a model of self-efficacy and teacher discrimination on attendance and GPA. In this regard, we evaluated the unique and intersectional relations of race, gender, income, self-efficacy, and teacher discrimination with attendance and GPA. Three exogenous variables were race, gender, and income, while endogenous variables were self-efficacy, teacher discrimination, attendance, and GPA. Each of these exogenous variables was hypothesized to be jointly and significantly associated with endogenous variables.

As a preliminary step to applying the structural equation model for the hypotheses, the possible clustering effects were first examined to identify whether the variances used were distorted due to clustering effects or not; 704 students from 32 different schools might lead to cluster effects. The intraclass correlation coefficients (ICC) were, therefore, examined to identify possible clustering effects. All ICC values were too small to include any allowance for the clustering effect in this study. Second, three exogenous variables (race/ethnicity, gender, income), and four endogenous variables (self-efficacy, teacher discrimination, school attendance, GPA) were selected after conducting factor analysis.

The path diagram for the hypothesized model is depicted in Figure 1 and shows students’ characteristics, teacher discrimination, academic self-efficacy, absences, and GPA. Based on the variance–covariance matrix constructed to model the academic self-efficacy and teacher discrimination, the data were examined to test whether the hypothesized model was a good fit for the data using the maximum likelihood estimation (MLE) function in AMOS 26.

To test the model fit of the data, three model fit indices and a chi-square test result were recorded. Root mean square estimate of approximation (RMSEA), comparative fit index (CFI), and standardized root mean square of residual (SRMR) were applied using the following criteria for good fit: RMSEA < 0.05, CFI > 0.95, and TLI > 0.95. The final model, including only significant paths, is shown in Figure 2.

7. Results

Overall model fit. The hypothesized model shown in Figure 1 was fitted to the survey data. Using the MLE in AMOS 26, the model estimation terminated normally within the default convergent criterion (Chi-square = 186.105, df = 72, p = 0.000; RMSEA = 0.047, CI = (0.039, 0.056); CFI = 0.97; TLI = 0.955). These model fit values exceed the criteria recommended by Hu–Bentler [58], which confirms that the hypothesized model globally fits our data. Significant correlations between variables can be observed in the following

Table 1. Parameter estimation (standardized regression weights) for the modified model.

Path			Estimates	Standard Error	Critical Ratio	P Label
Teacher discrimination	<---	Race	0.366	0.062	5.882	***
Teacher discrimination	<---	Gender	0.155	0.059	2.628	0.009
q196	<---	Teacher discrimination	1
q195	<---	Teacher discrimination	0.948	0.047	20.024	***
q197	<---	Teacher discrimination	1.134	0.043	26.28	***
q198	<---	Teacher discrimination	0.996	0.039	25.262	***
q200	<---	Teacher discrimination	0.663	0.036	18.481	***
q178	<---	Self-efficacy	0.832	0.054	15.287	***
q170	<---	Self-efficacy	0.917	0.067	13.641	***
q174	<---	Self-efficacy	1
q173	<---	Self-efficacy	0.975	0.06	16.243	***
q171	<---	Self-efficacy	0.916	0.05	18.466	***
School Absence	<---	Income	−0.736	0.085	−8.628	***
GPA	<---	School Absence	−0.022	0.003	−8.316	***
GPA	<---	Self-efficacy	0.321	0.03	10.577	***
GPA	<---	Gender	−0.469	0.052	−8.967	***
GPA	<---	Income	0.032	0.006	5.057	***
GPA	<---	Teacher discrimination	−0.137	0.035	−3.922	***
GPA	<---	Race	−0.369	0.048	−7.618	***

Note: *** p < 0.001.

, where the beta estimates with their standard errors are summarized.

The estimates for the modified model are presented in Figure 2. All estimates for the model shown in Figure 2 were significant at the p < 0.05 level.

Race, gender, and income had significant direct associations with performance in GPA simultaneously, but the direction of the associations differed. Significant negative direct associations were observed simultaneously among race, gender, and the GPA and indirectly through discrimination in the context of teacher–student relationship: those students who were either African American or male were more likely to have lower GPA scores. While no direct association was observed among race, gender, income, and self-efficacy, a significant negative association was observed between discrimination and self-efficacy when observed through teacher–student interaction. Moreover, self-efficacy was significantly associated with academic performance.

Following Hair and colleagues’ study [29], the mediation test results show that the mediation effect is present when the total effects of the exogenous variable on endogenous are greater than the absolute value of 0.8, partial mediation between 0.2 and 0.8, and no mediation below 0.2, respectively (see

Table 2. Testing the mediation effect.

Dependent Variable	Regression	Independent Variable	Mediator	Total Effect	Mediation Effect
GPA	<---	Race	Teacher discrimination	−0.22 + (0.1 * − 0.13) = −0.24	Partial Mediation
GPA	<---	Gender	Teacher discrimination	−0.27 (0.23 * − 0.13) = −0.30	Partial Mediation
GPA	<---	Income	Absences	0.16 + (−0.32 * − 0.25) = 0.24	Partial Mediation

Notes: * p < 0.05.

). Overall values show discrimination through teacher–student interaction partially mediates the effect of race and gender on GPA, respectively. Similarly, absences/attendance showed a partial mediation effect between GPA and income/class.

8. Discussion

The study used an intersectionality framework to answer the hypotheses: (1) race, gender, and income/class have both independent and collective direct associations with discrimination through teacher–student relationship and academic outcomes (academic self-efficacy, attendance, and performance in GPA), and (2) race, gender, and income/class have indirect associations with attendance and performance through discrimination within teacher–student interactions and academic self-efficacy. The role of discrimination within the teacher–student interactions and academic self-efficacy as mediators in relation to students’ race, gender, income and GPA was the primary foci. While the hypothesized model confirmed previous works that established associations among race, gender, income, and academic outcomes, it also showed discrimination activated through the teacher–student relationship as a mechanism by which structural forces such as racism and sexism simultaneously and intersectionally operate to affect student’s academic performance. Furthermore, income/class seems embedded in race and gender and operates to directly affect students’ school attendance, indirectly affecting academic performance.

This study confirms teacher-based discrimination negatively affects African American youth academically [11,57]. In the current study, African American males were more likely to experience discrimination in their interaction with teachers, which tended to affect their academic performance. As seen in Table 1, African American males were associated with lower academic performance, suggesting that race and gender intersect to pattern young Black males’ academic outcomes differently [59,60]. On the surface, it may seem to suggest that Black males have a lower academic performance at the individual level. However, a deeper examination, as displayed in Figure 2, helps to provide some context. The findings suggest racialized experiences that African American male students encounter through their interaction with teachers is one pathway to understand Black males’ lower performance outcomes observed here.

Contrary to previous studies which found associations among race, gender, income/class and self-efficacy [17,47], we did not find such relationships. However, consistent with these studies, we found that academic self-efficacy was positively associated with academic performance. One explanation for the lack of association between race, gender, and income/class and academic self-efficacy may be the cross-sectional nature of our analysis. These findings confirm the contention by Marks and Garcia Coll [7] integrated model of the developmental competences for the development of minority children: that students’ academic experiences and academic outcomes are linked to their social position variables, and that these social stratification variables, although distinct, operate intersectionally to influence students’ overall academic experiences and outcomes [16].

9. Implications

These findings highlight three major implications. African American students especially males, regardless of income/class, may benefit from more positive and race-conscious interactions with teachers. By embedding teacher-based discrimination awareness and positive race relations in professional development, teachers may build more positive intentional connections to support African American students broadly and males in particular. Building teachers’ racial competencies regarding the reality that they are agents of racialized institutions that are designed to systematically sustain the racial contract upholding the white racial worldview [29,61,62] will strengthen teachers’ positive relationships with racialized students. Because schools are a subsystem of the racialized system, teachers exist within and mediate racialized structures, interpret these structures, and create meanings that direct behaviors [63,64]. Therefore, when teachers are cast as the source of racial discrimination instead of as mediators in the racialization process, it masks and distorts the structures of white supremacy behind educational policies and practices that teachers are expected to execute in the post-Brown vs. Board of Education era. A race-conscious framework can provide teachers with the analytic framework to understand the role of the racialized system to recruit them as its agent. That awareness is important for teachers to engage in practices of transgression against structural racism that empower students through engagement practices that recognize and leverage the cultural and community assets that students of color and those from working-class background bring to the learning context [65].

The second implication is the potential impact of income/class on students’ overall academic performance. The number of school absences, in the model, could be much detrimental to students in lower-income levels, regardless of race. Attention to income/class is critical to absenteeism policies and truancy interventions. In this regard, given that race and income are strongly linked [2,66], applying an intersectional framework to examine their combined effect on potential discrimination in the context of teacher–student interaction is important to support African American students. In this case, student absenteeism that on the surface may indicate an individual behavior, may in fact signal a form of communication related to the impact of structural stereotypical narratives around race, gender, and working-class status that may be influencing their interaction with teachers. Instead of directing attention to family or the individual child, it would be helpful for educators to examine their institutional practices and culture and how such institutional practices may be orienting teachers to recreate and perpetuate such stereotypical meanings directly or indirectly toward students. Additionally, practicing listening sessions with students who might be disengaged in order to learn from them about how the activation of certain institutional policies and practices by teachers harm the teacher–student relationship is an important way for teachers to ensure that they leverage their role as institutional agents to empower students, and not to contribute to their disengagement.

The third implication is the inverse relationship between teacher-based discrimination and academic self-efficacy, and the association between self-efficacy and academic performance. While academic self-efficacy is noted as an important pathway for academic performance among African American youth [67], teachers are critical to the development of academic self-efficacy [50]. These findings also imply that practices by other school professionals and agents of youth-serving institutions [19] that bolster higher academic self-efficacy are crucial to academic performance. Policies and practices that deactivate discrimination within the teacher-student interactional context are likely to support the development of strong academic self-efficacy to positively affect academic performance. That is, policies and practices that recognize students as active co-constructors in the designing of learning context and development and delivery of pedagogical practices can reduce the potential for discrimination. Additionally, educational practices that decenter hierarchical thinking by recognizing and leveraging the distinct cultural and navigational capital [68] that various youth bring to the learning context can strengthen student–teacher relationship and reduce the potential for discrimination.

10. Limitations

Some limitations need to be acknowledged. First, the analyses used cross-sectional data which were collected in 1993. Therefore, the findings should be interpreted within the context of these data. We expected race to be directly associated with academic self-efficacy, but it was not. While we speculated that this might have been due to the cross-sectional nature of the data, racial–ethnic identity content and processes may show such association. However, in the current study, we did not examine racial–ethnic identity content and its process. Future research that uses longitudinal data and examines the association among racial–ethnic identity content and process, teacher-based discrimination, academic self-efficacy, and academic performance will make an important contribution to the field. Additionally, the data used for the current analysis were collected in a single county in the northeastern part of the United States. The generalization of the findings, therefore, should be with caution.

11. Conclusions

The current study shows that schools are actors in the production and perpetuation of structural inequities, and that the teacher–student interactions may be one mechanism to understand and interrupt them. Operating from these racially constructed narratives that intersect with gender, and class, educators may have lower expectations for African American students. Additionally, educators may discipline and grade African American students more harshly, track them to lower-ability classes, and regard them as criminals [1,18,69,70]. Although these discriminatory experiences are likely to be experienced by both boys and girls [71], our findings, which are consistent with prior work [59,72] suggest that from an intersectional lens, these experiences can be more detrimental for African American males, as well as those from lower income backgrounds relative to academic experiences and performance. Furthermore, as shown in Figure 2, income/class had an independent association with student attendance and an indirect association with academic performance through attendance or absenteeism. Students from higher-income families were less likely to have attendance problems, which in turn affected their performance. From an intersectional framework, income/class additively affected academic performance—indirectly through attendance—while simultaneously intersecting with race, priming students—African American males in particular— for teacher-based discrimination, resulting in unequal academic performance as compared to whites. That means in a context where income/class may be salient, it is a critical factor to understand the differences in students’ school attendance concerns in relation to academic performance. While the extant literature has conceptualized the discriminatory practices that occur within teacher-student engagement as teacher-based discrimination, from a structural perspective, it seems that teachers may be enacting and perpetuating the expectations of the institutions/organizations. Future research should investigate the extent to which teachers are mediating structural forces that shape students’ differential access to educational resources. Such work will help us to better understand the root causes of structural inequalities and the mechanisms by which they are embodied in students’ educational experiences and outcomes to develop effective response strategies.