The Implications of Intersectionality on Southeast Asian Female Students’ Educational Outcomes in the United States: A Critical Quantitative Intersectionality Analysis

Significance of This Study

This study specifically challenged the model minority stereotype related to Asian populations and differentiated the school experiences and educational outcomes of Southeast Asian students in the United States. The model minority stereotype reveals a larger systemic failure to consider the specificities within the Asian population, which limits the provision of adequate support for Southeast Asian females to realize their full potential through their future academic careers. By focusing on Southeast Asian female students’ experiences in school, this study challenged the model minority myth and identified the heterogeneity among Asian and Pacific Islander students in the United States. Furthermore, this study provided more visibility to typically understudied groups (i.e., Southeast Asian females) in education policy. This study revealed critical findings to help policymakers design better support for Southeast Asian female students so that these students do not feel like “strangers in a strange land” (Bell, 1970, p. 540). Finally, using a U.S. national sample to identity the intersectionality of Southeast Asian female student identities in education, this study addressed gaps in the literature on intersectionality, which is typically focused on Black and Hispanic students. Thus, this study broadened the research perspective for students living in the margins of margins (i.e., Southeast Asian females). It also provided policy and leadership implications by answering critical questions related to the complexity of students’ experiences, particularly for Southeast Asian female students in the United States. Ultimately, it aims to contribute to our understanding of efforts aimed at improving social justice—a morally crucial goal.

The next section specifically discusses the implementation of the theoretical and methodological framework of critical quantitative intersectionality as well as the core premises of intersectionality. It also reviews literature focusing on Southeast Asian students’ experiences and outcomes.

Critical Quantitative Intersectionality

Studies conducted from a critical perspective often employ a qualitative approach due to its strengths in fully describing the nuances of lived experiences among marginalized people based on unique historical, cultural, and organizational contexts. In particular, critical researchers using qualitative approaches can thoroughly explore the meaning of students’ multiple and less visible social constructs (e.g., sexual orientation) by actually talking with students about their identities through diverse qualitative research methods. For example, Kozol (2012) described the lives of poor female and racial or ethnic minority youths based on ethnographic research. He provided a rich, detailed illustration of struggles in the lives of people with multiple marginalized social categorizations (e.g., female, poor, Black, or Hispanic) through striking narratives. In particular, Kozol observed “how human beings devalue other people’s lives, how numbness and destructiveness are universalized, and how human pity is at length extinguished and the shunning of the vulnerable can come in time to be perceived as natural behavior” (p. 186). As such, critical studies using qualitative methodologies are well-suited to reveal the processes of amplification in privileging or oppressing specific populations based on different social categorizations.

Even admitting the relative predominance of qualitative methodologies in critical approaches, critical quantitative researchers argue that different methods can answer diverse critical questions and have a powerful impact in educational policy to realize social justice (e.g., Covarrubias et al., 2018; Gillborn et al., 2018; López, Erwin, Binder, & Chavez, 2018; Stage, 2007; Teranishi, 2007). This is because policymakers typically have a strong desire to rely on numbers from research when designing educational policies (Covarrubias & Vélez, 2013).

Much of the work conducted by critical scholars using quantitative methods is based on endeavors to apply quantitative methods guided by critical race theory (CRT) that extensively focus on the impact of race and racism. For example, Covarrubias and Vélez (2013) used the term critical race quantitative intersectionality (CRQI), which they defined as

an explanatory framework and methodological approach that utilizes quantitative methods to account for the material impact of race and racism at its intersection with other forms of subordination and works toward identifying and challenging oppression at this intersection in hopes of achieving social justice for students of color, their families, and their communities. (p. 276)

Other scholars have similarly utilized different terms, such as QuantCrit (Gillborn et al., 2018), QuanCrit (López et al., 2018), and CritQuant (Sullivan, Larke, & Webb-Hasan, 2010), to fuse CRT tenets into quantitative methods for the goals of social justice and racial and ethnic equity.

The framework of critical quantitative intersectionality (CQI) used in this study is based on the similar positionality of the previously mentioned critical frameworks utilizing quantitative methods (i.e., CRQI, QuantCrit, and QuanCrit). In particular, this study shares four principles of QuantCrit proposed by Gillborn and his colleagues (2018): (a) numbers are not neutral; (b) categories are neither “natural” nor given; (c) voice and insight: data cannot “speak for itself”; and (d) using numbers for social justice. In particular, being “critical” in the framework of critical quantitative intersectionality is based on these principles. Challenging the neutrality of quantitative data is useful for criticizing the ways in which privileged groups of people protect their power (e.g., the model minority stereotype of Southeast Asian students). In addition, as noted, the social categorizations (i.e., race, ethnicity, gender, and SES) used in this study are neither preexisting fixed categories nor material things, but social constructs that historically oppress and separate particular groups. Thus, the interpretation of these social categorizations should not use causation (e.g., deficit-informed approach) (Zuberi, 2001), but rather the association between the social classification and educational outcomes, which reveals critical inequality and the operation of embedded oppressions. Identifying diverse social and educational inequalities associated with Southeast Asian girls’ different multiple social categorizations based on the use of numbers will be important for realizing social justice and educational equity.

Although CQI shares the four main principles of the previously identified critical quantitative frameworks, the main difference compared to previous critical approaches using quantitative methods is the centrality of multiple intersecting oppressions. In other words, CRQI or QuantCrit places mainly race and racism at the core of the frameworks, as CRT is a guiding theoretical approach; indeed, CRT scholars have emphasized that subordination and institutionalized racism are intrinsic to most American institutions (Bell, 1995, 2008; Solórzano & Solórzano, 1995; Yosso & Solórzano, 2005). This critical perspective in CRT using the concepts of power and oppression challenges ideas deeply rooted and “enmeshed in the fabric of [American] social order” (Ladson-Billings, 1999, p. 212) and provides important knowledge for eliminating the oppression of students of color as well as transforming unjust educational systems in the United States. Although CRT also explores the effects of multiple oppressions on the lived experiences of people of color, the centrality of racism in the theory often overlooks the core premises of intersectionality (e.g., multiplicity, simultaneity, power relations, and social contexts). Thus, CRT argues that diverse oppressions are attached to different social categorizations in the experiences of people of color, but cannot explain how multiple oppressions operate in creating unique experiences of people with multiple marginalized social categorizations. To complement this limitation of CRT and previous critical race perspectives using quantitative methods (i.e., CRQI, QuantCrit), the CQI framework places core premises of intersectionality theory at its heart. A combination of the intersectionality theory and the critical quantitative approach will be useful for understanding the complexity of Southeast Asian female students’ educational experiences and providing policy implications for this particular student group. The next section specifically discusses the core premises of intersectionality that can shed light on Southeast Asian female students’ experiences in particular, which differentiates CQI from previous quantitative approaches.

Intersectionality

The foundational concepts of intersectionality have their origins in “U.S. third world feminism” (Sandoval, 1991, p. 1), which seeks to differentiate experiences of women of color from those of White women (e.g., Black feminism). In particular, U.S. third-world feminists challenge the converging systems of oppressions (e.g., racism, sexism, colonialism, capitalism) that women of color experience. Sandoval (1991) identified these forms of consciousness in multiple oppressions in U.S. third-world feminism using the term “differential consciousness,” which represents “the strategy of another form of oppositional ideology that functions on an altogether different register” (p. 2). For example, Black feminist scholars emphasize that lives of women of color (mainly Black women) are distinct from those of White women or Black men due to inextricably bound pressures from sexism and racism (e.g., Collins, 1998; Crenshaw, 1991; Dill, 1983; hooks, 1981, 1992, 2000). Collins (1998), as a Black feminist scholar, articulated the experiences of Black women through the lens of violence, which ties oppression to race or ethnicity, gender, and social class. Since the 1960s, this different consciousness of U.S. third-world feminist intellectuals has contributed to the amplification of the essential ideas of intersectionality, which have also appeared in critical texts, such as bell hooks’ Ain’t I a Woman: Black Women and Feminism, Toni Cade Bambara’s The Black Woman, and the Combahee River Collective’s A Black Feminist Statement. Within the stream of U.S. third-world feminist thoughts, Kimberlé Crenshaw (1991) created the umbrella term intersectionality to build a coalition among the study areas of race or ethnicity, SES, and gender as an interdisciplinary endeavor and to include diverse social categorizations and contexts beyond Black women (Dhamoon, 2010).

Southeast Asian Students

The schooling experiences and academic achievements of Southeast Asian students, such as Vietnamese, Hmong, Cambodian, and Lao students, are often masked in academic research because the data for Southeast Asian students are frequently aggregated in the general Asian population (Covarrubias & Liou, 2014; Lowe, 1996; Ngo & Lee, 2007; Teranishi, 2007). Scholars focusing on Southeast Asian students have demonstrated that schooling experiences and educational outcomes of Southeast Asian students are significantly different from those of Black, Hispanic, and overall Asian students (Ngo, 2013; Ngo & Lee, 2007). Scholars have explained racial inequities and uniqueness in schooling experiences and educational outcomes of Southeast Asian students mainly through cultural characteristics from qualitative studies. For example, Timm, Chiang, and Finn (1998) found that Hmong students in American schools typically value kinship and cooperation over individualism. Timm et al. argued that the cultural dissonance between Southeast Asian students’ home cultural characteristics (i.e., preferring to work with others and having external guidance as well as focusing on social cues) and school culture (e.g., independent learning) often creates disadvantages for Hmong students. Furthermore, focusing on Cambodian students, Rumbaut (1989) claimed that Theravada Buddhism, leading to “an adaptive style that is more passive and reactive, comparatively less pragmatic and more fatalistic” (p. 181), may cause inequalities in the educational outcomes of Cambodian students.

Scholars have found that Southeast Asian students’ schooling experiences and educational outcomes are closely related to the social construct of gender. In particular, studies have demonstrated that patriarchal norms devaluing females as well as early marriage and childbearing patterns are salient factors affecting the distinct experiences and outcomes of Southeast Asian students. For example, Goldberg (1999) found that Cambodian girls face significant cultural pressure from their family members to marry and have children, which leads to higher dropout rates. In addition, researchers have demonstrated that the cultural expectation for Hmong girls is that they have to take care of their younger siblings and do other household tasks (S. J. Lee, 2001). These unique cultural expectations about gender roles might negatively affect their schooling experiences and educational outcomes (Ngo & Lee, 2007). As these studies demonstrate, the gender categorization of Southeast Asian students is inextricably linked to racial or ethnic categorization, thereby creating a unique particularity in their experiences and educational outcomes.

Studies have also identified structural barriers and racism experienced by individual students in schools as critical factors influencing Southeast Asian students’ schooling experiences and educational outcomes. Studies focusing on the structural barriers in schools experienced by Southeast Asian students have addressed the issues of teachers’ lack of knowledge about students, tracking into lower level courses, and teachers’ low expectations (DeVoe, 1996; Ngo & Lee, 2007). Furthermore, Kiang and Kaplan (1994) reported that all the Vietnamese students they interviewed indicated that they have experienced harassment and discrimination; one surprising example of such discrimination was the limited use of spaces in school classrooms, hallways, and cafeterias for Vietnamese students due to acts of exclusion. These qualitative studies show that Southeast Asian students also experience structural barriers and individualized racism. Yet more quantitative studies are needed to explore and provide rigorous and objective evidence of Southeast Asian students’ structural barriers in schools.

Research Gaps

Although the intersectionality of multiple marginalized social categorizations, primarily as an analytical framework or tool, has been utilized in education research, insufficient attention has been focused on the intersectionality of Southeast Asian female students’ race or ethnicity, gender, and SES affecting their educational outcomes in schools. For example, Dantley et al. (2008) pointed out that researchers tend not to adopt a broader intersectionality approach; researchers instead are described as “a feminist who mainly addresses gender issues or an African American who primarily does research on racism” (p. 125). Indeed, specific research gaps in critical quantitative intersectionality in education have been identified.

First, critical quantitative researchers consistently criticize essentializing the lived experiences of the Asian population and emphasize the importance of ethnic categorization in differentiating Southeast Asian students’ experiences (Covarrubias & Liou, 2014; Teranishi, 2007). Nevertheless, we still have limited knowledge about Southeast Asian female students from quantitative studies due to the data limitation of missing ethnic background variables in large administrated datasets (e.g., Covarrubias & Liou, 2014). Furthermore, studies disaggregating Asian students by ethnicity, class, or immigration status are often based on only simple descriptive statistics (e.g., Teranishi, 2007). Although simple descriptive statistics have shown the importance of considering multiple social categorizations when exploring Southeast Asian female students (e.g., Teranishi, 2007), more diverse quantitative explorations are needed to improve the educational outcomes for this particular student population. For example, quantitative intersectionality studies often omit different levels of analysis (e.g., organization, nation) to examine how schooling experiences and outcomes of students with multiple social categorizations intersect with organizational factors or contextual factors in society. The use of intersectionality focusing only on individual social categorizations undermines the use of intersectionality to provide practical knowledge for educational leadership to alleviate educational inequalities in schools.

Second, previous frameworks of quantitative research using a critical lens (e.g., CRQI or QuantCrit) are typically limited in exploring how systems of multiple discrimination and structural oppression produce differences for Southeast Asian women. As noted, this is because they mainly place race and racism at the core of the frameworks rather than multiple oppressions and marginalizations simultaneously. In this study, the critical quantitative intersectionality analysis instead aims to identify patterns of structural inequality in Southeast Asian female students’ educational outcomes, which is the product of multiple oppressions simultaneously. In particular, McCall’s (2005) intercategorical (across groups) and intracategorical (within groups) approach may be applicable to challenge assumptions of universal experiences and unpack complexity in educational outcomes of Southeast Asian female students within and across groups.

This study based on the critical quantitative intersectionality framework proposes two main research questions focusing on educational outcomes tied to the multifaceted social constructs (i.e., race or ethnicity, gender, and SES) of Southeast Asian high school girls in the United States:

Data Source

This study uses restricted-use national longitudinal data provided by the National Center for Education Statistics (NCES) High School Longitudinal Studies 2009 (HSLS:09). As the most recent national-level longitudinal study, HSLS:09 has tracked sample populations of high school students from the beginning of high school into postsecondary education, the workforce, and beyond, especially emphasizing STEM education. The HSLS:09 dataset includes a nationally representative sample gathered from more than 23,000 ninth-grade students in 944 schools since 2009. An average of 25 ninth-grade students per school were randomly selected from sampled schools. In addition, a stratified, two-stage random sample design was used to acquire the sample schools, including both public and private schools. The stratified two-stage random sample design of HSLS:09, which is often called complex sampling design (Lumley, 2011; Rabe-Hesketh & Skrondal, 2006), typically creates two critical issues for analysis: (a) nonindependence among units because of the nonsimple random sampling design and (b) disproportionate sampling resulting in unequal selection probabilities (Hahs-Vaughn, McWayne, Bulotsky-Shearer, Wen, & Faria, 2011). In order to address these complex sampling issues, this study employs balanced repeated replication (BRR) methods to ensure that the results are representative of the population and calculation of the variance and parameter estimates (see Hahs-Vaughn et al., 2011, for specific details and explanation about replication methods).

The HSLS:09 dataset offers specific advantages for this study, including rich contextual information about students’ ethnic backgrounds as well as organizational information from the school level in the restricted-use data. Furthermore, the HSLS:09 consists of five sub-data sets based on different data collection waves: base year (2009), first follow-up (2012), 2013 update (2013), high school transcripts (2013–2014), and second follow-up (2016). Of these data sets, this study uses the base-year data collected in the fall term in 2009 and focuses on ninth graders’ educational outcomes as a powerful predictor for their future educational outcomes, including graduation (Allensworth & Easton, 2005; McCallumore & Sparapani, 2010), 11th-grade achievement (Easton, Johnson, & Sartain, 2017), and college enrollment (Easton et al., 2017).

Social Categorizations

This study focuses on intersections of three different aspects of students’ social categorizations: race or ethnicity, gender, and SES. In terms of race or ethnicity, five dummy variables (i.e., Black, Hispanic, American Indian/Alaska Native, Southeast Asian, and Asian students who are not Southeast Asian students) were used separately to compare each student of a different race or ethnicity to White students (reference group). In addition, this study uses the Asian origin variable (S1ASIANOR) for specific ethnicity among the Asian group, which is only available from the restricted dataset, to create the categories Southeast Asian and Asians who are not Southeast Asian. Using White students as the reference is not a result of “privileging” them or maintaining the patterns of power related to racial categorizations; rather, this approach aims to reveal patterns of racial or ethnic educational inequality compared to the privileged racial group, which in turn adds to our critical understanding for educational equity. Although this study specifically focuses on Southeast Asian female students, it also includes other minoritized student groups in the analyses in order to identify the uniqueness of their experiences compared to those of all race or ethnicity groups. Furthermore, quantitative intersectionality scholars typically include all racial and ethnic minority students in analyses to identify structural inequality based on racial categorizations (e.g., Covarrubias, 2011; Irizarry, 2015). In terms of gender, the female variable is used by setting male students as the reference group based on data from the student questionnaire. The school-provided sampling roster or the parent questionnaire supplemented the data if this information was missing. Similar to the racial categorization, this study coded the dominant gender group in STEM (i.e., males) as a reference only for the purpose of coding and the quantitative analysis. Finally, this study used the SES index score that was calculated by NCES in the HSLS:09 dataset. The SES index score indicates a student’s family position in a vertical social hierarchy. A lower SES index score indicates a relatively lower socioeconomic status for a student’s family and vice versa. Quantitative analyses utilizing these three social categorizations in this study should not be interpreted as causality (e.g., race effect), but as statements of associations between the social categorizations and outcome variables. Furthermore, these associations related to multiple social categorizations represent oppressive forces (e.g., racism, sexism, capitalism, colonialism) (Zuberi, 2008).

School Characteristics

This research explores how the patterns of the convergence of students’ multiple identities on their experiences and characteristics differ according to each school’s climate, region, and demographic composition (i.e., percentages of free or reduced lunch eligible [FRL] students and students of color). These data were obtained from the HSLS:09 school-level dataset.

Dependent Variables: Educational Outcomes

This study uses two dependent variables to measure educational outcomes: standardized mathematics achievement scores and intention to enter higher education. First, this study uses ninth graders’ math standardized theta scores for math achievement scores. The math achievement score (X1TXMTSCOR) is a continuous variable and a ratio measurement using a 100-point test. The math achievement score was used as the only curriculum-related variable because previous studies have shown the association between high school math achievement and students’ future academic success, such as college success (e.g., Claesens & Engel, 2013; J. Lee, 2012). Another issue is the model minority myth suggesting that all Asians are good at math. Thus, this study challenges the model minority myth and stereotype in STEM subjects by exploring whether these assumptions and myths apply to Southeast Asian females’ math achievement scores in the same way across racial and ethnic groups and within the Southeast Asian group. This study acknowledges that achievement scores of other subjects can also reveal a part of the significant educational inequality that Southeast Asian female students experience. For example, reading or writing scores may be critical for illustrating educational inequalities because language proficiency might be critical for academic achievement in these subjects if students come from immigrant families whose first language is not English. However, due to the data available in the HSLS:09 dataset, this study focused only on math achievement scores as a curriculum-related dependent variable.

Second, students’ intentions to pursue postsecondary education are also obtained from the questionnaire (X1STUEDEXPCT) and are taken from the question asking how far in school ninth graders think they will get. Dichotomous coding is used for the 10-point categorical answers: high school diploma/General Educational Development (GED) or less than high school served as the reference group compared to all categories at or above a bachelor’s degree.

Research Models

This study used three statistical research techniques to answer the proposed research questions: multiple regression, logistic regression, and linear mixed effect modeling (LMM). In particular, this study used multiple regression and logistic regression in order to answer Research Question 1, examining how the convergence (i.e., intersectionality) of race or ethnicity, gender, and SES is associated with the educational outcomes of Southeast Asian female students. In addition, LMM was employed to answer Research Question 2, which explored the associations between the intersectionality of race or ethnicity, gender, SES and educational outcomes based on schooling context.

Research Question 1

Research Question 1 examines the patterns of educational outcomes in relation to different convergences of race or ethnicity, gender, and SES. In order to identify patterns of educational inequality in educational outcomes associated with multiple intersecting social categorizations, this study compared base models without interaction terms (i.e., intersectionality) to full models with interaction terms.

Intersectionality and Students’ Math Achievement

Table 1 presents the results of the multiple regression analyses of ninth-grade students’ mathematics achievement scores. Comparing overall model fits between Model 1 without interaction terms and Model 2 with interaction terms found that the values of deviance, AIC, and BIC are smaller in Model 2. This result led to the conclusion that Model 2 provides a better fit for explaining the relationship between math achievement scores and students’ multiple social categorizations.

OPEN IN VIEWER

Table 1

Results of Fitting Regression Models Predicting Mathematics Achievement

	Model 1	Model 2 a
Socioeconomic status (SES)	5.26*** (0.13)	5.86*** (0.23)
Female	0.37 (0.21)	0.36 (0.24)
Race or ethnicity
American Indian/Alaska Native	−6.39*** (1.55)	−5.46*** (1.42)
Black	−4.21*** (0.41)	−5.02*** (0.56)
Hispanic	−0.48 (0.28)	−0.63 (0.41)
Other Asian/Pacific Islander	3.69*** (0.59)	3.64*** (0.96)
Southeast Asian	3.30*** (0.53)	3.25*** (0.70)
Intersectionality
SES × Female		−0.63* (0.29)
SES × Black		−1.33** (0.52)
SES × Hispanic		−1.18* (0.56)
(Intercept)	50.81*** (0.19)	50.77*** (0.22)
Deviance (–2Loglikelihood)	925,839.1	144,337.8
AIC	925,921.1	144,387.8
BIC	926,246.3	144,585.5

Note. For all models, the reference group for race or ethnicity is White student, and the reference group for gender is male. Standard errors are in parentheses.

a. The results of Model 2 report only significant interaction terms among all 16 interaction terms.

*p < .05. **p < .01. ***p < .001.

Math achievement scores of Southeast Asian students were significantly higher than those of White students (β = 3.25, p < .001) and other race or ethnicity groups, except Other Asian/Pacific Islanders, regardless of gender. Furthermore, the association between SES and Southeast Asian students’ math achievement scores (β = 5.86, p < .001) was not statistically different from those of White, American Indian/Native American, and Other Asian/Pacific Islander students. However, the statistically significant interaction between SES and female (β = −0.63, p < .05) indicates that gender categorization moderates the impact of SES on Southeast Asian female students’ mathematics achievement compared to their male counterparts. In particular, the impact of SES on Southeast Asian female students’ mathematics achievement is smaller than the average impact of SES on Southeast Asian male students’ mathematics achievement. This gender difference in the association between SES and math achievement scores among Southeast Asian students, which indicates that SES matters less for females than for male counterparts, hold true for other race or ethnic groups. However, SES matters less for the achievement of Black (β = −1.33, p < .01) and Hispanic female students (β = −1.18, p < .05) than it does for Southeast Asian female students. This finding indicates Southeast Asian female students’ higher inequality across SES compared to their Black and Hispanic counterparts.

Intersectionality and Students’ Intention to Enter Higher Education

Table 2 reports results of logistic regression that uses a binary dependent variable of students’ intention to enter higher education predicted by students’ multiple social categorizations. A seven-predictor main-effect logistic model (Model 3) was fitted to the data to examine the relationship between the likelihood that a student in ninth grade intends to enter higher education and his or her race or ethnicity, gender, and SES. By extension, Model 4 included 16 interaction terms to test the research hypothesis that the intersectionality of students’ three social categorizations is related to students’ intention to enter higher education. Comparing overall model fits between Model 3 and Model 4 based on the AIC, BIC, and deviance concluded that Model 4 provides a better fit to explain the relationship between a student’s intention to enter higher education and multiple social categorizations, including the different intersectionalities.

OPEN IN VIEWER

Table 2

Logistic Regression Analysis of Students’ Intention to Enter Higher Education Predicted by Their Multiple Social Categorizations

	Model 3		Model 4 a
Predictor	β	e β (odds ratio)	β	e β (odds ratio)
Socioeconomic status (SES)	0.73*** (0.03)	2.08	0.85*** (0.05)	2.34
Female	0.28*** (0.04)	1.32	0.34*** (0.05)	1.41
Race or ethnicity
American Indian/Alaska Native	−0.63* (0.25)	0.53	−0.62* (0.27)	0.54
Black	0.25** (0.10)	1.29	0.27** (0.11)	1.31
Hispanic	−0.07(0.07)	0.93	−0.16(0.11)	0.85
Other Asian/Pacific Islander	0.14(0.10)	1.15	−0.01(0.16)	0.99
Southeast Asian	−0.18(0.15)	0.84	0.22(0.15)	1.24
Intersectionality
Southeast Asian × Female			−0.80*** (0.25)	0.45
Hispanic × SES			−0.44** (0.17)	0.65
Hispanic × SES × Female			0.55** (0.19)	1.74
(Intercept)	−0.19*** (0.03)	NA	−0.15*** (0.04)	NA
Deviance (–2Loglikelihood)	25,940.0		25,858.3
AIC	25,996.0		25,906.3
BIC	26,019.2		26,006.1

Note. For all models, the reference group for race or ethnicity is White student, and the reference group for gender is male. Standard errors are in parentheses.

a. The results of Model 4 report only significant interaction terms among all 16 interaction terms.

*p < .05. **p < .01. ***p < .001.

In Model 4, the statistically significant coefficient of the interaction between Southeast Asian and Female (β = −0.80, p < .001) is the difference between the log-odds ratio comparing Southeast Asians to non–Southeast Asian males and the log-odds ratio comparing Southeast Asians to non–Southeast Asian females. In particular, the exponentiated regression coefficient of the interaction between Southeast Asian and Female, exp(–0.80) = 0.45, indicates that the odds of intending to enter higher education for a Southeast Asian female student are 0.45 times the odds for the other three groups (Southeast Asian males, males who are not Southeast Asian, and females who are not Southeast Asian).

Due to the complexity involved in the interpretation of the odds ratio including multiple interactions, researchers often use predicted probabilities of the dependent variables based on the employed statistical model (e.g., Peng, Lee, & Ingersoll, 2002). Based on the results of Model 4, Table 3 shows the logits, exponentiations, and probabilities of intending to enter higher education of different student groups when a student’s SES score is zero, for the purpose of illustration. In Table 3, the odds of intending to enter higher education for a Southeast Asian female student is 0.544 times the odds for the other three groups (Southeast Asian males, males who are not Southeast Asian, and females who are not Southeast Asian). Furthermore, the probability that Southeast Asian females with average SES intend to enter higher education institutions is low at probability = .353; thus, the probability of not intending to enter higher education is .647. This intentionality for higher education among Southeast Asian female students is remarkably lower than that of Southeast Asian males (probability = .462). Furthermore, this intention of Southeast Asian female students was also the lowest among all female students.

OPEN IN VIEWER

Table 3

Logits, Exponentiations, and Probabilities of Intending to Enter Higher Education of Different Student Groups

Gender	Race or Ethnicity	SES	Logit	Exp(Logit)	Probability
Male	White, Other Asian/Pacific Islander, Southeast Asian	0	−0.151	0.860	.462
Female	White, Other Asian/Pacific Islander	0	0.191	1.210	.548
Male	Black	0	0.117	1.124	.529
Female	Black	0	0.459	1.582	.613
Male	Hispanic	0	−0.151	0.860	.462
Female	Hispanic	0	0.191	1.210	.548
Male	American Indian/Alaska Native	0	−0.767	0.464	.317
Female	American Indian/Alaska Native	0	−0.425	0.654	.395
Female	Southeast Asian	0	−0.608	0.544	.353

Using the same method of calculating probabilities for intending to enter higher education of different student groups, Figure 1 shows the differences among student groups based on five different SES index scores (+2SD, +1SD, average, –1SD, –2SD). The left panel of Figure 1 focuses on expected probabilities for males. Whites, Other Asian/Pacific Islanders, and Southeast Asian males show the same expected probabilities of intending to enter higher education. The right panel of Figure 1 focuses on expected probabilities for females. In this figure, Whites and Other Asian/Pacific Islander females show the same expected probabilities of intending to enter higher education.

OPEN IN VIEWER

Figure 1.

Top panel: Predicted probability of intending to enter higher education across SES index scores among males. Bottom panel: Predicted probability of intending to enter higher education across SES index scores among females and Southeast Asian males.

From Figure 1, on average, the expected probability of the intention to pursue higher education of Southeast Asian female students was the lowest among females and lower than Southeast Asian males across all SES index scores. Compared to this finding of Southeast Asian students, the expected probability of a Black student intending to enter higher education regardless of SES index scores is higher than that of Southeast Asian students as well as other races or ethnicities students for both males (left panel of Figure 1) and females (right panel of Figure 1).

Focusing on the association between SES and students’ intention to enter higher education, the higher SES index scores show the greater expected probability to pursue higher education regardless of race or ethnicity and gender overall. On the one hand, the exponentiated regression coefficient of SES for Southeast Asian students, which was not statistically different from that of other race or ethnicity groups, exp(0.85) = 2.34, p < .001, creates parallel lines across race or ethnicity groups according to different SES. This result indicates that the relative positions of students’ intention to enter higher education based on different racial and ethnic groups are retained across SES. On the other hand, the significant interaction terms related to Hispanic students creates unique patterns in the intention to enter higher education of Hispanic students across different SES index scores. In particular, among male students, the expected probability for Hispanic students’ intention to enter higher education was highest among students with the lowest SES scores (i.e., SES index score = 2SD below average). Because of the positive interaction term of Hispanic, Female, and SES, the expected probabilities for Hispanic females’ intention to enter higher education were remarkably different from the lowest SES (2SD below average) to the highest SES (2SD above average) compared to other race or ethnic groups. In particular, the intention of Hispanic females to attend higher education institutions lagged behind that of Black, White, and Other Asian Pacific Islander females with lower SES (i.e., 1SD and 2SD below average). However, the odds increased significantly for Hispanic females from higher SES (i.e., 1SD and 2SD above average), outpacing White and Other Asian females.

Research Question 2

Research Question 2 examines the relationship between the intersectionality of race or ethnicity, gender, and SES on students’ educational outcomes and organizational characteristics. This study calculated intraclass correlation coefficients (ICCs) for two dependent variables showing the degree of dependence of observations. ICCs are typically used to check whether LMM is necessary for modeling the nested data (Raudenbush & Bryk, 2002). Organizational research often uses a standard for the ICC of greater than .10 to account for organizational differences (Vogt, 2011). An ICC of less than .05 is typically too small to address the between-group variance (Vogt, Gardner, & Haeffele, 2012). The ICC of mathematics achievement score was .21, suggesting that 21% of the variance in math achievement scores occurs at the school level and 79% occurs at the individual level. However, the ICCs for a student’s intention to pursue higher education showed that less than 10% of variance across schools exist (ICC = .04), indicating that more than 90% of the variances in these variables are at the individual student level. Considering the results of ICCs and the purpose of Research Question 2, this study explored the association between school characteristics and intersectionality, with only math achievement showing moderate to large values of ICCs (i.e., over .10) and including any significant intersectionality in the previous regression analyses.

Intersectionality in Math Achievement and School Characteristics

Table 4 reports the results of LMM predicting mathematics achievement scores with student-level and school-level predictors. The parameters for student-level predictors provide the estimated slopes in the school that are coded zero for all school-level variables (i.e., urban, average percentages of FRL and students of color, and average school climate). Note that those continuous measures for school and student characteristics were centered at the grand-mean for a meaningful interpretation of the intercept. In particular, the estimated intercept for the regression of the random slope for SES (β_5j) indicates the estimated slope of the regression line for the regression of the math score on the SES in urban schools with average percentages of FRL students and students of color as well as average school climate ( γ ^ ₅₀ = 5.089). In school-level predictors, the estimated regression coefficients are for the regression of the random intercept on school characteristics.

OPEN IN VIEWER

Table 4

Parameter Estimates for Math Achievement Based on LMM

	β	SE
Level 1: Student-level predictors
American Indian/Alaska Native, γ ^ 10	−3.218	1.706
Black, γ ^ 20	−2.795***	0.602
Other Asian/Pacific Islander, γ ^ 30	2.835***	0.698
Southeast Asian, γ ^ 40	2.489***	0.678
Socioeconomic status (SES), γ ^ 50	5.089***	0.291
Intersectionality between Female × SES, γ ^ 60	−0.738*	0.310
Intersectionality between Hispanic × SES, γ ^ 70	−0.707	0.478
Intersectionality between Black × SES, γ ^ 80	−2.106**	0.708
Level 2: School-level predictors
Suburban, γ ^ 01	−0.552	0.296
Rural, γ ^ 02	−0.741*	0.303
Percentage of free or reduced lunch eligible students (PFRL), γ ^ 03	−0.052***	0.006
Percentage of students of color (PSOC), γ ^ 04	0.013	0.051
School climate, γ ^ 05	0.919	0.536
Cross-level interaction
Intersectionality (Hispanic × SES) × PSOC, γ ^ 74	−0.271*	0.105
Intercept, γ ^ 00	51.785***	0.230

*p < .05. **p < .01. ***p < .001.

The cross-level interaction effect between organizational characteristics and intersectionality among multiple social categorizations was not statistically significant for Southeast Asian female students. This result indicates that the school organizational characteristics used in this study did not mediate or differentiate the intersectionalities related to Southeast Asian female students. In other words, the patterns in educational outcomes for Southeast Asian female students held regardless of schooling context. The only significant cross-level interaction effect was observed among Hispanic students ( γ ^ ₇₄ = −0.271, p < .05), which illustrates the relationship between the intersectionality of Hispanic and SES in math achievement and school organizational characteristics. In particular, the coefficient of cross-level interaction indicates that the slope of the regression line in the regression of math scores on the intersectionality between Hispanic and SES differs significantly based on the percentage of students of color in a school. Due to the negative and statistically significant coefficient, the cross-level interaction indicates that a higher percentage of students of color in a school is negatively associated with the effect of SES for Hispanic students. This finding indicates that a low-income Hispanic student who attends a school with a higher percentage of students of color will do better than other low-income Hispanic peers in a whiter setting.

Intersectionality Focusing on Southeast Asian Females

As the idea of intersectionality is historically and theoretically rooted in feminist theories, the primary objective of research based on intersectionality typically focuses on differences and commonalities in women’s lives. By differentiating the experiences of women of color from White women, intersectionality studies seek to identify diverse patterns of inequalities originating from multiple social categorizations. The findings of this study showed that math achievement scores of Southeast Asian students were significantly higher than those of other race or ethnic groups, except Other Asian/Pacific Islanders, regardless of gender. However, even though the math achievement of Southeast Asian females was not significantly different from their male counterparts, their intention to pursue higher education was significantly lower than that of Southeast Asian males, as well as being the lowest among all female students. By using the intersectionality framework, this study was able to identify the nuanced distinctiveness in the intention to enter higher education among Southeast Asian students.

Jones’s (2000) three-level categories of oppressions (particularly racism) are helpful for identifying the potential underlying reasons for the findings of this study vis-à-vis Southeast Asian high school girls. In particular, Jones classified three-levels of racism as (a) institutionalized racism, (b) personally mediated racism, and (c) internalized racism. First, institutionalized oppressions indicate normative, legalized, and often manifested disadvantages through differential access to opportunities of society. In addition, personally mediated oppressions mean prejudices, stereotypes, and assumptions regarding the abilities and intentions of others based on multiple social categorizations (e.g., model minority stereotypes). Finally, internalized oppression (e.g., stereotype threat) suggests the “acceptance by members of the stigmatized [social locations] of negative messages about their own abilities and intrinsic worth” (Jones, 2000, p. 1213).

Applying the last category of internalized oppression to Southeast Asian females’ educational inequality, school psychology scholars might argue that specific internalized psychological characteristic in play is “stereotype threat” (Steele & Aronson, 1995). Scholars have explained stereotype threat as “when a student perceives that (s)he could be viewed through the lens of a negative stereotype and lowers academic engagement and performance as a result” (Egalite, Kisida, & Winters, 2015, p. 45). However, identifying the reason for the low intentionality of Southeast Asian female students only from students’ psychological factors or deficiencies can be based on the deficit-thinking model. In particular, Valencia (1997) conceptualized and criticized the deficit thinking in schooling as follows:

Deficit thinking is a person-centered explanation of school failure among individuals as linked to group membership (typically, the combination of racial/ethnic minority status and economic disadvantagement). The deficit thinking framework holds that poor schooling performance is rooted in students’ alleged cognitive and motivational deficits, while institutional structures and inequitable schooling arrangements that exclude students from learning are held exculpatory. (p. 9)

Scholars have criticized the deficit-oriented approach (e.g., stereotype threat) as failing to account for oppressions based on multiple social categorizations and offering counterproductive policies and practices for educational success of marginalized students (Valencia, 1997).

Rather, as intersectionality and CRT scholars argue, Jones’s (2000) first category of institutionalized oppressions is significant for explaining inequality in Southeast Asian female students’ intentionality related to higher education. That is, different intersecting institutionalized oppressions and privileges based on the unique race or ethnicity, cultural beliefs related to gender categorization, colonization, and indigeneity of Southeast Asian females may create significant inequality. In particular, White (1991) used the term middle ground to describe the spaces of multiple social categorizations created by complex relationships of power. White defined the middle ground as “the place in between: in between cultures, peoples, and in between empires and the nonstate world of villagers.… On the middle ground, diverse peoples adjust their differences through what amounts to a process of creative, and often expedient, misunderstandings” (p. x). Based on their unique historical backgrounds as immigrants or refugees, Southeast Asian female students may live within in-between worlds (DeLeon, 2010; Ngo, 2009). The concept of middle ground suggests that people living in the spaces between cultures and multiple social categorizations (i.e., Southeast Asian females) often experience dynamic inequality and social division in terms of their relationships with each other. For example, educational policy and practice confirming postcolonial superiority in the United States make Southeast Asian students struggle to adapt to existence in the middle ground, which is between the ethnic homelands culture of their parents and the colonized culture (Ngo, 2013). Of course, even when institutionalized multiple oppressions are critical external forces affecting Southeast Asian female students’ lower intentionality for higher education, more mediated oppression also plays a role. For example, Khalifa, Bashir-Ali, Abdi, and Arnold (2014) argued that colonizers utilize a variety of stereotypes toward people of color and indigenous people (i.e., model minority stereotypes) as a tool for arranging their power and normalizing their hegemonic positionality.

Furthermore, the unique cultural norms and values for the roles of girls can amplify the external forces for Southeast Asian female students in the conceptual space of middle ground based on racial or ethnic and gender categorizations. In particular, Walker-Moffat (1995) found that parents’ relatively lower educational expectations for their daughters create particular challenges for Hmong girls. Furthermore, the cultural pressure for Cambodian girls to comply with traditional gender norms in their home culture (e.g., early marriage, having a baby) is a significant factor for their educational outcomes (Ngo & Lee, 2007). Although patriarchal structures at home and school might also affect White, Latina, and Black students, the particular positionality of Southeast Asian females in the middle ground based on multiple social categorizations can create their unique lived experiences and inequality in their pursuit of higher education.

Intersectionality Focusing on SES

Although intersectionality studies focusing on SES are still underexplored (Knapp, 2005), scholars have emphasized the influence of class in the interplay of diverse social categorizations. This study also demonstrated the influence of SES in creating different intersectionalities. In particular, in terms of a student’s math achievement, the influence of SES was different between males and females among Southeast Asian students. This study found that SES matters less for the achievement of Southeast Asian females than it does for both Southeast Asian males and for Whites overall. Furthermore, this study found higher SES Hispanic students are not experiencing the same benefits from SES in math achievement that other race or ethnic groups with higher SES do. This finding is consistent with O’Connor’s (2009) finding that SES significantly interrelates with college enrollment for students with Hispanic origin. She demonstrated that Hispanic students from a higher SES background are likely to have fewer enrollment resources for 4-year colleges than their White counterparts from a higher SES background. In addition, she argued, “the Hispanic community has been less likely than the black and white communities to translate parental SES into educational achievement of the next generation, and to engender upward mobility” (p. 138). Similarly, the findings in the current study may suggest that Hispanic or Black parents from a high SES background might be limited in their access to the financial, intersocial, and informational resources available to contribute to their children’s achievement in math. Another potential explanation of this finding might originate from neighborhood effects (e.g., Sharkey, 2013; Sirin, 2005). That is, Squires (2017) found that the concentration of affluent White population is more significantly prominent than that of affluent Black population in the United States. Due to stronger neighborhood concentration of White population from higher SES, Black and Hispanic students from higher SES might live in less concentrated neighborhoods, thereby not having similar neighborhood effects from resources and social networks.

Intersectionality and School Organizational Characteristics

Intersectionality scholars often emphasize the importance of organizational contexts and the role of organizations in creating different life opportunities (Núñez, 2014). Incorporating organizational factors into the interplay among individual-level intersectionalities can illuminate “a comprehensive picture, providing the best chance for an effective diagnosis and ultimately an effective prescription [for educational inequity]” (Hancock, 2007, p. 73). Although this study found that Hispanic students experienced different class inequalities based on the percentage of students of color in a school, there was no significant association between school characteristics and inequality that Southeast Asian female students experienced in educational outcomes. This finding shows that the school characteristics used in this study (i.e., community type, percentage of FRL students, percentage of students of color, and school climate) do not mediate or resolve inequalities in educational outcomes (i.e., math achievement). These patterns held regardless of schooling contexts. In addition to this finding, critical quantitative researchers should continue exploring what other school characteristics (e.g., leadership, school-level policy) could mediate inequities across school organizations through rigorous quantitative studies. With such concerted efforts, policymakers and school leaders should be implementing policy strategies and exercising leadership to address inequities that Southeast Asian females are experiencing, which will be specifically discussed below.

Implications for Policy and Leadership

One important component in intersectionality studies is the pursuit of transformative policy efforts to realize social justice (e.g., Collins, 2009; Dill & Zambrana, 2009). Intersectionality scholars emphasize such transformative actions to meet diverse and unique needs of students originating from their positionality on the intersections among race or ethnicity, gender, and SES. Furthermore, intersectionality thinkers highlight the importance of school or university leaders’ leadership in creating more inclusive organizations (e.g., Gooden, 2015; Patel, 2016). In particular, this study found that the expected probability of Southeast Asian female students pursuing higher education was the lowest among females and significantly lower than that of Southeast Asian males across all SES index scores. Ngo (2006) argued that Southeast Asian students’ pursuit of education is closely related with the needs of their families and they are often marginalized in education policy due to the model minority myth based on the story of Asian Americans’ success in the United States. Education policy from a unidimensional approach to improving college access might have only a very limited effect for Southeast Asian female students. In particular, typical education policies meant to encourage the pursuit of postsecondary education options will not work for Southeast Asian girls. Rather, this study demonstrated that the intersection of race or ethnicity and gender should point policymakers to a strategy that is different from what conventional wisdom would recommend.

The findings of this study suggest the need for educational strategies that are unique to Southeast Asian girls and different from those employed for other girls or for Southeast Asian boys. In particular, different solutions can support Southeast Asian female students based on Jones’s (2000) categories of oppressions. For example, policymakers and educational leaders may provide counseling services to remove stereotype threats that Southeast Asian females have about their academic pursuits for higher education (internalized oppression). Furthermore, policymakers might use professional development opportunities and capacity building for educators, school counselors, and leaders to engage in ways of addressing personally mediated oppressions (e.g., tackling the myth of the model minority stereotype). Utilizing only these two approaches, however, will not be enough to change the conditions that create Southeast Asian females’ structural inequality in educational outcomes. Thus—and most importantly—in addition to these strategies, policymakers and educational leaders as social justice leaders should focus on addressing multiple forms of structural oppressions in their organizations and society, which inhibit Southeast Asian female students from realizing their potential in schools. As Capper (2015) appropriately emphasized, policymakers and school leaders “must guard against the ways that unifying policies and practices across differences can reproduce racism, [classism, sexism, and the corresponding intersecting multiple oppressions]” (p. 822). Using social justice–oriented inquiries, researchers should also engage in destroying multiple oppressions and advancing equity in education. Critical quantitative researchers, in particular, utilize numbers and statistical inferences to reveal these multiple structural inequalities that are deeply associated with power relations between different social categorizations. In this regard, critical quantitative researchers have an essential and powerful role in creating conditions of equity that attend to intersectionality: They can inform and alarm policymakers and school leaders about structural inequalities as well as realize educational equity and social justice.

There are key limitations of this research that need to be addressed. First, this study was not able to reveal mutually constructing systems of power that Southeast Asian students experience in schools as this study used existing surveys. Furthermore, using fixed categories of diverse identities in surveys limits the ability to illustrate the complexities related to unfixed identities. For example, this study used fixed binary gender categorizations (males versus females). This is because the dataset used in this study as well as other nationally representative quantitative datasets (e.g., Census, National Center for Education Statistics) typically do not specify other gender categorizations including transgender and gender nonconforming (TGNC) (Rider, McMorris, Gower, Coleman, & Eisenberg, 2018). Thus, the quantified gender categorization in this study is limited in revealing inequalities among TGNC students, which is beyond the scope of this study. Furthermore, although the number of Americans who classify themselves as mixed race is increasing according to recent U.S. Census reports (Aumer, Hatfield, Swann, & Frey, 2011), this study has the limitation that it is not able to consider multiracial status in this analysis. Second, this study delimits the exploration of intersectionality by focusing on three identities: race or ethnicity, gender, and socioeconomic status. However, intersectional frameworks suggest expanding the perspective beyond these identities to other marginalized student social categorizations (e.g., age, sexual orientation, language).

In order to address these limitations and expand the knowledge related to intersectionality, future research is warranted as follows. First, future research needs to utilize qualitative anecdotes to complement the quantitative findings of this study and enrich the knowledge of nuances and complexities in Southeast Asian female students’ schooling experiences based on the intersectionality of race or ethnicity, gender, and SES. Furthermore, future research should expand intersectionality inquiries by including other social categorizations that can perpetuate marginalization and exclusion in education (e.g., immigration status, home language, age). In particular, scholars have emphasized the importance of language as a critical means of social reproduction and marginalization resulting from structured inequity against students from diverse linguistic backgrounds (S. J. Lee, 2009; Ricento & Wiley, 2002). Thus, future investigations focusing on the intersectionality related to Southeast Asian students’ home language use (or limited English proficiency) can show the underexplored patterns of educational inequity and suggest adequate policy alternatives to support students who need more support. Future research should also explore how macro-level policies (e.g., immigration policy, language policy) create different patterns related to intersectionality among Southeast Asian female students’ multiple social categorizations. For example, the Trump administration’s anti-immigration policies might be associated with the effect of intersectionality among immigration status and race or ethnicity on Southeast Asian female students’ lives. Taking into account the political climate and policy stream can unearth broader interlocking oppressions and systems of power that change over time and differ by specific geographical regions.