Invited Reflection

Framing Intersectionality as “Contested”

When psychology researchers discuss methodological approaches to intersectionality research, they often define intersectionality in terms of the “core features” of the theory. That is, scholars (e.g., Cole, 2009; Else-Quest & Hyde, 2016a) focus on those features of the theory for which some consensus has been achieved across scholars identified as experts on the theory, or those who are given credit to have originated the theory. These core features are often used to divide what is considered intersectionality from what is not.

Else-Quest and Hyde identify core features that align with some scholars’ perspectives (e.g., McCall, 2005), but not all (e.g., Alexander-Floyd, 2012; Hancock, 2013), and this discrepancy must lead us to contextualize their recommendations. For example, Else-Quest and Hyde claim that one essential element that makes research intersectional is that it focuses on the experiences and meanings of simultaneous multiple group memberships. This description implies that all members of multiple groups, including those in positions of privilege, can be considered for intersectional analyses, which is consistent with some key intersectionality scholars (e.g., Collins, 1990). However, other scholars argue that intersectionality should focus centrally on individuals who are multiply marginalized (e.g., Alexander-Floyd, 2012).

Another essential element that Else-Quest and Hyde (2016b) identify is that social categories, such as class, ability, and sexual orientation, are features of both the individual and the social context. However, some scholars argue that, within the context of intersectionality, social categories are fundamentally social structures, not features of an individual’s identity (e.g., Hancock, 2013; Marecek, 2016). While Else-Quest and Hyde (2016a) do emphasize that “there is no single, dominant gatekeeper or litmus test of intersectionality research” (p. 157) and that there are disagreements over what is meant by engaging in an intersectional approach, the features that undergird Else-Quest and Hyde’s (2016b) methodological recommendations are those that they frame as “common assumptions” (p. 320).

While emphasizing the core features provides a clear road map for researchers, it implicitly conveys that only those features are important considerations in integrating intersectionality into research and that other considerations are not. I challenge this assumption by arguing that the elements of intersectionality that are not agreed upon must also be considered when making methodological choices. I argue that a focus on the contested areas of intersectionality (1) communicates its nuanced, complex nature and (2) preserves its radical potential.

First, communicating the nuanced, complex nature of intersectionality is critical both to its usefulness and to its endurance as a theory. Open-endedness allows a theory to withstand multiple interpretations, generates academic debates, and provokes further thinking on the topic (Davis, 2008). Furthermore, solely emphasizing intersectionality’s shared or common features portrays it as narrow and limited in scope, which encourages criticism of intersectionality as a theory (Collins, 2015). Instead of criticizing intersectionality, Collins encourages scholarship to be directed toward critiquing the structures of inequality that intersectionality was created to address in the first place, citing Tomlinson (2013) to demonstrate this misplaced focus:

Many critics approach intersectionality carelessly, however, through meta-commentary and complaint and through recommendations to bring its radical critique under control by advocating recourse to specific disciplinary methods—without acknowledging that such methods may have long been criticized for their service to dominant discourse. Critics assume that their task is to critique intersectionality, not to foster intersectionality’s ability to critique subordination. (p. 996)

Second, communicating intersectionality’s contested areas is critical to preserving its radical potential. Conflicting interpretations of intersectionality help to ensure that an epistemologically dominant viewpoint deployed by researchers does not silence marginalized voices (Davis, 2008). For example, Bilge’s (2013) and Alexander-Floyd’s (2012) analyses of the “whitening” of intersectionality provide a critical check on research that undermines intersectionality’s focus of providing voices to marginalized groups. Similarly, Puar’s (2007) analysis of how intersectionality scholarship has privileged U.S. Black women’s voices over other women of color provides a check on ethnocentric perspectives of intersectionality.

Contested Areas of Intersectionality

To demonstrate why it is necessary to address the contested nature of intersectionality, specifically within the context of research methods and techniques, I examine two recommendations for how to integrate intersectionality into quantitative methods and techniques (Else-Quest & Hyde, 2016b): (1) factorial designs with main effects and an interaction term and (2) representations of intersectionality as person and stimulus variables. For each of these two examples, I demonstrate that framing intersectionality as contested leads to better understanding of the benefits and drawbacks of the methodological choices made in quantitative research.

First, Else-Quest and Hyde (2016b) explain that study designs employing analysis of variance (ANOVA) can test hypotheses relevant to intersectionality by examining multiple main effects and statistical interactions. They propose that social categories can be experienced additively, multiplicatively, and/or intersectionally. According to this quantitative feminist empiricist approach, whether or not social categories are intersectional or additive is an empirical question that can be answered via a research investigation. Indeed, some research has begun to address this debate by empirically exploring whether people perceive experiences with systems of oppression as integrated or separated. For example, Remedios and Snyder (2015) have explicitly examined power structures, that is, whether women of color perceive their experiences with racism and sexism intersectionally or as separate entities.

However, there is disagreement over whether social categories such as gender and race can be meaningfully separated into distinct entities, or if they are meaningless when considered apart. Crenshaw (1991) created the label of “political intersectionality” to refer to being situated in at least two subordinated groups that have, at times, competing political interests. When feminism fails to address racism, resistance strategies of feminists will replicate oppression of people of color, and when antiracist strategies fail to address patriarchy, resistance strategies of antiracism will replicate oppression of women. Political, legal, and organizational efforts have historically forced individuals from multiply marginalized groups to align with one group that only represents part of who they are. For example, Black lesbian women working within LGB rights campaigns may feel silenced and the campaign might subvert their needs as women of color. According to Crenshaw, antiracism and feminist strategies need to be fully integrated to empower women of color. In arguing against viewing race and sex separately, MacKinnon (2013) provides an especially vivid warning: “As a recent court put it, sex and race can ‘fuse inextricably’ so that ‘made flesh in a person, they indivisibly intermingle’” (p. 1020).

The problem with the argument that “additive versus intersectional” is simply a hypothesis to be tested is that it implicitly puts the additive and intersectional conceptualizations on equal grounds. In other words, treating additive versus intersectional explanations as equivalent in a statistical sense implicitly communicates that each are equivalent as political positions (Warner, Shields, & Settles, 2016). Researchers using factorial designs, then, need to be aware that factorial designs can implicitly invalidate an intersectional approach, even when researchers explicitly support intersectionality. This example shows the limitations that can occur when intersectionality is adapted to fit with a discipline’s methodological and epistemological traditions; a factorial approach can (re)produce knowledge and can remarginalize multiply marginalized people, especially women of color (Alexander-Floyd, 2012; Cho et al., 2013). In fact, some scholars have observed that it was the traveling of intersectionality into empirical projects that first led researchers to perceive that social categories could be experienced separately (Carbado & Gulati, 2013; Hancock, 2016).

A second recommendation that Else-Quest and Hyde (2016b) make for quantitative research on intersectionality is to offer two different frames for social categories. First, they describe person variables, which are variables that focus on the experiences of individuals in intersecting social categories. For example, a person variable might be the experiences that a Latino gay man has as a token within queer communities. Next, they describe stimulus variables, which are variables that affect how other people respond to individuals of intersecting social categories. They describe how these types of variables would involve materials such as photos or videos of people with differing intersections of social identities, and they would have participants react to these stimulus materials.

This recommendation conveys a particular stance in another contested area of intersectionality: whether to focus on intersections of identities or whether the primary focus should be on political and structural inequalities (e.g., Cho et al., 2013; Warner et al., 2016). As Marecek (2016) recently observed, Crenshaw’s (1991) and Collins’ (1990) work originally conceptualized intersectionality as characterizing the intersecting systems of social stratification, not as a feature of one’s individual identity. Hancock (2013) also observed that the original work on intersectionality stemmed from challenging “legal structures of power and social movements’ strategies for pursuing remedies” (p. 277), not empirical investigations of individual attributes that a person possesses. Most of intersectionality’s challenges to social policy address problems that are “causally complex” and thus cannot be reduced to individual experiences (Hancock, 2013). According to this view, it is the politics that surround the lives of women of color, rather than those lives themselves, that constitute the heart of intersectionality.

However, as Crenshaw herself notes, the theory has developed beyond her original conceptualization (Cho et al., 2013). Many scholars, especially feminist psychologists, argue that there is a place for considering intersectionality, as it pertains to subjective experience, individual identity, and person perception (e.g., Bowleg, 2008; Cole, 2009; Else-Quest & Hyde, 2016a). Some researchers have suggested conceptualizing social categories, both as features of the individual experience and as larger social structures (Warner et al., 2016). In addressing this debate, Collins (2015) analyzed the frequency of different types of intersectionality scholarship and found that one of the most popular topics is how intersecting identities produce distinct lived experiences for individuals and social groups. While not discounting this as a worthy research topic, she emphasizes that much less attention is paid to the associations between complex social structures of inequality and social justice efforts, and those areas should not be neglected. Collins highlights the need to research local, small-scale social justice efforts achieved by society’s overlooked populations. As an example, she describes Clay’s (2012) research on youth activism in Oakland, CA, in which young people of color used intersectional frameworks to challenge persistent inequalities that they faced in their school systems.

Applying this contested area to the context of quantitative research methods in psychology, available methods favor the conceptualization of intersectionality as a feature of individual experience. Both of the frames Else-Quest and Hyde (2016b) provide, person variables and stimulus variables, are more compatible with intersections of identities, rather than intersections of social structures of inequality, as they refer to demographic variables of participants and also individuals’ perceptions. To be able to stimulate structural change, Hancock (2013) argues that intersectionality researchers need to measure social categories in terms of how they actually operate within the structure that is challenged. To do that, researchers need to move beyond person and stimulus variables and instead utilize variables that directly assess the social structure. For example, Hancock employs the concept of “racialized subjectivity” (Lee, 2007) to avoid the reliance on self-reported demographic variables such as “race” or “ethnicity.” Racialized subjectivity involves all of the components that create a racialized experience. In addition to self-reported identity, it involves assessing aspects of the social structures that are out of a person’s direct control, such as the ethnic composition of their neighborhood or their experiences with institutionalized racism. Hancock uses all of these factors to create discrete categories that differ in the degree of racialized subjectivity.

Hancock (2013) suggests that the analytic tool of Boolean truth tables can determine the factors that create a particular outcome. This technique allows for the creation of variables, such as Hancock’s racialized subjectivity, since it does not rely on statistical methods that require corrections for multicollinearity, nor does it require that levels of a variable be mutually exclusive. In contrast, ANOVA relies on an assumption of independence of cases; for example, if “socioeconomic status” and “race” are identified as factors, they are treated in the model as independent from one another. Statistical techniques that allow for nonindependence of categories are more compatible with an approach that recognizes the interconnectedness of systems of oppression.

Further, when using Boolean truth tables, the mathematical representation of the explanatory factors represents them as fused rather than as separable, testing different combinations of the presence and absence of factors using dummy codes to determine which model best predicts the outcome; for example, the presence or absence of racialized subjectivity, evangelical orientation, ideological orientation, culture of religiosity, and gender assignment in predicting attitudes toward gay marriage. Thus, Hancock introduces a type of variable that goes beyond person or stimulus variables and instead integrates the social structural factors that affect the lived reality of individuals (such as their subjective attitudes toward gay marriage).

Else-Quest and Hyde’s (2016b) recommendation of person-centered methods would be most compatible with Hancock’s recommendation that social categories should be measured in terms of how they actually operate within that structure. For example, Else-Quest and Hyde’s discussion of latent class analysis, in which a population is organized on the basis of behaviors or characteristics, puts the focus on the impacts of the social structure on the social category, rather than on discrete social category membership. They categorized adolescents’ experiences of “multiple” versus “singular” forms of identity-based bullying, not based on demographic information (e.g., self-reported race and gender), but rather by analyzing adolescents’ patterns of responding to items about experiences of discrimination.

Future Directions of Quantitative Approaches to Intersectionality

Researchers who employ quantitative methods in psychology are currently in a difficult situation, because tools have not yet been developed to better incorporate intersectionality (Shields & Dicicco, 2011), but researchers still want deep engagement with intersectionality theory. Else-Quest and Hyde (2016b) state that a main purpose of their work is to stimulate the innovation of quantitative methods within intersectionality research, rather than to solely utilize available methods. Until new methods are created, however, researchers are left to acknowledge and address the limitations of the existing techniques. While researchers shouldn’t necessarily reject a methodological technique simply because it is at odds with an interpretation of intersectionality, it is also important to avoid a relativistic stance in thinking that all contested interpretations of intersectionality are created equal. In other words, one danger in recognizing the multiplicity of intersectionality interpretations is that researchers could be led to think that “anything goes.”

Consistent with this view, Else-Quest and Hyde (2016b) caution against believing that any single method in and of itself could capture intersectionality: “…we are hopeful that researchers will go beyond a specific method to incorporate the complexity and depth of intersectionality as a critical theory and approach” (p. 321). They emphasize that, in considering any method, readers should examine theoretical and epistemological issues concerning intersectionality theory, especially since both research and policy decisions have inadvertently functioned to remarginalize social groups (May, 2015).

How does a researcher avoid a relativistic stance on the contested areas of intersectionality? Bilge (2013) warns that problematic scholarship is that which “depoliticizes” intersectionality or inhibits its potential as a mechanism for social justice. Depoliticization occurs when researchers seize control of intersectionality at the expense of individuals who hold less powerful positions, support disciplinary practices that have historically marginalized certain social groups, turn intersectionality into a contemplative exercise rather than as a tool for social justice, and fail to give feminists of color credit for the creation of the theory.

The question remains, though, as to how one conducts quantitative research in a way that is true to intersectionality, despite the particular limitations that the research method creates. To resolve this issue, researchers who use quantitative methods need to acknowledge the limitations in their manuscripts. One key question that researchers have to ask themselves is: Does my research actually work toward institutional and social change by challenging dominant knowledge production? (Bilge, 2013; Cho et al., 2013). In addition, researchers who use quantitative methods can assume responsibility outside the context of the manuscript itself to engage in efforts to support intersectionality. For example, quantitative researchers, whose careers have benefited from their reliance on mainstream methodological approaches, could engage in efforts to support researchers who conduct more marginal forms of research (Cho et al., 2013). Mentoring, networking, and collaborating are also strategies to bring those at the margins of psychological research to the center (Shields & Dicicco, 2011). In addition, researchers could directly use their findings to influence public policy (e.g., Warner et al., 2016), rather than abandoning them after they have been published in an academic journal. As Collins (2015) asserts, “intersectionality is not simply a field of study to be mastered or an analytical strategy for understanding; rather, intersectionality as critical praxis sheds light on the doing of social justice work” (p. 16). A steadfast engagement with social justice efforts is central to psychology’s development of quantitative methods that fully engage with intersectionality theory.