Where Identities Converge: The Importance of Intersectionality in Eating Disorders Research

Abstract

Disparities in eating disorder (ED) risk, diagnosis, and treatment for those who occupy multiple marginalized social identities (e.g., combined racial/ethnic and sexual minority statuses), underscore the need for advancing multicultural research in the ED field. In this article, we argue that intersectionality-informed approaches, which examine the ways in which one’s multiple social identities interact to inform risk for ED outcomes, offer an established framework for identifying frequently underserved individuals who may be at greatest risk for EDs. We highlight preliminary intersectional research in EDs and offer suggestions for further progression. In particular, we encourage future intersectionality-informed research to incorporate a broader range of social identities (e.g., age, ability status), consider the ways in which these identities may be dimensional and fluid, and embrace strengths-based approaches to illuminate dimensions of identity that may serve as protective factors. To support such research, we describe quantitative and qualitative methods for pursuing questions of intersectionality in ED investigations. Given the success of intersectionality-informed research in other areas of psychopathology and its relevance to ED as suggested by initial research, the continued pursuit of these approaches in EDs has high potential to improve identification and treatment for patients who have too often been overlooked.

Given disparities in eating disorder (ED) diagnosis and treatment for marginalized groups (e.g., Sonneville & Lipson, 2018) and emerging evidence of compounded ED risk among populations who occupy multiple marginalized social identities (e.g., Calzo, Blashill, Brown, & Argenal, 2017), the need for greater attention to vulnerable intersecting groups is imperative. We propose that application of intersectionality-informed approaches to ED research provides a powerful framework to address this need.

The construct of intersectionality is rooted in the scholarship of Black feminists and critical race theorists (e.g., Beale, 1970; Crenshaw, 1993), who recognized that multiple social identities (i.e., race, class, and gender) operate both independently and interactively to determine risk of discrimination, disadvantage, and disparity (Cole, 2009). The intersectional approach offers an alternative to more traditional models of risk, which propose that holding multiple social identities associated with increased disease risk (e.g., holding female gender and racial/ethnic minority identities) would increase the risk for negative outcomes incrementally (Rouhani, 2014). In contrast, the intersectional approach acknowledges that disease risk for an individual who lies at the intersection of two high-risk identities may be greater than the sum of the risk born by each identity in isolation. In other words, this approach allows for the possibility that the effect of one identity (e.g., gender) may depend upon the existence of another identity (e.g., race/ethnicity).

Notably, governing professional bodies (e.g., American Psychological Association) have identified attention to intersectionality as the primary guideline for multicultural research and treatment within psychology (Clauss-Ehlers, Chiriboga, Hunter, Roysircar, & Tummala-Narra, 2019). Thus, for example, in the depression and trauma literature, researchers have utilized intersectionality-informed approaches to clarify culturally-relevant social categories that confer disease risk via structural and individual inequities, and identified specific subpopulations at elevated risk for disease development (e.g., Bryant-Davis, 2019; Patil, Porche, Shippen, Dallenbach, & Fortuna, 2018). These findings then informed targeted interventions and clinical practices by focusing strategies on the culturally-relevant risk factors faced by identified groups (Bryant-Davis, 2019). As culturally-adapted interventions have been shown to produce greater reduction and remittance of symptoms across a range of psychological disturbances (Hall, Ibaraki, Huang, Marti, & Stice, 2016), continued work of this nature is critical for improving outcomes in underserved and underrecognized populations that may be overlooked by approaches that fail to recognize intersecting identities.

Intersectional approaches are likely to bear similarly meaningful results within the ED field. Researchers focused on clarification of ED risk in frequently-underrepresented groups (e.g., racial/ethnic minorities, economically disadvantaged) have drawn attention to the unique experiences and elevated rates of eating pathology among those who hold multiple intersecting identities, with accumulating research suggesting that social identities may interact to modulate ED risk and amplify health disparities among intersectional groups (e.g., Austin, Nelson, Birkett, Calzo, & Everett, 2013). However, with notable exceptions, the ED field is lacking research conducted with an explicit foundation in intersectional theory. This oversight implicitly contributes to known disparities in ED identification and treatment in marginalized populations (e.g., Sonneville & Lipson, 2018) and impedes the identification of groups at uniquely high risk due to their specific sociocultural location. Below, we present examples from the scarce existing intersectionality-informed ED research, which highlight the applicability of intersectionality within the field. Further, we articulate directions for future investigation and provide methodological guidance for engaging in intersectionality-informed research. We argue that moving beyond individual cultural considerations to the complexities of intersectional identities is a critical and logical next step toward improving ED assessment, research, and intervention outcomes.

Extant Intersectional Research in EDs

While the ED field has lagged behind other areas of mental health in terms of research grounded in intersectionality, notable exceptions exist (e.g., Austin et al., 2013; Beccia et al., 2019; Calzo et al., 2017; McEntee, Serier, Smith, & Smith, 2020; Rodgers, Watts, Austin, Haines, & Neumark‐Sztainer, 2017), which have sought to identify whether compounded risk for EDs occurs at the intersections of gender, sexual orientation, racial/ethnic identity, and weight status. This research has increased the field’s awareness of highly vulnerable groups that may benefit from tailored and targeted resources. As examples, Asian boys of higher weight status may experience greater risk for dieting than expected based on the additive contributions of their racial identity and weight status (Rodgers et al., 2017), and U.S. Hispanic/Latina adolescent girls may experience greater risk for purging than expected based on the additive contributions of their ethnic and gender identities (Beccia et al., 2019).

Future Research Directions from an Intersectional Perspective

Although the literature reviewed above represents an important step in a positive direction, moving forward, a number of limitations bears acknowledgement. First, measures used to assess EDs in males and racial/ethnic minorities (who, historically, have been overlooked in the ED field) were primarily developed and validated with White female participants, and may thus fail to capture presentations in other groups (e.g., Murray et al., 2017). In addition, existing research has focused on a small number of identities, most commonly gender, race/ethnicity, weight status, and sexual orientation. While these dimensions are important and individually associated with increased ED risk (e.g., Bodell, Racine, & Wildes, 2016; Calzo et al., 2017; Murray, 2017; Rodgers, Berry, & Franko, 2018), expanding research to other dimensions is indispensable. While not an exhaustive list, additional dimensions might include socioeconomic status and class, age, country of origin, acculturation, acculturative stress, immigration status, ethnic identity, and ability, among others.

A second limitation is that this research has mostly viewed intersecting identities as “smaller groups within groups” (e.g., gender stratified by broad race/ethnicity categories presented as homogenous). It has yet to integrate a more holistic conceptualization of identities as multiple layers of experience that are informed and shaped by one another in a dynamic process. Moving forward, research should move away from categorical conceptualizations to better capture the nuances of identity as some identities may be more/less visible or salient in different contexts or at different times (Törngren, Irastorza, & Rodríguez-García, 2019). For example, the heightened risk for ED conferred by higher weight among individuals identifying as female may be modulated by concurrent identities such as sexual orientation or age, and the degree to which these dimensions are salient across development and social contexts may vary.

Third, the extant literature is limited by an exclusive focus on intersecting identities as a source of risk and a comparative neglect of the way these identities may also serve as protective factors. Strengths-based approaches focusing on the identification of positive or protective aspects of minority identity such as ethnic identity (Rodgers et al., 2018), cultural and religious group affiliation, and resilience (Walton & Oyewuwo-Gassikia, 2017) may help to identify moderators of ED risk that could be used to inform culturally-adapted interventions.

Theoretical Considerations

Intersectional approaches encompass those that emphasize the experiences related to intertwined dimensions of identity, and the way in which identities may be fluid, and more/less salient across contexts. Intersectional lenses, however, can be usefully combined with other theoretical approaches. Some that have proven most useful include feminist and critical approaches. These theories highlight how systemic forces shape identities that are then related to inequality, power, and privilege (e.g., Beale, 1970; Crenshaw, 1993). Such theories are well suited for such inquiry as, for example, socially-constructed, historical and political hierarchies and practices influence mental health disparities related to racial/ethnic minority identities and others (e.g., Williams, Priest, & Anderson, 2016). Importantly, intersectional approaches recognize concurrent membership to different groups. Thus, for example, given the global population increase of multiracial/multiethnic individuals and of those identifying as multiracial/multiethnic, burgeoning conceptual frameworks of majority and minority identity “mixedness” can inform future research (Törngren et al., 2019), and be extended to consider other facets of identity.

Importantly, in intersectional research pertaining to EDs, specific attention should be given to weight, shape, and appearance as dimensions of visible identity. Weight stigma and appearance-based discrimination are pervasive, even among ED professionals (e.g., Puhl, Latner, King, & Luedicke, 2014). How external and internalized weight stigma intersect across weight status and other identities, increase disordered eating risk (e.g., Puhl, Moss-Racusin, & Schwartz, 2007), and impact treatment remain important facets for research within an intersectional lens. Feminist and critical frameworks have been developed to help conceptualize these dimensions, including weight inclusive frameworks that may prove particularly useful (Tylka et al., 2014). Furthermore, strengths-based approaches to body image and eating behaviors grounded in positive psychology may also be useful tools (Tylka & Wilcox, 2006; Webb, Wood-Barcalow, & Tylka, 2015).

Methodology and Statistical Approaches

Both quantitative and qualitative approaches may yield rich intersectional findings.

Quantitative approaches.

One means of approaching questions of intersectionality includes leveraging large datasets with sufficient sample size and sampling diversity to support intersectional-based analyses (e.g., Youth Behavior Risk Survey, Centers for Disease Control and Prevention, 2017; National Longitudinal Study of Adolescent to Adult Health, Harris et al., 2009; National Comorbidity Survey Replication, Hudson, Hiripi, Pope, & Kessler, 2007; New Zealand Mental Health Survey, Wells et al., 2006). A true intersectional approach incorporates both the main effects for each individual social identity (e.g., gender, ethnicity) and their cross-product interaction terms (e.g., gender × ethnicity) into prediction models. A more nuanced consideration includes the scale on which interaction is assessed. In linear regression models, cross-product terms yield interaction tests on an additive scale, while in logistic regression models, cross-product terms yield interaction tests on a multiplicative scale (VanderWeele & Knol, 2014). Given that intersectionality theory aims to test whether multiply marginalized individuals experience risk that is greater than the sum (versus product) of the independent risks for each marginalized identity, assessment of interaction on the additive scale is most relevant for intersectionality research (Bauer, 2014; Jackson, Williams, & VanderWeele, 2016). Assessing dichotomous outcomes on an additive scale is more challenging logistically, but software programs and code are available to facilitate such analyses (Mathur & VanderWeele, 2018; VanderWeele & Knol, 2014; VanderWeele & Tchetgen, 2014), and proved fruitful (e.g., Beccia et al., 2019). Notably, in an adequately powered sample, if both of two identities are associated with greater ED risk and interaction does not exist on the multiplicative scale, interaction must exist on the additive scale (Greenland, Lash, & Rothman, 2008). This would provide evidence in support of intersectionality theory, whereas it might be missed with only the multiplicative scale. Therefore, additive scale interactions are the preferred approach.

In addition to large epidemiological datasets, we recommend greater collaboration across multiple laboratories to obtain adequate samples of diverse, often marginalized populations. Readily accessible, shared toolkits (e.g., the Social Determinants of Health PhenX Toolkit, https://www.phenxtoolkit.org/collections/view/6) provide researchers access to culturally-sensitive standardized measures of intersectionality-related individual and structural health factors (e.g., race/ethnicity, education level), which can be administered across multiple labs or data-collection platforms to build large uniform datasets.

Qualitative and mixed-methods approaches.

In addition to the quantitative methods described above, other approaches that are grounded in or build on different epistemologies afford valuable strengths for intersectional research. Qualitative research that captures the voices and lived experience of disenfranchised or vulnerable groups has been shown to increase understanding of EDs in groups with intersecting minority statuses (Neumark-Sztainer, Story, Faibisch, Ohlson, & Adamiak, 1999; Piran, 2016). Such research allows both for a more detailed understanding of the factors and mechanisms at play, and the development of measures and models of risk that are richer and experience-informed. In addition, mixed-method approaches that include quantitative components may yield some of the most useful data.

Still, caution is advised in this new area of research. It is critical to minimize our own biases to ensure we are asking the right questions in the right ways. Developing community partnerships with individuals and groups to help identify, frame, and potentially ask the needed questions would be beneficial (Jagosh et al., 2015). Such partnerships would also help in accessing potentially difficult to reach populations (e.g., Becker, Middlemass, Taylor, Johnson, & Gomez, 2017) to give voice to multiply marginalized groups. Ideally, this will be an iterative process. Qualitative research can also provide insight to areas where ED screening/assessment tools may need to be adapted to increase accuracy and sensitivity among individual intersectional groups.

The Future of ED Research Includes Intersectionality

Investigating EDs through an intersectional lens is needed to determine which intersecting identities confer highest risk, through which mechanisms, and how EDs may present differently by group. Increasing recognition and specificity around intersectional identities will provide further avenues of research to create and/or adapt prevention and treatment to higher-risk groups. Taking an intersectional approach requires a conceptual shift (Cole, 2009), but we risk missing individuals who need care, misguiding or misinterpreting research with marginalized populations, and potentially contributing to structural inequalities by not conducting intersectionality-informed research. Likely, we will need to expand our assessment and outreach mechanisms to reach those of multiply marginalized identities and be creative in considering treatment and retention options, including developing partnerships with community groups/leaders within populations of interest and broadening the scope of our research and treatment beyond traditional research and clinic spaces.