Abstract
The number of individuals identifying as multiracial in the United States (US) has significantly increased in the past few decades, yet they are rarely the focus of study in eating disorders (ED) research. The current study is among the first to examine prevalence estimates of ED pathology across several distinct multiracial groups, to contrast prevalence estimates of ED pathology in each multiracial group with those among the corresponding monoracial identities, and to investigate these findings intersectionally with gender identity. Data from 145,379 US students, 11,433 of whom were multiracial, were collected from 199 US colleges and universities participating in the Healthy Minds Study between 2014–2019. Elevated ED pathology was defined as a score ≥2 on the SCOFF. Multiracial individuals identifying as American Indian/Alaskan Native and Hispanic/Latinx exhibited the highest prevalence estimates of elevated ED pathology (41.4% compared to 23.5% in the full sample). This group, as well as some other doubly marginalized groups (African American/Black and Hispanic/Latinx; African American/Black and Asian American/Asian), exhibited higher prevalence of elevated ED pathology than expected based on the observed prevalence estimates in their corresponding monoracial groups. Across gender identities, greater than expected prevalence estimates of elevated ED pathology were observed among multiracial individuals identifying as African American/Black and White and lower than expected prevalence estimates were observed among multiracial individuals identifying as Middle Eastern/Arab/Arab American and White. These results have important implications for understanding ED pathology in multiracial individuals and should inform intervention and treatment efforts to support individuals from these underserved groups.
Keywords: race/ethnicity, multiracial, biracial, eating disorder, gender identity, intersectionality
The number of people identifying with more than one racial/ethnic group (hereafter referred to as multiracial) in the United States (US) has significantly increased in the past few decades due, in part, to changes in state/federal laws and US census procedures. Until 1967, interracial marriage was illegal in the US (Livingston & Brown, 2017), potentially reducing the number of actual or reported multiracial births. Further, prior to 2000, the US census required multiracial individuals to select only one racial identity, resulting in an underestimation of the US multiracial population (Jones & Bullock, 2012). Data from the most recent US censuses, both of which allowed respondents to endorse multiple racial/ethnic identities, indicate that the multiracial population increased by one-third between 2000 and 2010 (Jones & Bullock, 2012). According to more recent estimates, multiracial individuals comprise 6.9% of US adults, and 14.3% of US infants (Livingston & Brown, 2017; Parker et al., 2015). Indeed, individuals who identify with multiple racial/ethnic identities are growing rapidly, with an expected 200% population increase in the US by the year 2060 (Vespa et al., 2020).
Processes related to race/ethnicity and mental health risk suggest avenues by which multiracial individuals may be at higher or lower risk for poor mental health outcomes. For example, factors related to identity development (i.e., the process by which one develops a sense of oneself within the context of potentially competing cultural demands), identity integration (i.e., developing a sense that one’s multiple cultural identities are compatible), and acculturation (i.e., the process of learning and adapting to the values, attitudes, and practices of the majority or dominant culture) moderate the risk for poor mental health outcomes among racial/ethnic minorities (Shih & Sanchez, 2005). These processes might be heightened among multiracial individuals, who are likely balancing more than one racial/ethnic identity. Further, minority stress and discrimination, including potential lateral discrimination (i.e., discrimination from members of shared monoracial backgrounds), may be prominent among multiracial individuals (Shih & Sanchez, 2005) and increase risk for eating disorder (ED) pathology through pathways related to stigmatization and negative affect (Rodgers et al., 2017).
Disorder-specific risk factors for ED pathology are also likely to have relevance for, and vary across, distinct multiracial groups. For example, overvaluation of body weight and shape represents a core driver of ED pathology (DuBois et al., 2017; Fairburn et al., 2003). However, differences in acceptance of appearance ideals across certain monoracial groups (i.e., those identifying with a single race/ethnicity) and pressures to achieve dominant appearance ideals appear to be differentially related to ED risk across such groups (Cheng et al., 2019). Moreover, critical sociocultural approaches to ED pathology suggest that dominant US appearance ideals, which include fair skin and Eurocentric features, may uniquely contribute to body dissatisfaction among racial/ethnic groups whose physical features more strongly vary from these ideals (Jankowski et al., 2017). Therefore, the degree to which one’s multiracial background confers greater physical similarity to the dominant appearance ideal may influence body dissatisfaction and ED risk. Further, perceptions of conflicting appearance ideals and pressures stemming from one’s multiple racial/ethnic identities may accentuate this risk (Franko et al., 2012). Thus, individual differences in the endorsement of dominant ideals and the perception of their attainability may create specific risk and protective pathways for ED pathology among multiracial individuals.
Despite the growing multiracial population, there is a dearth of research examining ED pathology across distinct multiracial groups. Among monoracial groups, however, consistent with differences in ED-specific risk factors, existing empirical evidence suggests ED behaviors also vary by race/ethnicity. For example, some research indicates that individuals who identify as non-Hispanic Black (Marques et al., 2011) and as Hispanic/Latinx experience higher rates of EDs characterized by binge-eating/purging patterns than their non-Hispanic White peers (Marques et al., 2011; Swanson et al., 2011). Elevated rates of disordered eating behaviors have also been observed among Asian and Asian American groups (Rodgers et al., 2018). Given this, it is possible that levels of ED pathology may also vary across distinct multiracial groups.
Unfortunately, multiracial individuals are rarely the focus of study in ED research and are often excluded from analyses, possibly due to low sample sizes. When included in studies, multiracial individuals are often grouped together, potentially obscuring meaningful group differences in eating attitudes and behaviors, and contributing to mixed findings regarding levels of ED risk and prevalence (e.g., Goel et al., 2020; Ivezaj et al., 2010; le Grange et al., 2006). Importantly, research in other domains of psychological functioning indicates that health outcomes vary considerably across different multiracial groups (Campbell & Eggerling-Boeck, 2006; Charmaraman et al., 2014), highlighting the need for similar work to clarify the extent to which ED prevalence differs across distinct multiracial groups. These findings could be used to identify highly vulnerable groups and to appropriately tailor and orient intervention resources.
Importantly, intersectional approaches, which originated in the work of Black feminists and critical race theorists (e.g., Beale, 1970; Crenshaw, 1993), have demonstrated great utility for considering the ways in which one’s multiple identities may operate in concert to increase or decrease risk for outcomes of interest, including ED pathology (Beccia et al., 2019; Burke et al., 2020). The intersectional lens acknowledges that the impact of one aspect of identity on illness risk may be moderated by other elements of one’s identity. Thus, intersectional approaches offer a meaningful framework to explore the influence of holding multiple racial/ethnic identities on ED pathology. Further, although evidence indicates higher prevalence of ED pathology among women compared to men (Hudson et al., 2007) and suggests that the risk associated with holding a female gender identity may be compounded by also holding a minority racial/ethnic identity for some monoracial groups (Beccia et al., 2019), we are not aware of any studies examining the intersecting roles of unique multiracial identities and gender identity on ED prevalence.
To address this gap in the literature, the current study will be among the first to examine prevalence estimates (hereafter referred to as prevalence) of ED pathology across several distinct multiracial groups. Considering evidence that particular multiracial groups are likely to affiliate more with one race versus another (Campbell, 2007), this study will also contrast prevalence of ED pathology in each multiracial group with those among the corresponding monoracial backgrounds. Given known differences in prevalence of ED pathology across gender identity (Hudson et al., 2007), intersectional analyses by gender will also be conducted.
2. Method
2.1. Study Design
The Healthy Minds Study (HMS) is an annual, web-based survey about mental health in undergraduate and graduate students (Eisenberg & Lipson, 2019). Five waves of data were used for the present study, collected from 199 US colleges/universities between 2014–2019. For institutions that participated more than once across these waves (n=25), only data from the most recent wave were used. Institutional enrollment was voluntary. At larger institutions, a random 4,000-student sample was invited to participate, and all students were invited at smaller institutions. Students were recruited via email and informed that they were eligible to win one of ten $100 or two $500 gift cards regardless of participation. Students had to be ≥18 years old to participate. All respondents provided informed consent. Research was approved by Institutional Review Boards at participating institutions.
Response rates were 23% in 2014-2015, 27% in 2015-2016, 23% in 2016-2017 and 2017-2018, and 16% in 2018-2019. To account for non-response bias, sample probability weights were constructed based on gender, race/ethnicity, academic level, and grade point average, administrative data obtained from participating institutions for all students invited to participate in the study. These data were utilized to construct response weights, equal to 1 divided by the predicted probability of survey response, using a logistic regression to estimate predicted response probability based on these variables. Thus, weights were larger for respondents with underrepresented characteristics, ensuring estimates represented the full undergraduate and graduate student population in terms of these characteristics.
2.2. Participants
Due to inadequate statistical power for planned interaction analyses, groups with fewer than 100 respondents each and those identifying with more than two racial/ethnic identities were not included in analyses. Utilizing data from respondents who provided information on all variables of interest and represented monoracial or multiracial groups with at least 100 respondents each, the analytic sample includes 145,379 undergraduate and graduate students. Sample characteristics are reported in Table 1.
Table 1.
Sample characteristics
| % (n) | |
|---|---|
| Number of racial/ethnic identities | |
| One (“monoracial”) | 92.3 (133,946) |
| Two (“multiracial”) | 7.7 (11,433) |
| Gender identity | |
| Cisgender male | 40.7 (44,388) |
| Cisgender female | 56.8 (98,008) |
| Gender minority | 2.4 (2,983) |
| Age | |
| 18–22 years | 68.2 (96,524) |
| 23–25 years | 12.0 (19,300) |
| 26–30 years | 9.1 (15,563) |
| 31+ years | 10.7 (13,992) |
| Degree level | |
| Undergraduate | 84.2 (108,802) |
| Graduate | 15.8 (32,390) |
| First-generation college student | |
| No | 63.0 (96,055) |
| Yes | 37.0 (45,601) |
| Positive SCOFF screen | 23.5 (34,735) |
Note. First-generation college student indicates that neither parent has a bachelor’s degree. Frequencies represent observed counts; percentages are weighted to account for non-response.
2.3. Measures
2.3.1. Racial/ethnic identity.
Respondents self-identified as White, African American/Black, Hispanic/Latinx, Asian American/Asian, Native Hawaiian/Pacific Islander, American Indian/Alaskan Native, Middle Eastern/Arab/Arab American, and/or as “other” racial/ethnic identity. Respondents could select as many response options as were applicable. In the 2015-2016 survey, race and ethnicity were assessed via separate questions, with Hispanic/Latinx assessed via the ethnicity question. Other survey years, race and ethnicity were assessed simultaneously via one race/ethnicity question. If participants selected “other” as a response (n = 2,977), they could write in their racial/ethnic identity. Certain “other” responses were recoded back into the original racial/ethnic identity mappings (n = 288). More specifically, participants who selected “other” and wrote in a racial/ethnic identity that matched the identity language used within the original question were recoded into the original identity that their response mapped onto (i.e., if an individual wrote in “Latina”, they would be recoded into the “Hispanic/Latinx” race/ethnicity identity). Participants who selected “other” but indicated their identity status as affiliated with a country, region, or category outside the exact answer choices provided (e.g., “American,” “North Indian,” “Native American,” etc.)1, or, indicated they were “biracial” or “multiracial” without writing in their individual identity statuses remained coded within the “other” identity (n = 2,689).
2.3.2. Gender identity.
Gender identity was defined by the following three categories: cisgender men, cisgender women, and gender minorities. The gender minority category was comprised of students who identified as transgender men, transgender women, genderqueer/gender non-conforming, and other gender identity.
2.3.3. Eating disorder pathology.
ED pathology was assessed with the five-item SCOFF (Morgan et al., 1999), which is a widely used ED screening tool. Item response options are dichotomous (yes=1, no=0), with total scores ranging from 0 to 5. The cut-off for a positive screen (i.e., likely ED diagnosis) was ≥2 affirmative responses, which has been determined to yield the optimal trade-off between sensitivity and specificity (Mond et al., 2008).
2.4. Statistical Analysis
All analyses were conducted in Stata 16.0 and incorporated sample probability weights to account for non-response. Observed prevalence for positive SCOFF screens were computed for monoracial and multiracial groups. Expected prevalence for multiracial groups assuming additive risk based on each monoracial identity were predicted via log-binomial regression models among all monoracial respondents. Expected prevalence estimates assuming additive risk were calculated as the exponentiated sum of the intercept and each monoracial identity coefficient from the log-binomial regression models. Linear binomial models were conducted to test for differences between the observed and expected prevalence for multiracial groups by including interaction terms between the two corresponding racial/ethnic identities, with a significant interaction indicating a departure from additivity (Naimi & Whitcomb, 2020). For example, to test the difference between observed and expected prevalence for multiracial African American/Black and White respondents, the following three predictor terms were included in the model: (1) African American/Black, (2) White, and (3) African American/Black x White. All other monoracial identities served as the reference group. A positive interaction indicated that the observed prevalence was greater than the expected prevalence, whereas a negative interaction indicated that the observed prevalence was lower than the expected prevalence. Analyses were conducted in the full sample, and also stratified by gender identity. Three-way interaction terms with gender identity were tested among cisgender participants, but inadequate statistical power precluded our ability to examine interactions among gender minorities.
3. Results
Full-sample prevalence of positive SCOFF screens by racial/ethnic identity are reported in Table 2. When comparing observed prevalence to expected prevalence assuming additive risk based on each monoracial identity, multiracial groups identifying as African American/Black & White (interaction B [Bint]=0.12, p<.001), American Indian/Alaskan Native & White (Bint=0.04, p=.04), African American/Black & Hispanic/Latinx (Bint=0.05, p=.02), African American/Black & Asian American/Asian (Bint=0.13, p=.002), and American Indian/Alaskan Native & Hispanic/Latinx (Bint=0.13, p=.005) each demonstrated greater prevalence than expected. Conversely, multiracial groups identifying as Asian American/Asian & White (Bint=−0.05, p<.001), Middle Eastern/Arab/Arab American & White (Bint=−0.07, p<.001), and Asian American/Asian & Hispanic/Latinx (Bint=−0.07, p=.03) demonstrated lower prevalence than expected assuming additive risk based on each monoracial identity.
Table 2.
Prevalence of positive SCOFF screens by racial/ethnic identity in full sample
| Positive SCOFF Screen |
||||
|---|---|---|---|---|
| N | Observed Prevalence, % (n) | Expected Prevalence Based on Each Monoracial Identity, % | Obs. versus Exp. Prevalence | |
| Monoracial groups | ||||
| White | 98,062 | 22.6 (22,424) | -- | -- |
| Asian American/Asian | 17,585 | 28.2 (4,931) | -- | -- |
| Hispanic/Latinx | 8,441 | 27.9 (2,283) | -- | -- |
| African American/Black | 7,379 | 18.4 (1,404) | -- | -- |
| Middle Eastern/Arab/Arab American | 1,818 | 33.0 (581) | -- | -- |
| American Indian/Alaskan Native | 472 | 23.8 (134) | -- | -- |
| Native Hawaiian/Pacific Islander | 189 | 33.9 (68) | -- | -- |
| Multiracial groups | ||||
| Hispanic/Latinx & White | 4,366 | 25.3 (1,124) | 25.3 | n.s. |
| Asian American/Asian & White | 2,278 | 22.2 (511) | 26.7 | Obs. < Exp.*** |
| African American/Black & White | 1,241 | 24.4 (314) | 14.6 | Obs. > Exp.*** |
| American Indian/Alaskan Native & White | 1,194 | 25.2 (312) | 20.9 | Obs. > Exp.* |
| Middle Eastern/Arab/Arab American & White | 795 | 23.4 (199) | 29.5 | Obs. < Exp.*** |
| African American/Black & Hispanic/Latinx | 464 | 27.5 (125) | 21.8 | Obs. > Exp.* |
| Native Hawaiian/Pacific Islander & White | 248 | 32.2 (69) | 29.9 | n.s. |
| Asian American/Asian & Hispanic/Latinx | 230 | 27.1 (60) | 35.2 | Obs. < Exp.* |
| Asian American/Asian & Native Hawaiian/Pacific Islander | 212 | 33.8 (76) | 42.0 | n.s. |
| African American/Black & American Indian/Alaskan Native | 145 | 16.5 (33) | 18.3 | n.s. |
| African American/Black & Asian American/Asian | 135 | 36.1 (43) | 22.3 | Obs. > Exp.** |
| American Indian/Alaskan Native & Hispanic/Latinx | 125 | 41.4 (44) | 28.7 | Obs. > Exp.** |
Note. Obs. = observed; Exp. = expected; n.s. = not significant. Far-right column indicates whether observed prevalence of positive SCOFF screens in multiracial group differs from expected prevalence assuming additive risk based on each corresponding monoracial identity,
p < .05
p < .01
p < .001.
Frequencies represent observed counts; percentages are weighted to account for non-response.
Prevalence of positive SCOFF screens by racial/ethnic identity among cisgender men are reported in Table 3. When comparing observed prevalence to expected prevalence assuming additive risk based on each monoracial identity, multiracial groups identifying as African American/Black & White (Bint=0.06, p=.001), African American/Black & Hispanic/Latinx (Bint=0.09, p=.02), and Native Hawaiian/Pacific Islander & White (Bint=0.16, p=.047) each demonstrated greater prevalence than expected. Conversely, multiracial respondents identifying as Middle Eastern/Arab/Arab American & White (Bint=−0.08, p=.004) demonstrated lower prevalence than expected assuming additive risk based on each monoracial identity.
Table 3.
Prevalence of positive SCOFF screens by racial/ethnic identity among cisgender men
| Positive SCOFF Screen |
||||
|---|---|---|---|---|
| N | Observed Prevalence, % (n) | Expected Prevalence Based on Each Monoracial Identity, % | Obs. versus Exp. Prevalence | |
| Monoracial groups | ||||
| White | 29,586 | 14.3 (3,909) | -- | -- |
| Asian American/Asian | 6,053 | 20.6 (1,146) | -- | -- |
| Hispanic/Latinx | 2,482 | 22.0 (494) | -- | -- |
| African American/Black | 2,011 | 12.4 (227) | -- | -- |
| Middle Eastern/Arab/Arab American | 736 | 28.4 (183) | -- | -- |
| American Indian/Alaskan Native | 134 | 16.3 (24) | -- | -- |
| Native Hawaiian/Pacific Islander | 50 | 19.4 (11) | -- | -- |
| Multiracial groups | ||||
| Hispanic/Latinx & White | 1,338 | 18.4 (210) | 17.2 | n.s. |
| Asian American/Asian & White | 667 | 14.1 (90) | 16.2 | n.s. |
| African American/Black & White | 295 | 11.3 (39) | 8.2 | Obs. > Exp.** |
| American Indian/Alaskan Native & White | 348 | 16.4 (49) | 12.1 | n.s. |
| Middle Eastern/Arab/Arab American & White | 285 | 15.8 (46) | 21.9 | Obs. < Exp.** |
| African American/Black & Hispanic/Latinx | 113 | 27.8 (30) | 17.7 | Obs. > Exp.* |
| Native Hawaiian/Pacific Islander & White | 60 | 30.2 (15) | 14.5 | Obs. > Exp.* |
| Asian American/Asian & Hispanic/Latinx | 63 | 19.4 (10) | 31.3 | n.s. |
| Asian American/Asian & Native Hawaiian/Pacific Islander | 67 | 32.3 (21) | 26.7 | n.s. |
| African American/Black & American Indian/Alaskan Native | 27 | 9.6 (5) | 12.7 | n.s. |
| African American/Black & Asian American/Asian | 36 | 27.2 (8) | 16.8 | n.s. |
| American Indian/Alaskan Native & Hispanic/Latinx | 37 | 29.8 (10) | 23.5 | n.s. |
Note. Obs. = observed; Exp. = expected; n.s. = not significant. Far-right column indicates whether observed prevalence of positive SCOFF screens in multiracial group differs from expected prevalence assuming additive risk based on each corresponding monoracial identity,
p < .05
p < .01
p < .001.
Frequencies represent observed counts; percentages are weighted to account for non-response.
Prevalence of positive SCOFF screens by racial/ethnic identity among cisgender women are reported in Table 4. When comparing observed prevalence to expected prevalence assuming additive risk based on each monoracial identity, multiracial groups identifying as African American/Black & White (Bint=0.13, p<.001) and American Indian/Alaskan Native & Hispanic/Latinx (Bint=0.17, p=.001) each demonstrated greater prevalence than expected. Conversely, multiracial groups identifying as Asian American/Asian & White (Bint=−0.08, p<.001), Middle Eastern/Arab/Arab American & White (Bint=−0.06, p=.02), and Asian American/Asian & Native Hawaiian/Pacific Islander (Bint=−0.12, p<.048) demonstrated lower prevalence than expected assuming additive risk based on each monoracial identity.
Table 4.
Prevalence of positive SCOFF screens by racial/ethnic identity among cisgender women
| Positive SCOFF Screen |
||||
|---|---|---|---|---|
| N | Observed Prevalence, % (n) | Expected Prevalence Based on Each Monoracial Identity, % | Obs. versus Exp. Prevalence | |
| Monoracial groups | ||||
| White | 66,310 | 28.0 (17,766) | -- | -- |
| Asian American/Asian | 11,314 | 33.9 (3,704) | -- | -- |
| Hispanic/Latinx | 5,814 | 31.4 (1,739) | -- | -- |
| African American/Black | 5,268 | 22.0 (1,139) | -- | -- |
| Middle Eastern/Arab/Arab American | 1,067 | 37.6 (391) | -- | -- |
| American Indian/Alaskan Native | 332 | 28.0 (109) | -- | -- |
| Native Hawaiian/Pacific Islander | 135 | 41.9 (54) | -- | -- |
| Multiracial groups | ||||
| Hispanic/Latinx & White | 2,914 | 29.1 (873) | 29.9 | n.s. |
| Asian American/Asian & White | 1,539 | 26.3 (394) | 34.4 | Obs. < Exp.*** |
| African American/Black & White | 903 | 30.1 (258) | 18.6 | Obs. > Exp.*** |
| American Indian/Alaskan Native & White | 810 | 29.6 (247) | 26.1 | n.s. |
| Middle Eastern/Arab/Arab American & White | 492 | 29.6 (150) | 35.5 | Obs. < Exp.* |
| African American/Black & Hispanic/Latinx | 338 | 25.7 (87) | 24.0 | n.s. |
| Native Hawaiian/Pacific Islander & White | 179 | 35.6 (53) | 39.2 | n.s. |
| Asian American/Asian & Hispanic/Latinx | 162 | 33.6 (49) | 38.7 | n.s. |
| Asian American/Asian & Native Hawaiian/Pacific Islander | 139 | 36.3 (54) | 51.0 | Obs. < Exp.* |
| African American/Black & American Indian/Alaskan Native | 112 | 17.1 (26) | 21.2 | n.s. |
| African American/Black & Asian American/Asian | 93 | 37.5 (34) | 26.3 | n.s. |
| American Indian/Alaskan Native & Hispanic/Latinx | 87 | 48.6 (33) | 31.1 | Obs. > Exp.** |
Note. Obs. = observed; Exp. = expected; n.s. = not significant. Far-right column indicates whether observed prevalence of positive SCOFF screens in multiracial group differs from expected prevalence assuming additive risk based on each corresponding monoracial identity,
p < .05
p < .01
p < .001.
Frequencies represent observed counts; percentages are weighted to account for non-response.
The discrepancies between observed and expected prevalence of positive SCOFF screens differed significantly between cisgender men and women among multiracial Asian American/Asian & White (p=.007), Native Hawaiian/Pacific Islander & White (p=.03), and African American/Black & White groups (p=.003). Multiracial Asian American/Asian & White respondents demonstrated lower prevalence than expected in cisgender women but not men, whereas multiracial Native Hawaiian/Pacific Islander & White respondents demonstrated greater prevalence than expected in cisgender men but not women. Multiracial African American/Black & White respondents demonstrated greater prevalence than expected in both cisgender men and women, but the discrepancy between observed and expected prevalence was greater among cisgender women than men.
Prevalence of positive SCOFF screens by racial/ethnic identity among gender minorities are reported in Table 5. However, comparisons between observed and expected prevalence were not formally tested due to small group sizes.
Table 5.
Prevalence of positive SCOFF screens by racial/ethnic identity among gender minorities
| Positive SCOFF Screen |
|||
|---|---|---|---|
| N | Observed Prevalence, % (n) | Expected Prevalence Based on Each Monoracial Identity, % | |
| Monoracial groups | |||
| White | 2,166 | 35.5 (749) | -- |
| Asian American/Asian | 218 | 38.2 (81) | -- |
| Hispanic/Latinx | 145 | 34.0 (50) | -- |
| African American/Black | 100 | 35.6 (38) | -- |
| Middle Eastern/Arab/Arab American | 15 | 58.5 (7) | -- |
| American Indian/Alaskan Native | 6 | 19.7 (1) | -- |
| Native Hawaiian/Pacific Islander | 4 | 66.6 (3) | -- |
| Multiracial groups | |||
| Hispanic/Latinx & White | 114 | 38.2 (41) | 31.8 |
| Asian American/Asian & White | 72 | 45.8 (27) | 37.9 |
| African American/Black & White | 43 | 43.3 (17) | 34.1 |
| American Indian/Alaskan Native & White | 36 | 53.6 (16) | 18.9 |
| Middle Eastern/Arab/Arab American & White | 18 | 20.3 (3) | 57.9 |
| African American/Black & Hispanic/Latinx | 13 | 73.5 (8) | 33.8 |
| Native Hawaiian/Pacific Islander & White | 9 | 3.5 (1) | 64.8 |
| Asian American/Asian & Hispanic/Latinx | 5 | 11.7 (1) | 36.4 |
| Asian American/Asian & Native Hawaiian/Pacific Islander | 6 | 3.0 (1) | 71.5 |
| African American/Black & American Indian/Alaskan Native | 6 | 41.4 (2) | 19.6 |
| African American/Black & Asian American/Asian | 6 | 60.8 (1) | 38.2 |
| American Indian/Alaskan Native & Hispanic/Latinx | 1 | 100.0 (1) | 18.7 |
Note. Comparisons between observed and expected prevalence estimates were not formally tested due to small group sizes. Frequencies represent observed counts; percentages are weighted to account for non-response.
4. Discussion
Given the gap in available research focusing on multiracial individuals, and the theoretical and budding empirical support for specific and sometimes elevated mental health risk among such individuals, the present study investigated ED prevalence among a large sample of multiracial individuals. Specifically, the objectives of the present study were to examine prevalence of clinically-significant ED pathology across several distinct multiracial groups, compare the prevalence against those in the corresponding monoracial groups, and investigate differences by gender identity. Overall, our findings confirm the necessity of increased research attention to multiracial individuals and the importance of improving our understanding of the specific patterns of ED prevalence associated with different multiracial identities. These findings have important implications for orienting future research efforts, intervention, and resource allocation.
In support of considering the heterogeneity among multiracial groups, substantial differences in clinically-significant ED pathology by multiracial groups emerged. Specifically, though small, the multiracial group identifying as American Indian/Alaskan Native and Hispanic/Latinx exhibited the highest prevalence of clinically-significant ED pathology across all multiracial and monoracial groups examined, and was greater than expected based on the prevalence observed in monoracial American Indian/Alaskan Native and Hispanic/Latinx groups. In other words, the observed levels of ED pathology in multiracial American Indian/Alaskan Native and Hispanic/Latinx individuals were higher than the additive combination of the risk conferred by both of those groups together. This provides evidence that this dual identity has its own specific implications for ED pathology, above and beyond that associated with separate American Indian/Alaskan Native and Hispanic/Latinx identities. Similar patterns were observed for some other doubly marginalized groups — those identifying as African American/Black and Hispanic/Latinx and those identifying as African American/Black and Asian American/Asian — who also exhibited higher prevalence of clinically-significant ED pathology than expected, whereas others did not. In multiracial individuals identifying as White and racial/ethnic minority, higher than expected prevalence of clinically-significant ED pathology were observed for some groups (African American/Black and White; American Indian/Alaskan Native and White) whose corresponding minority monoracial groups demonstrated relatively low or average prevalence of clinically-significant ED pathology.
Although these findings support the theory that multiple marginalized identities may confer greater risk of negative outcomes when combined, other groups with dual identities emerged as possessing lower levels of ED prevalence that might have been expected based on each single identity. Thus, lower than expected prevalence of clinically-significant ED pathology were observed for some groups (Asian American/Asian and White; Middle Eastern/Arab/Arab American and White) whose corresponding minority monoracial groups demonstrated relatively high prevalence of clinically-significant ED pathology. Although some of these patterns differed by gender identity, findings of higher than expected prevalence among multiracial individuals identifying as African American/Black and White and lower than expected prevalence among multiracial individuals identifying as Middle Eastern/Arab/Arab American and White were consistent across cisgender men and women. Overall, these findings emphasize the importance of considering intersecting racial/ethnic identities and examining distinct multiracial groups separately in ED research.
Together, these findings have important implications. First, the current study adds to the scant data on prevalence of clinically-significant ED pathology among multiracial individuals by reporting on varying prevalence within different multiracial groups. Given the paucity of data in this area, this is an important and significant contribution. In addition, the finding that some combined identities potentiate ED risk, whereas others may reduce the risk associated with individual identities, supports the utility of intersectional approaches as a method of clarifying the complex ways in which held racial/ethnic and gender identities may modulate ED risk. Although previous research has broadly suggested that racial/ethnic minority identities are associated with elevated ED risk (Rodgers et al., 2018), our findings point to the critical need to develop a more fine-grained appreciation of the ways in which identities combine to influence ED outcomes. Importantly, however, and despite the overarching theoretical models outlined previously, the present findings shed no light on the mechanisms underpinning differing levels of ED prevalence. Increasing our understanding of the pathways through which multiracial individuals may be protected or more vulnerable to ED pathology is an important future direction.
Although highlighting the need for a more detailed and nuanced understanding of multiracial identity related to ED prevalence, our findings overall suggest structural factors and inequalities related to identity (visible and or privately held) likely contribute to levels of ED prevalence among multiracial individuals. The present findings should be approached through a lens of health equity, with a view to better identifying groups that may be most vulnerable, and continuing to work towards better understanding these differences, identifying systemic factors that contribute to them, and advocating for appropriate allocation of resources.
The current study possesses notable strengths including the large overall sample size, which facilitated an examination of ED pathology among distinct multiracial groups who may commonly be overlooked in smaller datasets. Further, ED pathology was assessed via the SCOFF, which is a validated measure of disordered eating symptoms (Morgan et al., 1999). Using an undergraduate and graduate student sample represents a notable strength, as multiracial individuals in the US are, on average, significantly younger than the general population (Livingston & Brown, 2017; Parker et al., 2015), and the university years represent an important developmental period with regard to ED risk (Grammer et al., 2020) and identity growth and formation (Kaufman & Feldman, 2004). Although sample size limited our ability to formally examine observed versus anticipated prevalence of ED pathology among gender minorities, clarification of the ways in which gender identity may influence ED prevalence among multiracial/ethnic groups is a considerable strength. Limitations of the present study suggest important avenues for future work. Although prior research suggests that ethnic identity, discriminatory stress, acculturative stress, appearance pressures, and appearance ideal internalization may impact ED risk among ethnic minorities (Cheng et al., 2019; Higgins Neyland & Bardone-Cone, 2017; Kwan et al., 2018; Ordaz et al., 2018; Rakhkovskaya & Warren, 2014), we were unable to consider the potential moderating role of these experiences in the current study. This would be a valuable goal for future research, in addition to broadening the investigation of risk for multiracial individuals in similarly aged non-college populations and across the developmental spectrum more broadly. Finally, although this study organized participants into groups identified by their multiracial background, significant heterogeneity within those groups exists. Notably, intersectionality approaches encourage clarification of the ways in which multiple aspects of identity (e.g., age, sexual orientation, weight status) may synergistically influence risk. Although power constraints and sample sizes for subgroups limited our ability to examine additional facets of identity or those with more than two racial/ethnic identities in the current study, future work is needed to explore the impact of additional domains and aspects of identity and experience on ED risk in multiracial individuals.
In sum, the current study represents the first known attempt to clarify prevalence of ED pathology among distinct multiracial groups, and to examine how these multiracial identities intersect with gender identity. Observed variation in clinically-significant ED pathology across examined groups underscores the importance of distinguishing between unique multiracial identities, as some identities appear to confer heightened risk for disordered eating, whereas other identities may offer protection. Findings from the current study may elucidate mixed results from previous research that collapsed all multiracial individuals into a single category (Goel et al., 2020; Ivezaj et al., 2010; le Grange et al., 2006). As the number of individuals identifying as multiracial will rise (Vespa et al., 2020), and research indicates that culturally-adapted interventions result in improved efficacy (Hall et al., 2016), continued work is needed to identity the key mechanisms underlying increased ED pathology risk for certain multiracial identities. Such mechanisms might then be targeted in culturally-adapted prevention and treatment efforts.
Acknowledgments
Funding Statement:
This research was supported in part by the National Institute of Mental Health (grant numbers T32MH082761 (VMH) and K01MH121515 (SKL)) and the William T. Grant Scholars Program (SKL).
Footnotes
Conflicts of interest: The authors have no conflicts to declare.
We chose this conservative approach to honor participants’ implicit desire to be identified in a way other than what was available through the survey response choices. In this way, we have allowed their chosen ways of identifying to remain versus assigning them an identity they did not assign themselves.
References
- Beale F (1970). Double jeopardy: To be black and female. In Cade T (Ed.), The black woman: An anthology (pp. 90–100). Signet. [Google Scholar]
- Beccia AL, Baek J, Jesdale WM, Austin SB, Forrester S, Curtin C, & Lapane KL (2019). Risk of disordered eating at the intersection of gender and racial/ethnic identity among U.S. High school students. Eating Behaviors, 34, 101299. 10.1016/j.eatbeh.2019.05.002 [DOI] [PubMed] [Google Scholar]
- Burke NL, Schaefer LM, Hazzard VM, & Rodgers RF (2020). Where identities converge: The importance of intersectionality in eating disorders research. International Journal of Eating Disorders, 53(10), 1605–1609. 10.1002/eat.23371 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Campbell ME (2007). Thinking outside the (black) box: Measuring black and multiracial identification on surveys. Social Science Research, 36(3), 921–944. 10.1016/j.ssresearch.2006.07.001 [DOI] [Google Scholar]
- Campbell ME, & Eggerling-Boeck J (2006). “What about the children?” The psychological and social well-being of multiracial adolescents. The Sociological Quarterly, 47(1), 147–173. 10.1111/j.1533-8525.2006.00041.x [DOI] [Google Scholar]
- Charmaraman L, Woo M, Quach A, & Erkut S (2014). How have researchers studied multiracial populations? A content and methodological review of 20 years of research. Cultural Diversity & Ethnic Minority Psychology, 20(3), 336–352. 10.1037/a0035437 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cheng ZH, Perko VL, Fuller-Marashi L, Gau JM, & Stice E (2019). Ethnic differences in eating disorder prevalence, risk factors, and predictive effects of risk factors among young women. Eating Behaviors, 32, 23–30. 10.1016/j.eatbeh.2018.11.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Crenshaw K (1993). Demarginalizing the intersection of race and sex: A black feminist critique of antidiscrimination doctrine, feminist theory and antiracist politics. In Weisbert DK (Ed.), Feminist Legal Theory: Foundations (pp. 383–395). Temple University Press. (Original work published 1989) [Google Scholar]
- DuBois RH, Rodgers RF, Franko DL, Eddy KT, & Thomas JJ (2017). A network analysis investigation of the cognitive-behavioral theory of eating disorders. Behaviour Research and Therapy, 97, 213–221. 10.1016/j.brat.2017.08.004 [DOI] [PubMed] [Google Scholar]
- Eisenberg D, & Lipson SK (2019). The Healthy Minds Study. [DOI] [PubMed] [Google Scholar]
- Fairburn CG, Cooper Z, & Shafran R (2003). Cognitive behaviour therapy for eating disorders: A “transdiagnostic” theory and treatment. Behaviour Research and Therapy, 41(5), 509–528. 10.1016/s0005-7967(02)00088-8 [DOI] [PubMed] [Google Scholar]
- Franko DL, Coen EJ, Roehrig JP, Rodgers RF, Jenkins A, Lovering ME, & Dela Cruz S (2012). Considering J.Lo and Ugly Betty: A qualitative examination of risk factors and prevention targets for body dissatisfaction, eating disorders, and obesity in young Latina women. Body Image, 9(3), 381–387. 10.1016/j.bodyim.2012.04.003 [DOI] [PubMed] [Google Scholar]
- Goel NJ, Burnette CB, & Mazzeo SE (2020). Racial and ethnic differences in the association between parent-oriented perfectionism and disordered eating in college women. International Journal of Eating Disorders, 53(2), 191–200. 10.1002/eat.23179 [DOI] [PubMed] [Google Scholar]
- Grammer AC, Fitzsimmons-Craft EE, Laing O, Pietro BD, & Wilfley DE (2020). Eating disorders on college campuses in the United States: Current insight on screening, prevention, and treatment. Current Psychopharmacology, 9(2), 91–102. 10.2174/2211556009999200416153022 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hall GC, Ibaraki AY, Huang ER, Marti CN, & Stice E (2016). A meta-analysis of cultural adaptations of psychological interventions. Behavior Therapy, 47(6), 993–1014. 10.1016/j.beth.2016.09.005 [DOI] [PubMed] [Google Scholar]
- Higgins Neyland MK, & Bardone-Cone AM (2017). Tests of escape theory of binge eating among latinas. Cultural Diversity & Ethnic Minority Psychology, 23(3), 373–381. 10.1037/cdp0000130 [DOI] [PubMed] [Google Scholar]
- Hudson JI, Hiripi E, Pope HG Jr., & Kessler RC (2007). The prevalence and correlates of eating disorders in the national comorbidity survey replication. Biological Psychiatry, 61(3), 348–358. 10.1016/j.biopsych.2006.03.040 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ivezaj V, Saules KK, Hoodin F, Alschuler K, Angelella NE, Collings AS, Saunders-Scott D, & Wiedemann AA (2010). The relationship between binge eating and weight status on depression, anxiety, and body image among a diverse college sample: A focus on bi/multiracial women. Eating Behaviors, 11(1), 18–24. 10.1016/j.eatbeh.2009.08.003 [DOI] [PubMed] [Google Scholar]
- Jankowski GS, Tshuma S, Tshuma S, & Hylton M (2017). Light except Lupita: The representation of black women in magazines. Psychology of Women Section Review. http://eprints.leedsbeckett.ac.uk/4016/ [Google Scholar]
- Jones NA, & Bullock J (2012). The two or more races population: 2010 (2010 Census Briefs, C2010BR-13). United States Census Bureau. https://www.census.gov/prod/cen2010/briefs/c2010br-13.pdf [Google Scholar]
- Kaufman P, & Feldman KA (2004). Forming identities in college: A sociological approach. Research in Higher Education, 45(5), 463–496. 10.1023/B:RIHE.0000032325.56126.29 [DOI] [Google Scholar]
- Kwan MY, Gordon KH, & Minnich AM (2018). An examination of the relationships between acculturative stress, perceived discrimination, and eating disorder symptoms among ethnic minority college students. Eating Behaviors, 28, 25–31. 10.1016/j.eatbeh.2017.12.003 [DOI] [PubMed] [Google Scholar]
- le Grange D, Louw J, Russell B, Nel T, & Silkstone C (2006). Eating attitudes and behaviours in South African adolescents and young adults. Transcultural Psychiatry, 43(3), 401–417. 10.1177/1363461506066984 [DOI] [PubMed] [Google Scholar]
- Livingston G, & Brown A (2017). Intermarriage in the U.S. 50 years after Loving v. Virginia. Pew Research Center. https://www.pewsocialtrends.org/wp-content/uploads/sites/3/2017/05/Intermarriage-May-2017-Full-Report.pdf [Google Scholar]
- Marques L, Alegria M, Becker AE, Chen CN, Fang A, Chosak A, & Diniz JB (2011). Comparative prevalence, correlates of impairment, and service utilization for eating disorders across us ethnic groups: Implications for reducing ethnic disparities in health care access for eating disorders. International Journal of Eating Disorders, 44(5), 412–420. 10.1002/eat.20787 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mond JM, Myers TC, Crosby RD, Hay PJ, Rodgers B, Morgan JF, Hubert Lacey J, & Mitchell JE (2008). Screening for eating disorders in primary care: EDE-Q versus SCOFF. Behaviour Research and Therapy, 46(5), 612–622. 10.1016/j.brat.2008.02.003 [DOI] [PubMed] [Google Scholar]
- Morgan JF, Reid F, & Lacey JH (1999). The SCOFF questionnaire: Assessment of a new screening tool for eating disorders. BMJ (Clinical research ed.), 319(7223), 1467–1468. 10.1136/bmj.319.7223.1467 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Naimi AI, & Whitcomb BW (2020). Estimating risk ratios and risk differences using regression. American Journal of Epidemiology, 189(6), 508–510. 10.1093/aje/kwaa044 [DOI] [PubMed] [Google Scholar]
- Ordaz DL, Schaefer LM, Choquette E, Schueler J, Wallace L, & Thompson JK (2018). Thinness pressures in ethnically diverse college women in the United States. Body Image, 24, 1–4. 10.1016/j.bodyim.2017.11.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Parker K, Horowitz JM, Morin R, & Lopez MH (2015). Multiracial in america: Proud, diverse, and growing in numbers https://www.pewsocialtrends.org/2015/06/11/multiracial-in-america/
- Rakhkovskaya LM, & Warren CS (2014). Ethnic identity, thin-ideal internalization, and eating pathology in ethnically diverse college women. Body Image, 11(4), 438–445. 10.1016/j.bodyim.2014.07.003 [DOI] [PubMed] [Google Scholar]
- Rodgers RF, Berry R, & Franko DL (2018). Eating disorders in ethnic minorities: An update. Current Psychiatry Reports, 20(10), 90. 10.1007/s11920-018-0938-3 [DOI] [PubMed] [Google Scholar]
- Rodgers RF, Watts AW, Austin SB, Haines J, & Neumark-Sztainer D (2017). Disordered eating in ethnic minority adolescents with overweight. International Journal of Eating Disorders, 50(6), 665–671. 10.1002/eat.22652 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Shih M, & Sanchez DT (2005). Perspectives and research on the positive and negative implications of having multiple racial identities. Psychological Bulletin, 131(4), 569–591. 10.1037/0033-2909.131.4.569 [DOI] [PubMed] [Google Scholar]
- Swanson SA, Crow SJ, Le Grange D, Swendsen J, & Merikangas KR (2011). Prevalence and correlates of eating disorders in adolescents. Results from the national comorbidity survey replication adolescent supplement. Archives of General Psychiatry, 68(7), 714–723. 10.1001/archgenpsychiatry.2011.22 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Vespa J, Medina L, & Armstrong DM (2020). Demographic turning points for the United States: Population projections for 2020 to 2060 (Current Population Reports, P25-1144). United States Census Bureau. https://www.census.gov/content/dam/Census/library/publications/2020/demo/p25-1144.pdf [Google Scholar]