Cultural consonance, body image, and disordered eating among young South Korean men

Abstract

Eating disorders are serious mental illnesses, but little research explores non-Western men’s cultural experiences of body image and what affects their risks of disordered eating. Drawing on data collected over 17 months (August 2019 to January 2021) of fieldwork in Seoul, South Korea, the lens of intersectionality is employed alongside multiple regression and moderation analysis to understand how two axes of identity which emerged as important from the ethnography—sexual identity and university prestige—shape the ways in which young Korean men’s cultural consonance with their local model of the ideal male body, influenced heavily by the kkonminam (flower boy), relates to risk for developing an eating disorder. Among young Korean men, intersections of university prestige and sexual identity frame embodiment of cultural models of male body image as a strategy for the making and maintenance of social relations and the advancement of social status in a precarious neoliberal economy.

1. Introduction

Eating disorders are among the most serious mental illnesses (Arcelus et al., 2011), and can result in debilitating, long-term medical complications (Cass et al., 2020). Underdiagnosis and undertreatment due to racialized and gendered biases (Becker et al., 2010; Räisänen and Hunt, 2014) and barriers created by for-profit health insurance companies in the US (Lester, 2019), lead to increased illness severity and death (Khalsa et al., 2017).

Cultural biases are also evident in this literature (Anderson-Fye, 2018). A recent bibliometric analysis of articles published in the specialty journal Body Image found a preponderance of Anglophone samples. Moreover, the smallest cluster of keywords surrounded “ethnicity” and focused primarily on African-Americans in the US (Andersen and Swami, 2021). The field, overall, has a “Western” bias, focusing particularly on US, UK, and Australian Anglophone populations. Across this literature, “culture” is treated almost exclusively as a monolithic categorical variable in which cross-cultural variability in body dissatisfaction and/or disordered eating is attributed to membership in that cultural group (often collapsed to nationality; e.g., McCabe et al., 2020), while problematically centering a static and reified “Western” culture and its global exportation as causative of the global expansion of eating disorders (Lee, 2004; Lester, 2004). Little attention is paid to how eating disorders in East, South, and Southeast Asia, for example, historically have manifested without body dissatisfaction, instead emerging in relation to a fear of loss of control over one’s own life (Lee et al., 1993). Further, body ideals are generally not ends in and of themselves, but rather shape fields of “symbolic body capital” (Anderson-Fye and Brewis, 2017) in which conformity to local body ideals can be “exchanged” for economic, social, and cultural mobility.

Only 1% of eating disorders research focuses on men specifically (Murray et al., 2016). Consequently, research and treatment strategies often assume men’s experiences of body image and eating disorders are the same as women’s, resulting in invalid and unreliable findings (Strother et al., 2014) and potentially self-stigma-invoking interventions (Bunnell, 2021). Research also focuses on muscle dysmorphia (Pope et al., 2000) and muscularity-oriented disordered eating (Cunningham et al., 2021), despite “Western” cultural attitudes toward muscularity and masculinity not being universal (Monocello, 2022b). Reasons for differences in disordered eating as related to sexual orientation are also not well understood (Convertino et al., 2021).

Further, in anthropology, despite some examination of gender differences in body size across wealth (Dos-Santos et al., 2001; Hruschka, 2017) and more recent examinations of male body ideals and fat stigma across societies (Anderson-Fye et al., 2017; Monocello, 2020) there has been little sustained engagement with men’s subjective experiences of body image and disordered eating across cultures. Greater attention is needed globally and across genders to body image and eating disorders through theories and methods that integrate culture and the diverse ways it may be embodied in a community into research design. An intersectional lens (Crenshaw, 1989; Hill Collins and Bilge, 2016) applied to the theories of cultural models (D’Andrade, 1995), cultural consensus (Romney et al., 1986), and cultural consonance (Dressler, 2018), is proposed as one framework for doing so.

1.1. Intersectionality

Intersectionality does not refer to an additive model of oppression through “checking-off” identity-axes but rather attends to lived experiences emerging from these axes’ synergy (Crenshaw, 1989). In employing the lens of intersectionality, how multiple forms of oppression and privilege interact to create experiences irreducible to a sum of the component parts is examined (Hill Collins and Bilge, 2016). For example, Crenshaw (1989) demonstrates neither gender-related laws (for which the fundamental status of “woman” is “white woman”) nor race-related laws (for which the fundamental status of “African-American” was “African-American man”) sufficiently address African-American women’s legal discrimination claims in the US. Analyses centering the most-privileged members of a given axis of experience fail to account for those with multiple, intersecting histories of marginalization.

Some have criticized intersectionality as being overfixated on “Western” articulations of race, class, gender, sexuality, and nation, to the point of ethnocentrism; such criticisms mistakenly position the concept’s past employment as exhaustive of its potential (Tomlinson, 2013; Carbado, 2013). Further, as MacKinnon (2013, p. 1023) notes, axes of identity are not themselves constitutive of the social processes which result in inequalities, but are rather these processes’ “ossified outcomes.” However, once ossified, intersecting identities structure interactions in varying contexts which result in differences in health outcomes (Polos et al., 2021).

McClure (2017, 2020), through ethnographic research attending to how intersectionality shapes the embodiment of body ideals, argues that African-American girls and women have been considered an “exception” within larger trends in the United States showing women to be at increasing risk of disordered eating, viewed as immune to body image concerns and endorsing body positivity instead. This narrative of body positivity obscures how African-American girls and women internalize body image messages. In contrast with dominant (i.e., white, female) narratives of body dissatisfaction, African-American women tend not to experience fat itself as discrediting; rather, fat is perceived negatively when it compounds racialized and gendered social exclusion by preventing engagement in practices of self-presentation that would facilitate alternative avenues toward social inclusion, such as if their bodies were too large for professional or “respectable” clothing available in most stores. Social treatment based on “multiple material otherness” in terms of race, gender, class, and body size intersect to predict variation in African-American girls’ long-term commitment to physical activity, with repercussions for future health outcomes (McClure, 2020).

1.2. A cognitive theory of culture

Drawing on McClure’s (2017, 2020) methodological and theoretical insights, this study approaches intersectional cultural and power dynamics involved in disordered eating risk as instantiated in cultural models (D’Andrade, 1995). “Cultural models” derive from a cognitive theory of culture which posits culture as whatever an individual needs to know in order to live within a particular social group (Goodenough, 1957). Cultural models encode knowledge within a cultural domain, a “sphere of discourse” (Weller and Romney, 1988) about which people tend to think in similar ways despite individual lived experiences shaping that knowledge (D’Andrade, 1992). Cultural models also are employed in everyday decision-making and action without being explicitly brought to consciousness; cultural models work because one need not think about them in daily life (D’Andrade, 1992; Strauss and Quinn, 1997).

Cultural consensus theory (Romney et al., 1986) argues that cultural models are cultural because they are “shared,” meaning that while these models exist in individuals’ minds and pervade their bodies, people can coexist and anticipate others’ beliefs and behaviors in communities because they, overall, if imperfectly, “know” what is expected of one another. Individual knowledge of the model, that is, the degree to which an individual tends to answer similarly to every other individual in a sample, is known as “cultural competence” (Romney et al., 1986). “Residual agreement” (Boster, 1986; Henderson et al., 2022) refers to patterns of disagreement within a cultural domain; in other words, people tend to disagree in similar ways within a given cultural group.

Finally, cultural consonance (Dressler, 2018) refers to an individual’s ability to approximate in their own life the elements of a given cultural domain. Importantly, the experience of cultural consonance is generally understood to be related not to self- and social-comparison against the most successful members of society (those engaged in conspicuous consumption; cf. Sweet, 2010) but rather to what Veblen (1899) referred to as a “common standard of decency,” whatever is needed to be considered in a given cultural group as having a normal life (Dressler, 2018). Berlant (2007) has described the affective element of this experience as “aspirational normalcy,” especially evident among those whose intersectional circumstances preclude their full engagement in their society. Consequently, cultural consonance repeatedly has been found to predict multiple physical and mental health outcomes across societies (Dressler, 2018).

1.3. South Korea

Nationally representative data collected in 2001 found 0.2% lifetime prevalence of eating disorders among Korean men (Cho et al., 2007). In a 2015 study, 3.7% of adolescent boys exhibited disordered weight control behaviors, and single-food diets were argued to be a Korea-specific symptom (Y. Kim et al., 2018). Further, adolescent boys with same-sex sexual experience were more likely than other boys to engage in disordered eating (Yu et al., 2018).

However, Thomas et al. (2016) found that epidemiological studies often struggle to identify eating disorders in South Korea. While national health insurance records show 4–5 men per 100,000 per year receiving treatment for eating disorders (Lee et al., 2021), stigmas against mental illness and concerns about data privacy (Grinker and Cho, 2013) lead many to pay out-of-pocket for mental healthcare and would not be counted in these numbers. Further, the number of people across Asia who are treated for eating disorders is far fewer than those experiencing them (e.g., Chua et al., 2022), suggesting that these numbers vastly underestimate the problem.

A major barrier to understanding eating disorders among men in South Korea is that Koreans’ cultural models of the male body tend to differ from those of the predominantly white, “Western” assumptions that underlie men’s eating disorders research (Monocello and Dressler, 2020). Koreans foreground an ideal male image known as the kkonminam (꽃미남), or “flower boy,” which is marked by a thin (not bulky) defined physique (called “small muscle”), attention to fashion trends, and makeup and cosmetics use (Maliangkay, 2013), as well as a distaste for highly muscled bodies (Monocello, 2022b). Importantly, these body ideals are part of the model of hegemonic masculinity in South Korea (Connell, 2005; Jung, 2011; Woo, 2019), and do not conflict with local heteronormative ideologies (Lim, 2008). Rather, they are seen as projecting self-confidence, professional competence, regard for others’ comfort, and desire to maintain social cohesion (An, 2019).

Male beauty has long been valued among Koreans, going back at least as far as the Silla-era (57 BCE–935 CE) with the hwarang (화랑 elite poet-warriors famous for their beauty) and persisting through Chosŏnera yangban (양반 the ruling elite, 1392–1910) men’s sartorial behaviors (Gale, 1975 [1898]; Rutt, 1961). As kkonminam grooming, fashion, and body-type preferences (re)emerged in popularity in post-dictatorship South Korea, they came to be understood to be “natural” and “instinctive” elements of chagi kwalli (자기 관리self-maintenance) practices under the precepts of oemojisangju-ŭi (외모지상주의 “lookism,” or discrimination based on appearance; Hong, 2007), rather than indicative of nondominant sexual and/or gender identity (Lim, 2008). Recent work suggests that Korean men and women are similarly subject to lookism and experience anxieties related to appearance discrimination (Lim, 2015).

Greater appreciation of larger, muscled bodies began to emerge as this research took place (2019–2021). Nonetheless, “attractive” features were most closely associated with kkonminam, rather than the more muscular chimsŭngnam (짐승남 beastly man). An emerging, hybrid masculinity in the form of the helch’ang (헬창gym-whore), attempts to combine facial beauty and fashionability with heavier musculature. However, even many of those pursuing a helch’ang form expressed concerns about becoming “too” muscular and looking uneducated, unprofessional, or gang-involved (Monocello, 2022b).

Despite the popularity of “boy love” media, male/male “shipping” among domestic and international KPop fans, homoerotic fanservice sometimes performed by idols themselves, and sŭk’insip (스킨십 intimacy built through physical touch) frequently engaged by boys and young men, South Korean society heavily stigmatizes homosexual identity. One study found queer people to be the most distrusted group in South Korea, due to their “abnormality” (Kim, 2004). While homosexuality is not illegal among civilians, it is during military service (which is mandatory for all Korean men) and military leadership has been known to persecute gay soldiers (Gitzen, 2020). They also have no legal protections (Na, 2014), and the social, psychological, and economic costs of revealing their sexuality can be severe. Moreover, the difficulties of maintaining long-term relationships under these conditions and their inability to live up to expectations of filial piety by having sons of their own creates further stresses (Cho, 2020).

These factors also interdigitate with career and education. For one, successful completion of military service is a prerequisite for men’s gainful employment, and their career advancement is often implicitly tied with their practice of the heteronormative and chrononormative family (i.e., marriage and parenthood at the culturally “right” ages; Cho, 2020). Engagement in appearance-enhancing behaviors also affects men’s career prospects (Holliday and Elfving-Hwang, 2012), although there is conflicting evidence whether employed (Han et al., 2018) or unemployed men (Lim and Paek, 2016) feel more pressure to lose weight. Importantly, appearance-maintenance is about more than just weight or BMI; for Koreans, it encompasses the entirety of one’s self-presentation (Monocello, 2020).

Moreover, stable jobs at major companies needed to support a household are increasingly rare since the International Monetary Fund’s post-1997 neoliberal restructuring of the Korean economy, and such jobs favor graduates of top universities through both societywide valorization of graduates of these universities, in line with Korea’s historical valuation of education as a sign of moral worth and authority to lead in society, and the hakbŏl (학벌, university cliques) which circumscribe hiring (and also marriage) opportunities (Byun et al., 2020). The top three universities are Seoul National University (SNU), Korea University, and Yonsei University, known collectively as SKY Universities. Other top universities include Sungkyunkwan University and KAIST. Attending one of these universities requires successfully navigating the hypercompetitive Korean education system, utilizing cram schools and private tutoring financially available only to the upper classes and the middle classes willing and able to go into debt to support it, thus reproducing social inequalities through educational hierarchies (Abelmann et al., 2009). Anderson-Fye et al. (2017) found that university education, as a proxy for upward mobility, predicted attitudes toward bodies in developing economies; it was not being elite but the potential for becoming elite that made the difference. Lim and Paek (2016) did not find a relationship between educational attainment and body size among Korean men; however, over 70% of South Korean adults pursue postsecondary education, such that university prestige in this developed economy may be more important than education itself.

Rather than attribute variation in risk of disordered eating to the category of culture, this paper employs the theory and method of cultural consonance, framed by the concept of intersectionality, to systematically analyze intracultural variation in individuals’ relationship with local cultural models as a risk factor for disordered eating. In taking this radically emic, ethnographically grounded cognitive anthropological approach, intersectional identities are explored as ossifying ossifications of local histories and meaning systems (explained above) which frame men’s approaches to their embodiment of local male body ideals and risk of disordered eating.

2. Methods

This research was approved by the Institutional Review Board of the University of Alabama in 2019 (19-OR-140-ME). This paper reports results from the third phase of a multiphase project conducted during ethnographic field research in the Seoul Capital Area of South Korea between August 2019 and January 2021. Prior to the emergence of the novel coronavirus pandemic in Korea in February 2020, the first author attended the gym four to five times per week. He also participated in language exchange, attended plays and performances, went to noraebang (노래방 karaoke), bars, and cafes, and spent holidays with friends. Ethnographic fieldnotes were taken throughout the research period, alongside participant-observation and formal and informal interviews.

The first author remained in Seoul until January 2021, although most of the data was collected via phone, video conferencing, or online questionnaire. Participants tended to prefer the questionnaire format, since they were able to respond asynchronously during commutes, breaks, or before bed. The first author remained reachable by KakaoTalk (the most popular messaging app in South Korea) for any questions or issues. While this paper focuses particularly on the quantitative data, with a deep analysis of the interviews and ethnography beyond its scope, decisions about the directions taken in methods and analysis drew directly from the ethnography across all phases of research.

2.1. Sampling

Participants were recruited using an informal respondent-driven sampling approach (Heckathorn, 1997), in which participants from previous phases in the research were asked to participate in this survey, and then ask their friends and acquaintances to contact the researcher if they wanted to participate themselves. These initial seeds included established contacts and students purposively selected for attending elite and non-elite universities. This was necessary because people tended to be reticent to provide personal and contact information to strangers (cults are known to gather contact information by engaging people in surveys on social issues), and the coronavirus pandemic prevented in-person rapport-building activities. Most referrals resulted in two to three further participants.

The survey was originally written in English and then translated and back-translated by two fluent speakers of Korean and English. Between March 2020 and January 2021, 118 young Koreans (103 men and 15 women) between the ages of 18 and 35 were asked to participate in a survey about male body image. Women were included for purposes of describing and verifying the cultural model because they are heavily involved in its co-construction; however, they were not asked questions relating men’s body ideals to their own bodies as the questions were irrelevant to their subjective experiences of their body image. Three men were excluded for not finishing the survey. Participants who completed the survey were compensated with 5000 Korean won (about USD$4.40). Because of local pandemic restrictions, individual discomfort with meeting in public, and participants’ preference due to convenience (e.g., they could respond during their commute rather than schedule time outside of work or school), the survey was conducted on Qualtrics (2020).

2.2. Demographics

Participants were asked to indicate their gender (male, female, other [please describe]); year, month, and date of birth; birthplace; height (in centimeters); weight (in kilograms); university of attendance (self-report); academic major (self-report); relationship status (single, in a relationship, married, divorced); occupation (self-report); parents’ occupation (self-report); military service status (in the future, in progress, complete, exempt); sexual identity (heterosexual, bisexual, homosexual, other [please describe]), and experiences abroad.

Body mass index (BMI) was calculated using participants’ self-reported height and weight. Women declined to report their weight, so BMI was not calculated for them. Participants refused to answer questions about household income in previous phases (Monocello, 2022b), so income was not collected.

University of attendance was further dichotomized to Low-Prestige University (0) and High-Prestige University (1), based on local assessments of university prestige. High-Prestige universities included the SKY universities, Sungkyunkwan University, and KAIST. Some consider all Seoul-based universities to be high-prestige; however, this research was performed in Seoul, where the Seoul/non-Seoul dichotomy is less important than intra-Seoul distinctions are. Moreover, KAIST, while not based in Seoul, is regarded as a highly prestigious university throughout Korea. Participants also tended to agree that university prestige drops precipitously after these five.

2.3. Rating task

Participants were presented with a list of 42 features identified through cultural domain analysis (Borgatti, 1999; see Monocello, 2022b for a description of the cultural domain analysis used in this research). They were asked to indicate the extent to which they understood people in Korea to find each item attractive (maeryŏkchokida 매력적이다) in men on a four-point Likert scale including 0 (highly unattractive), 1 (somewhat unattractive), 2 (somewhat attractive), and 3 (highly attractive).

Responses to the rating task were then analyzed using cultural consensus analysis (Romney et al., 1986). While not itself a factor analysis, cultural consensus analysis uses factor analytic methods to solve for three variables: cultural consensus, cultural competence, and residual agreement. In cultural consensus analysis, cases and items are transposed to calculate the similarities between people (Romney et al., 1987) and patterns of intracultural variability (Henderson et al., 2022). It also produces a “cultural answer key,” averages of all respondents’ ratings of each item weighted by the degree to which respondents agreed with every other respondent.

2.4. Cultural consonance scale

For cultural consonance analysis (Dressler, 2018), men were presented with the same list of 42 items used in the rating task and asked to think about what another, typical Korean person would say about their appearance, indicating the extent to which each item was represented in their own bodies. The intent was to divorce their response from their own opinions about their bodies and reduce the potential for cultural proscriptions against self-aggrandizement to depress cultural consonance scores. Participants responded on a four-point Likert scale including 0 (not at all represented), 1 (slightly represented), 2 (represented), and 3 (highly represented). Cultural consonance scores were calculated as a sum of all of the items scored greater than 1.5 in the cultural answer key. Women were not presented with the scale as it was not relevant to their experiences.

2.5. Male Body Attitudes Scale

Men were asked to respond to the Male Body Attitudes Scale (MBAS; Tylka et al., 2005), a 24-item psychological scale designed to examine the extent to which men are dissatisfied with their body image, also known as body dissatisfaction. Respondents indicate the frequency with which they feel each statement on a 6-point scale from 1 (never) to 6 (always). Responses are averaged; higher values indicate greater body dissatisfaction. Women were not presented with the scale as it was not relevant to their experiences. Cronbach’s alpha was 0.88 for this sample.

2.6. Eating Attitudes Test

Men and women were asked to respond to the Eating Attitudes Test (EAT-26; Garner et al., 1982; validated Korean translation: Rhee et al., 1998), a 26-item psychological scale which assesses risk for disordered eating. Respondents indicate the frequency of experience of each statement on a 6-point scale from 1 (always) to 6 (never). Items rated 1 (always) receive a score of 3, 2 (usually) receive a score of 2, 3 (often) receive a score of 1, and 4 (sometimes), 5 (rarely), and 6 (never) receive a score of 0. Scores are summed; higher scores indicate greater risk of disordered eating. Among “Western” samples, scores above 20 indicate a likely eating disorder. Among Korean men, the threshold is 18 (Rhee et al., 1998). While the specific ways South Korean men manifest eating disorders may not be reflected fully in the instrument (e.g., “I cut my food into small pieces:” most Korean meals do not present the opportunity to cut one’s own food), the overall score points to eating- and body image-related distress in line with globalized/glocalized expectations (Watters, 2010). Cronbach’s alpha was 0.87 for this sample.

3. Results

3.1. Descriptive statistics

Table 1 presents the descriptive statistics. All participants identified as male or female. Men were older and taller than women. Men’s average height was 175.9 ± 4.8 cm and average weight was 73.0 ± 9.4 kg, for an average BMI of 23.5 ± 2.5 kg/m². Regarding sexual identity, 78.2% of the sample identified as heterosexual, including 79.0% of men and 73.2% of women. Thirty percent of men and 26.7% of women attended a High-Prestige university, and 68% of men had completed their military service. MBAS scores were, on average, 3.45 ± 0.81. EAT-26 scores were, on average, 10.6 ± 9.3, and 21% of men had an EAT-26 score greater than 18 (i.e., high risk for clinically disordered eating).

Table 1.

Sample characteristics.

Variables	Male (n=100)			Female (n=15)			Total (n=115)
	Mean	(SD)	[Range]	Mean	(SD)	[Range]	Mean	(SD)	[Range]
Korean Age (years)	26	(3)	[20–34]	23	(3)	[20–28]	26	(3)	[20–34]
Height (cm)	175.9	(4.9)	[164–187]	161.8	(3.9)	[156–168]	174.2	(6.6)	[156–187]
Weight (kg)	72.9	(9.3)	[51–100]
BMI (kg/m2)	23.5	(2.5)	[18.2–29.5]
Cultural Competence	0.76	(0.22)	[−0.45–0.96]	0.82	(0.09)	[0.57–0.90]	0.77	(0.21)	[−0.45–0.96]
Cultural Consonance	35.6	(12.3)	[0.09–0.65]
MBAS	3.45	(0.81)	[1.88–5.21]
EAT-26	10.6	(9.3)	[0–36]	9.4	(5.9)	[2–21]	10.5	(8.9)	[0–36]
%High Risk (EAT-26>18)	21.0%			6.7%			20.5%
%Heterosexual	79.0%			73.3%			78.3%
%High-Prestige University	30.0%			26.7%			29.6%
%Completed Military Service	68.0%

3.2. Cultural consensus analysis and residual agreement analysis

Cultural consensus analysis of the rating task confirmed a single cultural model of “attractiveness” within the sample. The ratio between the eigenvalues of the first and second factors was 16.7, explaining 67.064% of the variance, indicating that young Korean adults tend to judge features as attractive or unattractive in similar ways. The average cultural competence was 0.77 ± 0.21 (Men: 0.76 ± 0.22; Women: 0.82 ± 0.09), indicating that both men and women strongly agreed with each other on these judgments. Table 2 presents the cultural answer key.

Table 2.

Cultural answer key for the dimension of attractiveness.

Item	Aggregate Cultural Answer Key
Dress well	2.65
180 cm	2.63
Neat and efficient	2.59
Wide shoulders	2.58
Tall	2.55
Small muscle	2.49
8-head body	2.46
Long legs	2.42
Abs	2.36
Angular features	2.34
Robust	2.34
Flawless skin	2.30
Strong	2.29
Small head	2.26
Styled hair	2.26
Straight teeth	2.22
High nose	2.21
Small face	2.19
V-line jaw	2.15
Big eyes	2.14
Thin	2.14
Big muscles	2.11
White skin	2.10
Thick eyebrows	2.08
Big hands	2.08
Dimples	1.94
Double eyelid	1.81
Tanned skin	1.75
Brown hair	1.67
Makeup	1.28
Small eyes	1.21
Glasses	1.19
Colorful hair	1.11
Beard	0.94
Body hair	0.91
Flat nose	0.87
Short	0.73
Acne	0.51
Dress poorly	0.47
Stomach comes out	0.47
Fat	0.42
Balding	0.19

Consistent with a monocentric model (Henderson et al., 2022), participants tend to cluster heavily toward a high cultural competence score (the X axis) with little variation in residual agreement (the Y axis) (Fig. 1).

Fig. 1.

Cultural competence versus residual agreement.

3.3. Cultural consonance analysis

Cultural consonance scores were calculated as a sum of individual men’s self-reports of the extent to which each feature deemed positive (with a score greater than 1.5 in the cultural answer key) was represented in his own body. For example, if a participant said that he was tall (3, highly represented) and fat (3, highly represented), only his rating of the “tall” item was included in the Cultural Consonance with Attractiveness score. The average score was 35.6 ± 12.3 (range = 9–65).

Pearson correlations were calculated between all variables. Low-Prestige University was the reference group for University Prestige. Sexual identity was dichotomized to heterosexual (0) and gay/bisexual (1) to account for the homophobia experienced by both latter identities and retain the bisexual-identified men (n = 5) in the analysis. Higher risk of disordered eating was associated with identifying as gay or bisexual and reporting higher body dissatisfaction. Cultural consonance also trended toward a relationship with disordered eating. Cultural consonance and body dissatisfaction were not related (Table 3).

Table 3.

Bivariate correlation matrix of variables.

	Variable	1	2	3	4	5	6	7	8
1	Korean Age
2	Height (cm)	−0.069
3	Weight (kg)	−0.012	0.604**
4	BMI (kg/m^2)	0.023	0.223*	0.910**
5	University Prestige†	0.052	−0.109	−0.163	−0.141
6	Sexual Identity‡	0.065	−0.095	−0.279**	−0.283**	0.198*
7	Cultural Consonance	0.081	0.258**	0.064	−0.053	−0.283**	−0.025
8	MBAS Score	0.220*	−0.175	0.142	0.270**	0.111	0.228*	−0.134
9	EAT-26 Score	0.076	0.010	0.209*	0.260**	0.002	0.335**	0.183	0.516**

^*p < 0.05,

^**p < 0.01,

^***p < 0.001.

^†Low Prestige University (0), High Prestige University (1).

^‡Heterosexual (0), Gay or Bisexual (1).

During interviews and informal conversations, some participants noted that “SNU guys don’t care about how they look,” pointing to the less relative importance appearance plays in life success for those with high prestige educations (compared to those from lower-prestige institutions) and the social and cultural capital which results from it. At the same time, a SKY university professor once told the first author that a colleague from a non-SKY university had recently remarked that male students at the SKY university campus were more handsome than the male students at his own. Moreover, some gay men from these high-prestige institutions intimated that “no gay man has ever been impressed by my degrees;” others said that they could not see themselves dating someone outside their hakbŏl, out of pride and because those with lesser educational capital would find their degrees intimidating. Regardless of education, men spoke about the need to maintain their appearance to keep their friends comfortable and not embarrass them, and gay men were particularly emphatic about the connection between appearance and having relationships at all.

From aforementioned interviews, conversations, ethnographic observations, and literature suggesting that eating disorders risk varies along axes of education (e.g., Katzman et al., 2004) and sexual identity (e.g., Yu et al., 2018), it was hypothesized that intersections of sexual identity and university prestige would affect how cultural consonance bears on disordered eating. Based on intersectional variation in how men spoke about men’s appearances, a three-way interaction effect of cultural consonance × sexual identity × university prestige was then included in a multiple regression model with disordered eating as the dependent variable, controlling for BMI and body dissatisfaction (Table 4).

Table 4.

Multiple regression of variables to the EAT-26.

Variables [95% CI]	Model 1	Model 2	Model 3	Model 4
(Constant)	0.004 [−0.175,0.184]	−0.194 [−0.413,0.025]	−0.137 [−0.364,0.090]	−0.134 [−0.355,0.087]
BMI (kg/m^2)	0.152 [−0.054,0.358]	0.296** [0.094,0.498]	0.273** [0.071,0.474]	0.246** [0.048,0.444]
Male Body Attitudes Scale (MBAS)	0.502*** [0.316,0.687]	0.427*** [0.248,0.606]	0.442** [0.256,0.620]	0.486*** [0.309,0.663]
Sexual Identity†		0.817*** [0.387,1.275]	0.593** [0.024,1.164]	0.615* [0.060,1.171]
University Prestige‡		0.007 [−0.370,0.384]	0.037 [−0.432,0.506]	−0.125 [−0.602,0.352]
Cultural Consonance: Attractiveness		0.278** [0.105,0.452]	0.239* [0.033,0.445]	0.300** [0.092,0.502]
University Prestige × Cultural Consonance: Attractiveness			0.447* [0.020,0.873]	0.060 [−0.467,0.587]
Sexual Identity × Cultural Consonance: Attractiveness			−0.348 [−0.779, 0.082]	−0.780** [−1.334,−0.226]
University Prestige × Sexual Identity			0.364 [−0.486,1.215]	0.532 [−0.310,1.373]
University Prestige × Sexual Identity × Cultural Consonance: Attractiveness				1.039* [0.170,1.909]
N	100	100	100	100
df	97	94	91	90
R2	0.282	0.432	0.473	0.504
F	19.015	14.276	10.223	10.167
p	<0.001	<0.001	<0.001	<0.001

^*p < 0.05,

^**p < 0.01,

^***p < 0.001.

^†Low Prestige University (0), High Prestige University (1).

^‡Heterosexual (0), Gay or Bisexual (1).

Variables were entered in steps controlled by the investigator. In the first model, only the MBAS was significantly associated with disordered eating. Upon the addition of the cultural consonance, university prestige, and sexual identity variables, BMI, MBAS, sexual identity, and cultural consonance were all significantly associated with disordered eating. In Model 3, these variables remained significant, with the addition of the two-way interaction effect of university prestige × cultural consonance. Finally, in Model 4, BMI, MBAS, cultural consonance, sexual identity × cultural consonance, and the three-way interaction effect of university prestige × sexual identity × cultural consonance were all statistically significant. Based on leverage values and studentized deleted residuals, there were no highly influential cases.

Fig. 2 presents visually the interaction effect. High and Low Cultural Consonance are defined as one standard deviation above and below the mean solely for the purposes of visualization; testing of the interaction effects employed the full range of variation in cultural consonance. Gay and bisexual men evidence higher disordered eating than heterosexual men overall (p < 0.01), although this is not true for gay and bisexual men at lower prestige universities (n = 11). For this latter group, risk of higher disordered eating declines with increasing cultural consonance (p < 0.01). For heterosexual men at lower prestige universities (n = 59), risk of eating disorders increases with higher cultural consonance, as well as for men at higher prestige universities regardless of sexual identity (heterosexual n = 20, gay/bisexual n = 10). The association of cultural consonance and risk of eating disorders is somewhat stronger among low-prestige university heterosexual men than among high prestige university men, and in the high prestige university the association of cultural consonance and risk of eating disorders does not differ between gay/bisexual and heterosexual men. The pattern of the association of cultural consonance and risk of eating disorders thus differs by university prestige and sexual identity (p < 0.05).

Fig. 2.

Three-way interaction effect of cultural consonance, sexual identity, and university prestige.

4. Discussion

Understanding that culture should not be treated as a categorical variable in body image and eating disorders research (Lester, 2004; Monocello, 2022a), this paper employed the theory and method of cultural consonance with an intersectional lens to explore the relationship between culture and disordered eating. These data show that a young Korean man’s ability to approximate in his own body the features locally designated as ideal is significantly related to disordered eating, showing that culture predicts disordered eating beyond traditional, individual-level risk factors like BMI (a rough proxy of adiposity, and only one component of body image) and body dissatisfaction. In other words, culture matters in predicting disordered eating among young men, likely because appearance-based social and structural conditions affect embodiment regardless of individuals’ feelings about their own bodies (see also McClure, 2020).

The cultural answer key (Table 2) replicates and expands data on preferences in presentation of self from a previous study (Monocello and Dressler, 2020; see also Hong, 2007), demonstrating concerns with ornamental aspects of the body, especially being tall, thin, and having “pretty” features like big eyes, a V-line jaw, and the eight-head body proportion. With the contemporary rise in popularity of helch’ang masculinity, big muscles were also gaining in acceptance, although less than “small muscles” more typical of kkonminam-style presentation (see Monocello, 2022b). Items like “strong” and “robust,” while sometimes associated with big muscles, more typically referred to embodied vitality and contrasted with a timid demeanor. Others, like dressing well, looking neat and efficient, having styled hair, and flawless skin, demonstrated engagement in chagi kwalli.

Except for the subgroup of gay and bisexual men without a high-prestige education (see below), there is a direct association between higher cultural consonance in the cultural model of the ideal male body and higher disordered eating. The overall pattern of men with greater cultural consonance evidencing greater disordered eating runs counter to the general associations of measures of cultural consonance with better health outcomes (see Dressler, 2018).

Other researchers have also found such exceptions. In Sweet’s (2010) study of African American adolescents in Chicago, socioeconomic status moderated the association of cultural consonance in social status (the number of status goods owned) and blood pressure. Higher cultural consonance was associated with lower blood pressure among adolescents with higher parental socioeconomic status, while higher cultural consonance was associated with higher blood pressure among adolescents with lower parental socioeconomic status. This latter effect was attributed to stresses of participating in conspicuous consumption (Veblen, 1899) under conditions of economic deprivation.

Another example is Snodgrass, Dengah, and Lacy’s (2014) finding that problematic internet usage was associated with “cultural dissonance,” the difference between cultural consonance in the real world and cultural consonance in World of Warcraft. A simultaneous increase in in-game cultural consonance and decrease in real-world cultural consonance was associated with greater problematic internet usage as the gap widened. Increasing in-game cultural consonance was facilitated by problematic internet usage, and the positive psychological effects of in-game consonance further exacerbated the behavior.

Taking an intersectionality approach to interpreting these interaction effects “account [s] for, rather than control [s] for, structural inequalities in the relationship between social contexts in health” (Polos et al., 2021, p. 7), unveiling how axes of oppression and privilege (Carbado, 2013) synergistically shape men’s engagement with their bodies as media of social success and embeddedness (An, 2019) beyond a simple additive model. This approach also attends to the heterogenous paths taken between variables and outcomes rather than assuming uniformity in a population (McClure, 2020): these are not just men, but Korean men with varying intersectional relationships to disordered eating through axes of body, education, and sexuality.

Among the heterosexual men with high-prestige educational backgrounds, pressures to maintain an appearance in line with local male body ideals largely pertain to maintaining the image of their social status. On the job market, their educations speak for themselves and their school cliques (hakbŏl) ease access to high-powered, stable, and lucrative careers; in the romantic marketplace, their hakbŏl circumscribes their pool of potential partners to those of similar status. Here, men’s pursuit of cultural consonance in male body image relates less to potential for socioeconomic success and comfort than as a strategy for getting further ahead among the upper echelons of Korean society (Lim, 2005).

For the gay and bisexual men with high-prestige educational backgrounds, pressures to achieve or maintain an appearance in line with local male body ideals related to an ability to maintain status and to find romantic or sexual partners. Many expressed that their economic futures were largely secure, owing to the social status conferred by their educational backgrounds, as long as their sexualities were kept secret. However, these gay and bisexual men often cannot utilize their hakbŏl to find partners in the ways that their heterosexual counterparts can. While some have trusted friends who will set them up on blind dates (sogaeting 소개팅), others rely heavily on apps like Tinder, Grindr, and Jack’d, which center their appearances, especially their bodies, as the primary initiator of interaction, reinforcing investment in the pursuit of cultural consonance in male body image.

For heterosexual men without high-prestige educational backgrounds, pressures to achieve or maintain cultural consonance with local male body ideals compensate for deficiencies in educational prestige in the intensely competitive Korean job market. Korean men widely recognized that their appearance served as part of their job applications and romantic/sexual prospects, with more attractive men having an edge over less attractive men, especially in the cutthroat competition of Korea’s occupational marketplace. Under Koreans’ model of lookism maintaining an attractive appearance communicates a man’s engagement in chagi kwalli (self-maintenance) practices on which an employer can capitalize; coworkers and customers are also more likely to feel comfortable with and confident in the abilities of more handsome men, facilitating business relationships (An, 2019; Hong, 2007; Monocello, 2020). Women also said they felt men who engaged in chagi kwalli signaled that they would be more attentive and caring partners in addition to the earning power that their fashion sense and skincare regimens implied. Kim et al. (2018) found a comparable pattern among male and female adolescents, for whom attending vocational high schools rather than general high schools also led to a higher risk of disordered eating. However, they attributed the difference to variation in school resources to engage in prevention programs rather than stresses associated with inequalities in current and potential social and cultural capital related to education. The data from the present study suggest that disordered eating for this subgroup in comparison to the high-prestige education subgroups is attributable to their greater investment in symbolic body capital toward social-cultural-economic success in the absence of prestigious educational networks.

In contrast to the previous three cases, the gay and bisexual men without high-prestige educational backgrounds experienced cultural consonance as reducing disordered eating. For these men, on top of professional and romantic considerations, pressures to achieve or maintain an appearance in line with local male body ideals were related to feelings of a right to be part of society at all. As the most distrusted group in South Korea (Kim, 2004), not only were they concerned with career prospects, they also spoke about fitting in, fear of ostracism should they be “outed,” and chŏng (정 rapport; deeply-felt connectedness). Those with higher cultural consonance were more easily able to establish chŏng with others, providing a sense of coherence and social embeddedness. On the other end, some of those with lower cultural consonance described their “unattractive” embodiment as a barrier to having these relationships altogether, thwarting their aspirations to normalcy (An, 2019; Berlant, 2007; Cho, 2020); their disordered eating emerged as a way to assuage their intense loneliness, to exert control over something in their lives, or to transform their bodies to be more acceptable in society.

Notably, in contrast with Snodgrass et al. (2014), these young Korean men are not necessarily pursuing these body ideals at the expense of more broadly accepted models of social success, but in pursuit of them. Moreover, unlike the adolescents in Sweet’s (2010) study, these outcomes result not from the stresses of mismatches between culturally defined lifestyle goals and economic resources but from strategies to maximize long-term economic and social capital when traditional, dominant pathways to success are inaccessible or precarious. Different intersections of experience then instigate heterogenous engagements with cultural models of the ideal male body as a means to the end of a theorized cultural consonance in lifestyle, ultimately leading to diverse manifestations of disordered eating as well.

In this sample, disordered eating appears to emerge intersectionally in two patterns in relation to cultural consonance with the ideal male body. In one pattern, disordered eating may develop as a strategy for exerting control over one’s life, or attempting to cope with it, when multiple, intersecting forms of “otherness” (here, nondominant sexual identity and educational background) prevent full engagement and embodiment in society. In the other, more widespread pattern, under conditions in which one’s being is not fundamentally discreditable, disordered eating may be utilized alongside other strategies to maximize one’s symbolic body capital (Anderson-Fye and Brewis, 2017) in line with local ideals for exchange on both the job and romantic marketplaces.

4.1. Limitations

The data are limited by a nonrandom sample, potentially consisting of participants already interested in body image or experiencing disordered eating. While the main effect of cultural consonance on eating disorders is robust, there are potential problems with testing 2- and 3-way interactions in such a small sample. Data “sparseness” in the quadrants that generate those interactions leads to low statistical power, along with the potential for capitalizing on sample-specific and non-replicable findings; therefore, interaction effects must be interpreted with caution. However, a theory of intersectionality anticipates such interaction effects, and ethnographic observation strongly suggests the validity of these specific intersections. Moreover, the interactions themselves are statistically reliable, even without employing sampling theory and estimating errors using bootstrap methods. Thus, the different associations of cultural consonance and risk of eating disorders by university prestige and sexual orientation suggest reliable patterns and further insight to the process.

These findings could be expanded through a more granular approach to examining educational prestige, taking into consideration that local cultural models of university ranking are often more ordinal than continuous and that those who do not pursue postsecondary education may also interact with their bodies in different ways. More research is also needed into the effects of sexual orientation across identity, attraction, and behavior (Gibb et al., 2021) on disordered eating at these intersections within this particular cultural context. Finally, rigorous qualitative research is necessary for unpacking how Korean men’s embodiment (or not) of hierarchized cultural models of the ideal male body, shaped by their masculinities and muscularities (Monocello, 2022b), comes to bear on their risk of disordered eating.

5. Conclusion

Although eating disorders are dangerous mental illnesses and increasing across genders around the world, the relationships of culture, male gender, body image, and eating disorders remains largely unproblematized in the literature. The present study’s findings add to this literature, especially with respect to the importance of intersectionality revealed by the substantial interaction effects between cultural consonance, university prestige, and sexual identity. These data show that conceptualizing cultural consonance as an intersectional axis in concert with sexual identity and university prestige explains an additional 22.4% of the variation in disordered eating outcomes in this sample. Moreover, it demonstrates that there is not a singular “cultural” way in which young Korean men engage with their bodies that puts them at risk of disordered eating; rather, the intersectional pathways (see McClure, 2020) between cultural consonance and salutary or deleterious disordered eating outcomes vary depending on culturally laden indices of social status, privilege, and oppression. Here, intersections of university prestige and sexual identity structure engagement with embodiment of cultural models of male body image as a strategy for the making and maintenance of social relations and the advancement of social status in a precarious neoliberal economy. Therefore, a framework combining cultural consonance and intersectionality is important for understanding heterogeneity in cross-cultural body image and eating disorders research moving forward.

Data availability

Data will be made available on request.