Introduction

Eurocentric standards of beauty reign supreme in the United States (and indeed in the world), placing increased value on features typically associated with whiteness such as light skin and straight hair, which reinforces a system of racial stratification whereby white people are privileged at the expense of black people and other people of color. One latent consequence of this preference for whiteness is intra-racial stratification among black people that also relies on skin color, affording increased privileges to those deemed closer to phenotypic whiteness. This is important because being perceived as more attractive has significant social benefits including higher income (Frieze et al. 1991), higher chances of being hired for a job (Hosoda et al. 2003). Studies consistently show that black people with more of the aforementioned white-like features, particularly skin tone, are deemed more attractive by both blacks and whites and are subsequently afforded the benefits of being attractive. But a preference for white-like features may only be part of the story. Net of skin tone, the near-universal disdain for blackness may also result in people being perceived as more attractive because they do not appear to be “completely” black (to the extent that a person’s racial group can be completely endogamous) or simply because they do not identify as “completely” black.

Miscegenation is an integral, if taboo, part of United States history, and even before legal bans on the practice were lifted, the children of interracial unions benefitted materially from their multiracial heritage, especially if they were phenotypically white (Krauthamer 2013). Despite the obvious presence of multiracial citizens in America, the year 2000 Census represented the first time since the “Mulatto” category was eliminated as an official racial designation in 1930 that the United States government recognized people of multiple races. Prior to 2000, respondents were forced to choose only one racial category. (Note that even self-identification was relatively recent on the Census. People were not allowed to select their own racial classification until 1960. Prior to that year, census takers were instructed as to how to classify people according to race.) The 2000 Census allowed people to select all categories that they felt applicable.

This official acknowledgement and measurement of multiracial Americans by the US Census led to a boom in studies seeking to explore the role that the more than 6 million self-identified multiracial people play in a racial structure that continues to rely on discrete racial categories (e.g., Harris and Sim 2002; Lee et al. 2003). Many studies conclude that multiracial black people—that is people whose racial background is a combination of black and other racial categories—either choose to discard their multiracial identities for a single-raced black identity, and even when they acknowledge their multiracial heritage, usually they more strongly identify with their blackness (Khanna 2010). I seek to contribute to this literature by exploring whether black people who identify as multiracial are perceived as more attractive than black people who only identify as black.

Multiraciality and Attractiveness

Multiraciality is a highly contested topic in the United States, particularly among Black Americans. Because the United States tends to use discrete racial categories with no intermediate racial groups (but for an argument on racial fluidity in the United States see Saperstein and Penner 2012) or neat placements for multiracial people, debates over which racial group to which a child of multiracial heritage belongs have raged throughout United States history. Since the Census removed the “mulatto” designation as an official racial category, the most iconic way that the State has handled multiracial black people is through the infamous “one-drop rule,” which became a blanket way to refer to the various state laws that dictated which various miniscule percentages of “black blood” legally made a person black. Though those statues no longer carry legal weight, the undergirding ideology of hypodecent—the idea that a child takes the identity of his or her more socially subordinate parent— remains prominent.

Multiracial people with black heritage often identify as black, citing how they are perceived as black by other people as a primary justification (Khanna 2010). But neither these self-identifications nor the external perceptions are static. They vary considerably by social context and the characteristics of the external person. For example, multiracial black Southerners are more likely to identify as black “only,” self-identification varies between school and home (Harris and Sim 2002), and people with conservative political attitudes are more likely to label racially ambiguous faces as black (Krosch et al. 2013).

These studies on self-identification hinge on the idea that self-identification is derived in part from people’s interpretations of external perceptions and social context, e.g., because multiracial people with black heritage think they are viewed as black rather than multiracial, they identify as black. I seek to take a different approach and examine how racial self-identification influences perception, not of race, but of attractiveness, which has also been shown to be an agent of stratification. People viewed as more attractive are afforded a variety of privileges including being viewed as more competent (Parks and Kennedy 2007; Ritts et al. 1992), having higher incomes (Frieze et al. 1991), and having increased chances of being hired (Hosoda et al. 2003; for an in-depth review of this literature see Frevert and Walker 2014). While studies have examined the intraracial effects of skin tone on attractiveness (e.g., Hill 2002), few have explored how other processes such as multiraciality affect perceptions of attractiveness.

Biracial Beauty Stereotype

Sims (2012) calls the popular idea that multiracial people are more attractive the “Biracial Beauty Stereotype.” Using this concept, Sims investigates whether people’s self-rated and externally perceived attractiveness affect whether they identify as biracial or monoracial when they have parents of different races. She hypothesizes that people who perceive themselves and are perceived as more attractive people would be more likely to embrace a biracial identity than to identify with a single race because of a belief that mixed race people are more attractive than monoracial people. For example, a woman who has one Asian-American parent and one black parent who perceives herself as attractive with be more likely to identify as multiracial than a woman of similar heritage who does not perceive herself to be as attractive. The latter would be more likely to identify as black only. She investigates people fitting into six combinations of two racial categories and found that even when controlling for skin tone, only people of black/Asian and black/Native-American heritage seemed to make decisions about their racial identity based on their attractiveness.

Sims’ study is a good starting point for the investigation of the perceived attractiveness of multiracial people, but the study suffers from theoretical deficiencies. The primary issue is her causal order. She asks whether attractiveness, both externally perceived and self-rated, influences multiracial identification, but I contend that the better question is whether multiracial identification influences externally perceived attractiveness. This is a better question for two reasons: 1) because externally perceived attractiveness is the measure by which benefits are conferred not self-rated attractiveness, which would be more salient if attraction precedes identity formation and 2) because individual identity choices are less important than relative aggregate external perceived attractiveness, which she loses in her study. My study reverses the causal direction and asks whether multiracial identification for black people positively influences how attractive they are perceived by others relative to black people who only identify with a single race.

Exoticizing, Identity, and Perception

Though anti-black attitudes associated with racial stratification shape the standards of beauty to fit white people’s typical characteristics to the detriment of black people, some black people and other people of color are still found to be attractive, not because they fit traditional beauty standards but because they deviate from them in ways that give them an air of exoticism such as having eyes reminiscent of an East Asian person or the long, dark straight hair of Native-American stereotypes. This perceived exoticism, though extremely subjective and objectifying, can have material returns for people of color, particularly women, through increased attractiveness (Brooks 2010; Frank 2002). In many cases exoticism is about deviating from the white standard of beauty but not so much as to be perceived as only black.

For example, in ethnographies of strip clubs Frank (2002) and Brooks (2010)) found that black women often found themselves marginalized by patrons who sought to interact with other non-white women. Multiracial, exotic-looking women (e.g., women with features that made them appear racially ambiguous) on the other hand, were sought after by these patrons and earned more income than women perceived as “only” black, even if they did not earn as much as their white counterparts. Because of this, black women learned, and were even told, to emphasize their multiracial heritage, or to pretend to be multiracial even if they actually were not, in an attempt to distance themselves from the “pure” blackness to a more exotic, acceptable, attractive and muted form of blackness. This provides support for my primary argument that identification changes one’s level of perceived attraction even when accounting for physical appearance. This latter part is especially salient as women with no multiracial heritage were able to increase their profiles through claiming this heritage.

Data and Methods

I used the National Longitudinal Study of Adolescent Health (hereafter Add Health) to conduct my study. Add Health is a well-known, nationally representative longitudinal survey that is ideal for this study because (1) it allows respondents to select multiple racial identifications, (2) it includes an interviewer coded rating of perceived attractiveness, and (3) it includes a skin-tone rating. Because I only examine black or multiracial persons who include black as one of their categories, my study only includes respondents who selected black as at least one of their racial identifications. There have been four waves of Add Health, and I used data drawn from Wave 1 and Wave 3. I created my race variable, which is a dichotomous variable for whether the respondent identifies as mixed race, from Wave 1 data because Wave 1 offers an “other” race category that is not offered in other waves. Omission of this category in subsequent waves may bias my results by disallowing people who seek to identify as mixed race from doing so. Because I use Wave 1 for my race variable and my dependent variable must be coded during that same period I also use interviewer-coded attractiveness, grooming, age, and gender from Wave 1. Attractiveness is an interviewer-coded rating of attractiveness on a progressive scale from 1 to 5.Footnote 1

My skin-tone variable and other phenotypic characteristics (e.g., eye color and hair color) are from Wave 3 because that is the only Wave in which those data were collected. Though some of my predictors of attractiveness (skin tone, eye color, and hair color) are measured at a different time than attractiveness, this should not bias my results because factors such as skin tone, eye color, and hair color will not likely vary much over time. Variables for characteristics such as weight and height would have proven to be more problematic because they constantly change, particularly during adolescence. Skin tone is also an interviewer-coded variable on a scale from 1 to 5 with 1 being “very dark skin” and 5 being “white” skin. I also included a variety of control variables from across the two waves that may also affect respondents’ attractiveness. See Table 1 for a full description of my variables.

Table 1 Variable descriptions
Full size table

In addition to my primary explanatory variables: mixed race and skin tone, I control for a variety of other characteristics that may affect the attractiveness rating of the respondent, including gender, age, how well groomed the respondent was at the interview, whether the respondent had ever been convicted of a crime, whether the respondent had ever been on welfare, whether the interviewer perceived the respondent as black, and other phenotypic characteristics (black hair, brown hair, blonde hair, red hair, bald, black eyes, brown eyes, blue eyes, hazel eyes, and green eyes).

The following tables offer a deeper view at the relationship between multiraciality and skin tone in this sample by showing a variety of disaggregated means.

Table 2 shows mean attractiveness and mean skin tone for non-mixed and mixed respondents separately. Multiracial respondents have both a higher average attractiveness rating and a substantially lighter average skin-tone rating. This underscores the importance of accounting for skin tone in the models to ensure that the higher average attractiveness of mixed race respondents isn’t driven completely by a positive effect of skin tone.

Table 2 Attractiveness and skin tone for nonmixed and mixed respondents
Full size table

Table 3 shows the mean attractiveness of non-mixed and mixed respondents separately for each skin-tone rating. For non-mixed respondents, attractiveness follows a predictable pattern, increasing stepwise as respondents’ progress from the darkest-brown to very light-brown skin. However, non-mixed respondents with “white” skin show a precipitous drop in mean attractiveness, which may be attributable to a few factors, the most obvious of which is small sample size. But because the data set codes “white” skin for respondents rather than “very light skin,” non-mixed black respondents with “white” skin may receive an attractiveness penalty for not appearing black enough. That is to say that if a person identifies as black, he or she may be perceived as less attractive if he/she has “white” skin because black people should have darker skin. That would be consistent with the theory driving this study, which is that one’s externally perceived attractiveness is influenced by expectations derived from one’s racial identification.

Table 3 Average attractiveness by skin tone
Full size table

Attractiveness ratings for mixed-race respondents fail to follow a similar stepwise pattern, which could be the result of sample size issues that result from this type of disaggregation or of unexamined covariates, such as other phenotypic features, that make the pattern seem random. But despite the lack of a clear pattern, it is worth noting that even the lowest average attractiveness rating for mixed respondents is higher than both the highest rating for non-mixed respondents and the highest cumulative attractiveness rating. But these data tables only tell a portion of the story; I must examine regressions in order to truly examine the relationship between multiraciality, skin tone, and attractiveness.

I used ordered logistic regression to estimate my models. The baseline model includes only the variable for mixed race and the non-phenotypic control variables to test whether mixed race predicts attractiveness. In the second model, I add the variables measuring phenotype: skin tone, hair color (I use black hair as the reference group), and eye color (I use black eyes as the reference group).

Results

Consistent with my argument, the results of my models reveal that mixed-race identity does increase perceived attractiveness even when controlling for phenotypic characteristics such as skin tone, hair color, and eye color. In the first model, the mixed variable was positive and significant, suggesting that black people who identify as multiracial are perceived as more attractive net of the covariates included in the model. But because mixed race people have, on average, lighter skin, it was vital to account for the potentially mediating effect of skin tone, so in the second model I included skin tone and the variables for hair color and eye color to test the extent to which the effect of mixed race is driven by the effect of phenotype. Though the coefficient for mixed race decreased slightly, the significance remained the same, suggesting that mixed race black people are perceived as more attractive even when controlling for the phenotypic characteristics measured here. Additionally, skin tone is also positive and significant, which not only confirms that lighter skin tone among blacks leads to higher perceived attractiveness but also that mixed race and skin tone, though related, are independent factors in determining perceived attractiveness (Table 4).Footnote 2

Table 4 Estimates for ordered logistic regression for perceived attractiveness
Full size table

Additionally, three of my control variables were also significant. Gender is positive and significant, suggesting that women are perceived as more attractive than men. Age is also positive and significant, suggesting that older people are perceived as more attractive than younger people, but this result must be interpreted cautiously. By the nature of this Add Health, the oldest respondents in the sample are 21 years old so this result actually suggests that older adolescents are perceived as more attractive than younger adolescents, which may be a result of increasing physical maturity. If I stretched this study across the full range of ages available, I would expect to find a curvilinear effect of age resembling a bell curve, with very young respondents perceived less attractive, attractiveness increasing and peaking somewhere between young adulthood and middle age, and then decreasing later in life.

The variable estimating whether the respondent was well-groomed during the interview is also positive and significant. Though the correlation between grooming and perceived attractiveness is clear I am unable to make a definitive claim about the causal direction of the relationship. Though the modeling strategy may offer the illusion that being well-groomed leads to higher attractiveness ratings, the causal direction could easily be reversed to argue that respondents who were perceived as more attractive were, in turn, rated more well-groomed.

Discussion

Racial stratification touches all parts of society, including attraction, meaning that white people, and white features, are deemed more attractive that black people and black features. This is the logic underlying colorism, which rewards black people (and other people of color) for appearing white. The results of the present study reveal that a similar process may be at work with multiracial individuals where they are rewarded simply because they do not identify as black, perhaps because they are deemed exotics as in the work of Frank (2002) and Brooks (2010).

But this study suffers from one primary limitation: the inability to control for phenotypic characteristics such as lip thickness, nose width, eye shape, hair texture, etc. It’s possible that the interviewers coding attractiveness are actually reading these other characteristics in addition to, or instead of, being heavily influenced by the respondent’s self-identification. Future research could address this limitation in one of two ways 1) collect these data on a nationally representative survey such as AddHealth or 2) a social psychological experiment where participants are asked to rate the attractiveness of a group of faces that are randomly assigned as “multiracial” or “black” from each participant.

Other future work should also examine which groups of multiracial people benefit the most from multiracial identity—relative to those who identify as monoracial and relative to other multiracials. This would offer more insight into the ways that exoticism works and who benefits from it.