Internet Daters' Body Type Preferences: Race-Ethnic and Gender Differences.

Employing a United States sample of 5,810 Yahoo heterosexual internet dating profiles, this study finds race-ethnicity and gender influence body type preferences for dates, with men and whites significantly more likely than women and non-whites to have such preferences. White males are more likely than non-white men to prefer to date thin and toned women, while African-American and Latino men are significantly more likely than white men to prefer female dates with thick or large bodies. Compatible with previous research showing non-whites have greater body satisfaction and are less influenced by mainstream media than whites, our findings suggest Latinos and African Americans negotiate dominant white idealizations of thin female bodies with their own cultures' greater acceptance of larger body types.

Introduction

This study compares differences in body type preferences for dates between African Americans, Asians, Latinos and whites in the United States. We want to determine whether men or women, whites or non-whites are more selective about dates’ body types, and the degree to which the dominant body ideal is accepted. With an intersectionality perspective, we address gender and race–ethnicity simultaneously. Using logistic and linear regressions we analyze a unique dataset compiled from the internet dating profiles of heterosexual males and females, African Americans, Asians, Latinos and whites in the United States, who are between the ages of 18–50. We expand on the work of several scholars who address the body as a site of inequality (e.g. Bordo 1999, 1993; Collins 2004; Foster 1995; Pipher 1994; Urla and Swedlund 2000; Wolf 1991) by examining the intersection of race–ethnicity and gender, as they affect body type preferences for potential partners.

There is mixed evidence regarding racial–ethnic differences in body type preferences. Some studies suggest that culture plays a significant role in body type preferences with African Americans and Latinos more accepting of heavier body types than whites (e.g. Crandall and Martinez 1996; Desmond et al. 1989; Massara and Stunkard 1979; Powell and Kahn 1995; Shaw 2005), while other studies show little or no significant racial–ethnic differences in either male or female body type preferences (Allison et al. 1993; Altabe 1998; Cachelin et al. 1998). All of these studies that assess body type preferences are methodologically limited. Many studies include ratings of silhouette figures (e.g. Allison et al. 1993; Cachelin et al. 1998), a data set of a particular age population such as college students (e.g. Altabe 1998; Demarest and Allen 2000) or adolescents (Crandall and Martinez 1996), a specific gender sample (Cash and Henry 1995; Poran 2002), or a particular community population (e.g. Cachelin et al. 1998). The methods vary significantly. For example, several of these studies do not control for the age and education level of respondents (see Cachelin 2001), and this may account for the differences in their findings. Most importantly, they do not test racial–ethnic and gender body type preferences of daters in an actual setting. In contrast, our study provides a rare opportunity to examine people’s stated preferences in a real-life situation. It is the first to systematically test the extent to which African-American, Asian, Latino, and white men and women of different ages, and in an actual dating context, may similarly prefer dominant portrayals of ideal bodies; that is thin women and muscular men, in dating preferences. By investigating internet daters’ body type preferences, this study examines the importance of race–ethnicity and gender in determining the extent and type of body type preferences that may or may not adhere to culturally prescribed standards of beauty in date selection. We focus our review of the literature exclusively on those studies that address the United States, unless otherwise noted, since that is the location of our sample of daters.

Gendered Body Type Preferences

To begin, we are interested in whether men or women are more likely to state specific body type preferences for a date. There is an extensive literature on sex differences in mate preferences showing that men place greater value than women on the physical attractiveness of an ideal mate (e.g. Buss 1989; Feingold 1990, 1992; Goodwin 1990; Hill 1945; Smith and Waldorf 1990; Sprecher, Sullivan and Hatfield 1994; Townsend 1989). Although they may do so, there is scant empirical data showing the extent to which men have more specific body type preferences than do women. While not the focus of their college student speed daters study, Kurzban and Weeden (2007, p.626) find men significantly more likely than women to “express a desire for specific body types, typically for thinner bodies over heavier ones.”

While both men and women are held to standards of attractiveness and body ideals, these remain much more demanding for women than for men (Bordo 1993; Urla and Swedlund 2000). In mainstream popular culture, there is more of an emphasis on women’s appearance than on men’s (Bordo 1993; Lazarus 1987; Pipher 1994). Analyses of magazine covers show 78% of popular women’s magazine covers contain a message regarding a woman’s appearance while no popular men’s magazine covers contain a message regarding bodily appearance (Malkin, Wornian, and Chrisler 1999). Women also suffer harsher social consequences for violating standards of beautiful bodies than men (Cash and Roy 1999; Stake and Lauer 1987). Overweight female college students are less likely to be currently dating, more likely to date less frequently overall, and to perceive their dates as less satisfied when compared to overweight male college students (Stake and Lauer 1987). Similarly, men’s personal ads are significantly more likely than are women’s to specify the desire for a particular body type (Miller et al. 2000; Smith and Waldorf 1990). To be certain, images of muscular men have gained popularity (Spitzer, Henderson, and Zivian 1999; Trujillo 1995), but representations of men are far less constrained (Hanke 1998; Nixon 1997) because men also gain status from power, wealth or prestige (Hanke 1998). In sum, although their study is limited to college student speed daters, Kurzban’s and Weeden’s (2007) finding that men have more body type preferences than women, as well as the implications from the aforementioned studies, strongly suggest that men are more concerned than are women with the body type of a potential date. Therefore, we expect:

Men will be more likely than women to indicate specific preferences for dates’ body types.

We test this hypothesis with two logistic regression models. The first model examines the effect of race–ethnicity and gender on the likelihood a dater indicates specific preferences for a date’s body type. The second model examines the main effects of race–ethnicity and gender while adding controls for how selective a dater is overall as well as a dater’s own age, own level of education, region of the country, own body type and racial–ethnic preference for dates. We use logistic regression analyses to test this hypothesis because this method allows us to examine the main effects of gender and race–ethnicity on body type preferences while also controlling for potential racial–ethnic group differences, such as daters’ own body type, level of education, and racial–ethnic preference for dates. Past research has identified these factors as important predictors of marital partner choice or body type preferences. Although our focus is on racial–ethnic and gender differences, we also consider the influence of other factors that have typically been studied in conjunction with mate selection including age, level of education and geographic region (see Kalmijn 1998 for a comprehensive review of intermediaries in marriage trends).

Education is another important control for various reasons. First, there are racial–ethnic variations in education on a national level (Kane 1994), which are also present in our sample. Additionally, proponents of social exchange theory consider mate selection as a gendered exchange process whereby men exchange their educational attainment and higher income potential for women’s youth and beauty (e.g. Bolig et al. 1984; Cicerello and Sheehan 1995). As such, it may be that more educated men feel more entitled to make limited or different demands about dates’ bodies than men with less education. Another important exchange factor of this nature is income, which will not be included in our analysis due to a non-response rate in the final sample of over 50%. However, education and income are highly correlated (Ellwood and Kane 2000), so the dater’s reported level of education will serve as a proxy for income.

We also control for age and region. Men and women differ in their age preferences for mates; men generally prefer younger women and women prefer older men (Wiederman 1993). It might also be the case that body and beauty norms vary by the region of the country in which one lives. For example, Georgia has higher obesity rates than California (CDC 2007), so a dater in Atlanta may be more open to larger body types and less open to smaller body types than a dater in L.A. The daters in our sample are from within 50 miles of four U.S. cities: Atlanta, Chicago, Los Angeles and New York City. We selected these four regions to ensure diversity in our sample, as these are geographically dispersed and have different racial–ethnic compositions. Since we expect that body type preferences will vary by race–ethnicity, it is important to control for region to be sure we are testing racial–ethnic differences and not simply regional differences.

We also control for dater selectivity by measuring the number of specific preferences a dater makes. It is important to control for general selectivity because body type preference could be indicative of a picky personality rather than a strong preference or concern specifically for potential dates’ body types. Further, if a dater skips a question about his or her dating preferences, the answer is automatically set to “no preference.” As such, controlling for selectivity will mediate the effects of having daters who did not put much effort into filling out their profiles or who do not wish to specify their preferences.

Among our sample there are significant racial–ethnic differences with regard to own body type so it is particularly important to control for a dater’s own body type. Past research finds one's own body type influences the body types desired in potential dates (Cachelin et al. 2002). Daters could indicate their own body type as being one of the following: Slim, Slender, Average, Fit, Thick, A few extra, Large, Curvy or Voluptuous. We collapsed these into four broad categories: Small (Slim or Slender), Average (Average), Athletic (Fit or Athletic) and Large (Thick, A few extra, Large, Curvy or Voluptuous).

Daters’ racial–ethnic preferences for potential dates will also be controlled since the racialization of beauty norms likely leads to daters having different expectations of bodies that are tied to their racial preferences for dates. For example, Prasso’s (2005, p.149) analysis of websites promoting Asian women partners concludes that there is a “sexual fascination with Asian women” who are viewed as both “a wildcat in bed”, and “a sweet and innocent little doll.” Additionally, historically rooted sexual stereotypes of Latinas as “hot-blooded, tempestuous, and hypersexual” persist, with a recent emphasis on their curvaceous bodies and big butts (Mendible 2007, p.1).

Mainstream Popular Culture and Racial–Cultural Influences

Body type preferences are socially derived; ideals of attractiveness and the body vary by culture (Crandall and Martinez 1996; Shaw 2005) and throughout history (Mulvey et al. 1998). We assert that body type preferences are driven by mainstream popular culture as well as other racial–cultural influences. The media is largely dominated by white control and/or hegemonic images that appear on television, in film and in magazines. Such images provide a standard for all that may be negotiated but must be engaged because of its pervasiveness and its association to structures of power and domination. It is important to note that while we refer to these body types as “dominant” and “idealized,” we do recognize that popular culture is fragmented and there are varied images of beauty portrayed. There are many different images of beauty in various segments of popular culture so that even if dominant hegemonic images persist, individuals are still able to choose among a variety of images. For example, the internet and non-white magazines provide resources that may promote opposing images.

As we stated at the outset, study results are inconclusive regarding the extent to which non-whites adhere to dominant standards of beauty. Allison et al. (1993), in their comparison of silhouette body type preferences of African American, Asian, Latino and white men, find no significant differences between race–ethnicity and body type preferences. Similarly, the community study of African American, Asian, Latino and white dieters of Cachelin et al. (1998) show no racial–ethnic differences in ratings of attractive male and female shapes. In contrast, studies show that white men prefer thin bodies for women (Greenberg and LaPorte 1996) and that they feel they would be ridiculed for dating a woman with a heavier than the ideal body type (Powell and Kahn 1995), although some studies show they equally prefer thin and average bodies (Furnham and Radley 1989; Furnham, Hester and Weir 1990). Additional studies confirm racial–ethnic differences such that African-American men are more accepting of heavier women, but white men choose thinner silhouette figures as ideal female body types, and hope their girlfriends will lose weight significantly more often than African-American men (Greenberg and LaPorte 1996). Compared to white men, African-American men prefer larger body types for women and attribute fewer negative and more positive personality traits and qualities to obese women (Jackson and McGill 1996).

Other studies show that non-whites are heavily influenced by their respective racial–ethnic cultural standards of beauty (e.g. Crandall and Martinez 1996; Desmond et al. 1989; Massara and Strunkard 1979; Shaw 2005) that mediate how they understand, interpret and make sense of media (Duke 2000; Hunt 1997). For example, race–ethnicity affects what parts of teen magazines young girls focus on, such that African-American girls are less likely than white girls to read the sections focused on beauty tips (Duke 2000). Non-white minorities, particularly non-white minority women, do not accept the mainstream, white beauty and body ideals (see Barnett, Keel and Conoscenti 2001; Demarest and Allen 2000; Duke 2000), and are more accepting of heavier bodies defining beauty in terms of personality traits rather than as physical characteristics (Landrine, Klonoff and Brown-Collins 1992; Parker et al. 1995). Beauboeuf-Lafontant (2003) and Root (1990) find that African-American and Latina women have a lower incidence of eating disorders than do white women, and Schooler and colleagues (2004) conclude that exposure to mainstream (i.e. white dominated) media has no effect on African-American women’s body images but results in poorer self-esteem among white women. Yet, these findings have been contradicted by the work of Shaw and colleagues (2004), who find no significant ethnic differences in eating disturbances.

Some scholars suggest that non-whites have oppositional ways of interpreting the dominant cultural images of beauty or that people of color find sources other than mainstream mass media and popular culture to form their ideals of beauty (see Craig 2002; Duke 2000; Durham 1999). Racial–ethnic groups may be insolated by their culture. Latinos, for example, report higher desired weights for women than do whites (Winkleby et al. 1996), and it is suggested that their cultural standards of beauty differ from those of whites (Harris and Koehler 1992; Rittenbaugh 1982). Shapelier and heavier body types are not only accepted within the community, but more desired. Similarly, scholars claim African-American women are protected from mass media portrayals by their sub-culture (Molloy and Herzberger 1998), perceiving African-American men as rejecting the “white” standard of beauty by preferring larger body sizes (Milkie 1999). However, Poran’s (2006) in-depth interviews of 15 female African-American college students find they are not entirely insolated, and instead feel pressure to be thin, and to conform to the preferences of men of diverse ethnicities.

There are few studies assessing the preferred body types of heterosexual Asian Americans, or their preferences for opposite sex body types. The results of these studies are contradictory with some concluding that white women are less satisfied with their bodies than are Asian-American women (Akan and Grilo 1995; Altabe 1998), and other studies finding Asian-American women are as likely as are white women to have body concerns and weight dissatisfaction (Gluck and Geliebter 2002; Koff et al. 2001; Sanders and Heiss 1998). There is some evidence showing that acculturation and the acceptance of Western ideals is a major risk factor for eating disorders among Asian Americans (Davis and Katzman 1998; Hall 1995). Asian-American women exhibiting a stronger adherence to family norms are less likely to develop bulimia or anorexia. This suggests that Asian Americans, much like Latinos, and African Americans similarly experience conflicting messages between their culture and the dominant white culture.

Given the conflicting findings of these studies, it is difficult to predict the extent to which racial–cultural influences override the influences of the dominant culture. Still, we reason that in the absence of competing racial cultural imagery, whites will be more highly influenced than are non-whites by dominant cultural imagery.

Whites will be more likely than all other racial–ethnic groups (African-Americans, Asians, and Latinos) to indicate specific preferences for dates’ body types.

White men as compared to non-white men will be more desirous of the dominant, idealized (i.e. thin) female body type.

White women as compared to non-white women will be more desirous of the dominant, idealized (i.e. muscular) male body type.

We employ separate logistic regression models to test each hypothesis. We test Hypothesis 2a with two logistic regression models; the first model tests race–ethnic and gender differences without any controls and the second tests if these differences remain while adding controls for how selective a dater is overall as well as a dater’s own age, level of education, region of the country, body type and racial–ethnic preference for dates. To test hypotheses 2b and 2c we split the sample by gender so that one model includes all male daters, the other all female daters. In all models we include controls for how selective a dater is on other qualities as well as a dater’s own age, level of education, region of the country, body type and racial–ethnic preference for potential dates.

We suspect that within each gender group, whites will be more likely than all non-white groups (African-Americans, Asians and Latinos) to desire these idealized body types because they have fewer contending media outlets than do non-whites and have less reason to question these images since people of their same race group are represented. If we find that this is the case, it might reasonably suggest two different scenarios; it may be that African-American, Asian, and Latino men and women acknowledge a variety of bodies as attractive, and/or that non-whites are just as selective in preferring another body type that has not been identified here as “ideal.” We expect both will be the case. First we hypothesize that non-whites, both men and women, will be open to a larger variety of body types than their white counterparts:

Non-white men will be open to dating women with a wider variety of body types than will white men.

Non-white women will be open to dating men with a wider variety of body types than will white women.

We run logistic regressions separately for men and women to determine whether non-white daters are open to a greater number of body types than are white daters. In each of these regressions we restrict our sample to daters who indicate specific body type preferences. The dependent variable for each regression is the number of body types (ranging from 1 to 10) that a dater indicates as preferable for a potential date. Daters who indicate no preference for a date’s body are not included in this analysis; the key independent variable is race–ethnicity. In each model we also control for a dater’s overall selectivity, own age, own level of education, region of the country, own body type, racial–ethnic preference for potential dates and the number of other body type preferences a dater has.

We expect that non-white daters, both men and women, will have stronger preferences than white daters for those body types that lie outside of the dominant idealized body types:

Non-white men will be more open than white men to dating women with average, curvy, thick, extra, large, and voluptuous body types.

Non-white women will be more open than white women to dating men who have slim, slender, average, thick, curvy, extra, large and voluptuous body types.

Lastly, to examine these hypotheses (4a and 4b), that whites, both men and women, will be less open than their non-white counterparts to dating people with body types that fall outside of the culturally dominant ideal, we look at the propensity of daters to choose those body types not captured by the idealized types previously described. To test Hypothesis 4a, we restrict our sample to men who have body type preferences and run separate logistic regressions to examine the racial–ethnic differences in the propensity to choose each of the body type categories provided by Yahoo Personals: Slim, Slender, Average, Athletic, Fit, Thick, A few extra, Large, Curvy and Voluptuous. In all of these analyses we control for a dater’s overall selectivity, own age, own level of education, region of the country, own body type, racial–ethnic preference for potential dates and the number of other body type preferences a dater has. We repeat these methods, using the same set of controls, to test Hypothesis 4b, but restrict our sample to women who have body type preferences.

Method

Sample

To analyze daters’ body type preferences we use internet dating profiles of heterosexual daters in the United States. Daters are a particularly good gauge for measuring body type preferences since attractiveness is an important criterion of date selection (Kurzban and Weeden 2005; Stewart et al. 2000). These data are unique and allow us to analyze dating behavior in a natural setting. While employing internet dater preferences has not been a widely used approach, scholars in a variety of fields are beginning to examine internet, video, and speed daters as well as print advertisements to ascertain what traits are considered when choosing a potential partner (Kurzban and Weeden 2005; Sakai and Johnson 1997), and to locate differences in dating preferences and behaviors by race (Miller et al. 2004; Phua and Kaufman 2003; Yancey 2009) and gender (Bolig et al. 1984; Goode 1996; Lance 1998). A number of studies have specifically examined the phenomenon of internet dating (Fiore and Donath 2004; Madden and Lenhart 2006; McKenna et al. 2002).

We collected dating profiles from Yahoo Personals, a national, internet dating site that, at the time of data collection, did not charge a fee for this service. At the time of data collection, Yahoo Personals was also the most popular internet dating website in the United States (Madden and Lenhart 2006). Dating profiles were collected between September 2004 and May 2005 by logging into the website as a user. For all searches, an age category of 18–50 years old was selected, as those over the age of 50 are less likely to be internet users (Madden and Lenhart 2006). Geographic dating regions within 50 miles of Atlanta, Chicago, Los Angeles and New York City were also selected. These cities provide regional diversity, while also varying in racial composition and ideology. Further, by selecting a region spanning 50 miles from the metropolitan areas we hope to have a mix of both urban and suburban daters in our sample.

We selected men and women who self-identified as being one of four racial or ethnic categories: African American, Asian, Latino or white. Within each region/gender/racial–ethnic category, 200 participants were randomly selected, though there were fewer than 200 Asian female and Latina daters in Atlanta. Search terms were also entered for “man seeking women” and “woman seeking men;” as such, these data only include those with heterosexual dating preferences. Further, only profiles with photographs were selected in order to limit the possibility of misrepresentation by daters. The final data set includes 6,070 internet daters. This is fewer than the target sample size of 6,400 due to the smaller number of female Asian and Latina daters in Atlanta and the posting of duplicate profiles by the same dater. After deleting cases with missing values on the variables of interest for this study, the final sample size for this study is 5,810. The final sample is split almost evenly by region, race–ethnicity and gender (see Table 1).

Table 1

Sample characteristics by race–ethnicity and gender: means and percents*

		MEN				WOMEN
		African American	Asian	Hispanic/Latino	White	African American	Asian	Hispanic/Latina	White
Dependent Variables
Body Type Preference a (1= Body type preference for potential dates)	Daters with any body type preference for potential dates	77.29b	76.85b	85.95b	76.66	76.49	79.45	75.11	75.94
Ideal Female Body Type Preference a (1= Wants dates with thin or thin toned bodies)	Daters who exclusively prefer slim slender dates	5.31b	8.23b	7.29b	22.55	2.43	2.76	2.86	2.36
Ideal Male Body Type Preferencea (1= Wants dates with fit athletic bodies)	Daters who exclusively prefer fit athletic dates	1.20	0.98	2.25	1.16	8.78	10.12	6.44c	10.57
Independent Variables
Selectivity a	Mean percentage of characteristics on which daters made requests (out of 20 possible characterisitics)	33.98 (sd=19.50)	35.48 (sd=18.73)	32.84b (sd=19.48)	35.56 (sd=18.41)	53.47c (sd=19.24)	51.20c (sd=20.27)	51.52c (sd=19.88)	48.00 (sd=20.94)
Age (in years)		32.37b (sd=7.35)	29.98b (sd=6.56)	30.14b (sd=7.32)	33.66 (sd=6.98)	32.30 (sd=7.37)	30.76c (sd=7.33)	29.40c (sd=8.25)	32.64 (sd=8.50)
Education
≤ High Schoola (reference)	Respondent’s level of education	6.51	3.91	11.67b	4.64	3.11c	5.06c	16.74c	8.21
Some Collegea		42.90b	18.41	45.49b	21.01	43.92c	20.70c	43.49c	30.46
College Grad		39.58b	48.68	32.76b	48.20	36.76c	49.85c	30.62c	42.70
Post- Collegea		11.02b	29.01	10.08b	26.16	16.22	24.39c	9.16c	18.64
Region
Atlanta (reference)a	Metropolitan area in which the respondent dates	23.77	24.13	24.40	34.36	25.54	14.57c	19.31c	26.56
Chicago		25.10	26.92	23.87	34.87	26.35	28.68	26.61	25.31
Los Angeles		25.10	24.97	25.99	23.71	24.19	28.37c	26.61	22.67
New York		26.03	23.99	25.73	27.06	23.92	28.37	27.47	25.45
Body Type
Averagea (reference)	Self-reported body type of the respondent	21.91	35.98b	39.39b	21.39	25.27c	35.89	36.48	34.63
Smalla		4.52b	13.53b	6.23	7.22	15.54c	44.17c	18.47c	30.32
Athletica		59.36b	44.77b	40.98b	65.98	16.89c	15.49c	11.87c	23.64
Largea		14.21b	5.72	13.40b	5.41	42.30c	4.45c	32.90c	11.40
Race-Ethnic Preference
No Preferencea	Respondent's preference for potential dates' ethnicity	44.49	36.12	45.62	40.85	22.97c	26.69	25.32	27.54
White Onlya		0.80b	2.23b	0.93b	17.01	1.62c	18.25c	6.58c	46.87
Not Whitea		39.71b	22.18b	16.71b	3.99	58.65c	8.13c	25.18c	3.06
Both White and Non-White (reference)		15.01b	39.47	36.74	38.14	16.76c	46.93c	42.92c	22.53
N		753	717	754	776	740	652	699	719

These values consider all of the daters. The subset of daters who express body type preferences is similar to the entire sample. (Available upon request)

aWhen the entire sample is considered, there are significant gender differences (p≤.05) for this variable

bThere is a significant difference between this group and white male daters in this sample

cThere is a significant difference between this group and white female daters in this sample

This sample contains only people who willingly choose to date online. Therefore, some social groups or subcultures may be under-represented or excluded. These data are also likely to under-represent individuals of low socio-economic status and others who may not have as much access to or experience using computers. Racial–ethnic minorities are over-represented in the sense that there are a smaller proportion of non-whites in the U.S. population than in this sample. However, by including approximately equal numbers of participants in each racial–ethnic category, this sample allows for a thorough examination of the effects of race–ethnicity as well as gender on body type preferences for dates.

Yahoo Personals provides a wealth of self-reported socio-demographic information about each participant. The greatest benefit of using these data is that they examine daters’ actual stated preferences in a natural setting. The daters in our sample did not indicate their preferences on a questionnaire or in a laboratory; rather, these are their actual dating profiles created for the express purpose of actually meeting potential dates. This is important in two ways. First, most studies examine marital outcomes (e.g. Kalmijn 1993, 1994, 1998; South 1991) or present subjects with hypothetical situations of whom they are willing to date (e.g. Miller et al. 2004; Stewart et al. 2000; Thompson et al. 1996). Studying marital outcomes does not attend to the full range in which daters are willing to date; rather, they observe only one outcome. Further, in research settings, it is more likely that subjects will acquiesce and provide responses they feel are socially acceptable or that will please the researcher.

Second, by studying the actual preferences of internet daters, we can understand the parameters daters place on their dating markets without the constraints of other factors, such as proximity or the composition of their social group, which can often constrain daters off-line. The internet allows daters to vastly expand their dating pools so that they can assert preferences, which may not be realistic in some other settings. For example, in a racially homogenous community, those daters who are willing to date individuals of varied races and ethnicities will likely pair with someone of their own race–ethnicity. In this case, examining marital outcomes will not reveal the degree of heterogamy that daters will accept; their stated preferences, on the other hand, can.

Measure

Dependent Variables

Each dater is asked about his or her preferences for a date. The specific characteristics of the participant and his or her preferences for a potential date include, among others: gender, age, region of the country, race–ethnicity and highest level of education. In addition, there is a measure specific to a respondent’s own body type and preferred body type of a potential date. The main dependent variables are derived from daters’ preferred body types for potential dates. Body type is defined as it is on Yahoo Personals. There are ten categories for a date’s body type: Slim, Slender, Average, Athletic, Fit, Thick, A Few Extra Pounds, Large, Voluptuous, and Curvy. If a dater has no preference for a date’s body type she or he can alternately choose “Any” (i.e. “no preference”). “Any” is also the default category so that a dater who chooses not to indicate what body type preferences she or he has for a date is counted as having no preferences. Daters are provided the same options to describe their own body type and their body type preferences for a date, regardless of gender.

For all of these analyses the dependent variable is derived from the part of daters’ profiles where they check which body type(s) they are seeking in potential dates. The way this variable is constructed into the dependent variable varies for different analyses; each dependent variable will be discussed in detail below.

Body Type Preferences

To determine whether men or women, whites or non-whites are more selective about dates’ body types, the main dependent variable is whether or not the dater indicates any preference at all for the body type of a potential date. As is described above, daters have 11 body type categories they can select as a preference for a potential date. They can choose either “Any,” indicating that they don’t state a preference for potential dates’ body types, or they can select any combination of the other 10 categories. For this measure, a selection of “any” was coded as a “0,” indicating no preference; any other selection indicates some preference and was coded as “1.” This measure is used to examine Hypotheses 1 and 2a, which assert that men will be more likely than women and whites will be more likely than non-whites to have specific body type preferences for dates.

The Ideal Body

We expand our analyses to address the intersections of race–ethnicity and gender in regard to daters’ adherence to dominant beauty standards by focusing on those daters who do specify a body type preference. We examine the extent to which individuals of different race–ethnicities accept the dominant, idealized images of beautiful bodies by taking gender and race–ethnicity into consideration simultaneously. This body type is limited to thin or thin and toned bodies for women and fit athletic bodies for men.

Since we are interested in the degree to which the dominant body ideal is accepted, we construct a variable from the available body type categories to represent this ideal. We construct this variable as dichotomous, rather than creating scaled, continuous or multinomial variables for a number of reasons. First, since daters can choose any number of the ten body types, there are over 1,000 different possible combinations of body type preferences daters can indicate. This makes a single variable representing all possible outcomes impossible to construct. In addition, the body type categories are not necessarily distinct or hierarchically orderable, and so they cannot be organized in any meaningful way to create a scale.

The Female Body

The dominant portrayal of attractive female bodies is premised on an ideal of thinness (Bordo 1993; Silverstein et al. 1986; Urla and Swedlund 2000) so daters who chose any combination of the “Slim” and/or “Slender” selections, with all other body type preferences excluded, was coded “1.” Since the ideal female body has also been increasingly portrayed as muscular and toned (Bordo 1993), any time a preference to date “Slim” and/or “Slender” individuals who are also “Fit” and/ or “Athletic” is indicated, it is also coded as “1,” provided no other body type preferences are selected. We refer to the ideal female body type as ‘thin’ or ‘thin and toned.’

The Male Body

Another dichotomous variable represents the ideal male body type, which is portrayed as extremely fit and muscular (Bordo 1999; Connell 1995; Pope et al. 2000). Therefore, we code any combination of “Fit” and/or “Athletic” as “1” when no other body types are selected; all other selections are coded as “0.” In order to ensure that these empirical characterizations of the theoretically conceptualized body type are valid, we have two criteria. First, that they are popular body type preferences within this sample. Second, that they are gender-specific preferences; men should prefer the thin, toned body type for their dates more often than women; women should prefer dates with a fit athletic body type more frequently than men prefer dates with this body type. We do find that both of these criteria are met.

Further, of the 2,378 men in the sample with body type preferences, a substantial number, 85.6%, have date preferences that include thin or thin and toned women. Furthermore, 86.1% of the 2,155 women with body type preferences desire men with fit or athletic bodies. This establishes that what we identify as the ideal female body type and the ideal male body type are popular within this sample.

Independent Variables

Individuals’ racial–ethnic category is the main independent variable of concern. Daters are asked to identify themselves as belonging to one of the following categories: African American, Asian, Caucasian/white, East Indian, Hispanic/Latino, Middle Eastern, Native American, Pacific Islander, Inter-racial, other, or “I’ll tell you later” (i.e. no answer). Our sample consists of daters who self identified as African American, Asian, Hispanic/Latino or white. Race–ethnicity is coded into dummy variables for the purposes of the final logistic regression models. Racial–ethnic categories include African American, Asian, Hispanic/Latino and white. Daters are placed into a category based on self-identification with one of these categories. Each racial–ethnic group will be compared to the white group of daters, since the American cultural hegemony is premised on a white standard. Gender is also a key independent variable. Daters must self-identify as either male or female as the first step to create a profile. When the logistic regression analyses are not run separately for men and women, gender is included in the model and is coded as a dichotomous variable, with male coded as “1” and female as the excluded reference category.

Control Variables

As was previously described, in all analyses we control for those traits that past research has identified as important to mate and date selection. Specifically, we control for a dater’s overall selectivity, own age, own level of education, region of the country, own body type and racial–ethnic preference for potential dates. To control for daters’ selectivity, we calculate the number of items on which a dater indicated preferences as a percent, ranging from 0% to 100%. The dating profiles include 20 different items, such as eye color, hair color, income, religion and smoking and drinking habits, on which a dater can specify preferences for potential dates’ characteristics. On average, women had preferences on about ten different items and men had preferences on seven (see Table 1). The more items on which a dater indicates a preference, the higher the dater’s selectivity score.

We measure age as a continuous variable. All daters are between the ages of 18–50 years old. We collapse the categories for education present on the Yahoo dating profile into the following four groups: high school diploma or less, some college, college degree, post-college education. We include these in all regression analyses as dichotomous variables with “1” indicating that is the dater’s highest level of education and “0” indicating that it is not. The omitted reference group is those daters with a high school diploma or less.

We sampled daters from within 50 miles of four U.S. cities: Atlanta, Chicago, Los Angeles and New York. We measure each of these as dichotomous variables with “1” indicating a dater is from that region and “0” indicating she or he is not. In all analyses we use Atlanta as the omitted reference group.

We also control own body type. Daters are presented the same options to describe their own body types that they are given to describe dates’ body types, however, they can only choose one body type to describe their own body. We group these body types into four broader categories: Small, Average, Athletic and Large. All selections of Slim and Slender are coded as “Small;” Average is coded “Average;” Fit and Athletic are coded “Athletic;” and Extra, Large, Curvy and Voluptuous are coded as “Large.” We include these in all analyses as dichotomous variables; Average is the omitted reference category.

Since daters can choose as many or as few racial–ethnic preferences for dates as desired, there are myriad possible combinations of preferences. As such, the regression models do not control for preferences for dates of specific race–ethnicities (i.e. African-American, white, etc.). Instead, we create variables reflecting daters’ preferences to date non-whites. Constructing this variable in this way also avoids a problem of colinearity attached to representing each racial–ethnic group separately since, when daters in this sample wanted to date one minority group explicitly, their preference was often for dates of their own racial–ethnic group. Each dater falls into one of four groups: those who indicate no racial–ethnic preference for potential dates; those who prefer to only date whites; those who prefer to only date non-whites; and those who indicate preferences that include both whites and non-whites. These are constructed as dichotomous variables, and in all analyses the group of daters having preferences that include both white and non-white daters is the omitted reference group.

For our final set of analyses, which test Hypotheses 4a and 4b (see Table 5), we also add a control for the number of body type preferences a dater has. Since we run separate regressions for each body type here, we calculate the number of total preferences, excluding the body type of interest for each regression, in order to avoid problems of endogeneity.

Table 5

Odds ratios from logistic regression analyses of the effects of race–ethnicity on requesting each body typea

Men	Slim	Slender	Average	Athletic	Fit	Thick	Extra	Large	Voluptuous
Independent Variables
Race–ethnicity
African American	0.311***	0.207***	2.033***	0.745	0.322***	11.321***	2.544***	1.626***	1.830**
Asian	0.613*	0.748*	2.521***	0.751	0.991	1.045	1.220	1.757	0.653
Hispanic/Latino	0.407***	0.379***	1.707***	1.261	0.418***	2.397***	1.527**	3.244*	1.709*
White (reference)	–	–	–	–	–	–	–	–	–
N	2,378
Women	Slim	Slender	Average	Athletic	Fit	Thick	Extra	Large	Voluptousb
Independent Variables
Race–ethnicity
African American	1.209	1.366	0.902	1.380	1.125	0.956	0.862	1.309	–
Asian	0.730**	1.312	1.176	0.453***	1.470**	0.348***	0.597**	0.832	–
Hispanic/Latino	0.739†	1.946***	0.965	0.418***	1.555**	0.978	0.773	0.978	–
White (reference)	–	–	–	–	–	–	–	–	–
N	2,155

†p≤ .1 *p≤.05 **p≤.01 ***p≤.001

aAll analyses include the following controls: Selectivity, Age, Education ( High school, Some college, College graduate, Post-college), Region (Atlanta, Chicago, Los Angeles, New York), Own Body Type (Small, Average, Athletic, Large), Racial–Ethnic Preference (No preference, White only, Non-white, Both white and Non-white) and Number of Other Body Type Preferences, excluding the body type indicated by the dependent variable. Full tables are available upon request

bOnly 14 women responded they would date Voluptuous males, so these findings are inconclusive and are not presented

Results

Descriptive Findings

The results of initial descriptive analyses show that most daters in the sample express some sort of body type preference: 78% (N = 4,355) of all daters in the sample have body type preferences. Men and women are both likely to have some preference for body type, and among all race–ethnic groups, at least 75% of the daters in our sample express some sort of body type preference for dates. Among men, there are significant race–ethnic variations in having body type preferences but among women there are not. Employing chi square analysis or significance tests of differences between means, we examine gender differences as well as within gender group racial–ethnic differences between daters on each of these variables before including them in our models (Table 1). We examine within gender racial–ethnic differences by comparing each race–ethnic group to the white group of daters of the same gender. Men have significantly (p ≤ .05) fewer stated preferences for dates overall and tend to have more education than women. Compared to men, women are more likely (p ≤ .05) to describe their own bodies as Small or Large but less likely to describe them as Average or Athletic, more likely to indicate no preference for dates’ race–ethnicity, and more likely to exclude whites from their pool of eligible dates while also being less likely to want to date only whites. There is also a significant gender difference in the likelihood of residing in Atlanta, because there were fewer than 200 female Asian and Latina daters in Atlanta. Men are more likely than women have post-college education and, among those daters with specific body type preferences for dates, but not the entire sample of daters, men more often than women have a college degree (p ≤ .05).

Among male daters we find that the non-white groups of daters are, on average, younger than white male daters, differ in the way they describe their own bodies and differ in their race–ethnic preference for dates. Latino men also tend to state fewer overall preferences for dates and African-American and Latino daters have lower levels of education than do white men. Among the female daters, all non-white groups state more preferences overall for dates and have different racial–ethnic preferences than whites. African American and Latina female daters tend to have less education than white female daters, while the Asian women in the sample tend to have more. There are also racial–ethnic differences among female daters’ descriptions of their own body types when non-white groups are compared to whites: African-American women are less likely to describe their bodies as Small, Average or Athletic but more likely to describe them as Large; Asian women are more likely to describe their bodies as Small but less likely to describe them as Athletic or Large; Latinas are less likely to describe their bodies as Small or Athletic but more likely to describe them as Large.

Gender Differences

We first test Hypothesis 1, that men will be more likely than women to state a preference for potential dates’ body types (Table 2, models 1 and 2). We employ logistic regression analyses to test gender and racial–ethnic differences; the dependent variable is whether a dater has any preferences for dates’ body types. The independent variables are a dater’s own gender and race–ethnicity. We use two models to test Hypothesis 1; the first tests whether there are race–ethnic and gender differences and the second tests if these differences remain while also controlling for individual daters’ overall degree of selectivity, own age, own level of education, region of the country, own body type and racial–ethnic preferences for potential dates. We find strong support for Hypothesis 1: men are significantly more likely than women to have body type preferences (OR = 1.161, p ≤ .05). When controls are introduced this gender difference is even more pronounced (Table 2, model 2); men are over 3.5 times more likely than women to have body type preferences (OR = 3.766; p ≤ .001).

Table 2

Odds ratios from logistic regression analyses of the effects of gender and race–ethnicity on whether daters state body type preferences for potential dates

	All Daters		Men	Women
	Model 1	Model 2	Model 3	Model 4
Independent Variables
Male (vs. female)	1.161*	3.766***	---	---
Race-Ethnicity
African American	0.774**	0.839	0.895	0.886
Asian	0.826*	0.697**	0.554***	0.971
Hispanic/ Latino	0.732**	0.815†	0.934	0.780
White (reference)	---	---	---	---
Control Variables
Selectivity		1.086***	1.101***	1.079***
Age		1.000	1.008	0.992
Education
≤ High School (reference)		---	---	---
Some College		0.876	1.216	0.713
College Grad		0.858	1.555†	0.543**
Post- College		0.647**	1.282	0.375***
Region
Atlanta (reference)		---	---	---
Chicago		1.030	0.901	1.143
Los Angeles		0.867	0.735*	0.989
New York		0.941	0.851	1.078
Body Type
Average (reference)		---	---	---
Small		1.395**	1.710**	1.329*
Athletic		1.664***	1.512***	1.996***
Large		0.433***	0.368***	0.464***
Race-Ethnic Preference
No Preference		0.724**	0.635**	0.856
White Only		0.769	1.076	0.848
Not white		0.631***	0.476***	0.840
Both White and Non-White (reference)		---	---	---
N	5,810	5,810	3,000	2,810

†p≤ .1 *p≤.05 **p≤.01 ***p≤.001

Racial–Ethnic Differences

We find mixed support for Hypothesis 2a, which predicted that whites would be more likely than non-whites to indicate preferences for dates’ body types. We use logistic regression analyses to examine the effects of race–ethnicity on whether or not daters specify particular body type preferences for potential dates. Controlling only for gender, Table 2, model 1, shows that all groups of non-white daters in this sample are significantly less likely than white daters to indicate a body type preference for potential dates (p ≤ .05). This relationship remains statistically significant for Asian daters, and approaches significance for Latinos, while also controlling for other dater preferences and demographic characteristics, including: a dater’s overall selectivity, own age, own level of education, region of the country, own body type and racial–ethnic preference for potential dates (Table 2, model 2).

We also find that daters’ overall selectivity, own body type, level of education and racial–ethnic preferences for dates influence whether daters have body type preferences: the more selective daters are overall, the more likely they are to have body type preferences; compared to daters with Average bodies, those with Small and Athletic body types are more likely to have preferences and those with Large bodies are less likely; daters with no racial–ethnic preferences for dates and those who prefer to not date whites are less likely than those with preferences for white and non-white dates to have body type preferences.

Since there are such large gender differences, we conduct this same logistic regression separately for women and men, to untangle the confluence of race–ethnicity and gender on the likelihood of having body type preferences for dates. In this analysis, the dependent variable remains whether or not a dater specifies any body type preference, the independent variable is race–ethnicity and we control for a dater’s overall selectivity, own age, own level of education, region of the country, own body type and racial–ethnic preference for potential dates. These results are presented in Table 2. We also ran a logistic regression that included interaction effects of race–ethnicity and gender (not shown) and the findings of this analysis are similar to the regression models that are split by gender.

Once the sample is split by gender, racial–ethnic differences only persist between Asian men and white men; Asian men are about 45% less likely than white men (OR = .554; p ≤ .001) to indicate body type preferences. The effects of the control variables generally remain the same in terms of the significant effects own body type and selectivity. The significant effects of education persist only for women, as women with college and post-college degrees are less likely than women without high school degrees to express body type preferences. The effects of racial–ethnic preferences for dates remain only for men.

Wanted: Thin Women and Fit Men

We also examine if daters who do have body type preferences are likely to prefer dates who exhibit the dominant cultural idealizations of a beautiful body. Daters who indicate that they have no body type preferences are not included in the remaining analyses. The sample of daters in the remaining analysis is restricted to the 4,533 (2,378 men; 2,155 women) internet daters with stated preferences for potential dates’ body types. While most daters are interested in dates with the ideal body types, they are also open to dating individuals with other body types as well. Of the 1,865 women who indicate a preference for men with fit athletic bodies, only about 13% (N = 252) want to date men only if they have this body type; most women want dates with fit athletic bodies but also indicate preferences for dates with other body types. Of the 2,036 men with a preference for thin or thin and toned body types about 15% (N = 329) only prefer dates with these body types; most male daters choose the thin or thin and toned body type categories in conjunction with other body type categories. For the remaining analyses, we are concentrating on those who prefer dates with the culturally ideal body to the exclusion of all other body types in order to focus on how race–ethnicity is driving this specific standard of beauty.

Gender Differences

Before we can thoroughly examine differing preferences for dates with ideal body types, we need to be certain that the thin or thin and toned body type preference is strongly associated with male preferences and that the fit athletic body type is strongly associated with female preferences for male dates. Using logistic regression analyses, Table 3 presents four logistic regression that test gender and racial–ethnic differences in the desire to have dates with culturally ideal body types—i.e. thin women and fit men (Hypotheses 2b and 2c). Race–ethnicity is the key independent variable in all models and, when the entire sample is considered, gender is also included. All models also control for a dater’s overall selectivity, own age, own level of education, region of the country, own body type and racial–ethnic preference for potential dates. Table 3 presents two logistic regressions that include all daters who have body type preferences. These regressions test the odds that male daters will prefer these body types exclusively more than female daters when race–ethnicity, demographic characteristics, and other dater preferences are controlled. We find that men are almost five times more likely than women to prefer to only date those with thin or thin and toned bodies (OR = 4.933; p ≤ .001) and are significantly less likely than women to prefer only fit athletic dates (OR = .102; p ≤ .001). This indicates that the body types we define as culturally dominant are gender-specific preferences among these daters.

Table 3

Odds ratios from logistic regression analyses of the effects of gender and race–ethnicity on the likelihood that daters will prefer culturally idealized body types.

	Thin or Thin Toned Body Type Preference		Fit Athletic Body Type Preference
	All daters	Men Onlya	All daters	Women Only
Independent Variables
Male (vs. female)	4.933***		0.102***
Race-Ethnicity
African American	0.360***	0.253***	0.865	0.769
Asian	0.419***	0.399***	1.017	0.964
Hispanic/ Latino	0.610**	0.498***	1.017	0.739
White (reference)	---	---	---	---
Control Variables
Selectivity	1.010**	1.009*	1.010**	1.009*
Age	1.010	1.020*	0.986	0.987
Education
≤ High School (reference)	---	---	---	---
Some College	0.690	1.158	0.872	0.922
College Grad	1.109	1.806	0.864	0.877
Post- College	1.345	1.806	0.863	0.910
Region
Atlanta (reference)	---	---	---	---
Chicago	0.930	0.805	0.546**	0.545**
Los Angeles	1.277	1.195	1.016	1.008
New York	1.349	1.317	0.786	0.813
Body Type
Average (reference)	---	---	---	---
Small	6.920***	5.461*	2.993***	2.879***
Athletic	5.070***	4.755***	5.370***	5.229***
Large	0.467	---	0.827	0.772
Race-Ethnic Preference
No Preference	0.922	0.926	1.497*	1.447†
White Only	0.804	1.081	1.505*	1.332
Not white	0.934	1.026	1.934***	2.046***
Both White and Non-White (reference)	---	---	---	---
N	4,533	2,194	4,533	2,155

†p≤ .1 *p≤.05 **p≤.01 ***p≤.001

aMen who described their bodies as Large (N=184) were dropped from this analysis since this perfectly predicted

the outcome; no large men desired the thin or thin and toned body type exclusively

Within Gender Race–Ethnic Differences

For men, race–ethnicity does affect the likelihood that a dater will prefer the ideal thin or thin and toned female body type exclusively while for women it does not. Table 3 presents a logistic regression that includes only male daters with body type preferences which tests our prediction (Hypothesis 2b) that white men will be more likely than non-white men to state preferences for thin toned female dates. We control for the degree to which daters are selective, daters’ own age, own level of education, region of the country, own body type and racial–ethnic preference for potential dates. We find support for Hypothesis 2b; all groups of non-white men are significantly less likely than white men to have body type preferences for thin or thin and toned women. White male daters are about 75% (OR = .253; p ≤ .001) more likely than African-American men, 60% (OR = .399; p ≤ .001) more likely than Asian men and 50% (OR = .498; p ≤ .001) more likely than Latino men to prefer to date thin toned women.

Employing regression analysis and controlling for the degree to which daters are selective, daters’ own age, own level of education, region of the country, own body type and racial–ethnic preference for potential dates we test Hypothesis 2c which predicts, similarly, that white women will be more likely than non-white women to exclusively prefer the culturally idealized fit athletic male body type. This is tested in a logistic regression (Table 3) that includes only female daters who have body type preferences. We do not find support for Hypothesis 2c; white female daters are no more likely than any group of non-white female daters to desire male dates who have fit athletic bodies.

Own body type remains important in all analyses examining preferences for the idealized body types. Compared to same-gender daters with Average bodies, men with Small and Athletic bodies are notably more likely to exclusively prefer thin and toned dates (OR = 5.461, p ≤ .05; OR = 4.755, p ≤ .001). Similarly, Small and Average women are more likely to prefer the ideal male body exclusively (OR = 2.879, p ≤ .001; OR = 5.229, p ≤ .001).

We use linear regression analysis to examine our prediction that non-white men and women will be open to a wider variety of body types than will their white counterparts (Hypotheses 3a and 3b, respectively). The dependent variable is the total number of body type preferences a dater selects, race–ethnicity is the independent variable of interest, and we control for a dater’s overall selectivity, own age, own level of education, region of the country, own body type and racial–ethnic preference for potential dates. Support for hypotheses 3a and 3b varies and is dependent on the intersections of gender and race–ethnicity. The results for these regressions are presented in Table 4. We find that, when daters’ demographic characteristics and other date preferences are controlled, African-American male daters indicate significantly more while Latino men indicate fewer (.225, p ≤ .05) body types as desirable than do white male daters (.296, p ≤ .01). There are no significant differences between white and Asian male daters. African-American female daters have significantly more body type preferences than do white female daters (.247, p ≤ .05) while Asian females have fewer (.358, p ≤ .001). We find no significant differences in the number of body type preferences of white and Latina female daters.

Table 4

Coefficients from linear regressions of the effects of raceethnicity on the number of body type preferences daters havea

	Males	Females
Independent Variables
Race-Ethnicity
African American	0.296** (.10)	0.247* (.11)
Asian	-0.015 (.10)	-0.358*** (.10)
Hispanic/ Latino	-0.225* (.10)	-0.021 (.11)
White (reference)	---	---
Constant	4.905***	4.283***
R2	0.045	0.083
N	2,378	2,155

†p≤ .1 *p≤.05 **p≤.01 ***p≤.001

aAll analyses include the following controls: Selectivity, Age, Education (≤ Highschool, Some college, College graduate, Post-college), Region (Atlanta, Chicago, Los Angeles, New York), Own Body Type (Small, Average, Athletic, Large) and Racial-Ethnic Preference (No preference, White only, Non-white, Both white and Non-white). Full tables are available upon request

These results must be interpreted carefully, however, since the body type categories are not distinct or mutually exclusive. The body type categories are not organized in a meaningful hierarchical manner so certain choices might be considered synonymous by some daters but not others. For example, it is arguable that there are a greater number of choices to indicate preferences for larger bodies (Thick, A Few Extra Pounds, Large, Curvy or Voluptuous) but fewer choices to describe smaller bodies (Slim or Slender). As such, a dater who has a preference for larger dates might choose all five aforementioned categories whereas a dater with preferences for smaller dates might choose only the two later body type categories. This would make the first dater appear open to a wider variety of bodies than the second dater when, in fact, they are both just as selective—each preferring one body type. Other daters, however, might interpret meaningful differences between Slim and Slender or Extra and Voluptuous, etc. Further, different interpretations of the words Yahoo uses to describe bodies might vary along lines of race and ethnicity.

Our last set of hypotheses, 4a and 4b, assert that whites will be less open to dates with those body types not captured by the dominant ideal (i.e. thin for women and muscular for men). To test these hypotheses we run a separate logistic regression for each of the body type categories; these analyses are also split by gender (Table 5). In this analysis, the dependent variable is whether or not daters select each particular body type category; we include only daters with specific body type preferences (2,378 men; 2,155 women). Each analysis has a dichotomous dependent variable—whether or not the dater chooses each specific body type and the independent variable is a dater’s own race–ethnicity.

We run separate logistic regressions for each of the ten body type categories from which daters could choose. No daters in this sample indicate a preference for dates with Curvy bodies, so these results are not presented. Additionally, only 14 women responded they would date Voluptuous males, so these findings are inconclusive and are not presented. We do separate logistic regressions for each body type rather than making a scale because these body types are categorized in such a way that there is no clear hierarchy or order. Further, daters can choose as many or as few body types as desired, making a plethora of preference combinations possible. Since all daters are included in each logistic regression for each body type, we add a control for the number of other body type preferences each dater has to avoid problems of endogeneity. We also control a dater’s overall selectivity, own age, own level of education, region of the country, own body type, racial–ethnic preference for potential dates.

White men are significantly more likely than African-American, Asian and Latino men to prefer dates with Slim and Slender bodies and more likely than African-American and Latino men, but not Asian men, to prefer dates with Fit bodies. There is no statistically significant difference between white male daters and Asian, African-American or Latino male daters to prefer women with Athletic body types. All non-white male groups of daters in this sample are significantly more likely than white men to prefer women with Average body types.

While there was no significant difference between white and Asian men’s preferences for body types described as Thick, Extra, or Voluptuous, African-American and Latino men are significantly more likely than white men to prefer these body types. This last finding is particularly notable for the Thick and Large body types. Latino men are about 2.4 times (p ≤ .001) more likely than white men to prefer female dates with Thick bodies and about 3.2 times (p ≤ .05) more likely to prefer dates with large body types. African-American men are even more likely to prefer these particular body types, being 11 times (p ≤ .001) more likely than white men to prefer dates with Thick bodies and 10.6 times (p ≤ .001) more likely to prefer dates who describe their bodies as Large.

Hypothesis 4a predicts non-white men to be more selective than white men on body types that are not encompassed by the thin ideal. This is not supported when considering Asian male daters, as they are less selective or no different than white daters in most cases. However, when comparing African-American or Latino male daters to white male daters, there is support for Hypothesis 4a—African-American and Latino men have stronger preferences than white men for female body types that are not captured by the thin ideal. These results indicate that while white men are more selective than African-American and Latino men about having female dates with the culturally idealized thin or thin and toned body types, these non-white daters are also selective but not on the same body types: African-American men and Latinos in this sample are more interested in larger body types than are white men. Notably, these body types are not generally reflected positively in the mainstream popular culture.

We find less support for Hypothesis 4b, which predicts differences between white and non-white female daters’ preferences for male body types that fall outside of the fit athletic ideal. There are no significant differences between white women and African-American women in preferences for any of the body types. The greatest number of differences are between white female daters and Asian female daters, the latter being significantly less likely to prefer body types labeled as Slim (OR = .730, p ≤ .01), Athletic (OR = .453, p ≤ .001), Thick (OR = .348, p ≤ .001) and Extra (OR = .597, p ≤ .01) but more likely to prefer dates who are Fit (OR = 1.470, p ≤ .01). Latina daters are significantly less likely than white female daters to prefer men with Athletic body types (OR = .418, p ≤ .001) but more likely to prefer Slender (OR = 1.946, p ≤ .001) and Fit (OR=1.555, p ≤ .01) dates. The differences between white and non-white female daters are less pronounced than the race–ethnic differences we find for male daters and don’t follow any obvious pattern. Overall, we do not find strong support for Hypothesis 4b that non-white female daters would be more selective on body types that aren’t idealized in mass media.

Discussion

Who is more selective? We find that men are more likely than women to specify particular preferences for dates’ body types. However, it is important to consider that this might also be a reflection of differing gender role expectations. Men may feel more ease or entitlement to assert their expectations of dates’ bodies. Women might generally be more timid about expressing preferences, or feel it is not socially acceptable to violate notions of romance by placing too much value on the body.

Non-white female daters are no more likely than white female daters to have a preference for a date’s body type. When compared to white male daters, Asian men are the only group of non-white daters more likely to indicate they have no preference for a date’s body type. It is unclear why Asian men are different in their propensity to indicate body type preferences. However, controlling images of Asian men have been particularly emasculating and media images of Asian men are frequently desexualized (Chen 1999, 1996); their bodies have been portrayed as small and weak and even when they do have a masculine role, their sexuality is not addressed. For example, while Jackie Chan is an action hero, he is also comedic and never “gets the girl,” as is typical of Asian male portrayals in most American comedy and action films. The result of this may be that Asian men feel they will be less desired and therefore should be less choosy about the body of a potential date.

Adherence to Ideal Standards

African-American, Asian and Latino women are just as likely as white women to prefer the fit athletic body ideal for potential male dates. Previous research finds that African-American and Latina women are more accepting of heavier body types when discussing their own body, but we do not find that this translates to a more fluid definition of the ideal male body. This might be due to the fact that the ideal images of male bodies are more racially varied in the mass media. The fit athletic male is not only portrayed by white men in commercial advertisements but also by African-American men in images of athleticism and sport (Collins 2004; Jackson 2006; Messner 2002; Oates and Durham 2004). Though these images do not frequently portray Asian or Latino men, having at least a dual racial image of the ideal, muscular male body may create a more pervasive and universal body standard for men that is, in turn, consumed by a more diverse population of heterosexual female daters. This is an important finding as it suggests that the male body, much like that of the female body, is now viewed as an object, a victim of the commodification and limited acceptance of diversity in heterosexual and commercial relationships. These findings support current theory, which identify masculinity as a limiting and unrealistic ideal that all men are supposed to achieve (see Bordo 1998, 1999; Connell 1995; Kimmel 1994; Pope et al. 2000).

Of those Asian men who do state preferences for potential dates’ body types, they are significantly less likely to exclusively prefer the idealized thin toned body type and are about two and a half times more likely than white daters to indicate a preference for women with average body types. It is important, however, to note that “Average” may mean something different to daters of different race–ethnicities. Nonetheless, we believe Asian men are more likely to reject dominant ideals of beauty and to cultivate more realistic expectations of potential dates since, as already discussed, their bodies are either not represented or are represented negatively in the images they consume of the dominant culture’s ideals. Indeed, a recent study shows that Asian male college students are more likely than white men to view their bodies as smaller than the ideal (Barnett et al. 2001).

African-American men and Latino men are more likely than white male daters to prefer female body types described as Average, Thick, Extra, Large and Voluptuous. Only African-American men are open to greater variety of body types than white male daters. These findings suggest that the degree to which daters adhere to the dominant body standards is tied to their own race–ethnicity. However, we must be careful interpreting these results since the Yahoo Personals body type categories are not distinct or mutually exclusive categories. Also, while these results indicate various female body types as desirable, they do not suggest that women’s bodies are unconstrained by cultural standards of beauty. Men in this study are still much more likely than women to indicate that potential dates’ body types are important criteria for selecting potential dates. In fact, African-American and Latino men are very selective on other specific body types that are not captured by the thin or thin and toned body ideal. Most notable is the preference for women with larger body types.

One interpretation of literature which finds African-American and Latina women are more tolerant of excess weight on their own bodies suggests that African-American and Latino communities are more accepting and tolerant of body fat (Demarest and Allen 2000; Duke 2000; Poran 2002), and Asians and whites are not (Barnett et al. 2001). Our findings support the possibility of culturally specific ideals of beauty—some of which operate outside of the mainstream media and popular culture: African-American and Latino men are more accepting of larger female body types; African-American, Asian and Latino men are less interested than white men in the thin, toned female body; and African-American and Asian men generally choose significantly more body types as ideal than do white men.

That acceptance of, and preference for, dates with larger bodies exists for African-American and Latino men, but not women, highlights the importance of addressing intersections of race–ethnicity and gender. Our findings suggest that there are different cultural resources that influence non-white minorities in addition to mainstream images in the mass media, and that these portray a larger variety or a different variety of female beauty and bodies. This is endorsed by Craig’s (2002, p. 11) assertion that African-Americans have developed their own “semiautonomous social locations,” including schools, newspapers and social movements, which are aimed at fostering “critical readings of the dominant culture.” Perhaps, because dominant and idealized notions of beauty portray white women, rather than African-American women, there has been a chance to cultivate independent images of beauty within such semiautonomous social locations in the African-American community. However, since the images of the African-American male athlete loom large in popular culture, particularly in the age of televised sports, there has been less room to cultivate an image away from that of the dominant popular culture.

Limitations of the study

There are some limitations to using dating profiles as data for this research; however, we feel that the benefits greatly outweigh the costs. We are most concerned with the categorizations of race and ethnicity provided by the dating site because these categories conflate race and ethnicity, are overly broad, and gloss over much cultural and ethnic diversity. This is most notable for the categories “Asian” and “Latino,” which likely include the greatest diversity in terms of nationhood, ethnicity and culture. Within the category “Latino” there is racial diversity that we cannot distinguish within this sample of daters. We are also unable to discern the degree to which individual daters identify with racial–ethnic subcultures. This is particularly important for understandings of the body and beauty as perceptions of attractiveness vary by culture (Crandall and Martinez 1996; Shaw 2005).

Another concern with using online daters’ self reported profiles is that this information may be falsified—daters are likely trying to present themselves in the best possible light and might be willing to fabricate information. This, however, is not expected to have a grave effect on this study. Cornwell and Lundgren (2001) find that individuals are somewhat more likely to misrepresent themselves online than in person, but these trends do not vary significantly by gender. It is also unlikely that they will present false reports of their preferences for potential dates, since honestly answering these questions serves to filter out those individuals whom an online dater does not wish to date. If daters do lie, it is likely to be about their own body type, not the body type they desire of potential dates. To try to address this possibility and reduce deception, we only sample daters with pictures in order to limit the degree to which daters may misrepresent their own body types or other physical traits. Ellison et al. (2006) interviewed internet daters and found that the daters feel they are generally honest since they anticipate meeting the people who respond to their profiles in person. However, when they do lie it is usually about own weight or age, which they indicate misrepresenting only slightly, in order to fit into a different, but similar, category that they feel is more positive or will garner more dates. In response to this possibility, when using a dater’s own body type as a control variable, we classify broad, rather than specific, body type categories by grouping the 10 broad categories into four: Small, Average, Athletic and Large.

While the drawbacks to these data are important to note, the benefits of using these data far outweigh these costs. This is particularly true as internet dating becomes a more prevalent way to meet dates; over one-third of adults who use the internet and are seeking romantic partners have gone to dating websites (Madden and Lenhart 2006). Further, as marriage rates decline and non-marital partnerships increase (Bumpass et al. 1991; Schoen and Standish 2001), dating behavior may be a more accurate measure of partner preferences than marital outcomes.

Conclusion

Our findings show that race–ethnicity and gender influence body type preferences; men and whites are significantly more likely than women and non-whites to have such preferences. We believe white male daters’ body type preferences are influenced more by dominant media images of ideal body types than are those of non-white male daters because these images are predominately of whites; the images of the ideal female body are more constrained for women than for men; and other cultural resources compete to shape the preferences of non-whites. Although previous work employing silhouette figures, or select samples, provides inconsistent findings, our study of actual dating choices show that non-white men, particularly African-American and Latino men, are far less likely than are white men to prefer a date with the ideal thin body type. In general, men of color are far more open than white men to dating average women, and African Americans and Latinos, but not Asians, are significantly more likely to prefer a thick or heavy body type.

Our study has implications for a body of literature that shows African-American women and Latinas are less impacted by mainstream media than are white women with the former more satisfied with their bodies and less constrained to the thin ideal body type (e.g. Poran 2002). The fact that African-American men and Latinos are more open to dating a variety of body types, undoubtedly, places less pressure on these women to lose weight (Greenberg and LaPorte 1996; Hsu 1987; Levinson, Powell, and Steelman 1986). However, we suspect that Latinas and African-American women may be constrained in other ways, perhaps adhering to cultural standards of beauty outside of dominant cultural ideals or experiencing conflict between dominant ideal body types and those of their specific racial–ethnic group. Results of Poran’s (2006) focus group with African-American women college students find they are conflicted, feeling pressures to be thin by dominant cultural images while at the same time aware of the greater openness of African-American men to thicker body types. These conflicting expectations can create a no win situation in which African-American women receive approval for higher body weight by African-American men but disapproval from the dominant culture.

The fact that Asian men are less likely to prefer a thin date than are white men, but no more likely to prefer larger body types suggests that Asian men, like other non-white men likely subscribe to additional cultural prescriptions of body type desirability. While Asian men do not adhere as strongly as do white men to the beauty ideal, they are, nonetheless, no more accepting of heavy dates. This finding also has implications for a body of literature (Gluck and Geliebter 2002; Koff et al. 2001; Sanders and Heiss 1998) showing Asian women to have similar weight concerns, body dissatisfaction, and idealized underweight body images as white women. Since Asian men are more accepting of average bodies, but much less so of heavier bodies, it makes sense that Asian women would maintain concerns about weight. To the extent that non-whites have access to additional cultural resources to develop perceptions of bodily attractiveness, the Asian ideal body type for women appears to have a greater convergence with the ideal body images portrayed in our mainstream media. Recent studies of both Japanese (Mukai et al. 1998) and Chinese (Lee 1999) women find that both cultures adhere to a standard of beauty for women increasingly similar to that of the thin and fit idealized Western body. A comparison of Japanese college women with U.S. college women finds the former even more dissatisfied with their body than the latter (Mukai et al. 1998), and Lee’s (1999) qualitative study of Chinese women in Hong Kong shows they are increasingly concerned about their weight. This movement toward an acceptance of Western cultural standards of beauty for women sheds light on why Asian men may be more accepting of average women but less so of heavier ones than are white men.

Surprisingly, there are no significant racial–ethnic differences in women’s body type preferences as women are equally likely to desire the ideal, muscular man. Media studies suggest that in recent years through sports, non-white men, especially African-American men, have gained “symbolic access to masculinity as independent or respected” (Coltrane and Messineo 2000, p. 385). Athletic prowess provides a mechanism through which non-white men such as Kobe Bryant, Oscar de La Hoya, and Jet Li, gain media attention which may similarly influence white and non-white women’s desire for a date with a fit and muscular body type. However, we cannot account for why cultural expectations do not significantly intervene. In the absence of any empirical studies on non-white intracultural expectations of men’s body types, we surmise that there is a convergence between mainstream media images, and non-white resources in which male athleticism is highly valued. Overall though, we find women are less likely than men to have a body type preference, and we have argued that this may be explained by the media’s portrayal of a narrower range of “attractive” ideal body types for women as opposed to men.

In sum, our findings are compatible with previous research on body satisfaction and opposite-gender body type preferences, and support our argument that there is gender and racial–ethnic variation in daters’ body type preferences. It is clear that non-whites, especially African-Americans and Latinos, negotiate dominant white idealizations of thin female bodies with their own cultures’ greater acceptance of larger body types.