Abstract
The mere-exposure effect and exchange theory was tested in relation to openness to and approval of romantic interracial relationships. Levels of exposure were measured via self-reported proportions of acquaintances of particular races, television viewing, and family and friends having been in interracial relationships. Willingness to exchange attributes was measured by how strongly respondents agreed or disagreed with certain statements. Political orientation was also measured by asking how liberal or conservative respondents were. Results from tests revealed that the mere-exposure effect as it relates to openness to interracial relationships and influenced by the proportion of acquaintances of corresponding races was confirmed for Non-Asians (when broken down by race) and Latinos (when not broken down by race) but not for Whites. Furthermore, having family and friends in interracial relationships increased the incidence of respondents having been in interracial relationships but TV viewing had no significant effect nor did willingness to exchange attributes or political orientation.
Introduction
In light of new data indicating that interracial marriages are on the rise and that one-in-seven newly married couple is interracial (Passel, Wang and Taylor 2010), this study intends to explore the forces which would influence a population’s openness or lack of openness to engaging in romantic relationships with a partner of a different race or ethnicity. What are the factors that determine a population’s openness to intimate interracial relationships and another population’s lack of openness to intimate interracial relationships? Do the levels and types of exposure to certain races/ethnicities, preference for similarity in physical attractiveness per cost-benefit analyses, and political orientation associated with a population’s openness or lack of openness to engaging in romantic interracial relationships?
In testing these questions, the independent variables are the frequency of seeing certain races/ethnicities and interracial couples on television and daily life, the importance placed upon similarity in physical attractiveness and willingness to exchange desired attributes, and where participants place themselves on the political spectrum. The dependent variables are the personal openness to engaging in a romantic interracial relationship and approval of interracial relationships.
Literature Review
Zebrowitz, White and Weineke (2008) hypothesized that “either supraliminal or subliminal exposure to other-race faces would increase the likeability of a different set of other-race faces as compared with either exposure to own-race faces or no exposure to faces of either race (p. 262). The hypothesis was confirmed and is consistent with theories of the mere exposure effect: “Increasing White participants’ familiarity with an other-race facial prototype through exposure to Korean or Black faces increased the likeability of a different set of Korean or Black faces” (p. 271-272).
By extension, Bissell and Hays (2010) asked “how is television exposure and exposure to thin-ideal media related to appearance evaluations?” (p. 391). They explore cultivation theory which posits that “the exposure to and frequency of, messages disseminated through a medium influence an individual’s perceptions of the non-media world” (p. 390); sociocultural theory, which posits that “all activities of the mind are socially contextualized, and that learning is essentially a process of acquiring culture through social interaction” (p. 390); and social cognitive theory, which posits that “cognitive, affective, and biological events, behavioral patterns and environmental events all operate as interacting determinants that influence each other” (p. 391). In other words, increased exposure to television and the ideal of thinness in the media would correlate with an increased likeliness to evaluate others based on the same ideals (p. 397-398). Their hypothesis was confirmed: “…when the thin-ideal viewing index was considered with the appearance evaluations of celebrity models such as Ashley Tisdale, participants who spent the least amount of time viewing thin-ideal content had the least favorable appearance assessments. However, when the athletic models were considered, those viewing the most thin-ideal content had the least favorable assessments of the models” (p. 398).
Furthermore, Murstein, Merighi and Malloy (2001) hypothesized that in a hierarchical society blacks would have to be more attractive than whites in interracial relationships in order to offset the hierarchy: “…in interracial courting couples, the Black member would be more physically attractive than the White member” (p. 326). The hypothesis was partially confirmed in that third-party judges rated the black partners to be more attractive than their white counterparts and the “data partially support an exchange interpretation in that the objective disinterested measure of attractiveness was significant” (p. 330).
In addition, researchers Wilkins, Chan and Kaiser (2011) explored the effects of stereotypes on attractiveness, specifically the stereotype that Asians are feminine which is assumed to be desirable for Asian women but not Asian men. They hypothesized that “Whites will perceive women who look more Asian as being more desirable than those who have less prototypical appearances, and Whites will see prototypically Asian men as less attractive than their less racially prototypical counterparts” (p. 427). The hypothesis held true for Asian men: “…higher PP [Phenotypic Prototypicality] Asian men are perceived as possessing less masculine appearances than their lower PP counterparts. Furthermore, higher PP Asian men are perceived as being less attractive than lower PP Asian men” (p. 430).
Lastly, Yancy (2007), in Homogamy Over the Net: Using Internet Advertisements to Discover Who Interracially Dates, hypothesized that conservatives and the highly religious have a lower propensity to engage in interracial dating: ”To the degree there are important political and religious factors influencing individual racism, it can be expected that certain political and religious groups may have an overall lower propensity to interracially date” (p. 918). His hypothesis was confirmed: ”Conservative political ideology is inversely related to a willingness to date Blacks or to date regardless of race while religiosity is positively related to being only willing to date one’s own race” (p. 925).
In accordance with previous research, it is predicted that 1a) increased exposure to a certain race will increase liking of that race and willingness to engage in an intimate relationship with someone from that race, 1b) increased exposure to interracial relationships will have a positive association with approval of interracial relationships and increase the probability of engaging in interracial relationships, 2) increased exposure to media portrayals of racial beauty will be positively associated with increased influence by the portrayals with regards to openness to engaging in romantic relationships with certain races, 3) people seeking romantic relationships will tend to choose partners who at least match in physical attractiveness, 4) those who feel intelligence, personality, and/or success can compensate for it (or vice versa) will be more open to engaging in romantic interracial relationships, and 5) increased conservatism will negatively associate with openness to intimate interracial relationships. Although it may seem that there are five separate hypotheses, the first two are generally related to environmental influences, the remaining have more to do with individual preferences, and all tie into the theory that people make decisions based on cost-benefit analyses.
Methods
In testing the mere-exposure hypothesis and its relation to intimate interracial relationships, participants are to indicate what proportion of their acquaintances is of a particular race, what proportion of their friends has been in interracial relationships, and whether anyone in their families has ever been in an interracial relationship.
In order to measure the concept of media influence, and its effect on the judging of interracial relationships, participants are asked how many hours of television is watched, how often they see interracial couples on television, and whether the content is perceived to be favorably or unfavorable in regards to a certain race.
To test the exchange theory and matching hypothesis in regards to openness to intimate interracial relationships, respondents are asked how important they feel it is to find a good physical match and whether they would trade off looks for intelligence/personality/success.
For the concept of political orientation and extremism, respondents are to rate on a scale of one to seven how liberal or conservative they consider themselves to be with one being extremely liberal and seven being extremely conservative.
To test the dependent variables, i.e. the degree of openness to intimate interracial relationships, participants would answer whether they have ever been in an interracial relationship, and then to rate on a scale of one to seven how open they are to relationships with various races. They would also be asked how strongly they agree or disagree with the notions that 1) relationships between couples of the same race work better and 2) that couples of different races have the right to marry.
In order to account for variations among demographics, information regarding age, race/ethnicity, gender, educational attainment, and income will be asked.
The target population that this study intends to research is college students. The reason for this is that college students tend to be young and are at the stage where they begin to enter serious relationships or consider doing so. Thus, this population is ideal for this particular study which examines various preferences for intimate relationships. A group survey will be conducted along with individual surveys, where participants are to fill out the survey independently without having the questions read to them.
Surveys were administered to groups and individuals (Chambliss and Schutt 2010:174-176) and availability sampling was used (Chambliss and Schutt 2010:121-122). Permission to survey the class SOC 346: Race, Class, and Gender at California State University, Long Beach on Monday, April 2, 2012 was requested and granted by the professor on Wednesday, March 28, 2012. On April 2, at 3:20 PM in the LA5 building – Room 267, ten minutes before class would start, I introduced myself to the students that were present, informed them about my enrollment in SOC 355, debriefed about the general purpose of the survey, assured that anonymity would be provided, that they could skip any items, and requested their participation. As more students arrived, they were similarly debriefed individually. Students that arrived after 3:30 PM were not asked to participate since class had officially started. In total, 28 surveys were filled out and collected.
At around 1:00 PM on Wednesday, April 4th,2012, five students in the basement of the library at California State University, Long Beach were approached via convenience sampling and asked to participate in the survey research and debriefed in the same manner as mentioned above. Four agreed to participate and one declined. Four surveys were collected there.
On Thursday, April 5, 2012, permission to survey the class SOC 318: Case Studies in International Social Conflict was requested but denied.
On Monday, April 9, 2012, permission to survey the class SOC 464: Aging and Society on Wednesday, April 11, 2012 was requested and granted by the professor. On April 11, at 10:50 AM, ten minutes before class started and when at least half the students arrived, I introduced myself, debriefed those in attendance on the general purpose of the survey, assured them that confidentiality would be strictly respected, that they could skip any items, and requested their voluntary participation. As more students arrived, they were asked to participate and debriefed individually in the same manner as above. Students that came in after the professor began the class were not asked to participate. In total, 26 surveys were filled out and collected.
On the same day, at around 1:00 PM, thirteen students in the basement of the library at California State University, Long Beach were approached per convenience sampling and asked to participate in the survey research and debriefed accordingly. Ten agreed to participate and three declined. Nine surveys were collected there.
Overall, 68 surveys were filled out and collected. As mentioned, availability sampling was used, that is, classes and students are asked on a basis of convenience (Chambliss and Schutt 2010:122). The advantage of this sampling technique is its convenience but the disadvantage is its bias, that is, it may provide an over-representation or under-representation of certain population characteristics (Chambliss and Schutt 2010:114) and therefore limit generalizability. For example, the population characteristics of students in sociology classes may be unrepresentative of the population characteristics of students in engineering classes, and the population characteristics of students in the basement of the library may differ from that of students on the fourth floor.
Also, group sampling was the dominant method used for administering surveys with individual surveying complementing it. The advantage of the group method of surveying is the high response rate and cost-effectiveness. However, the disadvantages are that participants may feel coerced to respond to the survey – thus affecting honesty – and that comments by of the survey administrator may influence the responses (Chambliss and Schutt 2010:174-176). Considering these issues, efforts were made to indicate that participation is voluntary while maintaining neutrality and consistency.
Results
In the sample population, 74.2% were female (N=49), 25.8% were male (N=17); 41.2% were Latino/Hispanic (N=28), 25% were Asian/Pacific Islander (N=17), 16.2% were White/Caucasian (N=11), 7.4% were mixed (N=5), 5.9% were Black/African American (N=4), 2.9% were Other races (N=2), and 1.5% were Native American (N=1); 30.8% had household income levels at less than $20,000, 18.5% had household income levels at $20,000 to $40,000, 16.9% had household income levels at $40,000 to $60,000, 13.8% had household income levels at $80,000 to $100,000, 10.8% had household income levels at over $100,000, and 9.2% had household income levels at $60,000 to $80,000; 41.2% were seniors, 36.8% were juniors, 10.3% were sophomores, 7.4% were beyond senior level, and 4.4% were freshman; the mean age is 22.6119 years ranging from 18 to 53 with a standard deviation of 4.63503.
Univariate analyses revealed that, when respondents were asked what proportion of their acquaintances were White, 54.4% of respondents answered “Small proportion” (N=37), 20.6% answered “About half” (N=14), 13.2% answered “None” (N=9), and 11.8% answered “Most” (N=8); when respondents were asked what proportion of their acquaintances were Black, 67.6% answered “Small proportion” (N=46), 23.5% answered “None” (N=16), 5.9% answered “About half” (N=4), 1.5% answered “Most” (N=1), and 1.5% answered “All” (N=1); when asked what proportion of their acquaintances were Latino, 35.3% answered “Small proportion” (N=24), 22.1% answered “Most” (N=15), 20.6% answered “About half” (N=14), 13.2% answered “All” (N=9), and 8.8% answered “None” (N=6); when asked what proportion of their acquaintances were Asian, 51.5% answered “Small proportion” (N=35), 16.2% answered “About half” (N=11), 16.2% answered “Most” (N=11), 14.7% answered “None” (N=10), and 1.5% answered “All” (N=1); when asked what proportion of their acquaintances were Native American, 73.5% answered “None” (N=50) and 26.5% answered “Small proportion” (N=18); when respondents were asked what proportion of their friends have been in interracial relationships, 64.7% answered “Small proportion” (N=44), 16.2% answered “About half” (N=11), 16.2% answered “Most” (N=11), 1.5% answered “None” (N=1), and 1.5% answered “All” (N=1); and when asked if anyone in their families has ever been in an interracial relationship, 75% answered “Yes” (N=51), 22.1% answered “No” (N=15), and 2.9% answered “Don’t know” (N=2).
Furthermore, in regards to television viewing, 48.5 percent of respondents watch 1 to 3 hours of television daily (N=33); 38.2 percent watch less than an hour of television daily (N=26); 11.8% watch 4 to 6 hours daily (N=8), and 1.5% watch 7 or more hours daily (N=1.5%). Respondents were asked how often they see interracial couples on television on a scale of 1 to 7, with 1 meaning “Extremely seldom” and 7 meaning “Extremely frequently.” The mean is 2.9706 with a standard deviation of 1.51579. Respondents were then asked how they think interracial couples are portrayed on television. The mean is 3.9853 with a standard deviation of 1.45061. An index for the latter two variables was created which produced a mean of 3.4779 and a standard deviation of 1.31986.
To test the exchange theory, respondents were asked how strongly they agree or disagree with the following statements on a scale of 1 to 7, with 1 meaning “Strongly disagree” and 7 meaning “Strongly agree”: 1) “I prefer a romantic partner who is on my level in terms of physical attractiveness,” 2) “In choosing a romantic partner, I would overlook appearance if the person is intelligent,” 3) “In choosing a romantic partner, I would overlook appearance if the person is financially successful,” 4) “In choosing a romantic partner, I would overlook appearance if the person has an agreeable personality.” The means (and standard deviation) were 4.5294 (St. Dev. = 1.57850), 4.8382 (St. Dev. = 1.61717), 4.0588 (St. Dev. = 1.70935), and 5.7059 (St. Dev. = 1.49685), respectively. An index was created with the values for preference for same-level physical attractiveness reversed to reflect the direction of the other variables. The mean of the index was 4.5184 and the standard deviation was 1.12744.
For political views, respondents were asked how liberal or conservative they consider themselves to be on a scale of 1 to 7, with 1 meaning “Extremely liberal” and 7 meaning “Extremely conservative.” The mean was 3.1029 and the standard deviation was 1.33986.
In terms of the dependent variables, when respondents were asked if they have ever been in a romantic interracial relationship, 59.7% answered “Yes” (N=40) and 40.3% answered “No” (N=27).
Respondents were also asked, on a scale of 1 to 7, how open they were to dating or marrying someone who is White, Black, Latino, Asian, and Native American, respectively, with 1 meaning “Not open” and 7 meaning “Very open.” For whites, the mean was 6.2647 with a standard deviation of 1.39955. For Blacks, the mean was 4.9706 with a standard deviation of 1.97726. For Latinos, the mean was 5.6912 with a standard deviation of 1.83870. For Asians, the mean was 5.3284 with a standard deviation of 1.94148. For Native Americans, the mean was 5.0 with a standard deviation of 1.81742. An index was created to measure openness to interracial relationships in general on the same scale. The mean of the index is 6.2 and the standard deviation is .76942.
To measure the level of approval of interracial relationships, respondents were asked, on a scale of 1 to 7, how strongly they agreed or disagreed with the statements “Romantic relationships between couples of the same race tend to work better than romantic relationships between couples of different races” and “Two people that love each other should have the right to marry regardless of race,” with 1 meaning “Strongly disagree” and 7 meaning “Strongly agree.” The former statement was recoded to reflect the direction of the latter statement and the two were combined into an index. The mean for the former statement was 3.2273 (St. Dev. = 2.02139), the mean for the latter statement was 6.9118 (St. Dev. = .41377), and the mean for the index was 5.8409 (St. Dev = 1.07474).
When the 7-point scale for openness to a particular race was recoded with 1-2 meaning “Not open,” 3-5 meaning “Somewhat open,” and 6-7 meaning “Very open,” bivariate analyses reveal that for White males, 100% were “very open” to a relationship with someone who is White; for White females, 16.7% were “somewhat open” and 83.3% were “very open”. For Latino males, 100% were open to a relationship with someone who is White; for Latino females, 4.5% were “not open,” 18.2% were “somewhat open,” and 77.3% were “very open”. For Asian males, 25% were “not open” to a relationship with someone who is white, 25% were “somewhat open” and 50% were “very open.” The statistical significance is .645, i.e. not significant. Note that for this test and next tests, associations were run for White, Latino, and Asian since there is sufficient data for only these groups.
For Latino males, 100% were “very open” to a relationship with someone who is Latino; for females, 13.6% were “somewhat open” and 86.4% were “very open.” For White males, 40% were “somewhat open and 60% were “very open”; for White females, 33.3% were “not open” and 66.7% were “very open.” For Asian males, 25% were “not open” to a relationship with someone who is Latino and 75% were “very open”; for Asian females, 16.7% were “not open,” 66.7% were “somewhat open,” and 16.7% were “very open.” The statistical significance is .010 which is significant.
For Asian males, 25% were “somewhat open” to a relationship with someone who is Asian and 75% were “very open”; for Asian females, 33.3% were “somewhat open” to a relationship with someone who is Asian and 66.7% were “very open.” For White males, 20% were “not open” to a relationship with someone who is Asian, 20% were “somewhat open” and 60% were “very open”; for White females, 60% were “not open” to a relationship with someone who is Asian and 40% were “very open.” For Latino males, 40% were “not open” to a relationship with someone who is Asian, 20% were “somewhat open,” and 40% were “very open”; for Latino females 13.6% were “not open,” 31.8% were “somewhat open,” and 54.5% were “very open.” The statistical significance is .015, which is statistically significant.
To test for the relationship between openness to relationships with someone who is White and proportion of White acquaintances, crosstabs were run for Whites and Non-whites (see Table 7). The statistical significance for Whites is .065 and the significance for non-Whites is .828, both of which are not statistically significant.
In testing for an association between openness to relationships with someone who is Latino by proportion of Latino acquaintances, crosstabs were run for Latinos and Non-Latinos (see Table 8). The statistical significance for Latinos is .719 and the significance for non-Latinos is .346, both of which are not statistically significant.
Results from crosstabs also showed that the association between openness to relationships with someone who is Asian and the proportion of Asian acquaintances was not statistically significant for Asians (sig.=.725) but it is significant for non-Asians at .048 (see Table 9).
Additionally, ANOVAs were run to test the association between openness to relationships with a particular race and proportion of acquaintances of that particular race but without separating races (e.g. Latino and Non-Latino) and using the 7-point scale to measure openness. The significance of the association between openness to relationships with someone who is White and proportion of White acquaintances is .117, i.e. not statistically significant (see Table 10). The significance of the association between openness to relationships with someone who is Latino by the proportion of Latino acquaintances is .000, which is significant (see Table 11). And the significance of the association between openness to relationships with someone who is Asian and proportion of Asian acquaintances is .234, i.e. not significant (see Table 12).
Results from ANOVAs further showed that the association between approval of interracial relationships and family members having been in interracial relationships is significant (sig.=.017), that the relationship between respondents having been in interracial relationships and family members having been in interracial relationships is significant (sig.=.011), and the association between respondents having been in interracial relationships and the proportion of their friends having been in interracial relationships is statistically significant (sig.=.05). (see Tables 13-15)
However, crosstabs showed that the number of hours spent watching television did not significantly influence whether respondents have been in interracial relationships (sig.=.759). Furthermore, results from logistic regression indicated that there was no significant association between how often respondents saw interracial couples on television and how they were portrayed (as measured by the TV index) and the likelihood of having been in an interracial relationship. The regression coefficient was -.163 and the significance was .394. Even when controlling for race, sex, and household income, the coefficient produced was -.121 with a significance of .554 (see Table 16).
Linear regression was run for the exchange index and the index for openness to interracial relationships which revealed a regression coefficient of .031 and a significance of .954, i.e. not statistically significant. When controlling for race and sex, the coefficient was -.825 and the association was still not significant at .336.
Lastly, results from logistic regression showed that there was no significant association between how liberal or conservative respondents were and whether they have been in interracial relationships. The coefficient was -.160 without controls and -.028 with controls for race, sex, and household income. The significance without controls was .395 and .902 when controlling for race, sex, and household income, respectively.
Discussion
The mere-exposure effect as it relates to openness to a relationship with someone who is White and influenced by the proportion of acquaintances that are White – broken down by White and non-White – was not confirmed for Whites nor Non-Whites. The mere-exposure effect as it relates to openness to relationships with someone who is Latino as influenced by the proportion of Latino acquaintances – broken down by Latino and Non-Latino – was not confirmed for Latinos nor Non-Latinos. The mere-exposure effect as it relates to openness to relationships with someone who is Asian by the proportion of Asian acquaintances – broken down by Asian and Non-Asian – was not confirmed for Asians but it was confirmed for Non-Asians, i.e. as the proportion of Asian acquaintances of Non-Asians increased, openness to relationships with Asians increased.
When the association between openness to a relationship with a particular race and the proportion of acquaintances of that particular race was not broken down by race, i.e. exposure alone was tested, the hypothesis was confirmed in regards to openness to relationships with Latinos. In other words, as the proportion of Latino acquaintances of all races increased, openness to relationships with Latinos increased. However, such an association was not confirmed for Whites or Asians.
The mere-exposure effect as it relates to approval of interracial relationships and influenced by family members having been in interracial relationships was confirmed. Moreover, the mere-exposure effect as it relates to the likelihood of having been in an interracial relationship and influenced by family and friends having been in interracial relationships was also confirmed. That is, increased exposure to family and friends being in interracial relationships was positively associated with respondents’ approval of interracial relationships and themselves having been in interracial relationships.
Thus, even as the findings by Zebrowitz, White and Weineke (2008) show that the mere-exposure effect can be credited for reduced racial prejudice, here it is partially supported in terms of openness to interracial relationships as influenced by the proportion of acquaintances of particular races and fully supported in terms of having been in interracial relationships as influenced by family and friends in interracial relationships.
The study also found that the likelihood of having been in an interracial relationship was not influenced by the number of hours spent watching television, frequency of interracial couples on television as seen by respondents, nor how favorably or unfavorably respondents perceived the portrayal of interracial couples to be. Thus, in terms of the mere-exposure effect, the hypothesis is not confirmed. Ergo, although television viewing may influence how young girls perceive beauty as demonstrated by Bissell and Hays (2010), it appears to have had no significant influence on how respondents viewed interracial relationships or whether they have engaged in interracial relationships.
Exchange theory as it relates to openness to interracial relationships was not confirmed. That is, it was predicted that as respondents were more willing to trade off attributes, e.g. appearance for personality, they would be more open to interracial relationships. This refutes the research by Murstein, Merighi and Malloy (2001) which suggests that White partners exchange status for appearance while Black partners exchange appearance for status.
The present study also found that race and sex were strong predictors for openness to relationships with Latinos and Asians but not Whites. In other words, respondents tend to be open to relationships with White regardless of race and sex, e.g. with 100% of Latino males and 77.3% of Latino females being “very open” to relationships with Whites. On the other hand, for Latinos and Asians it was predicated by race and sex, e.g. with 60% of white females not being open to relationships with Asians whereas 60% of white males are “very open” to relationships with Asians. This reflects the observation by Wilkins, Chan and Kaiser (2011) that Asians are stereotyped as feminine which is desirable for females but not males.
Finally, it was expected that increased political liberalism and decreased political conservatism would be associated with higher likelihood of having been in an interracial relationship. In contrast to findings by Yancy (2007), this hypothesis was not confirmed.
In concluding, with 59.7% of respondents reporting that they have been in romantic interracial relationships, which complements the report by Pew Research (2010) that interracial marriages are on the rise, the influences found in the present study may contribute to understanding these trends. Given these results, it is also worth noting the limitations, that is, the sample was comparatively small and unrepresentative of the larger population, and there were a limited number of survey questions, specifically regarding media influence. Thus, more expansive research in the future would be required to more fully comprehend the dynamics of the relatively increasing openness to interracial relationships in the general populace.
References
Bissell, Kim, and Hal Hays. 2010. “Exploring the Influence of Mediated Beauty: Examining Individual and Social Factors in White and Black Adolescent Girls’ Appearance Evaluations.” Howard Journal Of Communications 21(4): 385-411.
Chambliss, Daniel F. and Russell K. Schutt. 2010. Making Sense of the Social World: Methods of Investigation. Thousand Oaks, CA: Pine Forge Press.
Murstein, Bernard I., Joseph R. Merighi, and Thomas E. Malloy. 1989. “Physical Attractiveness and Exchange Theory in Interracial Dating.” Journal Of Social Psychology 129(3): 325-334.
Passel, Jeffrey S., Wendy Wang, and Paul Taylor. 2010. “Marrying Out.” Pew Research Center (http://pewresearch.org/pubs/1616/american-marriage-interracial-interethnic).
Wilkins, Clara L., Joy F. Chan, and Cheryl R. Kaiser. 2011. “Racial Stereotypes and Interracial Attraction: Phenotypic Prototypicality and Perceived Attractiveness of Asians.” Cultural Diversity & Ethnic Minority Psychology 17(4): 427-431.
Yancey, George. 2007. “Homogamy over the net: Using Internet advertisements to discover who interracially dates.” Journal Of Social & Personal Relationships 24(6): 913-930.
Zebrowitz, Leslie A., Benjamin White, and Kristin Wieneke. 2008. “Mere
Exposure and Racial Prejudice: Exposure to Other-Race Faces Increases Liking for Strangers of That Race.” Social Cognition 26(3): 259-275.