Peer Influence and Attraction to Interracial Romantic Relationships

: The present research examined the effect of social influence on White, heterosexual individuals’ attraction to targets of varying races (White vs. Black) in two college student samples from the United States (one that leaned politically liberal and one that leaned politically conservative). Using a within-subjects experimental design, participants were given artificial peer evaluation data (positive, negative, or none) before providing ratings of attractiveness and dating interest for a series of targets. In both samples, positive information was associated with greater levels of attraction and dating interest than negative information, regardless of target race. Within the conservative sample, participants reported greater attraction toward and more dating interest in White targets relative to Black targets, while in the liberal sample, participants’ ratings of targets did not significantly differ from one another. These findings suggest that social influence can affect perceptions of attractiveness even in very different political climates.


Introduction
Interracial marriage has reached an all-time high in the United States, with couples of different racial backgrounds now making up 8.4% of all marriages [1].For comparison purposes, this is nearly triple the rate from 1980.Despite the fact that interracial relationships are increasingly common, a large segment of the population still opposes them.For instance, a recent, nationally representative survey found that approximately two out of every five Americans reported that they would not be comfortable with a family member marrying someone of another race [1].Consistent with this, members of interracial couples frequently report high levels of perceived disapproval from family, friends, and society at large as a result of their relationship [2] and, importantly, this disapproval predicts future breakup [3].Although many factors may contribute to disapproval of interracial dating, political variables appear to play a prominent role.As some support for this idea, not only is political conservatism associated with less interest in dating outside of one's race [4,5], but the lowest rates of interracial marriage tend to occur in the most politically "red" states [1].
To date, most of the research examining interracial relationships has focused on people's attitudes toward them and the outcomes associated with such involvements.As a result, we know surprisingly little about how these relationships actually develop or what attracts people to partners of different races in the first place (save for a few exceptions [4][5][6]).The goal of the present research was to study two variables that may affect romantic attraction in the context of interracial relationships: social influence and political climate.By exploring both the independent and interactive effects with these variables, we can potentially develop a better understanding of when and why attitudes toward this important social issue are likely to change.

Social Influence, Political Context, and Attraction Processes
Researchers have found social influence to be a reliable factor influencing perceptions of how people view potential romantic partners [7][8][9].For instance, one study found that both male and female participants' perceived desirability of a potential romantic partner was influenced by peer evaluations of several features of the target, including physical attractiveness, expressiveness, and earning potential [8].Specifically, participants reported greater attraction to romantic targets when their peers had evaluated those targets as being attractive, highly expressive, and having great earning potential; when peer evaluations of the target were low on these characteristics, participants found the target less attractive.In another set of studies, it was found that peer influence affected participants' judgments of a target's physical attractiveness and dating desirability [9].For instance, compared to a no information control condition, both male and female participants who were led to believe that their peers had evaluated a prospective romantic target positively expressed more interest in dating the target; when peers evaluated the target negatively, participants' dating interest decreased (Study 4).Taken together, the research in this area suggests that social influence processes play an important role in the development of initial attraction.
To date, the effect of peer influence has only been examined with regard to socially normative targets (i.e., targets that were of similar race and age to the participants).Consequently, it is unclear how far social influence goes in affecting attraction judgments.Specifically, it has yet to be established whether peer influence extends to perceptions of targets of a different race, or targets whose demographic characteristics vary in other ways.On the surface, it might seem intuitive that the effects of social influence would replicate in such cases.However, attraction processes may be fundamentally different when it comes to starting nontraditional relationships.For instance, in the case of interracial relationships, recent research has shown that political orientation is a strong predictor of desire to date someone from another race [5].Specifically, White conservatives' desire for Black romantic partners is significantly lower than that of White liberals.This may be because conservatism is a component of "symbolic racism", a belief system centered around the idea that discrimination is no longer an issue for Blacks, that Blacks are too demanding, and that they get more than they deserve [10].Of course, this is not meant to imply that all or even most conservatives are inherently racist.There may be other factors at play in explaining the link between conservatism and low interracial dating desire other than prejudice, such as reduced intergroup contact.Either way, it is clear that variables related to politics and prejudice should not be ignored in the context of interracial dating and we propose that such factors might moderate any potential effect of social influence on attraction to members of different races.

The Present Research
We conducted an experiment to examine the effect of social influence on attraction toward and interest in prospective interracial romantic partners, but also to explore whether these effects might be moderated by socio-political context and individual differences in prejudice.Using a within-subjects design, we manipulated social influence by presenting participants with either positive, negative, or no peer evaluation data before they provided ratings of attractiveness and willingness to date for several targets who varied in race (i.e., White vs. Black).To evaluate the potential moderating role of political climate, we collected data from two independent samples: students attending college in a politically conservative U.S. state, and students attending college in a politically liberal U.S. state.We then examined whether the social influence manipulation exerted different effects in each sample.We also assessed participants' own anti-Black attitudes for exploratory purposes in order to determine whether individual differences in racism would moderate the effects of social influence in a similar way.

Hypotheses
Consistent with research indicating that social influence affects attraction to socially normative targets [8,9], we hypothesized that social influence would also affect attraction toward and interest in dating different-race romantic targets.Specifically, we predicted that when participants were presented with positive peer evaluation data (i.e., when participants believed that their peers found a given target to be attractive), attraction and dating interest ratings would be higher than when negative information was given (i.e., when participants believed that their peers found a given target to be unattractive).When given no peer evaluation data, we tentatively expected attractiveness and dating interest ratings to fall in between the positive and negative conditions.We also predicted that this pattern of results would hold for both same and different race targets, (i.e., positive and negative social influence should affect Black and White targets in a similar way).
In light of previous research linking political orientation to interracial dating desire [4], we predicted a sample by target race interaction such that within the more conservative sample, a general prejudice effect should emerge in which participants would express more interest in dating same race targets than different race targets, irrespective of social influence condition.Within the more liberal sample, no such prejudice effect was expected, meaning that participants should rate same and different race targets as being of approximately equivalent attractiveness.
Furthermore, we hypothesized that a local social influence manipulation would not necessarily overcome broader social pressures against interracial dating or individual differences in racism.In other words, living within a socio-political context that has historically been opposed to interracial unions may make it more difficult to alter people's attitudes on this issue with a single experimental manipulation.Thus, we predicted a sample by social influence interaction, such that the social influence effects were expected to be more potent in the liberal sample than in the conservative sample.In other words, the positive and negative peer evaluation data should have more of an effect on attractiveness ratings in the liberal sample because participants from a conservative context may be more resistant to social influence attempts, particularly those that counter the prevailing social trends (in this case, positive peer evaluation data for racial minority targets).

Context of the Present Research
The current study focused specifically on White, heterosexual individuals' attraction toward and interest in dating targets of the other sex that were either White or Black.We focused on individuals with a heterosexual orientation here because our experience recruiting non-heterosexuals from college student subject pools typically yields insufficient numbers to examine them as an independent group, and this was indeed the case with both samples collected for the present research (only six non-heterosexuals were present in both samples combined).We opted to exclude these participants rather than collapse across sexual orientation because gay and lesbian couples appear to have different attitudes toward interracial dating.As some evidence of this, recent census data has found that gay couples are more likely to have an interracial composition than their heterosexual counterparts [11].
The racial context of the present research was restricted for similar reasons.For one thing, the student subject pools we had access to were disproportionately White, which meant that we were unable to recruit a sufficiently sized sample of any one racial minority group (for instance, only seven Black participants were present across both samples combined).Thus, we restricted our final sample to White participants because they were the only group for which we had enough statistical power to conduct analyses, not because they are the most interesting or even the most likely group to date interracially.We also opted to focus specifically on White-Black relationships because this pairing is the most relevant to existing research in this area [4] and also because it is often regarded as the ultimate taboo in interracial dating within the United States [12], the country in which this research was conducted.

Participants
College undergraduates from introductory psychology courses were recruited for a study of "social perceptions" at two large, public universities, each of which is the sole land grant university in its respective state (meaning that the vast majority of students are from that state).The first sample (N = 102) was collected in a state (Indiana) that is generally regarded as politically conservative.Most politicians elected in the state are Republican, including the governor and the majority of the state's Congressional delegation at the time of data collection.The second sample (N = 152) was collected in a state (Colorado) that tends to lean liberal and elects more Democratic candidates, including the governor and the majority of the Congressional delegation at the time of data collection.It is also worth noting that the percentage of new marriages that are interracial is almost twice as high (18.5%) in the more liberal state than in the more conservative state (10% [1]).Of course, this should not be taken to mean that all participants in the first sample (hereafter referred to as the "conservative sample") were necessarily conservative, just as not all participants in the second sample (hereafter referred to as the "liberal sample") were inherently liberal.In describing these samples as conservative and liberal we are simply seeking to label the broader political climate, not the political inclinations of the individual participants.
After excluding non-White and non-heterosexual participants, this final conservative sample consisted of 80 (31 men, 49 women) participants with a mean age was 18.94 (SD = 1.00).The final liberal sample consisted of 126 (39 men, 87 women) participants with a mean age was 19.23 (SD = 1.58).

Materials and Procedure
Participants in each of the two independent samples (i.e., conservative and liberal) rated a series of online photos featuring members of the other sex.The photos presented varied on two dimensions, containing the within subject manipulations of target race (i.e., White vs. Black) and the type of peer evaluation data provided (i.e., positive, negative, or none).Thus, the study was a 2 (target race: White vs. Black) X 3 (peer evaluation data: positive vs. negative vs. none) design.
All photos were obtained from an online database [13] and varied in terms of both target gender (i.e., male vs. female) and race (i.e., Black vs. White).A large pool of photos was initially taken from this database, but was subsequently whittled down to 24 through a pilot study.Based upon the results of our pretesting, we chose 12 male (6 White and 6 Black) and 12 female photos (6 White and 6 Black) that were reliably rated as being of approximately the same level of physical attractiveness (in the range of 3.5 to 4 on a scale of 1 to 10).All photos consisted of a college-age target, who was smiling, and looking directly at the camera.Only the target's head and shoulders were visible against a neutral background.The nature of these photos therefore minimizes any potential extraneous factors that could impact attractiveness ratings.In our pilot study, virtually all of the photos were rated as being slightly below average in terms of attractiveness.Given the difficulty of obtaining good quality photos of people of varying races that have such a high level of standardization, we utilized these photos in spite of the relatively low overall attractiveness levels.
Participants were informed that they would be evaluating the physical attractiveness of several persons presented in a series of online photos.They were also told that we had already obtained attractiveness ratings for some targets and that in cases where we had this information, they would be able to see it before making their own ratings.Our experimental manipulation thus consisted of the type of peer evaluation data (positive, negative, or none) that was made available to participants before they rated each photo.Positive evaluation data notified participants that other university students of the same gender and sexual orientation found the target to be relatively attractive (i.e., other students rated them between 5 and 6 on a 1-10 scale, with higher numbers indicating greater attractiveness).Negative evaluation data notified participants that other university students of the same gender and sexual orientation found the target to be relatively unattractive (i.e., other students rated them between 2 and 3 on a 1-10 scale, with lower numbers indicating less attractiveness).For photographs that received no peer evaluation data, participants were simply notified that no attractiveness information had been previously collected for the photo.In reality, all peer evaluation data was fabricated and was designed solely to manipulate social influence.
In total, each participant rated 12 photos of other sex targets over the course of the study, half of which were Black and half of which were White.Each form of peer evaluation data (positive, negative, none) was presented twice across the Black targets and twice across the White targets, meaning that all participants were exposed to all forms of social influence.Before rating the 12 critical photos, each participant was asked to evaluate two filler photos that contained no peer evaluation data in order to familiarize them with the task.
The presentation of the stimulus materials was loosely modeled after the popular Hot or Not website (www.hotornot.com).Participants first indicated which photoset (male or female) they would like to rate based upon their own pattern of sexual attraction.Each photo was then presented individually on a computer screen.For each photo, participants were asked to rate the attractiveness of the target and, separately, their interest in dating the target on a scale of 1 (not at all attractive/not at all interested) to 10 (extremely attractive/extremely interested).In cases where peer evaluation data was given, it appeared prominently above the photo in order to ensure that participants noticed it.All photos and peer evaluation data were presented in a random order.
After completing the photo ratings, participants filled out a series of questionnaires, including standard demographics and the 7-item Modern Racism Scale (MRS; alpha = .76 in each sample) [14].The MRS included items such as "discrimination against blacks is no longer a problem in the United States" and was scored such that higher numbers indicate higher levels of racism.A 5-item measure of political orientation was also included as a check to ensure that levels of conservatism did indeed differ across the samples (alpha = .92 in each sample) [15].The political orientation scale included items such as "the major national media are too liberal for my tastes" and was scored such that higher numbers indicate more conservatism.The MRS and political orientation items were rated on a scale ranging from 1 (strongly disagree) to 7 (strongly agree).Upon completion of these questionnaires, participants were fully debriefed.

Results
All analyses were conducted using PROC MIXED (mixed linear model) with SAS 9.3, allowing for the repeated measurements of attraction and dating interest for the targets.The degrees of freedom for the denominator were computed using the between and within method, and the covariance structure was specified to be unstructured.Predictors of the dependent variables included target race (Black vs. White), peer evaluation data (positive vs. negative vs. none), sample (politically liberal vs. conservative), and the participants' levels of modern racism.

Preliminary Analyses
Scores on the Modern Racism Scale and political orientation measure were compared for the two samples in order to confirm that the anticipated political characteristics of these samples did indeed differ.As expected, the mean racism score in the conservative sample (M = 2.93, SD = 1.07) was significantly higher than it was in the liberal sample (M = 2.34, SD = 0.93), t(204) = 4.18, p < .001.Likewise, levels of political conservatism were significantly higher in the conservative sample (M = 4.23, SD = 1.58) than they were in the liberal sample (M = 3.59, SD = 1.77), t(204) = 2.64, p < .01.

Social Influence Analyses
We began our primary analyses by combining the data from the two samples into a two-level stacked dataset, in which participants' ratings of attractiveness and dating interest for the twelve targets were nested within participant.See Table 1 for mean attractiveness ratings and Table 2 for mean dating interest ratings by sample, target race, and social influence condition.Using PROC MIXED in SAS 9.3, we first tested whether the peer evaluation data provided impacted participants' ratings of the dependent variables.Results indicated that peer evaluation data was significantly associated with ratings of attractiveness, F(2,506) = 83.84,p < .001,and dating interest, F(2,506) = 45.31,p < .001.Compared to participants who received no peer evaluation data, participants who received positive information rated the target as significantly more attractive (t = −11.87,Tukey-adjusted p < .001)and dateable (t = −8.39,Tukey-adjusted p < .001).Participants who received positive information also rated the target at significantly more attractive (t = −10.41,Tukey-adjusted p < .001)and dateable (t = −8.09,Tukey-adjusted p < .001)than did those who received negative information.However, participants who received no information provided statistically equivalent ratings for attractiveness (t = 1.46,Tukey-adjusted p = .31)and dating interest (t = 0.29, Tukey-adjusted p = .77)as those who received negative information.Next, we examined whether target race moderated the impact of peer evaluation data on the dependent variables.First, a model containing the main effects of peer evaluation data and target race revealed that target race was a significant predictor of attractiveness ratings, above and beyond the effect of peer evaluation data (F(1, 253) = 4.04, p = .05),such that White targets were viewed as more attractive than Black targets.The interaction between target race and social influence condition significantly predicted attractiveness ratings (F (2, 506) = 4.58, p = .01).Additionally, including this interaction term significantly improved model fit from the model with only the two main effects (-2 Log Likelihood = 11348.8vs. 11339.6, 2 = 9.2, p < .01).When given no information, Black targets were evaluated as less attractive than were White targets (t = 2.87, Tukey-adjusted p = .05).However, Black and White targets were evaluated equivalently when given negative information (t = −1.24,Tukey-adjusted p = .82)and when given positive information (t = 1.85,Tukey-adjusted p = .43).See Table 3 for the least square means from this analysis.These results indicate that in the two critical social influence conditions (i.e., positive vs. negative), target race did not moderate the effects of the peer evaluation data; target race only affected attractiveness ratings in the absence of peer evaluation data.Second, a model containing only the main effects of peer evaluation data and target race revealed that target race significantly predicted willingness to date the target, above and beyond the effect of peer evaluation data (F(1, 253) = 26.06,p < .001),such that participants reported greater willingness to date White targets than Black targets.The interaction between target race and social influence condition did not predict willingness to date the target (F(2, 506) = 1.86, p = .16);however, including this interaction term significantly improved model fit from the model with only a main effect, as was the case with attractiveness ratings (-2 Log Likelihood = 10692.9vs. 10696.6, 2 = 3.7, p < .05).Consequently, we examined the interactive effect.As with the attractiveness ratings, when given no information, participants were less interested in dating Black targets than White targets (t = 4.21, Tukey-adjusted p < .01),but there was no significant difference between the target races when participants were given negative information (t = 1.50, Tukey-adjusted p = .13).Unlike the attractiveness ratings, when given positive information, White targets were rated as significantly more dateable than were Black targets (t = 3.14, Tukey-adjusted p < .01).See Table 3 for the least square means from this analysis.These results indicate that target race only moderated the effects of peer evaluation data when that information was absent or when it was positive; when the information was negative, target race had no effect.These simple effects should be interpreted with caution, however, given the nonsignificant interaction term and the modest improvement in model fit afforded by the interaction.

Analyses with Conservatism and Modern Racism
We considered whether the nature of the sample further moderated the associations documented above.First, we examined the possible two-way interactions separately.The two-way interaction between sample (i.e., liberal vs. conservative) and social influence condition did not significantly predict ratings of attractiveness (F(2, 504) = 0.88, p = .41)or willingness to date the target (F(2, 504) = .76,p = .47),contrary to expectations.However, the two-way interaction between race of the target and the sample was significant for both ratings of attractiveness (F(1, 252) = 4.69, p = .03),and willingness to date the target (F(1, 252) = 7.42, p < .01.The nature of these interactions was such that Black and White targets did not significantly differ on attractiveness (respectively, M = 3.50, and M = 3.51; t(252) = .13,Tukey-adjusted p = .99)or dateablility (respectively, M = 2.44 and M = 2.57, t(252) = 2.17, Tukey-adjusted p = .13)in the liberal sample; in contrast, White targets were viewed as significantly more attractive than Black targets (respectively, M = 3.36 and M = 3.12; t(252) = 2.91, Tukey-adjusted p = .02)and more dateable than Black targets (respectively, M = 2.61 and M = 2.22, t(252) = 5.30, Tukey-adjusted p < .001) in the conservative sample.The three-way interaction between social influence condition, target race, and sample was not significant for attractiveness (F(2, 504) = 0.74, p = .48)or for willingness to date (F(2, 504) = 1.39, p = .25).Together, these results suggest that whereas the conservative sample indeed viewed Black targets as less attractive than White targets overall, they were influenced by the peer evaluation data in the same way as were those in the liberal sample, whose ratings of Black and White targets did not significantly differ.
For exploratory purposes, we also considered whether racism scores were associated with interest in interracial dating and whether they moderated the effects of social influence.Although they are distinct constructs, conservatism and racism share common variance [10] and might be expected to yield similar results in this context.If so, it would suggest that part of the reason political climate is linked to interracial dating interest may be attributable to regional differences in racism.Indeed, the pattern of results was largely similar when using the individual difference measure of modern racism as a predictor instead of sample.In this case, there was no significant interaction between modern racism scores and social influence condition both in predicting attractiveness (F(2, 2788) = 0.07, p = .93)and willingness to date the target (F(2, 2786) = 0.05, p = .95);however, the interaction between modern racism scores and target race was significant for both attractiveness (F(1, 2790) = 32.95,p < .001)and willingness to date (F(1, 2788) = 20.73,p < .001).The nature of these interactions were such that modern racism was unrelated to ratings of attractiveness (b = −0.06,t(252) = −0.74,p = .046)and dating interest (b = −0.01,t(252) = −0.20,p = .85)for White targets, but significantly and negatively associated with ratings of attractiveness (b = −0.35,t(252) = −4.26,p < .001)and dating interest (b = −0.22,t(252) = −3.07,p < .01)for Black targets.The three-way interaction between modern racism, social influence condition, and target race was not significant for attractiveness (F(2, 2782) = 1.95, p = .14)or for dating interest (F(2, 2780) = 1.51, p = .22).

Discussion
Across two very different college student samples, we found reliable effects of social influence on participants' attraction to prospective romantic targets.Specifically, participants found targets of both their own and another race to be more attractive and dateable when presented with positive rather than negative peer evaluation data, consistent with hypotheses.These findings are in line with past research on social influence and attraction [8,9] and suggest that, in developing our own standards of attractiveness, we appear to take into account the standards of others.However, the current work is a significant extension of past studies in this area in that it demonstrates that social influence effects extend to non-traditional romantic targets.When provided with no information, participants found White targets to be significantly more attractive than Black targets.However, social influence information largely led this difference to disappear; for the most part, White and Black targets were rated similarly regardless of whether positive or negative peer evaluation data was provided, which suggests that the standards of others are highly influential in determining individuals' own standards.
It could be argued, based upon previous research linking political orientation to interracial dating desire [4], that the effect of social influence would be moderated by the nature of the sample.Perhaps White individuals in politically conservative contexts, or individuals high in modern racism, would be less influenced by the standards of others when evaluating the attractiveness and dateability of a Black target.However, the hypothesized sample by social influence interactions did not reach significance, indicating that no group of participants was immune to the effects of social influence.Analyses with the modern racism variable revealed the same thing (i.e., regardless of level of racism, participants were affected by social influence in the same way).This speaks to how powerful social influence can be; even for those people who may be predisposed to evaluate members of certain races negatively, social influence information can still impact their evaluations.
That said, we found that in our more conservative sample, participants tended to rate same race targets as more attractive and dateable than different race targets, while in the liberal sample, there were no differences in these ratings based upon target race.Had the social influence manipulation been equally potent across samples, this effect might have disappeared.This finding thus suggests that the social influence manipulation did not completely override the effects of preexisting differences in racial attitudes and political climate.
Our predictions were not entirely supported with respect to where the no information control condition fell relative to the positive and negative conditions.We consistently found that attraction ratings were lower in the no information than in the positive condition, which was to be expected.In contrast, the no information condition did not differ reliably from the negative condition.There are several possible explanations for this.One is that the overall attractiveness of the targets was already below average to begin with, so there was more room for ratings to move up than to move down.Thus, if the baseline level of attractiveness had been higher to start, the negative information may have been more effective at pulling ratings down.Another possible explanation is that a lack of confidence emerged in the no information conditions.Specifically, when presented with no peer evaluation data, participants may have erred on the side of giving ratings that were too low rather than too high, perhaps because people see more social risk associated with providing a rating that is too high rather than too low.
The present research provides us with important insight into factors that may predispose individuals to pursue an interracial romance.Although we know that social network members' opinions and beliefs may exert an important influence on the quality of ongoing relationships [16,17], our findings suggest that social networks may play an equally important role in affecting relationship initiation, specifically when it comes to starting a relationship with someone who comes from a different demographic background.In addition, this work provides an important connection between the prejudice and relationships literatures in social psychology.The bridge between these areas has grown in recent years [2][3][4], but much remains to be learned as we continue to fuse research and theory on these topics.

Strengths and Limitations
The primary strengths of the present research reside in the experimental, within-subjects research design and the fact that the basic social influence findings were replicated across two dependent variables in two quite different student samples.We also tested two very different moderator variables (i.e., the socio-political context of the sample and individual differences in racism) and found converging effects.As a result, we can have a certain degree of confidence in the results and conclusions.
That said, there are some important limitations that should be noted.First, this study only examined White, heterosexual participants' perceptions of Black targets of the other sex, which means our ability to make generalizations is limited.For one thing, it would have been interesting to consider whether social influence and socio-political contexts have similar effects on Blacks' perceptions of White targets (i.e., do the effects extend equally to members of dominant and non-dominant groups?);however, we lacked statistical power to make this comparison.In addition, it would have been interesting to examine different types of interracial relationships because White-Black is just one of many possible racial combinations.Thus, whether these same results would hold in a larger and more diverse sample of participants, as well as for a wider range of targets is a worthwhile area of inquiry for future research.Although we would anticipate that the basic social influence effect would consistently emerge, the strength of such effects would likely vary when examining different racial combinations, given that social distance varies across racial groups and that certain interracial combinations appear to be more socially accepted [18], which means that some may be more difficult to manipulate than others.Likewise, the specific racial combinations that are considered more or less taboo varies cross-culturally, which suggests that culture may be another important moderator variable to consider.Related to this, participant gender is another moderator that should be addressed in future studies, given that some research has found that White women have more positive attitudes toward dating someone outside of their race than White men [19].Unfortunately, given that our sample was disproportionately female, we lacked statistical power to give adequate consideration to potential gender effects; however, it would be worth exploring whether men and women are equally susceptible to social influence effects of this nature in future research.
Second, as previously mentioned, the overall attractiveness of the targets used in this study was below average, which may have made it easier for positive rather than negative social influence effects to emerge.It would be useful for future studies to use targets with a higher overall level of attractiveness to further explore the potential effects of negative information.Third, the dependent measures in this study (i.e., attractiveness and dating interest ratings based only upon exposure to online photos) may not necessarily be a reliable indicator as to whether someone would want to start a relationship with a given target.Future research would therefore benefit from studying social influence effects in the context of a live interaction (e.g., through use of a speed-dating paradigm [4]), which would help to establish whether the effects observed in this study extend to what people actually do in a real life romantic context.Fourth, we focused our analyses on the broad state level measure of political conservatism as opposed to an individual difference measure.However, future researchers might wish to consider whether there are unique or potentially interactive effects of different conservatism measures.
Finally, this study only considered the roles of peer influence and socio-political context in attitudes toward interracial relationships.There are likely to be numerous other factors that affect interest in dating outside of one's race, such as intergroup contact.For example, research has found that dating members of other racial groups reduces intergroup anxiety [20], which would presumably increase the likelihood of further interracial dating.Likewise, positive media portrayals of interracial couples generate positive attitudes toward such relationships [21].In light of such findings, future research on peer influence and interracial relationships might give consideration to the potential moderating role that contact and such other factors might play.

Conclusions
The prevalence of interracial relationships has grown considerably over the past few decades [1].Despite this growth, relatively little research has addressed the question of why people are increasingly open to pursuing partners outside of their own race.The present research contributes to our understanding in this area by highlighting the important role that social influence plays in shaping our perceptions of attractiveness.Specifically, others' opinions and beliefs affect how we evaluate both traditional and non-traditional romantic targets.

Table 1 .
Mean attractiveness ratings by type of peer evaluation data, target race, and sample.

Table 2 .
Mean dating interest ratings by type of peer evaluation data, target race, and sample.
Note: Numbers in parentheses represent standard deviations.

Table 3 .
Least square means for ratings of attractiveness and willingness to date by type of peer evaluation data and target race.