Introduction

There is a relatively small, although rapidly expanding body of research on people’s beliefs about the concept of race. The majority of this work has examined how people’s beliefs about the concept of race affect their intergroup behavior. For example, past research has found that belief that race is a biological construct is related to racial stereotype endorsement (Wong et al. 2013) and lack of willingness to initiate friendships with out-group members (Williams and Eberhardt 2008). In this paper, the researcher further contributes to understanding how the concept of race affects intergroup behavior, specifically examining the relationship of belief in race as biological on people’s out-group social distance. In addition to the continued focus on how beliefs about race affect intergroup behavior, the researcher also expands this body of research in two relatively less-known directions: (1) by examining how early life experiences, specifically childhood exposure to diversity, influences the formation of beliefs about race as biological and (2) by examining how beliefs about race can be decreased through educational interventions.

One of the promising aspects of the study of people’s beliefs about race is that people’s cognitions are relatively susceptible to change, for example, through education. Thus, greater understanding of people’s cognitions about race including how they form, how they affect their group relations, and how they can be changed has great implications for diversity-based education and ultimately for promoting positive intergroup relations. These three areas of inquiry are subsequently addressed in the three studies reported in this paper. Before describing each study in detail, some brief background is provided on the idea of race as biological and the relatively recent refutation of this myth.

Belief in Race as Biological

Both biological and social scientists have reached consensus that race is not a biologically meaningful construct, but rather that it is a sociopolitical idea, historically developed to rationalize atrocities against groups of people (e.g., enslavement of Africans, genocide of Native Americans; Marks 1996; Omi and Winant 1994; Smedley and Smedley 2005). Biological references to race in the USA appeared as early as the late seventeenth century, and these views increasingly postulated that racial groups possessed distinct, immutable trait characteristics; for example, blacks were naturally unintelligent but physically endowed. Whites used these myths to rationalize the enslavement of blacks; indeed, if they possessed inferior mental capacity but superior physical capacities, they were more suited to work on plantations rather than to manage them (Omi and Winant 1994; Smedley and Smedley 2005).

Today, there is overwhelming evidence from both social and biological sciences to suggest that racial groups are not biologically meaningful concepts; for example, at least 99.9 % of DNA is shared between any two randomly selected people, either within or across “racial” groups (Marks 1996; Omi and Winant 1994; Smedley and Smedley 2005). However, the refutation of the biological validity of race and the articulation of race as a sociopolitical idea are relatively recent in the literature, perhaps only becoming generally accepted within the past few decades. Furthermore, it is unclear to what extent the articulation of race as a sociopolitical construct within the sciences has become common knowledge to laypersons outside of the sciences. Research does suggest that people do continue to endorse beliefs in race as biological and that such views lead to negative intergroup outcomes such as greater prejudice toward out-groups (Jayaratne 2006; Williams and Eberhardt 2008).

Given that race is an idea and not a natural order, the idea that racial groups are biological must be learned. Psychologists use the term “essentialism” to refer to the tendency to think of social groups as being biological or naturally occurring entities, each with their own inherited and fixed traits (e.g., intelligence, laziness), and supports the idea that essentialism is a learned tendency (Medin 1989). Yet, the researcher is not aware of previous empirical studies within psychology about what early life factors may influence learned belief in race as biological. In Study 1, the researcher examined the possible influence of early life exposure to diversity on the belief in race as biological, before examining the intergroup affects of belief in race as biological (Study 2) and the extent to which belief in race as biological can be unlearned (Study 3).

Study 1

Study 1 examined how early life experience, and specifically the effects of exposure to diversity as a child, affects people’s belief in race as biological. Classic intergroup contact theory predicts that exposure to diversity (i.e., contact) with members of other racial groups leads to reductions in prejudice (Allport 1954). According to Allport, prejudice occurs initially as a result of oversimplifications and generalizations about others based on broad categorical assignments such as race. Prejudice decreases when race-based assumptions about others break down over time as people are exposed to out-group members and come to know them individually. Indeed, research has supported that exposure to diversity is related to less prejudice and stereotype endorsement (Bowman and Denson 2012; Tadmor et al. 2012).

Allport’s contact theory is extended here by proposing that exposure to diversity—particularly early on in life—will interfere with people’s ability to accept essentialized beliefs about race. People’s early life exposure to nuances and within-group differences (e.g., variations in skin color and personalities within presumed racial groups) will contradict learning or messages that racial groups are biologically distinct. The researcher is not aware of research that has directly tested the relationship between exposure to diversity and less essentialized beliefs about race. However, there is a small body of research that has found that exposure to diversity is related to less essentialism and more complex cognitive processes more generally (Pascarella et al. 2014; Schmid et al. 2012), supporting this hypothesized relationship. For example, in a study of 949 fourth-year college students across 17 schools, exposure to diversity in school was related to greater development of critical thinking skills and greater complexity in thinking and problem solving (Pascarella et al. 2014). These findings buttress the hypothesis in Study 1, given that understanding that race is not biological constitutes a relatively complex level of thought. For instance, understanding that race is not biological requires people to understand the historical development and rationale for the biological myth of race, arguments refuting the validity of biological views of race, and to simultaneously understand the continued social significance of race.

However, exposure to diversity may not always lead to positive intergroup outcomes. Indeed, Allport (1954) himself has suggested that contact theory is only applicable under conditions where there is a balance of social structural power. Group threat theory suggests that in conditions in which resources are scarce, contact may actually lead to greater prejudice because the presence of out-groups may be perceived as a threat to consume limited resources (Blumer 1958). Similarly, group threat theory is extended here to suggest that for people in lower socioeconomic statuses, exposure to diversity may actually be related to greater belief in race as biological. When out-groups are perceived as posing a threat to consume resources, people may rationalize feelings of threat from other groups by relying on essentialized depictions of racial group members (e.g., “my feeling of threat to lose resource is reasonable because Asians are inherently greedy and blacks are inherently violent”). Thus, the research examined the potential moderating role of socioeconomic status on the relationship between exposure to diversity as a child and beliefs about race as biological.

Method

Participants

Participants in Study 1 included 155 people (87 females; 68 males) recruited from the researcher’s University and from a community sample of participants recruited via flyers posted throughout the city of Boston, specifically targeting lower socioeconomic neighborhoods and thus residents who may be underrepresented in a sample that relied only on college students. The goal was not necessarily to arrive at a US representative sample, but rather to have enough socioeconomic variability within the sample to enable statistical comparisons. All participation was voluntary, and participants signed electronic consent forms prior to their participation. Data collection was supported by a small internal grant from the author’s previous University and thus limited to approximately 150 participants. Participants in this sample self-identified racially as: Asian (10.3 %), black (7.1 %), Latino (6.5 %), multiracial (8.4 %), white (58.1 %), and Other (9.7 %). People of color, including anyone indicating affiliation with at least one non-white race, collectively comprised 32.3 % of the sample. Participants born outside of the USA comprised 26.4 % of the sample, and participants born in the USA comprised 73.6 % of the sample. Participants’ ages ranged from 18 to 64 (M = 28.81; SD = 9.90). Parental/caretaker’s educational status was used as a proxy for socioeconomic status (SES); participants whose parents/caretakers were college graduates (or higher) comprised 56.8 % of the sample, participants with parents/caretakers with some college or less comprised 32.3 % of the sample, and 10.9 % of the sample did not report their parents/caretakers' educational status.

Measures

Belief in race as a biological construct was measured by four items inquiring about the biological nature of race (Tawa and Kim 2011). Items included: (1) Different races are biologically/genetically different from one another; (2) black and white people basically have the same genetic makeup (reverse coded); (3) there are no biological differences in intelligence among the races (reverse coded); (4) there are genetic differences between the races, resulting in some races being physically stronger than others. With the current sample, the internal reliability for this scale was adequate (a = .64).

Childhood neighborhood diversity was constructed from items in which participants were asked to report cumulative percentages of the racial groups in the neighborhood in which they grew up (e.g., 10 % Asian, 10 % black, 10 % Latino, 65 % white, 5 % Native American). Childhood neighborhood composition was computed by using an entropy index (Quillian and Redd 2009), which provides a single value to reflect the imbalance or balance of racial group representation in one’s neighborhood; a score of 0 indicates no diversity at all (e.g., someone who grew up in a neighborhood that was 100 % white) and a maximal value is reached when the racial groups represented in one’s peer group are as equally balanced as possible. In the current study, only Asian, black, Latino, white, and Native American racial groups were available to participants for selection on the survey. Thus, a maximal entropy score of .70 would be achieved if one reported 20 % Asian, 20 % black, 20 % Latino, 20 % white, and 20 % Native American. As a case in comparison, a neighborhood composition of 50 % black and 50 % white would yield an entropy score of .30. Computationally, an entropy score is calculated by taking the negative of the log of the proportion of each group, multiplied by the proportion itself, and summing the products. An entropy index is calculated for each participant (i) and is represented by the following equation, where K is the number of racial groups available (in the case of this study, always 5), and p ki is the percentage of racial group representation described for each group:

$${\text{Entropy}}_{i} = - \sum\limits_{k = 1}^{K} {p_{ki} \,\log \,p_{ki} }$$

Socioeconomic status was measured using parental educational attainment as a proxy. This decision was based on previous reviews of social science literature, which have found that parental educational attainment is the most commonly used measure of SES within the social sciences (Sirin 2005; White 1982). For example, in a meta-review of studies in the social sciences from 1990 to 2000, Sirin (2005) found that parental educational attainment was used significantly more often than parental occupational status, family income, access to free lunch programs, and access to neighborhood- and home-based resources. For this study, parental educational attainment was used to create a dichotomous variable for “SES,” in which SES comprised people with parents/caretakers without college degrees (low SES = 0) and high SES comprised people with parents/caretakers with college degrees or more (high SES = 1).

Preliminary Analyses

Prior to examining our primary hypothesis, we examined the effects of participant demographic variables on belief in race as biological (RB) and childhood neighborhood diversity (referred to henceforth as diversity). Gender did not significantly affect RB (F = .63; p = .88; d = .13) or diversity (F = .18; p = .67; d = .07). Racial minority status was significantly related to RB (F = 2.05; p = .01; d = .25); people of color had a mean RB of 12.00 (SD = 4.84) and whites had a mean RB of 10.27 (SD = 5.10). Racial minority status was not related to diversity (F = .38; p = .54; d = .11). Immigration status was significantly related to RB (F = 1.92; p = .02; d = .25); immigrants had mean RB of 13.25 (SD = 4.93) and US born participants had a mean RB of 9.85 (SD = 4.78). Immigration status was not significantly related to diversity (F = .95; p = .58; d = .18). Age was not significantly related to RB (r = −.12; p = .18; d = −.24) but was significantly related to diversity (r = -.18; p = .04; d = -.37). SES was not significantly related to RB (F = .50; p = .48; d = .12) or diversity (F = .26; p = .61; d = .08). Since none of the participant variables were significantly related to both RB and diversity, participant variables were not controlled for in further analyses.

Results

Primary Analyses

Childhood neighborhood diversity was not correlated with RB with the overall sample (r = .01; p = .95; d = .02). A multiple regression analysis was conducted to test the potential moderating effect of SES on the primary relation between neighborhood diversity and RB. To examine this potential moderator, an interaction term was created by multiplying the child neighborhood diversity variable with the dichotomized SES variable. Using a multiple regression analysis, neighborhood diversity (the independent variable) and SES (the moderator variable) were entered in the first step, and the interaction term was entered as the second step. The interaction term was significantly related to RB (t = −2.51; p = .01; see Table 1). We used regression coefficients to create “prototypical plots” for interpretation of the interaction (Singer and Willet 2003; see Fig. 1). As predicted, among high SES participants, greater levels of diversity were related to lower levels of belief in race as biological. Among low SES participants, higher levels of diversity were related to higher belief in race as biological.

Table 1 Moderation of SES on the relationship between early exposure to diversity and belief in race as biological
Full size table
Fig. 1
figure 1

SES moderation of the relationship between early exposure to diversity and belief in race as biological

Full size image

Discussion

Although childhood neighborhood diversity was not related to belief in race as biological for the whole sample, relationships emerged when moderated by SES. Within lower SES neighborhoods, presence of out-group members—whom may be perceived as competitors for limited resources—may evoke essentialist ideas about race which are used to rationalize feelings of threat. Comparatively, in higher SES neighborhoods, in which resources are more abundant, diversity may work to break down or challenge essential ideas about race. In turn, beliefs about race as biological may further perpetuate distances between racial groups, which is the focus of Study 2.

Study 2

In Study 2, the researcher examined the effects of belief in race as biological on social distances toward racial out-group members. Social distance reflects the extent to which people are likely to date, befriend, and/or share communities with other group members (Bogardus 1925; Rollock and Vrana 2005). It is hypothesized that people who believe race is biological will keep greater social distance from out-group members. In addition, the possible mediating roles of out-group discomfort and intergroup anxiety are considered as affective mechanisms through which belief in race as biological affects social distances.

This study contributes to an emerging body of research demonstrating the negative consequences of belief in race as biological on intergroup behavior. More specifically, belief in race as biological has been found to be related to lack of willingness to initiate friendships with out-group members (Williams and Eberhardt 2008), less actual diversity among friend groups (Williams and Eberhardt 2008), greater discomfort interacting with out-group members (Tawa and Kim 2011), heightened physiological arousal while discussing cultural experience (Chao et al. 2007), closed-mindedness (Tadmor et al. 2013), racial stereotype endorsement (Wong et al. 2013), greater tendency to engage in racial categorizing of others (Chao et al. 2013), worse memory for out-group members’ compared to in-group members’ faces (Gaither et al. 2013), and greater perceived differences between in-group and out-group members (No et al. 2008).Footnote 1

Consistent with previous findings of negative consequences of belief in race as biological on intergroup behavior, it is hypothesized that belief in race as biological will be related to greater social distance toward out-group members. The added contribution of this study to this literature is an effort to understand more precisely how a cognitive process such as belief in race as biological affects intergroup behavior (social distance). This study considers the mediating role of two affective mechanisms through which belief in race as biological may affect social distances, out-group discomfort, and intergroup anxiety. Intergroup anxiety is affective arousal related to anticipating intergroup interactions and is conceptualized as a stable, internal trait (i.e., xenophobia; Stephan and Stephan 1985). Out-group discomfort is also affective arousal experienced while anticipating intergroup interactions (Cole and Yip 2008); however, unlike trait-based intergroup anxiety, the authors conceptualize out-group discomfort as a more temporary state of arousal that is specifically related to situations in which one is a minority. In both cases, the affective arousal experienced can lead to distortions in information processing (e.g., overgeneralization) and inhibit behavior (e.g., minimizing interactions).

It is predicted that when people who endorse biological beliefs about race anticipate interactions with out-group members, affective arousal will activate racialized overgeneralizations about the group members with whom they are anticipating interaction (e.g., “I am anxious because there will be many black people at the party and they are aggressive”). In turn, it is expected that these overgeneralizations will lead to shortened or avoided interactions (i.e., greater social distance) toward out-group members. In these analyses, we examined both intergroup anxiety and out-group discomfort as mediators on the primary relationship between belief in race as biological and out-group social distance, in an attempt to tease apart trait- versus state-based arousal as the operating mechanism.

Method

Participants

Participants (N = 210) included 61 males, 148 females (1 unreported). Data collection for this study was also supported by a small internal grant from the author’s previous University and thus limited to approximately 200 participants. Participants were all monoracial and included: Asian (n = 53), black (n = 39), and white (n = 118). Immigrants comprised 16.1 % of the sample. Ages ranged from 18 to 64 (M = 24.48; SD = 8.28).

Measures

Belief in race as a biological construct was measured by the same scale used in Study 1 (Tawa and Kim 2011). With the current sample, the internal reliability for this scale was adequate (a = .58).

Out-group social distance was measured using an adaption of the Bogardus Social Distance scale (Bogardus 1925) developed by Rollock and Vrana (2005). In Rollock and Vrana’s (2005) operationalization, the researchers used Likert scales with which participants indicated the degree of intimacy felt toward specific racial groups (e.g., black, white). In the current study, social distance was assessed in relation only to blacks, Asians, and whites. Relationship intimacy ranged on a continuum from “would marry into” (1) to “would debar from my country” (7). Scores were totaled across levels of intimacy and then again totaled across out-group reference groups (e.g., for a white participant, out-group social distance was the total of social distance toward Asians and blacks).

Out-group discomfort was measured by a revised measure of Cole and Yip’s (2008) Out-group Comfort Scale (OCS) originally used to assess black students’ state-based comfort, specifically in interaction with whites and primarily when in white social contexts (e.g., primarily white college settings). A revised version (Cole, personal communication) uses race-neutral language and can be used to assess the out-group comfort levels of multiple racial groups. Sample items from the OCS include: “Socializing with people of other races is especially difficult for me,” and “In general, being around people of other races makes me nervous.” The OCS has been shown to have good predictive validity, for example, among a sample of 98 black freshmen, higher scores on the OCS taken in the fall semester predicted greater contact with peers across racial groups during the academic year. All items were reversed scored in order to measure out-group discomfort. With the current sample, the internal reliability for this scale was strong (a = .92).

Intergroup anxiety was measured by the Intergroup Anxiety scale (IGA; Stephan and Stephan 1985), which is widely used to assess trait-based affective arousal that a person tends to feel in cross-racial interactions. On the IGA, participants are asked to rate on a scale of 1–10 whether they would feel more or less Comfortable, Awkward, Self-conscious, Happy, Accepted, Confident, Irritated, Impatient, Defensive, Suspicious, and Careful when interacting with members of another racial group compared to their own. In the initial scale development study, the authors supported strong construct validity, finding positive associations between the IGA and less contact, greater stereotyping, and greater belief in dissimilarity with out-group members (Stephan and Stephan 1985). With the current sample, the internal reliability for this scale was strong (a = .88).

Results

Preliminary Analyses

Analyses of Variance were run to determine any potentially confounding effects of participants’ demographic statuses on the main analysis in this study and to aid interpretation of discussion. Participants’ race was not significantly related to belief in race as biological (RB; F = 1.09; p = .34; d = .05), but was significantly related to out-group social distance (F = 12.28; p < .01; d = .79). Asian participants had significantly higher out-group social distance (M = 28.60; SD = 12.42) than both black (M = 21.34; SD = 11.21) and white participants (M = 18.94; SD = 11.25). Participants’ race was also related to out-group discomfort (F = 3.95; p = .02; d = .34); Asian participants’ out-group discomfort (M = 30.08; SD = 11.04) was significantly higher than black participants’ out-group discomfort (M = 23.24; SD = 8.46). Participants’ race was not significantly related to intergroup anxiety (F = 2.11; p = .12; d = .21). Immigration status was not significantly related to RB (F = .24; p = .63; d = .09), out-group social distance (F = 2.49; p = .12; d = .30), out-group discomfort (F = .029; p = .866; d = .03), or intergroup anxiety (F = .461; p = .498; d = .13). Gender was not significantly related to RB (F = 1.32; p = .25; d = .18), but was marginally related to out-group social distance (F = 3.51; p = .06; d = .29). Males had an out-group social distance of 24.29 (SD = 13.57), and females had an out-group social distance of 20.79 (SD = 11.45). Gender was not significantly related to out-group discomfort (F = 2.25; p = .14; d = .23) or intergroup anxiety (F = .00; p = .99; d = .00). Age was not significantly related to RB (r = -.04; p = .60; d = −.08), out-group social distance (r = −.12; p = .10; d = −.24), out-group discomfort (r = −.10; p = .15; d = −.20), or intergroup anxiety (r = .04; p = .55; d = .08). Because none of the participant variables affected both the independent and dependent variables, participant variables were not controlled in further analyses.

Primary and Mediator Analyses

Pearson’s correlations were run to examine all possible relationships between each of the study variables. As expected, belief in race as biological was significantly related to out-group social distance (see Table 2).

Table 2 Correlation matrix for the variables: belief in race as biological, intergroup anxiety, out-group discomfort, and out-group social distance
Full size table

Mediation models were run using the bootstrapping method aided by PROCESS, an SPSS macro. Bootstrapping provides an estimate of the indirect path (i.e., the path from the predictor to the outcome, through the mediation variable) based on a sampling distribution of the products of the indirect path coefficients. Each of our analyses was based on 1000 resamples of the data set with a bias corrected 95 % confidence interval (Preacher and Hayes 2008). In this method, the indirect effect is considered significant at p < .05 if the provided confidence interval does not contain the value of 0. The PROCESS macro allows multiple mediators to be tested simultaneously in order to tease apart the potentially overlapping effect of confounded variables. Thus, for the first analysis examining out-group discomfort as a mediator on the relationship between belief in race as biological and out-group social distance, intergroup anxiety was included as a control variable. Likewise, for the second analysis examining intergroup anxiety as a mediator on the relationship between belief in race as biological and out-group social distance, out-group discomfort was included as a control variable. All variables were standardized before being entered into the analyses.

For the first analysis, the indirect path from belief in race as biological to out-group social distance through the out-group discomfort mediator was significant (indirect effect = .06, CI [.02, .13]). For the second analysis, the indirect path from belief in race as biological to out-group social distance through the intergroup anxiety mediator was not significant [indirect effect = .00, CI (−.03, .05)]. Out-group discomfort, but not intergroup anxiety, mediated the relationship between belief in race as biological and out-group social distance (see Fig. 2).

Fig. 2
figure 2

The meditation of out-group social distance on the relationship between belief in race as biological and out-group social distance. Note Standardized regression coefficient in parentheses indicates the relationship between belief in race as biological and out-group social distance, while out-group discomfort is included in the model

Full size image

Discussion

As expected, belief in race as biological was related to out-group social distance. People who believe race is biological appear to be less likely to date, befriend, or share communities with out-group members. Perhaps such people believe that interactions across racial groups are unnatural and may perceive such interactions as crossing biological boundaries. The aim for this analysis was to contribute to understanding how the cognitive process of belief in race as biological affects intergroup interactions; findings suggest that out-group discomfort—a state-based discomfort—may be one such affective mechanism. Belief in race as biological and out-group discomfort seem to work in a mutually reinforcing process whereby rigid beliefs about race trigger discomfort when anticipating situations in which there will be interaction across races, simultaneously, felt discomfort is rationalized as a function of essentialist beliefs (e.g., “I feel uncomfortable interacting across race because racial groups are naturally incompatible”). Belief in race as biological was not related to intergroup anxiety, a trait-based discomfort. Taken together, findings suggest that the affective arousal experienced while anticipating interacting with group members across presumed biological boundaries is context specific (rather than trait based) and may not be consistently experienced across contexts as a more trait-based model might predict. Perhaps belief in race as biological is more of a rationalizing ideology than an actively endorsed one that might be expected from trait-based anxiety; in other words, people may not so much actively triumph biological ideals about race, but rather in specific settings in which they feel affective discomfort, they may rationalize this discomfort with the belief that racial groups are biologically distinct.

Study 3

A relatively small body of research has examined how race as a biological idea is “learned,” specifically in relation to primary education (Donovan 2014; Morning 2008). Study 3 examined the extent to which beliefs about race as biological can be “unlearned”—in both the short term and long term—by training experiences (e.g., workshops) that specifically address beliefs about race. In addition, the researcher examined whether immigration status and social distance to out-groups would affect trajectories of unlearning beliefs about race as biological.

The research that has examined how belief in race as biological is learned has focused on the content of textbooks. In a comprehensive review of 80 of the most commonly used high school biology textbooks from 1952 to 2002, Morning (2008) found that 92 % of textbooks from 1952 to 1962 contained explicit definitions or categorizations of people into distinct races (e.g., a taxonomy of “races of man” p. 117), and this percentage declined to 43 % of textbooks from 1993 to 2002. However, Morning (2008) also found that while 0 % of biology textbooks from 1952 to 1962 contained discussion of race in the context of medical disorders (e.g., African Americans are more prone to sickle cell anemia than white/European Americans), this percentage increased dramatically to 93 % of textbooks from 1993 to 2002. Learning about racial differences in relation to medical conditions may engender more essentialist ideas about race (Smedley and Smedley 2005), particularly if they are not contextualized in counter arguments such as the idea that predispositions to diseases (e.g., sickle cell) actually affect clinal groups (i.e., people with shared indigenous ancestry), which is often confounded with “race” (Marks 1996).

Indeed, Donovan (2014) has directly tested the effects of reading racialized medical passages about race on grade students’ beliefs in race as biological. Forty-three eighth-grade students read either a racialized sample textbook passage (“The allele responsible for sickle cell anemia is particularly common among people of African descent; about 9 % of African Americans are heterozygous for this allele” Excerpt p. 470) or a non-racialized passage (“About 2 million Americans (0.6 %) are carriers of the allele responsible for sickle cell anemia. Around 72,000 people have the symptoms of the disease because they are homozygous.” Excerpt p. 470). Reading the racialized passage was significantly related to higher levels of belief in race as biological.

Just as beliefs about race can be learned, it is expected that beliefs about race as biological can be “unlearned” by training experiences (e.g., workshops) that specifically address arguments against the biological validity of race, the historical rationale for the myth of race, and the continued social significance of race today. It is expected, however, that learning trajectories about beliefs about race may be different for different subgroups of students. In the current sample, there is a large percentage of immigrant students, who may have different understandings about race given that the biology versus social construction of race has largely been a US-based discourse (Omi and Winant 1994). In addition, given the findings in Study 1 about the effects of early exposure to diversity on beliefs about race, it is possible that immigrant students come from more homogenous countries and have less early exposure to racial diversity than US-based students. Thus, the researcher examined whether there were different “unlearning” trajectories for immigrant versus US born students. In addition, given the implications of findings in Study 2 regarding the way in which people’s beliefs about race as biological may work as a way to rationalize out-group discomfort and distance, the researcher examined whether students with higher levels versus lower levels of out-group social distance had different unlearning trajectories related to beliefs in race as biological. The researcher’s prediction was that unlearning trajectories among students with higher levels of out-group social distance would be significantly more drastic (compared to those with low levels of out-group social distance) given that these students may have been using beliefs about race as biological to rationalize their distance to out-group members and may thus have a steeper “unlearning” curve.

Method

Participants

Thirty-one students from a private high school in the Boston area completed a short paper and pencil questionnaire, comprised of Likert-scaled items and a demographics questionnaire, prior to a workshop on race, immediately following, and again 6 weeks after the workshop. Participants’ mean scores were evaluated for changes over the 3 time periods. All students who were boarding at the high school were invited to participate in the study. Thirty-one students (20 females; 11 males) from a pool of approximately 75 boarding students completed the study. Students comprised a racially and ethnically diverse group, self-identifying as Asian (41.9 %), black (16.1 %), Latino, (6.5 %), white (12.9 %), and multiracial (19.4 %). Immigrant students comprised 45.2 % of the sample (n = 14), and US born students comprised 54.8 % (n = 17) of the sample. Immigrant students came from a number of different nations including Korea, Japan, Taiwan, Russia, and Germany. Among immigrants, the average age of immigration to the USA was 13.91 (SD = 3.02). The ages of the participants overall ranged from 15 and 19 (M = 15.97; SD = 1.43).

Procedure

Surveys were administered to participants 2 days prior to the workshop on race, immediately following the workshop, and again 6 weeks after the workshop. For the time 1 and time 3 administrations, the researcher and colleagues visited the students and administered the surveys in their dormitories during evening hours. Time 2 surveys were taken at the location of the workshop. The researcher and a team of 4 other colleagues with experience in both clinical psychology and cultural sensitivity training conducted the workshop. The workshop occurred during the course of the day and spanned a 5-hour period, which included a 15-min break and a 1-hour lunch break. The workshop included a didactic presentation about race and ethnicity, affinity groups (i.e., students met for closed discussions with same race peers), and large group discussions (see Tauric et al. 2013 for more detail about the workshop). The didactic presentation included explicit discussion about the biological myth of race including discussion about how the belief in race as biological developed (i.e., as a way to justify human atrocities such as genocide of Native Americans and enslavement of blacks) and arguments against the biological validity of race (e.g., at least 99.9 % of DNA is shared between any two randomly selected people, either within or across “racial” groups). In addition, students participated collectively in a “race sorting” task provided on the PBS website (http://www.pbs.org/race/002_SortingPeople/002_00-home.htm) which demonstrates to students the difficulty of categorizing people racially on the basis of physical appearance.

Measures

Belief in race as a biological construct was measured by the same scale used in Study 1 and Study 2 (Tawa and Kim 2011). With the current sample, the internal reliability for this scale was good at time 1 (a = .73) and time 2 (a = .72) and adequate at time 3 (a = .58).

Out-group social distance was measured by 3 questions constructed for this study, which the researcher felt would be briefer and more relevant for high school-aged participants compared to the social distance scale developed by Rollock and Vrana (2005). Each question was rated by participants in relation to five reference groups: Asian, black, Latino, white, Native American, and multiracial; thus, the scale was comprised of 15 responses overall. The questions were: “I would be comfortable at a party that was at least 75 % [insert reference group],” “I would be comfortable being friends with members of the following group [insert reference group]” and “I would be willing to date a member of the following group [insert reference group].” Similar to the approach in Study 2, out-group social distance scores were constructed by totaling distance scores to all out-group members (e.g., for an Asian participant, the total distance scores for black, Latino, white, Native American, and multiracial). With the current sample, internal reliability estimates for out-group discomfort were strong (a = .85).

Results

Preliminary Analyses

Relations were explored between participants’ immigration status and gender on each of the study variables: out-group social distance assessed at time 1, and belief in race as biological (RB) assessed at each time point. Racial subgroups were too small (ranged from 2 to 13) for statistical comparison. Immigrants (M = 44.15; SD = 11.58) had a higher social distance to out-groups compared to US born participants (M = 35.69; SD = 13.56), although this difference was marginal (F = 3.18; p = .09; d = .64). Immigration status was not related to RB at time 1 (F = 1.25; p = .27; d = .40), time 2 (F = .05; p = .83; d = .08), or time 3 (F = .13; p = .72; d = .13). Gender was not significantly related to out-group social distance (F = .17; p = .68; d = .16), or RB at time 1 (F = .52; p = .48; d = .27), time 2 (F = .25; p = .62; d = .19), or time 3 (F = .77; p = .39; d = .33).

Primary Analyses

A repeat measures analysis determined that changes in beliefs about race as biological across the three time periods was significant as a quadratic function (F = 6.73; p = .02), suggesting a significant “U” shaped trajectory. Belief in race as biological declined from time 1 (M = 4.02; SD = 1.13) to time 2 (M = 3.41; SD = .39) and then rose again slightly at time 3 (M = 3.52; SD = .69).

The researcher then examined how the RB trajectories were affected by participant differences in immigration status and high and low levels of out-group social distance. Out-group social distance was dichotomized into “high and low” levels by finding the median score and splitting the sample. Immigration status and the dichotomized out-group social distance variable were entered as a between-subjects factor. Only out-group social distance was significant as a between-subjects factor, and the interaction between RB and out-group social distance was significant as a quadratic function (F = 9.84; p < .01). This suggests that the trajectory of learning about RB for people high on out-group social distance was significantly different than those who reported being low on out-group social distance. For participants at high levels of out-group social distance, mean scores on RB were high at time 1 (M = 4.52; SD = 1.05), declined drastically at time 2 (M = 3.42; SD = .43), and rose again slightly at time 3 (M = 3.73; SD = .41). For participants with low levels of out-group social distance, mean scores on RB were relatively consistent across time 1 (M = 3.39; SD = .98), time 2 (M = 3.42; SD = .39), and time 3 (M = 3.34; SD = .90). These different trajectories are plotted in Fig. 3.

Fig. 3
figure 3

Unlearning trajectories about belief in race as biological for students with low and high levels of out-group social distance. Note OGSD = out-group social distance

Full size image

Discussion

Study 3 provides preliminary evidence that beliefs about race as biological can be reduced in both the short- and long term even with a relatively brief intervention. For the overall sample, there was a significant reduction in belief in race as biological immediately after the workshop. The gradual increase over the 6-week duration suggests that in the absence of sustained education about race, people may begin to resort to older beliefs. These patterns of change were particularly the case among people with higher levels of social distance to out-group members. This subset of students were likely to have previously used belief in race as biological to rationalize their distance from out-group members. Certainly these students’ greater decline in beliefs about race as biological reflects a steeper learning curve. Over the course of 6 weeks following the workshop, negative experiences with out-group members may have resulted in some students falling back on older rationalizations about biological distinctions in race. Yet, it seems that for other students, this was no longer an available heuristic as the mean scores of belief in race as biological did not rise to the level reported in time 1 (Fig. 2).

These findings offer some preliminary evidence that belief in race as biological can be relatively easily influenced through educational interventions. However, as a study with a relatively small sample size, future research should continue to more rigorously establish these unlearning effects particularly in lieu of some possible study confounds such as social desirability effects. It is possible, for example, that students scored lower on beliefs about race as biological after the workshop because they felt that it was the more socially correct response. To what extent these unlearnings are actually internalized and accepted by students remains to be seen through future research with more robust sample sizes, longer duration of assessment, and the inclusion of relevant control variables.

General Discussion

This paper contributes to a relatively small, but growing body of research on people’s beliefs about the concept of race. Study 2 supplements a major focus within this line of research; specifically, it adds to consistent findings that beliefs about race as a biological construct have negative implications for intergroup relations. Studies 1 and 3 direct future researchers toward further lines of inquiry related to beliefs about race; more specifically, future research could examine more thoroughly early life experiences contributing to beliefs about race as well as intervention strategies for “unlearning” beliefs about race as biological.

Moreover, across 3 studies, the empirical findings encourage a reconceptualization of beliefs about race as biological as relatively ephemeral, context specific, and perhaps more of a rationalizing ideology than a proactive one. Study 1 found that among lower SES participants, greater early life exposure to diversity was actually related to greater belief in race as biological. In diverse contexts in which there is greater competition for resources, people may rationalize feelings of threat from other groups (as threats to consume resources) through racialized attribution (e.g., Asian are inherently greedy). Consistently, in Study 2, out-group discomfort was found to be an affective mechanism through which beliefs about race as biological led to greater distance from out-groups. Unlike intergroup anxiety, which is conceptualized as a trait based, stable aversion to out-groups, out-group discomfort is conceptualized as momentary and specific to contexts in which one is a minority. It seems that people’s beliefs about race as biological become activated in specific settings in which people are a minority and may be used to rationalize avoidance or shortened interactions. Finally, Study 3 demonstrated that people’s beliefs about race as biological can be changed in both the short and long term, even with relatively brief intervention, again suggesting that beliefs about race are relatively ephemeral and subject to change. At least one study in the aforementioned body of research has also found evidence for a conceptualization of beliefs about race as a rationalizing ideology; Morton et al. (2009) found that people who endorse prejudicial views of out-group members tended to endorse biological beliefs about race as a way of rationalizing the exclusion of an out-group member, but no longer endorsed biological beliefs when a member of their in-group was the target of exclusion. To date, measure of beliefs about race, including the measure adopted in this study, appears to more directly assess stable and decontextualized beliefs about race. Future research should invest in developing a measure of beliefs about race that is context specific, and that captures the extent to which ideologies are endorsed proactively or as rationalizing ideologies.

The findings in this study should underscore to educators in particular, the pressing need to challenge students’ biological assumptions about race, particularly given the implications for students’ peer relations. It is concerning that today, an increasing number of biology textbooks promote—albeit indirectly—a biological conception of race in the context of medical disorders (Morton et al. 2009; Smedley and Smedley 2005). Although indirect, these messages about race do increase students’ beliefs in race as biological (Donovan 2014). Given the persistent findings in the literature and in this study about the negative affects of beliefs in race as biological on intergroup behaviors, educational curricula should be evaluated for the extent to which they promote essentialized beliefs about race. Moreover, educators and/or educational institutions should take it upon themselves to offer learning components (either within courses or as extracurricular workshops) that challenge the notion of race as biological. If beliefs about race are indeed more context specific and do function as rationalizing ideologies, then race relations interventions could simultaneously work to decrease beliefs in race as biological; by providing more positive in-group interactions, students will less frequently rely on biological heuristics about race. Study 3 was not able to determine the extent to which changes in beliefs about race were directly related to didactic learning about race or whether they were related to having positive cross-group interactions. While future research could work to tease apart these potentially independent contributing factors, it seems that promoting positive race relations and actively challenging beliefs about race as biological work in a “virtuous cycle” toward the end of improving peer relations across racial groups in college settings.

Although the strength of our findings in Study 3 is limited by a relatively small sample size, future research should continue to work to establish both the short- and long-term “unlearning processes” related to beliefs about race. Future research could work to establish that students are not only “unlearning” beliefs about race as biological, but also they are learning about the historical and continued significance of race as a social and political construct. Lastly, research should work toward the end of causally establishing that changes in beliefs about race do in fact relate to changes in quality and frequency of cross-racial interactions.