Shades of American Identity: Implicit Relations between Ethnic and National Identities

Abstract

The issue of ethnic diversity and national identity in an immigrant nation such as the USA is a recurrent topic of debate. We review and integrate research examining the extent to which the American identity is implicitly granted or denied to members of different ethnic groups. Consistently, European Americans are implicitly conceived of as being more American than African, Asian, Latino, and even Native Americans. This implicit American = White effect emerges when explicit knowledge or perceptions point in the opposite direction. The propensity to deny the American identity to members of ethnic minorities is particularly pronounced when targets (individuals or groups) are construed through the lenses of ethnic identities. Implicit ethnic–national associations fluctuate as a function of perceivers’ ethnic identity and political orientation, but also contextual or situational factors. The tendency to equate being American with being White accounts for the strength of national identification (among European Americans) and behavioral responses including hiring recommendations and voting intentions. The robust propensity to deny the American identity to ethnic minority groups reflects an exclusionary national identity.

Many countries face the challenge of balancing attachment to the nation and identifications with more specific ethnic groups. This issue has been at the center of long-standing debates in pluralistic or multiethnic states and will continue to garner attention as ethnic diversity increases in many nations. Social psychologists have developed insightful models to account for the dynamics between superordinate (e.g., national) and subgroup (e.g., ethnic) identities (Dovidio, Gaertner, & Saguy, 2009; Hornsey & Hogg, 2000; Mummendey & Wenzel, 1999). Adding to this literature, this paper focuses on a particular aspect of this complex puzzle and on a specific socio-historical context: we examine the relative inclusion or exclusion of ethnic groups in the national identity in the USA. The USA is an immigrant nation founded on shared values, as opposed to a shared ethnic identity, yet its population is composed of identifiable ethnic groups that vary in terms of numerical or social status, power, resources, length of immersion into mainstream culture, and conditions of immigration. In this context, what does it mean to be American¹ and who is perceived as embodying the national identity? To what extent is the American identity psychologically granted or denied to members of different ethnic groups? Is the national identity as readily ascribed to African, Asian, European, Latino, and Native Americans?

Recently, there has been a growing interest in social psychology and related disciplines for lay definitions of the American identity and their implications for ethnic relations or attitudes toward immigration (Citrin, Wong, & Duff, 2001; Deaux, 2008; Schildkraut, 2011a, 2011b; Stepick, Stepick, & Vanderkooy, 2011; Theiss-Morse, 2009; Wright, Citrin, & Wand, 2012). Although these issues vary depending on the socio-historical context, investigators in countries other than the USA have delved into similar questions (Pehrson & Green, 2010; Sibley, 2013). A core aspect of this work is to emphasize that lay conceptions of a national identity contribute to define its boundaries (Mukherjee, Molina, & Adams, 2012; Pehrson & Green, 2010; Sibley, 2013; Theiss-Morse, 2009). For example, characteristics deemed prototypical of a true American imply that individuals or groups who deviate from this prototype are seen as second-rank citizens. The American ethos is often defined in terms of civic values and emotional attachment to the nation rather than based on more nativist elements (Citrin, Haas, Muste, & Reingold, 1994; Devos & Banaji, 2005; Huynh, Devos, & Altman, 2014). Yet, behind the deliberate endorsement of inclusive principles lurks a more exclusionary national identity.

Capitalizing on recent conceptual and methodological advances in the study of implicit social cognition (Bargh, 2007; Gawronski & Payne, 2010; Wittenbrink & Schwarz, 2007), researchers have documented that the American identity is more strongly associated with European Americans than with ethnic minorities (Dasgupta & Yogeeswaran, 2011; Devos & Banaji, 2005). In other words, when perceptions are assessed at a level that escapes conscious control or awareness, a robust American = White effect emerges. Here, we review and integrate research on this topic. After a description of this effect, we briefly discuss factors that may potentially contribute to it. Then, we highlight striking fractures between the explicit and implicit giving of the American identity. We pursue with research showing that emphasizing or downplaying ethnic identities influences the implicit ascription of the American identity. Next, we cover sources of variability in implicit ethnic–national associations. Finally, we examine potential implications of the American = White effect. Although the discussion centers on findings based on implicit assessments of ethnic–national associations, we allude occasionally to results based on more traditional self-report measures. Comparing these two levels enriches our understanding of the dynamics coming into play and provides opportunities to highlight both differences and similarities between these two levels of responding.

Implicit American=White Effect

The difficulty of psychologically granting the American identity to ethnic minorities or the propensity to view members of ethnic minority groups as perpetual foreigners has been documented using a variety of implicit techniques. For example, the Implicit Association Test (IAT; Greenwald, McGhee, & Schwartz, 1998; Nosek, Greenwald, & Banaji, 2007) can easily be adapted to examine associations between ethnic and national identities (Devos & Banaji, 2005). In this case, the technique is set up to assess the direction and strength of associations between two ethnic groups (e.g., Asian Americans and European Americans) and the concepts American versus foreign. Ethnic stimuli can be faces of Asian Americans and European Americans. Symbols such as the American flag, a map of the USA, or the Capitol building can be used to represent the concept American and contrasted to similar foreign symbols. The technique involves comparing the ease with which participants categorize these stimuli under two configurations. In one set of trials, American symbols are paired, as quickly as possible, with Asian American faces, and foreign symbols with European American faces. In another set of trials, the two pairs of concepts are combined in the opposite way: American symbols are paired with European American faces, and foreign symbols are associated with Asian American faces. Typically, this second configuration is easier to perform (shorter response latencies) than the first one. Given that the role of conscious control is limited on this task (Cvencek, Greenwald, Brown, Gray, & Snowden, 2010), this pattern is consistent with the idea that European Americans are implicitly conceived of as being more American than Asian Americans. Indirectly, it also suggests that prominent symbols associated with the national identity (e.g., US flag and historical landmarks) may be divisive because they are inextricably linked with the dominant ethnic majority group (European Americans) and exclude or marginalize ethnic minority groups (Butz, 2009).

Evidence for a robust American=White effect has been documented for various interethnic comparisons. Consistently, European Americans are seen as more American than Latino Americans (Devos, Gavin, & Quintana, 2010), Asian Americans (Devos & Banaji, 2005; Devos & Heng, 2009), African Americans (Devos & Banaji, 2005; Rydell, Hamilton, & Devos, 2010), and even Native Americans (Nosek, Smyth, et al., 2007). This should not be taken as evidence for the notion that all ethnic minority groups were equally seen as not American. Lumping together these groups would neglect meaningful variations in the magnitude of the effect depending on the specific interethnic comparison. Indeed, the effect was particularly pronounced when European Americans were compared with Latino or Asian Americans; it was weaker when African or Native Americans served as the contrasting group. Other sources of variation, including perceivers’ ethnic identity, are discussed in another section of this paper.

Although initial demonstrations of the effect at the implicit level were based on studies relying on the IAT, convergent findings have been obtained with other techniques including the go/no-go association task (Devos & Ma, 2008; Yogeeswaran, Dasgupta, Adelman, Eccleston, & Parker, 2011), a sequential priming technique (Devos & Heng, 2009), and an inclusion categorization task (Rydell et al., 2010). Not only does this contribute to establishing that the effect is not limited to a specific technique, but it also provides a more fine-grained understanding of the locus of the effect. An important feature of the IAT is that it measures the relative strengths of associations between two pairs of concepts. To go back to the example described earlier, the IAT provides a relative assessment of the extent to which the concepts American versus foreign are associated with two ethnic groups (European Americans and Asian Americans). Given the comparative element built in the technique, it is not possible to separate the contribution of associations about European Americans versus Asian Americans or about the perceived Americanness versus foreignness of each ethnic group. Does the effect stem from a difficulty associating the American identity to Asian Americans or a proclivity to associate the concept foreign to this ethnic group? Do these two ethnic groups differ in the extent to which they are associated with the concept American or with the concept foreign? Studies relying on the IAT cannot tease apart these alternative accounts. Thus, it is important to keep in mind that the term American = White effect is a convenient shortcut to capture an effect that results from potentially distinct associations. The inherent limitations of comparative assessments should not necessarily be seen as flaws. Perceptions, attitudes, and judgments are often comparative in nature. Ascribing (or not) a trait or characteristic to a group often implies that it applies more (or less) to this group than to another group. Similarly, the ascription of the American identity to a target (group or individual) takes on meaning in relation to a concept or category that captures what is not American. For American respondents, the term foreign is often spontaneously used to characterize someone or something seen as not American. Studies relying on less comparative techniques provide some insights into the locus of the effect. For example, the go/no-go association task can be implemented such that respondents have to focus on the concept American and the role of the concept foreign is minimized. The effect obtained under this configuration suggests that White and non-White targets (groups or individuals) differ in the extent to which they are linked to the national identity and not just based on their relative foreignness (Devos & Ma, 2008).

Performances on implicit measures reflect a mix of automatic and controlled processes that can be isolated using a process dissociation approach (Payne & Stewart, 2007; Sherman, Klauer, & Allen, 2010). Borrowing this framework, Devos and Heng (2009) showed that the American=White effect can be conceptualized as an unintentional accessibility bias. From this perspective, individuals tend to assume that a White person is more likely to be American than an Asian person and rely on this assumption only in the event that they are unable to rely on appropriate information. In other words, individuals unintentionally use this heuristic to resolve ambiguity when controlled processes fail. A decomposition of performances on a sequential priming technique showed that the American=White effect was isolated into an automatic tendency to associate American with White (rather than with Asian). Increasing or decreasing the ability to rely on controlled processes did not alter this automatic bias.

Theoretical Foundations

The starting point for most of the work reviewed in this paper was an attempt to document a phenomenon (American=White effect) and to examine its boundary conditions, correlates, and ramifications rather than to test a specific theoretical account for the tendency to equate being American with being White. At the same time, this research was clearly inspired by a number of influential theoretical perspectives. In this section, we articulate various propositions that, in combination, provide a theoretical account for the American=White effect. This eclectic perspective will be expanded and nuanced as we cover more specific research questions in other sections.

Considering that implicit associations reflect knowledge acquired through repeated experiences within a particular socio-cultural context (Devos, 2008; Rudman, 2004), a variety of factors are likely to shape implicit ethnic–American associations including the relative numerical and social statuses of ethnic groups, their relative power, their symbolic and material resources, the length of immersion in American society, and the circumstances surrounding waves of immigration (e.g., immigration policies, economic opportunities, political or religious persecutions, or slavery). These factors are often inextricably linked. For example, length of immersion in American society and immigration policies are tied to asymmetries in terms of status or power. Given the pervasive ethnic hierarchy in the USA and the makeup of the general population, these factors contribute, in most cases, to fostering a strong connection between being White and being American.

According to social dominance theory (Pratto, Sidanius, & Levin, 2006; Sidanius & Petrocik, 2001; Sidanius & Pratto, 1999, 2012), the dominant ethnic group in a multiethnic society tends to be regarded as having ownership of the nation, its resources, and its symbols. Consequently, ethnic minority status entails an exclusion from the national identity. Consistent with this theoretical perspective, the American=White effect contributes to an ideological system that sustains group dominance. In line with the notion of ideological asymmetry (Sidanius & Pratto, 2012), the psychological and ideological forces underlying the effect work to the advantage of European Americans and to the disadvantage of ethnic minorities. The effect can be displayed across ethnic lines, but it is likely to have positive implications for the dominant White majority and negative implications for subordinate ethnic minorities. In the first case, the propensity to equate being American with being White reinforces the status of one’s ethnic group, and attachments to national and ethnic identities largely overlap. In the second case, the relative exclusion of one’s ethnic group goes hand in hand with a clear distinction, if not a conflict, between national and ethnic attachments. In that regard, the American=White effect should be thought of as one piece of a complex ideological puzzle that legitimizes ethnic-based hierarchies in the USA.

Research on social judgment and mental representations of categories has documented that some attributes may come to be perceived as expected or default values within a culture (Hegarty & Bruckmüller, 2013; Hegarty & Pratto, 2001; Smith & Zarate, 1992; Stroessner, 1996; Zarate & Smith, 1990). From this perspective, attributes of more frequently encountered or higher-status groups are conflated with attributes of larger social categories (Hegarty & Bruckmüller, 2013). Applied to the specific socio-cultural context considered here, the White racial identity in the USA operates as a norm or cultural expectation. The mental representation of the category Americans (i.e., the category norm) largely overlaps with attributes of the dominant White majority. As a result, lower-status ethnic minorities are seen as less prototypical of the national identity. Furthermore, when differences between ethnic groups need to be explained, attention shifts toward lower-status ethnic minorities. The various mechanisms and downstream consequences of this construction of the national identity contribute to the reproduction of status differences. Given that these processes largely function outside of conscious control and awareness, they perpetuate the notion of ethnic minority groups as second-class citizens. In other words, power and status asymmetries operate in a manner that cannot easily be noticed or challenged. The mechanisms underlying the construction of the national identity and asymmetrical positions in a social system mutually reinforce each other. Numerical majority, high social status, and dominance tend to be attributes associated with greater prototypicality. At the same time, once a group embodies the norm, it is seen as more powerful and higher in status (Bruckmüller & Abele, 2010).

The various factors shaping ethnic–national associations also point to the fact that the perceived prototypicality of ethnic groups cannot be reduced to a statistical basis (i.e., Whites represent the largest ethnic group in the USA) but is also likely to stem from differences in power and privileges between groups. In the contemporary USA, group-based hierarchies do not often manifest themselves through overt hostility or forceful oppression. At least on the surface, egalitarianism is one of the core components of the American identity (Devos & Banaji, 2005; Huynh et al., 2014). Below the surface, the tendency to equate being White with being American functions as a legitimizing ideology that perpetuates asymmetrical relations between the dominant White majority and subordinate ethnic minorities.

Dissociations between Implicit and Explicit Associations

Techniques assessing implicit social cognition are particularly useful when they reveal a different picture than self-reports. Several studies document a fracture between the implicit and explicit granting of the American identity. In the context of the Summer Olympics and considering a sport highly dominated by Black athletes such as track and field, would Black and White athletes competing for the USA be differentiated in terms of their association to the national identity (Devos & Banaji, 2005)? On explicit (self-report) measures, the American identity was more strongly associated with Black athletes than with White athletes. At the implicit level, Black athletes were less strongly associated with the American identity than White athletes. In the same vein, even when well-known Asian American celebrities (e.g., Connie Chung or Michael Chang) were contrasted to famous White Europeans (e.g., Elizabeth Hurley or Gérard Depardieu), the proclivity to associate American with White remained such that American symbols were more easily linked with names of White foreigners than with names of Asian American celebrities. In both cases, implicit associations strongly deviated from explicit perceptions or knowledge.

Native Americans represent a thought-provoking scenario for this topic of research. American Indians were living in the North American continent when the first European immigrants arrived. The term “Native American” itself stresses that members of this group were the original inhabitants of this land. From this perspective, Native Americans should be more strongly associated with the concept American than people of European descent. However, the fact that American Indians are a relatively small ethnic minority, have experienced the longest oppression by White people in Northern America, and are often depicted as “exotic others” may undermine the extent to which they are construed as typical Americans. A large data set collected through the Project Implicit website (https://implicit.harvard.edu) allowed us to determine whether American Indians were viewed as the “real Americans” or as “aliens in their own land” (Nosek, Smyth, et al., 2007). At the explicit level, American Indians were considered more American than Americans of European descent. At the implicit level, despite the fact that the American stimuli were not biased in favor of White America (e.g., pictures of natural scenes easily associated with American Indians such as Old Faithful, Niagara Falls, or Grand Canyon), the concept American was more strongly associated with White Americans than with Native Americans.

The striking dissociations between explicit and implicit associations reported in this section illustrate that statements or ideas that might be endorsed because they appear to be subjectively valid (e.g., “Native Americans were the original Americans; in that sense they are the true Americans”) might not align with what is produced based on associative knowledge that is being automatically activated (Gawronski & Bodenhausen, 2006). To a large extent, everyday images of prominent Americans – political and social leaders, celebrities, and a majority of the population – reinforce associations of America as White. As a result, the American identity might be associated with the ethnic identity of being White – so much so that even deliberate, conscious rejection cannot alter this association.

Emphasizing versus Downplaying Ethnic Identities

The extent to which individuals or groups are cognitively construed through ethnic lenses affects the ability to automatically ascribe the American identity to them. Let us consider a series of studies focusing on two famous actresses (Devos & Ma, 2008): Lucy Liu (well-known Asian American) and Kate Winslet (well-known White foreigner). When these two actresses were categorized based on their respective ethnic identity, the White person was seen as more American than the Asian person. When the same targets were categorized based on their personal identities, the effect was reduced although not always completely eradicated. Convergent findings were obtained when pictures either emphasized or de-emphasized their respective ethnic heritage. Pictures of Lucy Liu that portrayed her in a stereotypic Asian character magnified the American=White effect compared with pictures that depicted her in more neutral outfits.

Over the course of the 2008 US presidential campaign, questions surrounding the citizenship and national identity of Barack Obama were raised. To cast doubts on his attachment to the American identity, some opponents emphasized his foreign-sounding middle name (Hussein) and his ties with Indonesia. Others highlighted his failure to wear a flag pin and salute the flag on one occasion. Even well into his presidency, attempts to invalidate Obama’s citizenship continued. Underlying these reactions was a common theme: “Obama is not American”. These reactions might be symptomatic of the pervasive difficulty to grant the national identity to non-White Americans. In this context, Devos and Ma (2013) sought to document that the relative difficulty of seeing Obama as an American was a function of the extent to which he was construed as a Black person. As the campaign unfolded, they successively examined the extent to which Barack Obama was seen as more or less American than four White politicians: Tony Blair (then Prime Minister of the United Kingdom), Hillary Clinton (also running for President), and John McCain (the Republican opponent). Consistently, Obama was implicitly viewed as less American than McCain, Clinton, or even Blair when racial identities were stressed. This effect was attenuated when the politicians were categorized based on their personal identity or other relevant identities (e.g., gender, age, or political affiliation). In the context of the 2012 US presidential election, the picture changed in the sense that overall, Barack Obama was implicitly seen as more American than Mitt Romney, but highlighting racial identities still undermined the automatic association between the national identity and President Obama (Ma & Devos, 2013). In sum, data collected during the last two US presidential elections confirmed the findings of studies focusing on two specific actresses (Lucy Liu and Kate Winslet; Devos & Ma, 2008). In both cases, to the extent that the targets were categorized based on their ethnic or racial identities (in comparison with other dimensions), the person belonging to the ethnic or racial minority group was implicitly seen as less American than the White person.

Recent research documents complex reactions to the public or private expression of ethnic identity (Yogeeswaran et al., 2011). In particular at the implicit level, a double-standard was apparent: ethnic minorities were construed as less American when group members expressed their ethnic attachment publicly, while Whites were implicitly accepted as legitimate Americans regardless of the circumstances in which they expressed their ethnic identity. In the same vein, highlighting ethnic minorities’ allegiance to their ethnic group made them appear less American, whereas emphasizing ethnic minorities’ national service made them appear more American (Yogeeswaran, Dasgupta, & Gomez, 2012). As a whole, studies reviewed in this section stress that an emphasis on ethnic or racial identities may often implicitly disqualify members of ethnic minorities as American. This is not meant to advocate for a need to de-emphasize distinctions that are often impossible to set aside and also serve as important sources of identities (Park & Judd, 2005). More simply, they point to forms of exclusion that stem from ethnic categorizations and contrast with a definition of the American identity grounded in civic values.

Implicit Ethnic–National Associations as a Function of Perceivers’ Characteristics

Research reviewed in the previous section already points out that the propensity to equate being American with being White fluctuates. We now focus on other sources of variations in ethnic–national associations. First, implicit assumptions about who embodies the national identity are a function of the perceivers’ ethnic identity. The American=White effect is not exclusively displayed by European American individuals. In some cases, members of ethnic minorities also exhibit a greater tendency to associate the American identity with European Americans than with ethnic minorities. This is the case even when the ethnic minority group being considered is their own group. For example, Latino and Asian American respondents implicitly view their own ethnic group as being less American than Caucasian/European Americans, but not to the same extent than what is observed among Caucasian/European American respondents (Devos & Banaji, 2005; Devos et al., 2010). African American respondents typically do not differentiate their ethnic group and European Americans in terms of their implicit Americanness (Devos & Banaji, 2005). Despite the fact that their group represents an oppressed minority, Native Americans associate American landmarks more easily with their group than with people of European descent, showing the opposite of an American=White effect (Nosek, Smyth, et al., 2007).

Thus, ethnic–national associations cannot be reduced to a White hegemony effect. According to the ingroup project model (Mummendey & Wenzel, 1999; Wenzel, Mummendey, & Waldzus, 2007), group members often consider that the ingroup is more aligned with the superordinate identity than outgroups. This effect has been shown to operate at an automatic level (Bianchi, Mummendey, Steffens, & Yzerbyt, 2010). Two data collections conducted at diverse and urban universities isolate the impact of perceivers’ ethnic group membership on implicit ethnic–national associations (Devos, Stevens, & Anderson, 2009). African, Asian, Latino, and European American respondents were randomly assigned to an ethnic comparison that involved their ethnic group or not. For example, Latino participants completed an IAT comparing either Latino Americans and European Americans, Latino Americans and Asian Americans, or European Americans and Asian Americans. In the first two cases, the ingroup (Latino Americans) was involved in the comparison; it was compared either with the ethnic majority group (i.e., European Americans) or with another ethnic minority group (i.e., Asian Americans). We will refer to the third condition as the third-party perspective in the sense that participants compared two groups they did not belong to (i.e., European Americans and Asian Americans in the case of Latino participants). Although this third-party perspective should not be thought of as providing a neutral or objective perspective, it allowed us to isolate the crucial role of respondents’ ethnic group membership in patterns of implicit ethnic–national associations. Without going into the details of the fairly complex design of these two data collections, the findings clearly documented that, irrespective of their specific ethnic identity, individuals displayed a tendency to associate the American identity more strongly to an ingroup than to an outgroup. Based on the evidence reviewed so far, it is not surprising for White American respondents to do so. Thus, let us focus on evidence for this statement among members of ethnic minorities (Latino, Asian, and African American participants). In their case, evidence for an ingroup prototypicality effect clearly emerged when the ingroup was compared with another ethnic minority group. For example, Latino participants associated the American identity more strongly with their group than with Asian Americans, whereas Asian American participants displayed an effect in the opposite direction (i.e., associated the American identity more strongly with their group than with Latino Americans). In both cases, the effect was consistent with an ingroup prototypicality effect. Even when members of ethnic minorities compared their group with European Americans, a tendency to link the ingroup to the national identity emerged because the effect obtained was reliably weaker than the American = White effect displayed by participants who were not involved in the ethnic comparison (third-party perspective). For example, the tendency to associate the American identity to European Americans to a greater extent than to Asian Americans was less pronounced among Asian American respondents than among Latino respondents. Conversely, the American = White effect documented when European Americans were contrasted to Latino Americans was weaker among Latino respondents than among Asian American respondents. In addition, European Americans respondents always displayed a stronger American = White effect than members of ethnic minorities who were not involved in the interethnic comparison. In sum, the larger socio-cultural environment contributes to sustain the perception that the majority or dominant group in a society embodies the superordinate identity to a greater extent than minority groups or groups holding less power, but these perceptions fluctuate depending the perceivers’ position in this asymmetrical system (Waldzus, Mummendey, Wenzel, & Boettcher, 2004; Wenzel, Mummendey, Weber, & Waldzus, 2003). Ingroup prototypicality effects may stem from the relative accessibility of information about social groups (Waldzus et al., 2004) or a need to achieve or maintain a positive social identity (Wenzel et al., 2003), but perceptions of group prototypicality may also serve instrumental functions such as legitimizing or challenging status differences (Rosa & Waldzus, 2012; Sindic & Reicher, 2008; Weber, Mummendey, & Waldzus, 2002).

If the extent to which groups are assigned the American identity partially reflects group-based hierarchies, one would expect ethnic–American associations to covary with ideological or political placement (Jost, 2006; Jost, Glaser, Kruglanski, & Sulloway, 2003). Compared with liberals, conservatives tend to show implicit and explicit preferences for higher-status groups over lower-status groups. In line with this proposition, the magnitude of the implicit and explicit American = White effects increases as one moves from the liberal end to the conservative end of the political spectrum (Nosek, Smyth, et al., 2007). Half-way through the first term of Obama’s presidency, Crawford and Bhatia (2012) showed that political conservatism was reliably associated with the implicit and explicit beliefs that Barack Obama was not American. Thus, ideological motives are linked to the propensity to deny the American identity to non-White individuals.

Context Sensitivity of Implicit Ethnic–National Associations

In line with research stressing that implicit associations are not fixed or rigid (Dasgupta, 2009; Gawronski & Sritharan, 2010), several studies document the malleability of implicit ethnic–national associations. For example, valence affects the extent to which White Americans include African Americans into the national identity (Rydell et al., 2010). When positive information was received about African Americans, that group was more likely to be included into the national identity. In contrast, when negative information was received about African Americans, greater exclusion from the national identity ensued. The strength of the American=White association, although normally quite strong, varied depending on the valence (positive or negative) of information accessible about African Americans.

Practicing counter-stereotypic associations (Kawakami, Dovidio, Moll, Hermsen, & Russin, 2000; Plant, Peruche, & Butz, 2005) might also efficiently reduce the automatic tendency to associate the American identity more swiftly with Whites than with ethnic minorities (Devos & Heng, 2009). If individuals practiced responses counteracting the American=White linkage, meaning that the Asian ethnicity was more diagnostic of American than the White ethnicity, the automatic American=White accessibility bias was eroded.

Inducing a more complex representation of the social environment has been shown to reduce intergroup biases (Crisp & Hewstone, 2006, 2007). A largely neglected aspect of this proposition is to examine the impact of the number of groups being compared in a given context. In a series of studies (Devos, Hamilton, & Mohamed, 2012), we showed that the American identity was more readily ascribed to an ethnic minority group as the number of distinct ethnic groups (two, three, or four) being simultaneously considered increased. Exposure to diversity in and of itself was not sufficient to reduce the American=White effect, and the benefits of inducing a more complex representation of the environment was not linked to greater reliance on individuating processes. Instead, the American=White effect was attenuated when the setting moved away from simplistic dichotomous ethnic categorizations. In line with a multicultural perspective, inducing a more complex, yet categorical, mindset fostered a pluralistic and inclusive definition of the superordinate identity (Waldzus, Mummendey, & Wenzel, 2005; Waldzus, Mummendey, Wenzel, & Weber, 2003).

Research conducted in other countries also points to the fact that multiethnic states are not systematically characterized by a relative exclusion of minorities from the national identity. For example, similar studies conducted in New Zealand revealed a very different pattern (Sibley & Liu, 2007; Sibley, Liu, & Khan, 2008). Majority group members (Pakeha or New Zealanders of European descent) showed little evidence of implicit ingroup biases; they perceived their own ingroup and culture, and Maori peoples and culture, as equally representative of the nation. In contrast, minority group members (Maori or indigenous New Zealanders) perceived their ingroup and culture as being more closely associated with representations of the nation. As is the case in research conducted in the USA, the answer to the question of who embodies the national identity is contingent upon ethnic group membership, but the pattern deviates strikingly from what is documented in North America. Despite the fact that Maori are consistently disadvantaged, this ethnic group not only thinks of itself as being a part of the national identity, it is also included by members of the socially dominant group. The implicit inclusion of Maori displayed by Pakeha participants seems to be a function of the endorsement of a bicultural social policy reflecting support for the incorporation of symbolic aspects of Maori culture (Sibley, Liu, & Khan, 2010). In other words, depending on the extent to which minority cultures are symbolically included in the national identity, patterns of implicit ethnic–national associations are likely to fluctuate drastically. This being said, this relatively inclusive symbolization of the national identity coexists with a clear opposition to policies addressing current resources inequalities or providing reparation for past injustices (Sibley et al., 2010).

Variations across ethnic groups or national contexts are consistent with the notion that implicit associations are rooted in experiences, bear the mark of cultural socialization, and reflect socio-cultural realities. An important goal for future research is to examine more directly the impact of the socio-structural context on the interconnections between ethnic and national identities. Recent research has documented the influences of historical, structural, and political characteristics of national contexts on intergroup attitudes (Guimond et al., 2013; Ho et al., 2012; Pehrson & Green, 2010). Even among pluralistic societies, socio-structural factors might be strong determinants of the propensity to exclude or include various groups in the implicit or explicit definition of the national identity. At the same time, behind the variability in socio-structural realities (often linked to specific political models to deal with issue of diversity and integration), there might be some common social–psychological determinants that are shared across societies characterized by group-based hierarchies.

Implications of Implicit Ethnic–National Associations

Implicit perceptions of Americanness have important ramifications. First, assumptions about who is most American may have contrasting implications for the self-concept of majority and minority group members. Relying on the balanced identity design (Cvencek, Greenwald, & Meltzoff, 2012; Greenwald et al., 2002), studies document that the American=White effect accounts for an asymmetry in the strength of national identification (Devos & Banaji, 2005; Devos et al., 2010). For European American participants, the more they displayed an American=White effect, the more they themselves exhibited a strong national identification (self+American association). In contrast, for members of an ethnic minority group (Latino or Asian Americans), there were no reliable relations between the tendency to equate being American with being White and the strength of national identification. These data are strictly correlational and, as such, do not establish a causal impact of the American=White linkage. This being noted, the implicit American=White effect accounts for the merging of ethnic and national identifications among European Americans, whereas these two identifications are clearly more distinct for members of an ethnic minority group. These findings are consistent with research on the interconnections between ethnic and national identifications (Phinney, Cantu, & Kurtz, 1997; Rodriguez, Schwartz, & Whitbourne, 2010; Sidanius, Feshbach, Levin, & Pratto, 1997; Staerklé, Sidanius, Green, & Molina, 2010). At the same time, they add to the literature by providing evidence for this asymmetry at the implicit level and by identifying a key moderator. Pushing this analysis a step further, a follow-up study documented that the aforementioned pattern emerged when thoughts about ethnic hierarchy or inequalities were primed, but not when steps toward greater equality were emphasized (Devos et al., 2010). To use a terminology based on social dominance theory (Sinclair, Sidanius, & Levin, 1998), Caucasian Americans displayed a stronger national identification than Latino Americans in a hierarchy-enhancing context, but not in a hierarchy-attenuating context.

The fact that Asian and Latino Americans displayed the American=White effect at the implicit level but that the propensity for them to do so was not correlated with the strength of their implicit national identification points to a clear distinction between associations about ingroup and self. To put it bluntly, Asian and Latino respondents displayed patterns of responses that could be thought of as “my group may be less American, but I am an American”. The dissociation between these two levels could be taken as indirect evidence for the idea that members of ethnic minorities find ways to protect themselves from prevalent stereotypical assumptions about their group, despite the fact that they hold these stereotypical beliefs to some extent.

This being said, the propensity to be denied the American identity is not a trivial experience. Members of ethnic minorities are relatively aware that they are stereotyped as perpetual foreigners (Armenta et al., 2013; Barlow, Taylor, & Lambert, 2000; Cheryan & Monin, 2005; Huynh, Devos, & Smalarz, 2011). It is a common experience for them to have to contend with reactions implying or assuming that they are not American (Liang, Li, & Kim, 2004; Sue, Bucceri, Lin, Nadal, & Torino, 2007). The question “Where are you from?”, more routinely asked to non-White Americans than to White Americans, is often based on the assumption that the interlocutor is “not from here” and conveys that this person is not truly American. These common experiences not only result in an awareness of being denied the American identity but also account for a weakened sense of belonging to America, a perceived conflict between ethnic and national attachments, and lower well-being or psychological adjustments (Armenta et al., 2013; Huynh et al., 2011). Negative responses to identity denial are more likely to be displayed by individuals who have been reared in the USA for an extended period of time (e.g., second-generation rather than first-generation immigrants) (Wang, Minervino, & Cheryan, 2012). When their national identity is questioned, members of ethnic minorities may actively reassert their Americanness (Cheryan & Monin, 2005).

In addition to showing that implicit ethnic–American associations are linked to implicit self-definitions, several investigators have examined the extent to which they delineate opportunities and rights for members of various ethnic groups. For example, Yogeeswaran and Dasgupta (2010) tested whether implicit prototypes about who is authentically American predicted discriminatory behavior and judgments against Americans of non-European descent. As expected, the more participants held implicit beliefs that the prototypical American is White, the less willing they were to hire qualified Asian Americans in national security jobs. Interestingly, this relation did not hold in identical corporate jobs where national security was irrelevant. In a different domain, the propensity to harbor a race-based national prototype accounted for a more negative evaluation of an immigration policy when it was proposed by an Asian American, but not a White American, policy writer.

Research carried out in the context of recent presidential elections documents the importance of perceived Americanness in the political arena. For example, a longitudinal study showed that anti-egalitarian voters opposed Obama in part because of his perceived foreignness (Knowles, Lowery, & Schaumberg, 2009). This finding was based on self-reported perceptions and attitudes, but similar conclusions can be drawn from implicit assessments. There is now growing evidence that voting behavior or intentions are linked to implicit evaluations of political candidates or racial groups (Devos, 2011; Nosek, Graham, & Hawkins, 2010). In both 2008 and 2012, the extent to which the American identity was implicitly linked to candidates to the nation’s highest office predicted behavioral intentions toward them (Devos & Ma, 2013; Ma & Devos, 2013; Sheets, Domke, & Greenwald, 2011). Controlling for the effect of political orientation, individuals who had relative difficulty associating the American identity with Barack Obama were those who were less likely to vote for him or to actively support him rather than his opponent.

Conclusion

Defining what it means to be American inherently implies delineating the boundaries of the national identity. At the explicit level, various lay definitions of the American identity coexist, and ideals of inclusion and equality are strongly endorsed. In contrast, responses that are not under conscious control drastically deviate from these principles. The picture emerging from implicit assessments reveals a more exclusionary perspective. In the contemporary USA, more often than not, ethnic minorities are implicitly excluded from the national identity compared with European Americans. Even as the diversity of the US population increases and powerful symbols (e.g., an African American family living in the White House) point toward racial progress, many Americans continue to display, at the implicit level, an ethnocultural conception of the American identity aligned with the domination of White Americans. The enduring connection between the White racial identity and the American identity operates within the context of a broader ideological system as one of the mechanisms through which asymmetrical relations between ethnic groups are perpetuated. Large segments of the US population are excluded from being part of the nation state.

At the same time, perspectives on the American identity are not monolithic. Even at the implicit level, there is great variability and flexibility in the ascription of the American identity. The ease or difficulty to associate the American identity with various ethnic groups is a function of perceivers’ characteristics (e.g., ethnic identity or political orientation). The broader socio-cultural context and immediate aspects of the situation also affect the relative inclusion or exclusion of ethnic groups from the national identity. In some cases, these variations might cast doubts on the prospect of a pluralistic and multicultural nation. At first sight, the fact that emphasizing ethnic identities exacerbates the American=White effect and, conversely, that downplaying ethnic identities makes it easier to grant the American identity to non-White Americans could be seen as evidence for the virtue of color-blind or assimilationist perspectives. In our opinion, hastily reaching such conclusion might not be warranted. At least in some contexts, recognizing ethnic distinctions might not be antinomic with an inclusive national identity.

In contrast to claims of a post-racial America, ethnic distinctions continue to delineate the rights and opportunities of ethnic groups making up the US population. The present research reiterates that behind ethnic distinctions lurk asymmetries of power and status. Automatic and largely unconscious assumptions about what it means to be American imply that the national identity is less easily psychologically granted to ethnic minorities than to Americans of European descent. The unequal access to a national identity is a specific manifestation of a broader socio-structural reality.

Biographies

Thierry Devos is a Professor in the Department of Psychology at San Diego State University. He received his PhD in social sciences from the University of Lausanne, Switzerland. His research focuses on how social identities operate outside of conscious awareness and control. Specifically, he is interested in the implicit interconnections between ethnic and national identities, and in factors affecting the implicit academic self-concept of underrepresented groups. He has published more than 50 peer-reviewed articles or contributions to edited volumes. His research has been funded by the National Institute of Mental Health and the Swiss National Science Foundation.

Hafsa Mohamed is a Master’s student in the Department of Psychology at San Diego State University. Her current work focuses on the relationships between environmental complexity and implicit intergroup biases. In particular, she is interested in demonstrating the links between contextual diversity and implicit ethnic–American associations. On a broader level, her research interests include investigating the socio-cognitive processes that underlie attitudes, beliefs, and behavior related to diversity, ethnicity/race, gender, and multiculturalism, with an emphasis on how these processes can promote positive intergroup relations. She has presented her research at several local and national conferences. Her research has been funded by NIH-supported programs such as the Bridges to the Baccalaureate and the Minority Biomedical Research Support.