This is an open access article distributed under the terms of the Creative Commons
Attribution License (Each decade, the United States reimagines how to racially and ethnically classify its citizens. It is probably safe to assume that U.S. residents are classifying themselves on the basis of their “color or race” more than every ten years given the frequency with which other institutions, including education and workplace settings, collect such information. An overview of the country’s ever evolving classifications—the inclusion in 1870 of a separate “Chinese” description, the inclusion of Spanish/Hispanic origin beginning in 1980, allowing for the option to self-identify with the 2000 census—all reveal the constant tension inherent in reifying a concept mixed with historical, social, and physical features (see Begley, 2016).
The case of Arabs/Arab Americans offers a timely illumination on how the subjective meaning imbued in the common, seemingly predictable task of racial self-categorization leads to potential paradoxes. On standard demographic forms, often replicating the U.S. Census, citizens with Arab heritage have a clear designation: White racial background includes any person “having origins in any of the original peoples of Europe, the Middle East, or North Africa” (United States Census Bureau, 2018). Thus, a seventh-generation American with roots in what are today Germany and Sweden, a first-generation immigrant from Lebanon, and a second-generation Algerian American would all be categorized as White1. An apparent tension emerges—are people of European and Middle Eastern/North African descent actually similarly positioned and perceived in the American racial/ethnic landscape? Do descendants of both groups inherit the same advantages or histories of migration?
Data suggest that Americans of either European or Middle Eastern descent are not treated as interchangeable members of a single racial/ethnic group. It could be argued that particular forms of perceived exclusion are partially defining of being Arab in the U.S. In the early 2000s, residents of Arab descent (and those mistaken as Arabs such as Iranians or South Asians) experienced especially harsh treatment in the years following 9/11, though this treatment was only an exaggeration of historical discrimination (Cainkar, 2009; Huddy, Khatib, & Capelos, 2002). A recent audit study found Arabs received 40 percent fewer replies after responding to roommate-wanted advertisements compared to Whites (Gaddis & Ghoshal, 2015). How do Arabs themselves thus construct a racial/ethnic identity given their legal inclusion as ‘White’ with residents of European descent? This question is particularly salient given the evidence of malleability in the definition of racial groups. Historically, some marginalized groups (e.g. Irish and Jewish immigrants) could eventually be embraced as White (Roediger, 2017). Would Americans of Arab heritage today embrace membership into the White racial category? To these questions, we apply a social identity approach to understanding fluid racial categories and examine the ways that Arab Americans respond to the tenuous meaning of their racial/ethnic categorization.
A Social Identity Approach to Race Labels and Categorization
The social identity approach (SIA) posits that belonging to and identifying with groups are fundamental pillars in the development of a social self (Tajfel & Turner, 1986; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987). People intuitively understand themselves on the basis of meaningful differences in outcomes that occur based on shared characteristics that are foregrounded by particular social ecologies. Anthropologists point out that differences in geographic origin and phenotype are not named and established as the basis of ‘race’ until a meaningful (re-)organization of social life necessitates it. The social identity approach can thus explain why members of separate tribes may have sparred for resources on the African continent but would later, reflecting the intergroup context, see themselves as members of a single group in resisting against the newly constructed White American (Fields & Fields, 2012). Importantly, such a group would exist not only practically, but socially as well, such that they might even begin to understand themselves as part of a new black racial group. Evidence for this process occurring recently was shown among college students from 32 different countries, for whom perceived discrimination predicted identification with other international students, but not with their national group (Schmitt, Spears, & Branscombe, 2003).
Later work with the SIA established that identifying with important groups not only reflects social reality but also serves as a meaningful source of self. A vast literature has established that identifying with meaningful groups is good for people. For one, identifying with groups is a source of self-esteem (Jetten et al., 2015). Recent research has shown that, in addition to making people feel good, identifying with groups provides people with a sense of control in their lives (Greenaway et al., 2015).
The implications of identification for well-being are especially meaningful in the case of racial-ethnic minorities, whose groups are often stigmatized. In the sample of international students, the increased identification with other international students positively predicted self-esteem (Schmitt, Spears, & Branscombe, 2003). This work underpinning the rejection-identification model shows that experiencing discrimination promotes identification with the targeted group, and this increased identification buffers some of the harm of experiencing discrimination (Branscombe, Schmitt, & Harvey, 1999). This model replicated in a representative sample of Muslim and Christian Arab Americans (Hakim, Molina, & Branscombe, 2018).
Self-categorization builds from the initial premises of social identity to argue that in addition to producing the groups themselves as a basis for self-construction, the social context also influences with which of multiple groups an individual will self-categorize. This literature has been especially productive and relevant in understanding the ways in which people shift between subordinate and superordinate categorizations. A relevant example is nation (superordinate) and race/ethnicity (subordinate). Self-categorization theory argues that these overlapping identities are variably salient depending on the social context. Generally, discrimination would lead to stronger identification with the targeted group racial/ethnic subgroup, whereas participating in a national election would lead to stronger identification with the superordinate group.
Many modern nation-states with histories of immigration manage cultural diversity by affirming the existence of different racial-ethnic subgroups as equally important to the multicultural environment (Sears, Henry, & Kosterman, 2000). Subgroup respect captures this perception among ethnic/racial minorities, that their subgroup is affirmed or valued within broader society (Huo & Molina, 2006). According to this perspective, acknowledgment and appreciation of subgroups, rather than erasure, leads to stronger identification with the common group (such as a school or a country) among subgroup members (Hornsey & Hogg, 2000). This acknowledgement is communicated through interpersonal interactions as well as institutional patterns, of which the census form is a primary example.
Put differently, relevance of the social identity approach to the present research rests on the assumption that the act of racial labeling is determined by and expressed within social contexts, and thus varies depending on socially relevant factors. Combined with the fact that such identification is a source of well-being and resilience among marginalized groups, we argue that the mere act of self-categorization on a form has potentially important consequences for Arab Americans, whose proper location in such labeling forms is rendered ambiguous when considering the disparate outcomes Americans of European and Arab heritage generally experience.
One important hypothesis derived from self-categorization theory is that the practical ambiguity of Arab racial-ethnic location is interpreted based on perceptions of fit. Social identity theorists articulate two forms of fit in the categorization process, comparative and normative (Haslam et al., 2010). Comparative fit pertains to the outcome of a perceptual process called meta-contrast, whereby one is more likely to categorize as part of a group if they see less differences within said group than they see differences between said group and another group (Turner, 1985). For instance, a person of multiracial background would be more likely to identify as multiracial (and not with a single specific racial group) when the amount of similarity they share in some domain—e.g., the experience of being ambiguous in the eyes of perceivers—is greater within the multiracial category than between distinct racial categories. Thus, intergroup contexts will promote identification with groups more than intragroup contexts will (Haslam et al., 2010).
Perhaps more relevant to the present research on Arab-White distinctions is normative fit. Normative fit refers to the content that defines the group (Haslam et al., 2010). That is, beyond the existence of difference between groups, normative fit refers to the nature of such differences, from which people build shared expectations of what characteristics a person must hold to be part of a certain category. More specifically applied to the present case, to the extent that Arabs perceive themselves and are perceived as violating a normative understanding of what it means to be White, they should self-categorize as part of a separate group (Turner, Oakes, Haslam, & McGarty, 1994). Thus, the central hypothesis underlying the present work is that the more evidence from an Arab American’s experience that is incompatible with the content of White racial identity—e.g., lighter complexion or experiences afforded by White privilege—the more likely they are to categorize as something else.
Context and Racial Categorization
Understanding the ambiguity of Arab American racial/ethnic meaning benefits from examination of similar categorization and labeling processes of other groups. Latino identity shares many parallels to Arab American identity in that both can be legally classified as White yet often practically live as part of a racialized non-White group. Indeed, the U.S. Census (as of 2010) explicitly considers Hispanic, Latino, or Spanish origin as an ethnicity which can co-occur with whiteness. Several predictors of identifying as a minority or as not White among Latino participants include darker skin (Vargas, 2015; Wilton, Sanchez, & Chavez, 2013), greater Spanish fluency (Sanchez, Chavez, Good, & Wilton, 2012; Wilton, Sanchez, et al., 2013) and higher socioeconomic status (Vargas, 2015).
More research exists on the predictors of racial self-categorization among multiracial Americans. Much of this research generalizes the findings among Latino participants. For instance, darker skin predicts categorization as Black among Black-White biracials (Brunsma & Rockquemore, 2001) and among all multiracial Americans, categorization as White correlates positively with social class (Davenport, 2016). This literature has also demonstrated the role of ecological factors in predicting racial categorization. For instance, in one study, multiracial adolescents were more likely to write in as “Other” if they attended a predominantly minority-serving school (Harris, Ravert, & Sullivan, 2017).
While the results from the reviewed primary literature above is rooted in the social identity approach to varying degrees, in concert they clearly point to one of the approach’s core themes: Identification with even seemingly static group categories like ‘race’ is responsive to variations in context. Next, a brief overview of the historical shifts in Arab American racial/ethnic positioning will situate the previous discussion on context and race categorization within this paper’s group of interest.
Tracing Arab American Race/Ethnicity in the U.S. Over Time
The earliest Arabic-speaking immigrants to the U.S. preferred legal categorization as White. Their efforts relied on cultural evidence that they could adequately perform whiteness in the U.S. For one, the majority of immigrants were Christian, providing a sense of shared heritage in their new host communities throughout the U.S. (Naff, 1985). In addition, in their legal struggle to naturalize as citizens, these immigrants pointed to their economic self-sufficiency and origins in the purported birthplace of Western civilization (Gualtieri, 2009). Those among this first wave beginning at the turn of the 20th century were predominantly from the Mount Lebanon region, in what was then referred to as Syria (which today would encompass the countries of Syria, Lebanon, Jordan, and Israel/Palestine).
Racial/ethnic categorization of these arrivals was muddled by the fact that these lands were provinces in the Ottoman Empire, leading to their lumped grouping as “Turks” with various Arabs as well as with Greeks, Albanians, and Armenians (Naber, 2000). Eventually, they were referred to colloquially and in print media as “Syrians.” Importantly, however, while these immigrants maintained social and cultural connections to their Syrian heritage, the majority willingly preferred assimilation into mainstream White America (Suleiman & Abu-Laban, 1989).
Given the prevalence of legal racial discrimination, these Syrians sought to formally assert their whiteness through the courts. This was especially important given the restriction of citizenship. One South Carolina judge wrote in 1914 that though Syrians may be considered Caucasian given their geographic origins, they were not meant to be included among the group of free White persons to whom the privilege of citizenship was reserved, arguing that this designation referred only to those of European descent (Samhan, 1987). Eventually, by the 1950s, Arabs transitioned to legal recognition as White and psychologically self-categorized as such (Samhan, 1987; Suleiman & Abu-Laban, 1989).
This social and self-categorization of Arabs began to shift away from White with the second and third waves of Arab immigration. The second wave coincided with the end of the second World War, when Arabs previously living under Ottoman and then European rule began (European-prescribed) forms of self-governance. Thus, many of the Arab immigrants at this time brought with them stronger national attachments compared to their predecessors who brought more localized forms of identity (Naber, 2000). Many more Muslims also participated in this wave of immigration, compared to the predominantly Christian first wave.
Numerous scholars agree that identification as “Arab”—rather than with a specific nation—began with the third wave of immigration. With new legislation in the U.S. lifting limitation of geographic restrictions, these immigrants were much more diverse in their national and religious origins. These immigrants replaced the civic identification with the U.S. of their predecessors with a pan-Arab consciousness, especially in response to growing anti-Arab sentiment with the onset of the Arab-Israeli War of 1967. For these more recently arrived Arabs, the strong U.S. antagonism against the Arab side of the conflict summarized a hostile orientation that was manifest in media portrayals of Arabs as an enemy of the West. Similar to other groups responding to subordination, many Arabs thus adopted a strategic essentialism whereby the previously unwelcome conflation of their diversity by the public would be embraced and rearticulated for the sake of political struggle (Gualtieri, 2009; Spivak, 1988).
Thus, while Arab Americans originally fought for a place in the White racial category, today, Arab Americans find themselves in a society that ostensibly values racial/ethnic diversity but that in reality harms Arab Americans in many ways. Such harm has occurred through, for instance, pervasive social representations of Arabs as patriarchal violent extremists (Shaheen, 2003) and public support for extra security measures for Arabs and Muslims (Cainkar, 2009; Huddy, Khatib, & Capelos, 2002). From a SIA perspective, such evidence from an Arab American’s experience would indicate that they belong to a group distinct from that of a White American.
Indeed, one study among Arab Muslims in Metropolitan Chicago found that 63% believed that Arabs were not White (Cainkar, 2008). In the same study, the majority of interview participants made distinctions between their de jure and de facto racial identities. For instance, one participant lamented an employer’s position that he could not be considered for affirmative action because “they said you will be considered White. But of course in real life we are not” (Cainkar, 2008, p. 62). In general, the Arab Muslim participants translated their marginalization into psychological identification as non-White. The present research sets out to test the contextual and motivational factors determining the categorization of Arabs and Arab Americans at the edge of whiteness.
Study 1 – Predictors of Self-Categorization
The first study examined the factors influencing an Arab American’s choice of racial self-categorization as “White” or “Other”. Specifically, a cross-sectional approach using the regularly occurring racial categories provided by the Census in combination with several social-psychological factors would allow us to predict the likelihood that an Arab American would self-categorize as White or choose another category2. Based on the previous discussion on the role of normative fit, social context, and racial categorization, Study 1 tests the hypothesis that Arab Americans are more likely to self-categorize as non-White the more they identify with another group and the more they experience forms of racialized marginalization.
Method
Participants
We conducted secondary analyses on data obtained from a representative sample of 1,001 Arab Americans (57% female, ages 18 to 88, Mage = 43.64) living in the metropolitan area of Detroit, MI. The original purpose of the study, carried out by the Inter-university Consortium for Political and Social Research at the University of Michigan, was to examine Arab American experiences post-9/11 (see Baker et al., 2006). The original survey assessed variables relevant to the present investigation. Surveys were administered in-person3.
Measures
Perceptions of Discrimination
A composite measure included participants’ responses to a) five “yes or no” questions probing experiences with interpersonal discrimination specifically based on their racial/ethnic background (e.g. verbal insult or loss of employment) and b) one item measuring the perception the U.S. media is biased against Arab Americans (1 = biased in favor, 4 = biased against). Responses to these items were standardized and averaged.
Subgroup Respect
One item, “Arab Americans are not respected by broader American society” (1 = strongly disagree, 5 = strongly agree). This item was reverse-scored so that higher scores indicated more perceived respect towards Arab Americans.
Arab American Identification
One item, “I identify with other Arab Americans,” (1 = strongly disagree, 5 = strongly agree), assessed identification. This item very closely resembles a validated single-item measure of social identification (Postmes, Haslam, & Jans, 2013).
American Identification
Two items: “I feel at home in America” (1 = strongly disagree, 5 = strongly agree) and “I’m proud to be an American” (1 = not at all proud, 5 = very proud) were averaged, r = .26.
Complexion
An admittedly unusual measure, survey administrators rated each individual’s complexion on a scale from 1 (very light) to 5 (very dark).
Religion
Participants indicated their religious background, if any. The majority of the sample, consistent with the national Arab American population, was Christian (57%, coded as 0). Forty-two percent of the sample was Muslim (coded as 1). Fifteen participants were excluded because they identified with a third religion, or no religion at all.
Race
Participants self-categorized in terms of race using the racial categories designated by the U.S. Census with the important option to self-categorize as “Other” (coding strategy discussed below).
Demographics
We also took into account participant sex, age, and socioeconomic status.
Results
Our primary interest was in predicting the racial self-categorization of the participants. Table 1 below lists, in decreasing order, the frequency with which each racial category was chosen by the 1,001 participants.
Table 1
Frequencies of Racial Self-Categorization
| Racial self-categorization | N | % of total |
|---|---|---|
| White | 655 | 65.34 |
| Other | 294 | 29.37 |
| Asian | 43 | 4.30 |
| Black, African American, or Negro | 1 | .001 |
| American Indian or Alaska Native | 1 | .001 |
About two-thirds of the sample self-categorized in accordance with the U.S. Census definition of the White racial category. Conversely, about one-third of the sample chose to identify with some category other than that within which they are legally defined, mostly choosing a nonspecific “Other” category. What social psychological factors might be behind the different self-categorization choices, when everyone in the sample is legally part of the same race?
Predicting Likelihood to Identify as “Other”
Our primary analysis used the available, theoretically relevant, social psychological variables as predictors of the self-categorization decision. For conceptual and statistical clarity, we re-coded the self-categorization choices as either White or Other, thus collapsing all of the categories other than White into a single category to create a dichotomous outcome. Though there is surely valuable insight to be gained from understanding why someone would choose to identify as Asian (or African) rather than Other, there is not enough variability to quantitatively study these likelihoods. More importantly, conceptually, we see enough overlap between the non-White categories because what matters about the decision for the present study is not what particular categories the participants move towards but rather the choice at all to move from the White racial category.
With a dichotomous outcome, we submitted our important variables to a logistic regression whereby racial self-categorization (0 = White, 1 = Other) was regressed on the demographic variables, American identification, Arab American identification, perceptions of discrimination, complexion, and religious group membership. The results are presented in Table 2.
Table 2
Results of Logistic Regression Predicting Likelihood to Self-Categorize as Other (vs. White)
| Predictor | Estimate | SE | p | Odds |
|---|---|---|---|---|
| Age | -0.02 | 0.01 | 0.01 | .98 |
| Sex | -0.01 | 0.01 | 0.93 | 0.99 |
| American identification | -0.00 | 0.16 | 0.99 | 1.00 |
| Arab American identification | 0.19 | 0.11 | 0.08 | 1.21 |
| Discrimination | 0.30 | 0.14 | 0.04 | 1.35 |
| Complexion | 0.54 | 0.12 | < .001 | 1.72 |
| Religion (0 = Christian, 1 = Muslim) | 0.89 | 0.19 | < .001 | 2.46 |
Note. Odds column shows values indicating that likelihood that, with every unit increase of the variable and all else being equal, the participant would identify as Other.
The strongest predictor of increased likelihood to identify as Other was religious background, with Muslim Arab Americans being almost 2.5 times as likely as Christian Arab Americans to do so. Ethnic identification with other Arab Americans emerged as only a marginal predictor of increased likelihood, and American identification was unrelated to increased likelihood. In addition, more experiences of discrimination was positively associated with increased likelihood; for every unit increase in experiences of discrimination, participants were 1.35 times more likely to identify as “Other.” In addition, complexion was positively associated with increased likelihood; with every unit increase in skin tone darkness, participants were 1.72 times more likely to identify as Other. Age also emerged as a significant negative predictor or likelihood, indicating older participants were less likely to identify as “Other,” however the magnitude of the estimate and odds ratio are quite small as to be negligible.4
Discussion
This study measured racial self-categorization at the end of the survey with categories nearly identical to those used by the U.S. Census. Importantly, however, though the legal definition of White was available to participants, participants could also identify as “Other” and choose their own racial self-categorization. About one-third of the sample chose to identify as something other White (mostly as “Other,” but also some as Asian), providing variability to explore predictive factors in racial self-categorization.
In line with self-categorization hypothesis, the predictors that emerged as significant captured important aspects of lay perceptions of racial categories, social treatment, and symbolic belonging. One interpretation of experiences and perceptions of discrimination are as indicators of belonging and social position. In other words, White Americans of European heritage have historically wielded power in U.S. society and are not typically the victims of racial discrimination. That an Arab American would perceive and experience discrimination would, thus, be an indication that they are not truly a member of the White racial group, leading to increased likelihood to identify as “Other.” However, the present design cannot eliminate the reverse explanation that an Arab American’s inclination to identify as “Other” would promote greater awareness of discrimination.
Above and beyond experiences of discrimination, Muslims were significantly more likely than Christians to identify as Other. In this sample, Muslims reported higher levels of discrimination, and there may be other factors associated with Muslim heritage that lead to increased identification away from White. One explanation is historical, that Christian Arab Americans have a more established presence and an easier integration into broader U.S. society given their shared religious identification with the traditional normative religious heritage in the U.S. (Gualtieri, 2009). Relatedly, throughout their histories of immigration and especially at the time of data collection in post 9/11 America, Muslim Arab Americans were the targets of scrutiny that belied the public’s perception of this group as an outsider until proven otherwise (Detroit Arab American Study Team, 2009), which these participants may internalize and reflect onto the self-categorization task by indicating they do not at all feel they are considered White within their national context.
Of note is that the complexion variable was researcher-reported rather than self-reported, meaning that its significance as a predictor is not an indication of participants’ lay perceptions about prototypical racial phenotypes per se, but rather that perhaps another aspect of social experience beyond religious group membership and discrimination is not being captured among the present variables. It may simply be that, by Americans’ normative definition of whiteness, darker skin would lower the fit of an Arab target within the White category.
Study 2 – Effects of Self-Categorization
Study 2 varied the racial self-categorizations of Arab Americans by experimentally manipulating the available groups with which participants could identify. Instead of measuring how experiences predict the likelihood to racially self-categorize in different ways, the present study asks how the racial categorization of a group can influence those same perceptions about a group’s standing in society. Indeed, from a SCT perspective, self-categorization can dynamically shift such that the consequences of social categorization are also the predictors of categorizing in a certain way (Good, Chavez, & Sanchez, 2010).
More specifically, in Study 2 Arab American participants would be either prevented from identifying as anything but White or be given the opportunity to identify with a commonly proposed, more specific ethnic categorization. We were interested in testing whether the seemingly benign task of clicking a checkbox in a certain way (or simply its actual availability to do so) communicates to the participant something about how their group is perceived by society. Thus Study 2 tests the hypothesis that, to the extent that a demographic form is a manifestation of a broader public perception of the group’s standing, participants who could self-identify as Arab rather than White would report higher levels of subgroup respect and American identification. Conversely, participants who do not identify with White Americans yet must indicate their belonging to said group would be more likely to later express exaggerated levels of Arab identification and perceptions of discrimination. Study 2 also tests the possibility that such meanings conveyed by a demographic form may be more meaningful to participants based on social identification with the national group which the different racial/ethnic groups share.
Importantly, unlike in Study 1, perceptions of discrimination and subgroup respect would be measured independently. We presumed that a demographic form would for participants present an indication that their group is acknowledged as part of the multiethnic U.S. society, yet its absence would not be an instance of explicit negation or discrimination. We thus proposed it would be important to isolate these perceptions about the group’s place in society.
Method
Participants
Participants were undergraduate and graduate students recruited from Arab/Arab American student organizations at a large Midwestern university. After initial outreach and approval from group leaders, participants were encouraged to invite other students who would be eligible to participate. The final sample included 90 participants (Mage = 20.25), of whom 52.22% identified as women or female and 51.11% as coming from a Muslim background.
Ethnic Identity Manipulation
To mirror typical demographic questionnaires, and to avoid arousing suspicion with a stand-alone race/ethnicity question, the first page of the online survey asked participants to indicate their citizenship and then their race/ethnicity. After clicking an option to indicate citizenship, participants were randomly assigned to self-categorize using one of two sets of response options. In the White categorization condition, participants chose from the six racial groups recognized on the 2010 U.S. Census (presented in alphabetical order): American Indian or Alaska Native; Asian; Black or African American; Hispanic/Latino; Native Hawaiian or Pacific Islander; or White. In the MENA categorization condition, participants responded to the same list, with the addition of a Middle Eastern or North African option in the appropriate alphabetical position.
Measures
The remainder of the experiment was identical for participants in both conditions. For all dependent measures, participants responded to scales ranging from 1 (Strongly Disagree) to 7 (Strongly Agree).
Subgroup Respect
Three items assessed the extent to which participants felt that Americans respected Arabs as a group: “Most Americans respect what Arab Americans have accomplished in life,” “Most Americans value the opinions and ideas of most Arab Americans,” “Most Americans approve of how most Arab Americans live their lives,” α = .85.
American Identification
Three items assessed the extent to which participants identified with other Americans: “I identify with Americans,” “Being an American is an important part of how I see myself,” “I am glad to be an American,” α = .82.
Arab Identification
Three items, identical to those used to measure American identification, assessed the extent to which participants identified with other Arab Americans, α = .93.
Perceptions of Discrimination
Two items assessed the extent to which participants perceived Arab Americans as targets of discrimination in the U.S.: “Arab Americans as a group have been the victims of racism in broader Americans society,” “Prejudice and discrimination against Arab Americans exists today,” α = .91.
Resource Allocation
In the final measure, participants indicated how they would apportion a pool of compensation funds. Before the study began, during participant recruitment, the researchers and organization leaders had proposed two forms of compensation: money to support the efforts of the Arab Students’ Association on campus, and money for two gift cards for two randomly selected participants. To give them voice in this allocation, and to assess group-oriented resource allocation, participants used two sliding scales to divide the total sum between these forms of compensation. We operationalized group-based allocation as the proportion of funds allocated for the student organization.
Post-Experiment Open-Ended Responses
Following the primary dependent measures, two subsequent pages of the survey probed for participant-constructed, written reactions to the study and to the categorization of Arabs as White more generally. Whereas participants were naïve to the study’s focus on Arab categorization as White during the experiment, these follow-up items directly alluded to this phenomenon.
The first item referred participants to the race/ethnicity demographic options at the very beginning of the experiment (without alluding to the fact that half of all participants were randomly assigned to have a MENA option). Participants reported a) whether they typically check off as White or as ‘Other’ on forms that explicitly categorize Arabs as White, and b) their motivations for self-categorizing in that way.
The second page of items included two Likert statements and an open-ended questioning pertaining to subgroup respect. The two Likert items asked participants to rate the extent to which they agreed with statements expressing resistance to the categorization of Arabs as White. The first statement read “It bothers me that Arabs do not have a separate way of identifying on some forms other than 'White,’” and the second item read “Even if a form defines White identity as "having origins in any of the original peoples of Europe, the Middle East, or North Africa," I would try to find another way to identify.”
Results
We conducted a multivariate analysis of variance (MANOVA) with the categorization manipulation as the independent variable, and with subgroup respect, American identification, Arab American identification, perceptions of discrimination, and resource allocation as the dependent variables. The multivariate analysis showed no significant effect of the manipulation on any of the dependent variables, Wilks' λ = 0.98, F(5, 82) = .21, p = .93. Table 3 presents the means and correlations between all the variables.
Table 3
Means, Standard Deviations, and Correlations Between Variables in Study 1
| Variable | Correlations
|
||||||
|---|---|---|---|---|---|---|---|
| M | SD | 1 | 2 | 3 | 4 | 5 | |
| 1. Subgroup respect | 2.75 | 1.20 | – | ||||
| 2. American identification | 4.81 | 1.41 | 0.32* | – | |||
| 3. Arab American identification | 6.33 | 1.03 | 0.12 | .31** | – | ||
| 4. Perceptions of discrimination | 6.28 | 1.00 | -0.14 | -0.05 | 0.30** | – | |
| 5. Group-based allocation | 69.85 | 28.00 | -0.08 | 0.09 | 0.08 | 0.04 | – |
*p < .05. **p < .01.
We observed a ceiling effect for our measure of Arab American identification, and a paired-samples t-test showed that participants identified stronger with Arab Americans than with Americans as a national group, t(88) = 9.80, p < .001.
Since we observed no main effect consistent with the primary hypothesis—that an inclusive census form communicates greater respect—we next tested for the possibility of an interaction in predicting subgroup respect. Specifically, given the significant positive correlation between American identification and subgroup respect, we tested if the effect of the manipulation was significant for participants who highly identified as Americans. It may be the case that the effect of the manipulation depends on the meaningfulness of subgroup racial categorization process, which may vary depending on superordinate identification. We thus conducted multiple regression analyses with the categorization manipulation, American identification, and their interaction as simultaneous predictors of subgroup respect.
Consistent with the correlational analyses, the main effect of American identification was significant, β = .43, SE = .11, p < .001, and this effect was qualified by the Categorization manipulation × American identification interaction, β = -.37, SE = .17, p = .03. Simple slope analyses indicated that among participants in the White categorization condition, American identification did not predict subgroup respect, β = .06, SE = .12, p = .65. Conversely, for participants in the MENA categorization condition, greater American identification predicted perceived subgroup respect, β = .43, SE = .12, p < .001 (see Figure 1).
Figure 1
The Interactive Effect Between American Identification and the Categorization Manipulation on Subgroup Respect
Open-Ended Responses
The vast majority of participants, when directly probed, rejected their institutional classification as White. Participants were generally bothered by not having a separate way of identifying, M = 6.12, SD = 1.26; a single-sample t-test showed this average to be significantly higher than the scale midpoint of 4, t(75) = 14.60, p < .001, d = 1.68. Participants also reported trying to find other ways of identifying on a form, even if they were explicitly categorized as White, M = 5.36, SD = 1.85; a single-sample t-test showed this average to be significantly higher than the midpoint of 4, t(75) = 6.38, p < .001, d = .73.
Of the 59 participants who directly responded to the question of how they typically identified on forms that categorize Arabs as White, 68% (40 out of 59) reported categorizing themselves as “Other,” or another non-White category. For instance, a participant with Egyptian heritage reported that when they were young they would choose “African American,” before developing an understanding that “just because Egypt is in Africa, I don’t have the right to mark such. I hold [sic] mark other.” Another participant noted that one time they “filled out Asian, because I thought maybe I’m “West Asian” since my family is originally from Saudi Arabia.”5
The open-ended written responses revealed important themes contextualizing the participants’ choices to accept or reject Arab classification as White. To interpret qualitative responses, emergent themes were identified in an inductive manner to construct a framework through which the corpus of responses could be understood and categorized (Strauss & Corbin, 1997). These motivational themes were constructed based on choices to identify either as White or as Other. Sixteen out of 59 participants simply stated whether they typically categorize as White or Other without providing any motivational context. The themes from the remaining responses included 1) Arabness as a non-White culture and identity; 2) Whiteness as a privileged social position; and 3) Acknowledging legal classification. The coding scheme allowed for responses to corroborate more than one theme.
Eighteen participants grounded their choice to identify as Other by citing Arabness as a non-White culture and identity. This theme incorporated different aspects of heritage, representation, identity, diversity, culture, tradition, or simply in terms of “who I actually am” (P24). Some of these responses were couched in objectivity, as in noting that White is usually written as “white/Caucasian,” which I am not. “Other,” unfortunately described me better” (P16). Often this objectivity would note the ostensibly obvious non-White defining features of Arabness:
If "other" is in the list of options, I will certainly check it off. I do so because I do not believe I am white, despite Arabs often being grouped as such. I do not mean it with a negative implication, but rather simply as it is clear that Arabs have their own culture and identity and to be grouped as white seems to not acknowledge the Arab culture. (P71)
In other instances, to identify as White would negate a feeling of Arab pride and identification with other Arabs. For one, it would not make sense given that their family is “very Middle Eastern” (P74).
Thirteen participants referenced how their Arab heritage was incompatible with Whiteness as a privileged social position. These comments referenced various perceptions of how Arabs are viewed in the “imaginations” of White America, and made explicit reference to privileging/marginalizing social phenomena such as White privilege, advantage, acceptance, discriminatory treatment, and inclusion/exclusion. That is, for these respondents, Arabness could be defined not just as a distinct cultural group but rather as a position defined by its exclusion from White America. Several of these responses cited how it would not make sense to identify as White when most White Americans would not categorize Arabs as such. Beyond this appraisal of the normative construction of whiteness, several of these participants cited that their own experience did not feature the presumed advantages of White racial identity. Included within this theme were responses tracing differential treatment as a motivation to self-categorize as White, citing historical precedent and contemporary cause for concern:
I usually categorize myself as white. It seems safer. I would not include a MENA box on the census, because the government has a bad track record of abusing such information. If you look at historical precedent, you see how some German Americans were interned during World War I, how Japanese Americans were interned during World War II, and how Iranian students were subject to detainment or deportment in the era around the Iranian Hostage Crisis. Now, of course, under the Trump administration, we see continual attempts to infringe upon the civil liberties of Arab Americans. This will only become easier to do if we identify ourselves so easily for the government. In a better world, one with less anti-Arab racism and Islamophobia, I would support a MENA box with the hopes of gaining better political representation by showing that we ARE a significant population. However, I don't think it is a good idea in the current political environment. (P63)
In contrast to this form of intentional self-preservation, most other responses conceptualized the maintenance of their identity as not becoming “complacent,” (P73) “brainwashed,” (P28) or complicit in their own group’s “erasure” (P59).
Finally, eight participants described their motivations to categorize as White as a form of acknowledging legal categorization. Several of these responses explicitly referenced the Census designation or formal demographic questionnaires. Others referenced simply “going with the flow” (P46) given that the “intention is for Middle Eastern people to use that category” (P54).
Discussion
What does it mean for a group with a poorly-fitting racial/ethnic definition to be able to self-categorize in a more fitting way? On its own, the manipulation of alternative racial categorization options for Arab American participants did not influence perceptions of subgroup respect, American or Arab American identification, or perceptions of discrimination. However, the interaction effect is telling in its indication that the meaning of the possible categorization options is interpreted differently based on identification with the superordinate group. Specifically, the presence of the MENA category was especially indicative of subgroup respect for those Arabs who strongly identified as American; there was no such association for the participants only given the option to identify as White. Put differently, American identification among Arab Americans can be understood as the extent to which members of a subgroup want to and feel they can belong in a society that has egalitarian ambitions. And it is for these participants that the option to categorize as MENA is an indication that their group is respected by the broader society.
A compatible, additional interpretation can be offered with respect to the low-identified Americans expressing less subgroup respect. It may be that lower American identification overlaps with low Arab American identification (as indicated with the bivariate correlations), and thus the option to identify as MENA is not as meaningful. More tentatively, low American identification may belie a distrust of the racial categorization process, such that for these participants, the option to identify as MENA is especially unwelcome given its potential use for, among other purposes, governmental surveillance. For instance, though they are eventually categorized as White, respondents of Arab descent can write-in another racial/ethnic designation such as a nation of origin or simply ‘Arab.’ Interestingly, details gleaned from self-identified Arab respondents during the 2000 Census was shared in response to a request from the U.S. Department of Homeland Security (Clemetson, 2004).
General Discussion
Arab Americans have been characterized as “ambiguous insiders” (Naber, 2000) in the U.S. racial/ethnic imagination. They are, along with Americans of European descent, legally categorized as White. Few would dispute, however, that Americans of European and Arab descent receive qualitatively different treatment. The present studies sought to examine how Arab Americans decide whether they would like to be counted as White or something else. Importantly, the present studies sought to understand how the interplay of social experience and beliefs about groups interact to shape the psychosocial construction of group membership.
Results from the two studies generally supported the hypotheses that Arab American racial self-categorization depends upon psychological and experiential factors and not merely upon a recognition of how the state categorizes members of their group. Study 1 showed that Arabs are more likely to categorize as non-White (i.e., “Other”) if they experience discrimination, are Muslim, and have darker skin. Study 2 also demonstrated the affective component of the simple categorization task: Arab Americans who identified strongly with the superordinate American group perceived greater subgroup respect if they could self-categorize as MENA (vs. as White). The sample in Study 2 was limited by its recruitment of college students only, the results from which revealed a highly ethnically identified sample. This may also explain the discrepancy in the proportion of respondents choosing ‘White’ vs. ‘Other;’ the emerging adults in Study 2 may be in the midst of important identity-construction in concert with peers of similar second-generation backgrounds.
The methods and results speak to the potential of studying categorization processes by borrowing from institutional classifications (i.e., from the Census). That is, the studies deployed an everyday context assessing how people identify themselves on forms. Neither study drew extra attention to the act of categorization, only treating the categorization act as naturally occurring processes and analyzing self-categorization as an outcome (Study 1) or manipulation (Study 2).
We thus argue that this research contributes to the body of work that studies the reification of racial categories through familiar cultural products. Census forms do not merely tabulate where people live and what their backgrounds are; they are also affordances for people to “ask questions about themselves” (Cohn, 1987, p. 230), the answers to which promote social and individual definitions of race/ethnicity (Gualtieri, 2009). Nearly all participants in Study 2, regardless of their typical self-categorization in regular settings, spoke to some form of self-reflection, a way of asking questions about themselves that likely occurs with each instance of reporting their racial/ethnic background. Indeed, a growing qualitative literature illuminates the depth with which Arab Americans respondents reflect on the meaning of their race and ethnicity when asked about the categorization process (e.g., Cainkar, 2008).
The simple qualitative items in Study 2 produced themes consistent with existing literature showing Arab Americans partake in a unique form of racial/ethnic socialization via the Census categories. Just as discrimination was a strong predictor of categorizing as Other in Study 1, many Study 2 participants reported the basic illogic of identifying as White given their level of symbolic and practical marginalization. While the measure of discrimination in Study 1 incorporated practical forms of exclusion (hiring bias, slurs), Study 2 responses revealed the less tangible ways that social experience teaches Arabs that they are not White.
Altogether, the results point to the social and motivational factors underpinning (self-) categorization of Arabs as White, or not quite. Categorizing as part of a group is not simply a function of the available, institutionally recognized group labels; if it were, the self-categorization of Arabs would not interact with American identification or subjective experience. Instead, consistent with the social identity approach, categorizing as part of a group is a function of how the social world shapes lives on an intergroup basis, and people in turn understand themselves in these terms (Omi & Winant, 1994). In social psychological and historical terms, Arabs have, since their initial immigration, contested their racial categorization depending on the intergroup context. When they were mostly Christian and economically successful, Arabs in America fought for recognition as White. When they became increasingly Muslim and otherwise diverse, and in the wake of geopolitical events that fractured the harmony between their Arab and American selves, they learned of their Otherness and searched for new ways to racially/ethnically identify and categorize.
While the results may be used to reach different conclusions on the necessity of a MENA category in different settings, they uniformly point to the fact that, at the very least, Arab is a “lived race” (Harris & Sim, 2002) because people do understand themselves, others, and intergroup relations through this category, often in reference to other races, despite its regular absence from identification forms. As put by Jamal (2008):
Arabs neither are seen as white nor are they granted an officially defined minority status; rather, they stand outside all racial demarcations in an ambiguous, precarious position of Otherness compounded by existing policies and perceptions. Regardless, then, of the boxes Arabs check—whether white, black, or other—their racialization, which has resulted in a perception of Otherness, is real.
Future Directions
One aspect of the research in particular that deserves future consideration is religious group membership. Only Study 1 directly considered the role of religion and incorporated religious affiliation in analyses; Study 2 did not recruit a religiously diverse sample. Given the historical and contemporary trends that differentiate the experience of Muslim and Christian Arab Americans, further work should consider the way these group memberships influence Arab American self-understanding as well as considerations of Arab American categorization by White Americans and broader society. For instance, Muslim Arab Americans perceive more discrimination, and the relationship between discrimination and identification with their religious group is stronger than for Christian Arab Americans (Hakim, Molina, & Branscombe, 2018), The inclusion of religion in further research would be especially useful in merging the visible and symbolic components that we typically use in defining race, which further intersects with gender (since, for instance, some Muslim women wear a distinctive head covering and may be more visible and thus more likely to be racialized as non-White). Consistent with the present results, we would predict that veils (or other religiously inspired garments) would, like complexion, serve as a visual marker of non-White racialization and may lead to more experiences of perceptions of discrimination, which would further predict categorization as non-White by the targets and perceivers.
One straightforward next step would be to treat categorization of Arab Americans as a categorical outcome for European American participants. Only Study 1 among Arab Americans treated categorization as an outcome in this way, and further studies in this vein would also fall in line with methodological approaches in studies of the categorization of multiracial people. Such an outcome would invite the possibility for further moderators or mediators. For instance, an internal motivation to control prejudice was associated with higher likelihood to categorize Black-White multiracial people as multiracial (instead of as Black; Chen, Moons, Gaither, Hamilton, & Sherman, 2014). Another benefit of treating categorization as an outcome would allow for examining the role of perceived similarity as a potential mediator (as in Good, Chavez, & Sanchez, 2010).
Conclusion
Arab Americans occupy a peculiar in-betweenness in terms of race/ethnicity in the U.S. (Gualtieri, 2009). Arabs are, along with residents of European descent, legally categorized as White, yet many facets of their experience and social representations construct the group as an “Other.” Across two studies, Arab Americans confirmed themselves as an improper fit to the normative understanding of White.
Racial categorization is a sociocultural process that is not restricted to individual cognition. Because race is “a concept which signifies and symbolizes social conflicts and interests by referring to different types of human bodies” (Omi & Winant, 1994, p. 55), membership in these groups is fundamentally contingent upon ongoing, and shifting, intergroup relations. Despite some Arabs’ physical appearances allowing them to pass as White, Arabs may still currently lie outside imaginations of whiteness because rather than merely referencing purported differences in phenotypes, the racialization of Arabs occurs through socio-historically contingent intergroup dynamics.