Whom do White Americans consider to be White?Footnote 1 How White Americans draw boundaries of Whiteness is key to understanding the social and political positions of groups who reside at the cusp of Whiteness. In the United States, Whites operate as the dominant ethnoracial group and, as such, are a privileged identity group.Footnote 2 Whiteness bequeaths psychological and material benefits to those who fall into the category (Galonnier Reference Galonnier2015b). These benefits include but are not limited to better access to housing and credit (e.g., Thurston Reference Thurston2018), medical care (e.g., Hoberman Reference Hoberman2012), employment opportunities (e.g., Kalleberg, Reskin, and Hudson Reference Kalleberg, Reskin and Hudson2000), education (e.g., Farkas Reference Farkas2003), and even clean air (e.g., Ash and Boyce Reference Ash and Boyce2018). W. E. B. Du Bois (Reference Du Bois1935, 700), for example, refers to the “public and psychological wage” that Whites receive: “They were given public deference and titles of courtesy because they were [W]hite.” But at the center of the “wages of Whiteness” is the question, “Who is White?” The boundaries between those who are considered White, and those who are not, ensure that power and privilege are only reserved for those deemed White enough. White Americans themselves are key practitioners of inclusion into and exclusion out of Whiteness.
Indeed, to receive wages of Whiteness, individuals must be viewed as White not only by non-Whites but also by other Whites. Many of the benefits that come from being White depend on the social, economic, and political decisions of others and whether they perceive an individual as White. This is true in the case of credit and loans, medical care, hiring, regulatory decisions, and much more. Because Whiteness as a perceived social construct is ambiguous, understanding how people arrive at those categorizations is vital for understanding both the privileges and marginalization of those whose ethnoracial identities are unclear.
Racialization is one lens from which we can study how Whites operationalize Whiteness. This process occurs when social and political meaning is attributed to individuals based on traits, such as religion or region of origin and then, based on those attributions, individuals are assigned to a general category (Miles Reference Miles2004, 102). Scholars have written at length on the racialization of religious subjects, including the racialization of Muslims (Al-Saji Reference Al-Saji and Lee2014; Aziz Reference Aziz2022; Bayoumi Reference Bayoumi2006; Beydoun Reference Beydoun2013; Considine Reference Considine2017; Fourlas Reference Fourlas2015; Galonnier Reference Galonnier2015a; Garner and Selod Reference Garner and Selod2015; Jamal and Sinno Reference Jamal and S.H. Sinno2009; Meer Reference Meer2013) and those from the Middle East or North Africa (MENA; Ajrouch Reference Ajrouch2005; Ajrouch and Jamal Reference Ajrouch and Jamal2007; Awad, Hashem, and Nguyen Reference Awad, Hashem and Nguyen2021; Beydoun Reference Beydoun2013; Cainkar Reference Cainkar2009; Gualtieri Reference Gualtieri2009; Maghbouleh Reference Maghbouleh2017; Naber Reference Naber2000; Reference Naber2012). This article builds on this literature by empirically demonstrating, through a survey experiment, the relationship between religion and region and inclusion into Whiteness using the cases of Muslim and MENA identities. In this experiment, I show that White Americans use both country of origin and religious cues when operationalizing whom they consider White.
The case of MENA individuals is intriguing because they have been legally classified as White in the United States since 1909 (and codified as such in 1977). Yet, because of the racialization of MENA individuals and of Muslims (Bayoumi Reference Bayoumi2006; Beydoun Reference Beydoun2013; Lajevardi Reference Lajevardi2020; Maghbouleh Reference Maghbouleh2017; Maghbouleh, Schachter, and Flores Reference Maghbouleh, Schachter and Flores2022; Naber Reference Naber2000)—an important but by no means the only religious group of MENA individuals—the White label may neither be suitable nor used by society at large. At the same time, Muslim identity can also be understood at the margins of Whiteness. Muslim identity is a religious identity, but it is often treated as an ethnoracial identity. This means that MENA Muslims and White converts to Islam are both “racially White” while also being members of a racialized religious group—and, for MENA Muslims, members of a racialized ethnic group as well. Thus, both identities sit at the margin of Whiteness, yet they do not benefit from the practice of Whiteness.Footnote 3 Indeed, MENA and Muslim individuals have disproportionally faced hate crimes, increased government surveillance and profiling, discrimination at school, and higher rates of COVID relative to non-MENA Whites. And without a distinct MENA label, it is difficult to research this group to further understand the marginalization its members face.
Using court cases and an empirical study, I show that religion plays an equally important role as one’s ancestral background in understanding the racialization of individuals both into and away from Whiteness. Being Muslim decreases the chance that White Americans will assign someone as White, as does being from a MENA country, all else equal. Specifically, European Muslims are less likely to be rated as White relative to their Christian counterparts. Interestingly, however, although Islam has been racialized, the Russian Muslim is more likely to be classified as White relative to the Iranian Muslim. This suggests that country of origin still plays a role in the racialization of Muslims. But region alone is not the sole determinant of Whiteness, because Christian Iranians are more likely to be assigned White than Muslim Iranians. This indicates that religion and region are considered constitutive traits of boundaries of Whiteness. Moreover, this boundary extends into perceptions of skin pigmentation. Not only are White Americans less likely to assign Muslims, regardless of country of origin, as White but they also perceive them to be darker than their Christian counterparts, even when their skin is not darker.
Framework and Hypotheses
The following sections discuss the racialization of Islam, the racialization of MENA individuals, and the historical role of religion on MENA classification as White in the United States. From this discussion, I develop hypotheses for the empirical study.
Racialization is “the extension of racial meaning to a previously racially unclassified relationship, social practice, or group” (Omi and Winant Reference Omi and Winant, H2015, 111). The process of racialization undergirds social and political systems structured around race: hierarchies based on race cannot exist without raced classes of people. Investigating the process of racing people, or racialization, allows us to examine racial power structures. Yet, although the definition provided by Omi and Winant is useful, it can be limited in describing new racialization processes or re-racialization of previously racially classified groups (Garner and Selod Reference Garner and Selod2015).Footnote 4 For instance, MENA and Muslim individuals are often discussed in the literature as being subject to racialization and “racism targeting Muslims” (12), respectively. Yet, many Muslim individuals already have a racial classification such as Asian, Black, or White Muslims. Moreover, MENA individuals, regardless of religion, also have a racial classification: White. Given these considerations, in the context of this article, racialization can also mean the extension of new racial meanings to a group. Religion plays a distinct role in understanding contemporary race and ethnic politics and Whiteness, yet the inclusion of religion to understand race/ethnicity is not new. Scholars have written, at length, on the racialization of religious subjects.
The Racialization of Islam
Scholars have studied the racialization of religious groups, with a recent focus on the racialization of Muslims (Al-Saji Reference Al-Saji and Lee2014; Aziz Reference Aziz2022; Bayoumi Reference Bayoumi2006; Beydoun Reference Beydoun2013; Considine Reference Considine2017; Fourlas Reference Fourlas2015; Galonnier Reference Galonnier2015a; Garner and Selod Reference Garner and Selod2015; Jamal and Sinno Reference Jamal and S.H. Sinno2009; Meer Reference Meer2013). When a given religion, such as Islam, is racialized, “the religious beliefs and practices of the adherents are associated with cultural traits, which in turn are surrogates for biological traits” (Aziz Reference Aziz2022, 20). Thus, Islam may no longer be seen as a religious practice protected under the First Amendment (Aziz Reference Aziz2022; Garner and Selod, Reference Garner and Selod2015; Gotanda Reference Gotanda2011; Reference Gotanda2017). Rather, religious identity is seen as an immutable trait.
Religion was not always understood in these terms, however. In pre-medieval Europe, Jews and Muslims were discriminated against, but they could be “purified” through conversion (Meer Reference Meer2013; Thomas Reference Thomas2010). Throughout the medieval era, however, the differences between Christians and non-Christians came to be societally understood such that even conversion could not remove the bloodline of Jews and Muslims (Bayoumi Reference Bayoumi2006; Goldstein Reference Goldstein2006; Jordan Reference Jordan1974, 51; Maryks Reference Maryks2010; Meer Reference Meer2013; Soyer Reference Soyer2013; Thomas Reference Thomas2010).Footnote 5 The Purity of Blood Statutes (pureza [limpieza] de sangre) were implemented in fifteenth-century Spain to discriminate against Jewish (conversos) and Muslims (moriscos) converts to Catholics, claiming that even though they had converted, they were still of impure blood (Maryks Reference Maryks2010). Coinciding with the belief that religion was an immutable characteristic were beliefs that physical attributes were associated with religious subjects. Jews and Muslims were considered “black” and diseased, whereas Christians were “white” and pure (Meer Reference Meer2013; Thomas Reference Thomas2010). The essentialization of religious individuals’ corporal characteristics lies at the foundation of the racialization of religion. As Meer (Reference Meer2013, 389) states, “The category of race was co-constituted with religion, and our resurrection of this genealogy implicates the formation of race in the racialization of religious subjects.” The consequences of the racialization of religious subjects may be seen today in how non-Christians—particularly Muslims and Jewish individuals—are discriminated against in ways that mirror racism (Aziz Reference Aziz2022; Beydoun Reference Beydoun2018; Desmond-Harris Reference Desmond-Harris2017; Garner and Selod Reference Garner and Selod2015, Reference Garner and Selod2015; Said Reference Said1979; Selod and Embrick Reference Selod and Embrick2013). The racialization of religion means that religion acts as a trait that is fixed and biologically reproduced (Aziz Reference Aziz2022; Omi and Winant Reference Omi and Winant, H2015).
The immutability of religion calls into question how Muslim individuals’ other ethnoracial identities intersect and interact with their racialized Muslim identity. That is, if Muslim identity is immutable, then being from another race should not change how that individual is viewed in society. This leads me to ask these questions: How do we understand the racialization of those who may be perceived to be Muslim but are not Muslim, such as non-Muslim MENA and South Asian individuals? And how do we understand the racial identities of non-Brown Muslims—such as White American converts to Islam?Footnote 6 These questions are important because they call into question how Whiteness, as the dominant identity within the racial hierarchy, is understood. The racialization of religion can offer insight. As Aziz (Reference Aziz2022, 5) notes, “Whiteness is shaped as much by religious identity as it is by skin color, hair texture, facial features, and other phenotypical characteristics.” This has been the case for Jewish, Muslim, and even Mormon Americans (Aziz Reference Aziz2022; Goldstein Reference Goldstein2006; Moshin and Crosby Reference Moshin and Crosby2018; Reeve Reference Reeve2015); this is in stark contrast to much of the narrative surrounding religious freedoms and religion as a choice in the United States. If religion is a choice, then does it signal race?
The objective of this article is not to determine whether religion influences understandings of Whiteness but rather to build on the literature that has done so to empirically demonstrate the relationship between religion and country of origin on inclusion into (or exclusion out of) Whiteness using the case of MENA and Muslim identities. For instance, White Americans may consider religion singularly; that is, anyone Christian is White and anyone Muslim is not. This article experimentally tests the extent to which White Americans operationalize the boundary of Whiteness using religious cues to test the boundaries of racialization vis-à-vis White assignment for Muslims.
The Racialization of MENA Individuals
Often, US media and Americans, in general, discuss Muslim and MENA identities interchangeably. But the racialization of MENA identity out of Whiteness differed from the racialization of Muslims. Although many MENA individuals were legally classified as White by the courts, as discussed in detail later, the process of non-MENA racialization began in the mid-twentieth century and continued into the twenty-first century. Before 1945, most Arab immigrants entering the United States were Christian (Awad Reference Awad2010; Cainkar Reference Cainkar2009; Naber Reference Naber2000). After World War II, more Muslim Arabs emigrated to the United States than Christian Arabs. This difference is important, because “Christians have more easily built communities around white American Christians” (Naber Reference Naber2000, 42). And as the court cases suggest, Christian Arabs were typically ruled by the court to be White and thus admissible for naturalization. Muslims, in contrast, have been seen as outsiders in mainstream American culture (Kalkan, Layman, and Uslaner Reference Kalkan, Layman and Uslaner2009; Lajevardi Reference Lajevardi2020; Naber Reference Naber2000, 42; Spruyt and Elchardus Reference Spruyt and Elchardus2012). Thus, Muslim identity is one reason for the racialization of MENA individuals out of Whiteness, but it is not the only one.
Much of the racialization of MENA individuals out of White identity can be linked to the geopolitical landscape between the United States and the Middle East. For example, during the Arab-Israeli War of 1967, the United States took an active stance in support of Israel. Because of the US support of Israel and Israel’s struggles with Palestinians and other Arab countries, even identifying as Arab could be equated in American culture as being “anti-Israeli” (Naber Reference Naber2000). And although most Israelis are also White, the regional conflicts between Arabs and Israelis came to be understood as an ethnoreligious conflict, not merely a political conflict between two different racially White groups. The geopolitical conflicts between the MENA region and the United States continued throughout the 1970s and 1980s, exacerbating the racialization of MENA individuals out of Whiteness. The 1973 oil crisis occurred when Arab members of OPEC raised the cost of oil in retaliation against the US support of Israel during the Yom Kippur War; it caused many Americans to vilify Arabs within the United States. However, this racialization was not reserved for Arabs alone. The Iranian Revolution, the subsequent 1979 oil shock, and the Iranian hostage crisis also played a role in racializing Iranians, Islam, and the Middle East in general. Media portrayals of MENA individuals during that time shifted to portray them as terrorists and damaging to US security (Naber Reference Naber2012, 37).
By the 1990s, polls showed that Americans viewed Arabs as “religious fanatics” (Cainkar Reference Cainkar2009). This not only reflected the racialization of MENA individuals but also obfuscated the line between their ethnoracial category and religious affiliation. The distinction between MENA and Muslim identities was further confounded in post–9/11 America. There were increases in hate crimes toward Muslims, MENA individuals, and those perceived to be either Muslim or MENA (Abdelkarim Reference Abdelkarim2003; Ibish Reference Ibish2001; Ibish and Stewart Reference Ibish and Stewart2003; Kaplan Reference Kaplan2006).
As mentioned, Muslims and MENA individuals are not overlapping categories: even though the processes of their racialization were largely simultaneous and intertwined, they also differed. And although the media and society at large may use the labels “MENA” and “Muslims” interchangeably, research shows that Americans do distinguish between MENA individuals, Muslims, and MENA Muslims (Adida, Lo, and Platas Reference Adida, Lo and Platas2019; Calfano, Lajevardi, and Michelson Reference Calfano, Lajevardi and Michelson2021; d’Urso and Bonilla, Reference d’Urso2022). Thus, this question remains: If White Americans think of religion as a choice (Aziz Reference Aziz2022) and country of origin as an immutable characteristic, do White Americans only consider place of origin when determining Whiteness, regardless of religion? As stated, this article experimentally tests the extent to which White Americans operationalize the boundary of Whiteness using country of origin cues to test the boundaries of racialization vis-à-vis White assignment for MENA individuals.
The racialization of MENA individuals out of Whiteness throughout the twentieth and twenty-first centuries was in part due to their proximity to Muslim identity. In the next section, I show that MENA individuals’ status as White has always been contingent on religion by providing an historical account of MENA identification, paying careful attention to the role of religion.
The Role of Religion in MENA Classification as White
As mentioned, MENA individuals were legally classified as White as early as 1909 and then officially in 1977. Although many MENA Americans today are advocating for a state-sanctioned MENA category, as opposed to White, this is in direct contrast to the active role that MENA individuals took to be classified as White in the early twentieth century. As more people emigrated to the United States, they wanted legal benefits that came with being classified as White, including eligibility for citizenship through naturalization. Thus, MENA individuals petitioned federal courts to establish their race as White after the Naturalization Act of 1870 extended naturalization only to Whites and Blacks.
Court cases establishing how individuals were to be classified ethnoracially are referred to as “racial prerequisite” cases. Being White was one of the necessary qualifications—or prerequisites—for naturalization. Most of the racial prerequisite cases heard focused on how East Asians, South Asians, Southeast Asians, and MENA individuals were to be legally classified in the United States. Here, I only discuss the court cases regarding MENA individuals. Scholarship on these cases mainly addresses the petitioner’s country of origin, but using information from the judges’ decisions, newspaper articles, and secondary sources, I was able to include the petitioner’s religion as well. Table 1 includes MENA-related racial prerequisite cases that address religion.Footnote 7 It indicates that for MENA individuals, Whiteness—and thus the benefit of naturalization—is tied to religion. In fact, in every case when MENA petitioners were Christian, they were ultimately given citizenship through Whiteness.Footnote 8
MENA Prerequisite Cases
The early rulings for MENA petitioners were consistent.Footnote 9 Until 1925, all the cases heard involving MENA petitioners ultimately ruled to grant citizenship via Whiteness; it should be noted that all the petitioners in these cases were Christian. In fact, the petitioners’ Christianity was frequently mentioned both in media reports about the case and in the judges’ decisions. For example, in the 1909 Halladjian case, Judge Lowell writes, “A reasonable modesty may well remind Europeans that the origin of their letters was in Phoenicia, the origin of much of their art in Egypt, … and that the Christian religion, which most Europeans believe to have influenced their civilization and ideals, was born in Palestine” (In re Halladjian 1909, 840). Judge Lowell continues stating that if Syrians are “excluded from naturalization… it is hard to find a loophole for admitting the Hebrew” (839). For Lowell, no logic could argue that Jewish individuals could be admitted while Syrian Christians were excluded. Indeed, this fits with the immigration patterns at the time, with Muslim MENA immigration only beginning by the late 1920s (Gualtieri Reference Gualtieri2009; Naber Reference Naber2012). Indeed, table 1 shows that two cases heard after 1925 —In re Feroz Din (1928) and In re Ahmed Hassan (1942)—ruled that these MENA individuals, both Muslim, were not White
In the case of Ahmed Hassan, Judge Tuttle explains that Hassan is dark in complexion, but he would not deny his claim for Whiteness based solely on skin pigmentation. Rather, he argues that Arabs are not White because they are Muslim: “Apart from the dark skin of the Arabs, it is well known that they are a part of the Mohammedan world and that a wide gulf separates their culture from that of the predominately Christian peoples of Europe. It cannot be expected that as a class they would readily intermarry with our population and be assimilated into our civilization” (In re Ahmed Hassan 1942).
The only case in table 1 in which a Muslim is deemed White and thus is granted citizenship is the final one: Ex parte Mohriez (1944). Mohriez was a Saudi Arabian Muslim who petitioned for naturalization via Whiteness. Just two years after the case of Ahmed Hassan, in which Judge Tuttle explicitly stated that he did not believe Arabs—especially Muslim Arabs—were White, Mohriez was determined to be White. Scholars have suggested that this decision, in large part, was influenced by geopolitical motives and the need for the United States to maintain a good relationship with Saudi Arabia and ARAMCO. It is not clear Mohriez would have been ruled White otherwise (Bayoumi Reference Bayoumi2006; Beydoun Reference Beydoun2018).Footnote 10 Thus, geopolitics also played a role in the racing of MENA individuals into Whiteness.
Although evidence from these court cases is limited, these cases provided precedent for future cases, and the rulings provide insight into the ethnoracial assignment decisions of the judges. Scholars have analyzed these cases to understand the history of racial classification of MENA individuals in the United States (Bayoumi Reference Bayoumi2006; Gualtieri Reference Gualtieri2009; Lopez Reference Lopez1997; Tehranian Reference Tehranian2000). I expand on their research by including the petitioner’s religion.
The analysis of the racial prerequisite court cases involving MENA petitioners, with careful attention to their religion, reveals the pattern of White Americans operationalizing Whiteness through religious cues. However, do Whites operationalize Whiteness in the same way today? Because Whiteness provides psychological and material benefits to those who are included, understanding who counts as White influences the status, livelihood, security, and much more of individuals who are otherwise understood, legally, as White.
Hypotheses
Literature on the racialization of Muslims and MENA individuals shows the process of how these groups have been pushed out of Whiteness, despite the US legal classification. Moreover, the court cases indicate that conferring Whiteness on immigrants from the MENA region was largely contingent on their religion. But in what ways do these features operate on how White Americans assign Whiteness to others? That is, literature on the racialization of these groups and the courts’ decisions suggests that White Americans consider country of origin and religious traits together to maintain the boundaries of Whiteness. But it is not clear which trait matters more or whether White Americans consider these two traits in conjunction to assign Whiteness to others.
The first consideration is country of origin. Many individuals use country of origin as a heuristic for racial assignment. For example, an individual from Kenya would typically be assumed to be Black, and an individual from China would typically be assumed to be Asian.Footnote 11 For MENA individuals, the legal designation, as discussed, would be White.
H1: Respondents will be less likely to assign an individual who is from a MENA country—as compared to a country in Europe—as White, all else constant.
Here, I construct my hypothesis relative to a European base group. This is because most of the literature on Whiteness and White identification focuses either implicitly or explicitly on Europeans.
Drawing on historical literature and analyses, religion should also play a role. Even if an individual is from a “White” country of origin, being Muslim (as opposed to being Christian) may prompt respondents to view the individual as not White. Footnote 12 They use religion as a second-order racial signifier in the case of people from White countries of origin. In this case, that means, viewing Muslims as non-White (even when they are not from a MENA country).
H2: Respondents will be less likely to assign Muslims as White, relative to Christians, all else constant.
Here, I construct my hypothesis relative to a Christian base group. This is because much of the literature on White identification focuses on Christian Whites. Moreover, research has shown that Americans of different ethnic backgrounds tend to believe that being “American” means being White and Christian (Citrin and Sears Reference Citrin and Sears2014; Devos and Banaji Reference Devos and Banaji2005). It suggests that even when they are from a “White” country, Muslims will be less likely to be assigned as White, relative to their Christian counterparts.
I also expect an additive effect between the country of origin and religion. Consider the earlier discussion that Islam is treated as an ethnoreligious category and is conflated with being from the MENA region. That is, drawing from the literature on the racialization of Islam, those who are Muslim but not from a MENA country of origin should be racialized out of Whiteness at levels similar to that of other Muslims. In this case, MENA and Muslim identities might be understood as synonymous. If these two are entirely synonymous, then we would not expect to see any differences in evaluations between (1) those who are from the MENA region but are not Muslim, (2) those who are not from the MENA region but are Muslim, and (3) those who are both from the MENA region and are Muslim. Yet, religion plays a distinct role in assessments of racial assignment, as discussed in the literature and shown through historical court cases. Even for those from a MENA country, being Muslim will make that individual the least likely to be evaluated as White.
H3: There will be an additive effect where individuals who are both Muslim and from a MENA country will be least likely assigned White, relative to individuals who are only Muslim or only from a MENA country, all else constant.
This perception of racial assignment has obvious consequences, including positionality in the US racial hierarchy. But are people merely picking groups they think are socially acceptable, or do White Americans extrapolate from religion and country of origin to physical characteristics of individuals? In addition to hypotheses 1–3, I suggest that respondents will perceive the pigmentation of individuals differentially based on both country of origin and religion. Specifically, I hypothesize that perceived skin pigmentation of the hypothetical individual will align with how likely people are to identify the individual as White. Although there are several possibilities for checking the validity of hypotheses 1–3, I use skin pigmentation because of the work of Edward Said, particularly in Orientalism (Reference Said1979). He theorizes that the Occident is defined in contrast to the Orient, whose people are stereotyped as violent, backward, and darker. Research has shown this type of thinking, which is particularly influenced by the media, is still present and is directed particularly toward Muslims (Oskooii, Dana, and Barreto Reference Oskooii, Dana and Barreto2019). Testing perceived skin pigmentation is one way to capture these perceptions.
H4: Respondents’ indication of perceived skin pigmentation (i.e., how dark they perceive someone’s skin to be) will align with how likely they are to indicate a hypothetical individual is White (i.e., following from H1–H3), all else constant. Specifically:
H4.1: Respondents will rate an individual who is from a MENA country—as compared to a country in Europe—as having darker skin pigmentation, all else constant.
H4.2: Respondents will rate an individual who is Muslim as having darker skin pigmentation, relative to Christians, all else constant.
H4.3: There will be an additive effect where individuals who are both Muslim and from the MENA region will be rated as having the darkest skin pigmentation, relative to individuals who are only Muslim or only from the MENA, all else constant.
Data and Methods
To test the hypotheses, I used a survey experiment with an original design (see online appendix 2 for adherence to ethical research principles and practices). Respondents were told they were participating in a training module for individuals who recode data from the US Census. Respondents were then randomly assigned to one of four conditions in which they were presented with a hypothetical individual who erroneously wrote their country of origin under “some other race,” instead of selecting a racial category. Each condition varies by two factors: the country of origin given (Russia or Iran) and religion (Muslim or Christian).
The study was fielded from May 5 to May 9, 2020, through Lucid. It was a national quota-based sample of 1,091 respondents who identified themselves as non-Hispanic, non-MENA Whites. Respondents were first asked about their age to ensure they were old enough to participate and then their racial identification and gender. They were randomly assigned to one of four conditions but were not randomized based on their demographics. The groups were balanced across key sociodemographic variables. At the end, they were then asked questions about their socioeconomic backgrounds. Table 2 shows the experimental design, including the number of respondents in each condition. Descriptive statistics, balance tables, and models using demographic moderators are included in online appendix 4.
Experimental Design
An image of the same hypothetical individual—a composite of Black, White, Asian, and Latino faces from http://faceresearch.org (DeBruine and Jones Reference DeBruine and Jones2015)—was given to each respondent. I used the image of a man, because using a woman might yield different outcomes depending on whether she was wearing a hijab. Future work will consider the gendering of racial perceptions based on religion. For example, incorporating religious symbols such as the hijab may lead to differences in perceived race and skin pigmentation for Muslim women.
There were two dependent variables: White assignment and perceived skin pigmentation (figure 1). First, I asked respondents to reclassify the individual by selecting a racial category from the options provided. If the respondent reclassified the individual in the profile as White, that response was coded as 1. But if the respondent selected any other racial category, that response was coded as 0. As a result, values closer to 1 indicated a higher likelihood to be assigned White. Perceived skin pigmentation was measured from the skin pigmentation scale provided in the General Social Survey. The lightest pigmentation was coded as 1 and the darkest as 10. This value was not rescaled or recoded. Thus, higher values indicated darker perceived skin pigmentation. With the smallest cell having a sample size of 264, this design has enough power to detect effect sizes as small as Cohen’s d of 0.173.
Survey Questionnaire for Assignment Experiment
Justification of Design Choices
This design only includes two countries: Russia and Iran. This is an appropriate pairing because Americans view these two countries at similar levels of unfavorability (Country Ratings n.d.). In a Gallup survey conducted from February 3 to February 16, 2020, 43% of Americans indicated they felt “mostly unfavorable” toward Russia, and 44% of Americans indicated they felt “mostly unfavorable” toward Iran. The similarity ensured that negative sentiment about both countries did not differentially influence respondents’ racial evaluations. Moreover, although Iran has a smaller Christian population (around 1.5%) relative to the Russian Muslim population (around 6%), Iran is a better choice than a country like Syria: even though Syria has a larger Christian population (around 10%), respondents may imagine the Syrian as a refugee, which could lead to a confound in the treatment. I did not vary across different MENA countries of origin. As such, I selected a MENA country that would be a conservative test of White inclusion via White assignment. Iran is a conservative test because Iranians are ethnically Aryan, instead of Arab, and therefore might be seen as closer to being White. This would bias against my results. Thus, other MENA countries of origin might have larger differences relative to Russia. Even so, future work should seek to include different MENA countries of origin.
A second design consideration was the decision to include Middle Eastern, North African, or Arab American as one of the reclassification options. A supplemental study in online appendix 3 provides insight into why excluding that category may lead to measurement error.Footnote 13 In this study, I find that respondents are not more likely to select “White” over “Some Other Race” when no MENA option is given. That is, respondents do not necessarily migrate their classification of Iranians from MENA to White if the MENA option is not available. However, when MENA is given as an option, most respondents select it. More importantly, the key question of this article is not whether White Americans know Iran is in the Middle East or that Iranians are Middle Eastern versus White: it is about the role of religion in racial assignment. This means that a more conservative test of this theory would require MENA to be given as a response option. Are White Americans more likely to assign White inclusion to Iranians if they are Christian, even if MENA is available as an option? If so, the regional accuracy would be considered less important than a religious cue, which is an important finding regarding the role of religion in White assignment.
It is important to note that respondents may have been taken out of experimental realism when asked to assign their perceived pigmentation based on the black-and-white image. I argue this is not a critical limitation because respondents knew they were in a survey context and the task was hypothetical (i.e., they never signed up to work for the US Census as a coder). Although they may have tried to surmise the real intention of the survey, there is no reason to believe they would be systematically biased in any specific direction. That is, it was not clear whether the intent of the study related to stereotypes of individuals (i.e., Iranian Muslims are darker) or the accuracy of task completion. This means these potential responses, which would be randomly assigned across treatment groups, would only introduce nonbiased noise. Because I am interested in relative comparisons, the noise would only bias against a statistically significant result, rather than toward a specific outcome.
There was no pure control condition for either dependent variable: White assignment or perceived skin pigmentation. Because I was interested in relative comparisons across treatments, there was no need for a control. I showed that religion, alongside race, alters ethnoracial classifications. Moreover, including a condition that only focused on race might have introduced measurement error if respondents had inferred the individual’s religion and included it in their evaluation. Respondents were asked to select the race label they thought was most appropriate and to select a pigmentation for the individual. Except for the accompanying photo, the design does not disrupt the facsimile of a task completed outside the experimental setting. And although the Census Bureau does not ask about religion, only 25% of individuals are aware of this fact (“What Americans Know about the 2020 Census” 2020).
Results
The results indicate support for hypotheses 1, 2, and 3, and partial support for hypotheses 4.1–4.3. There is strong evidence to suggest that religion does play a distinct role in racial assignment—over and above country of origin. Moreover, Muslim and MENA are not synonymous with one another: individuals assess Iranian Christians and Russian Muslims as distinct from Iranian Muslims.
Who Is White?
Figure 2 shows data to support hypothesis 1 (the means for the relevant conditions with 95% confidence levels): evaluations of racial assignment differ based on country of origin, holding all else constant. Iranians are less likely to be selected as White than Russians, regardless of religion (see online appendix 5, table 9 for OLS regression results). The Russian individual was likely to be categorized as White by 71.7% of respondents; meanwhile, Iranians were classified as White only 51.1% of the time. This 20.6 percentage point difference is statistically significant at a level of p < 0.001. The Cohen’s d effect size of the difference between these two means is 0.43Footnote 14; this is considered a medium effect size. This rejects the null hypothesis that there would be no difference between the likelihood of White assignment based on country of origin and thus supports hypothesis 1.
Hypothesis 1
Figure 3 examines the role of religion in racial assignment, regardless of country of origin. Recall that hypothesis 2 predicts that Muslims will be less likely to be classified as White, relative to their Christian counterparts, holding country of origin constant; the results support this hypothesis. Christians were more likely than Muslims to be categorized as being White (see online appendix 5, table 10 for OLS regression results). Christians were categorized as White 69.3% of the time, whereas Muslims were classified as White by 53.5% of respondents. This 15.7 percentage point difference is statistically significant at a level of p < 0.001. The Cohen’s d effect size of the difference between these two means is 0.33, which is considered a medium to small effect size.
Hypothesis 2
Hypothesis 3 states there is an additive effect between country of origin and religion. Specifically, I hypothesize that the Russian Christian is the most likely to be assigned as White, and the Iranian Muslim is the least likely to be assigned White. Building from the historical cases, I argued that religion and country of origin would not be treated as redundant but as separate factors in determining racial assignment. Figure 4 plots White assignment for each treatment. We see that the Russian Christian is assigned White 80.9% of the time (see online appendix 5, table 11 for OLS regression results). In contrast, the Iranian Muslim is the least likely to be assigned White of the four conditions—only 44.3% of the time. This 36.5 percentage point difference between Russian Christian and Iranian Muslim is statistically significant at the p < 0.001 level. The Cohen’s d effect size of 0.82 is large. Although this finding may seem obvious, one must consider that, legally, both the Russian and Iranian are White and that Muslims are religious subjects and can be from any race.
Hypothesis 3
Next, I consider Muslim profiles conditional on country of origin. Iranian Muslims are less likely than Russian Muslims to be assigned White (see online appendix 6, table 15 for differences in means). The Iranian Muslim was assigned White 44.3% of the time versus 62.4% for the Russian Muslim. This 18.1 percentage point increase is statistically significant at a level of p < 0.001. The Cohen’s d effect size of 0.37 is medium to small. This suggests that country of origin still plays a role in White assignment. If White respondents considered Muslim identity to be sufficient for racial considerations, the difference between these two profiles would not be statistically significant.
Subsequently, I consider Iranian profiles conditional on religion. Iranian Muslims are less likely than Iranian Christians to be assigned White (see online appendix 6, table 15 for differences in means). Iranian Muslims were considered White 44.3% of the time versus 56.7% for Iranian Christians. This 12.4 percentage point increase in White assignment is statistically significant at p < 0.01. The Cohen’s d effect size of 0.27 is small. Recall that the task asked White respondents to reclassify an individual based on country of origin and religion. Importantly, one of the racial categories included in the list was “Middle Eastern, North African, or Arab American.” Even in the presence of this category option, respondents were more likely to classify the Iranian Christian as White relative to the Iranian Muslim, despite the fact that both are from Iran—a country in the MENA region. This underscores that religion plays a role in ethnoracial assignment decisions in addition to country of origin.
One of the most striking findings in figure 4 is that the mean White assignment between the Russian Muslim and the Iranian Christian is statistically indistinguishable (p = 0.251; see online appendix 6, table 15 for the difference in means). The Russian Muslim and the Iranian Christian are assigned White at similar levels at 62.4% and 57.6%, respectively. The Russian Muslim is less likely assigned White than the Russian Christian but more often than the Iranian Muslim. To be clear, this null effect suggests that country of origin and religion have additive effects. Religion matters nearly as much as geography for ethnoracial assignment decisions. This supports Hypothesis 3’s claim that country of origin and religion both influence White assignment. Moreover, if Muslim identity was racialized to completely encompass MENA identity, then we would not see any differences in White assignment between those who were Russian Muslim, Iranian Christian, or Iranian Muslim. However, we see that in evaluations of White assignment, White respondents disentangle, to a certain extent, the differences between country of origin and religion.
Who Is Lighter/Darker?
As a robustness test, I asked respondents to rate the perceived skin pigmentation of the hypothetical individuals to assess the connection between religion and region on perceptions of Whiteness. Thus, this study tests not only how White respondents assign ethnoracial labels to others but also how they perceive them. Recall that Hypothesis 4 states that evaluations of skin pigmentation would map onto the hypotheses for White assignment (H1–H3). Respondents’ White assignment would also map onto how light- or dark-skinned they perceived the individual’s image to be, which was the same across all treatments and thus showed no differences in perceived skin pigmentations.Footnote 15 The results of the experiment show that perceptions of skin pigmentation vary based on information about a hypothetical individual’s country of origin and skin color, which supports hypothesis 4.
Figure 5 shows the effect of country of origin on perceived skin pigmentation, holding religion constant. Russians were perceived as lighter than Iranians: 3.39 versus 3.57 on the skin pigmentation scale This 0.18-unit increase is statistically significant at the p < 0.1 level (p = 0.061; see online appendix 6, table 12 for OLS regression results). It should be noted that although all results are presented as two-tailed tests, the hypotheses are directional. In this case, I specified that Iranians are expected to be rated as darker. For a two-tailed hypothesis level, we would fail to reject the null hypothesis, but for a one-tailed hypothesis—which I theoretically grounded—we could reject the null hypothesis at p < 0.05. Therefore, I conclude this is a partial acceptance of hypothesis 4.1.
Hypothesis 4.1
Figure 6 shows that, holding country of origin constant, religion plays a role in perceived skin pigmentation. On average, Christians are rated lighter than their Muslim counterparts: 3.34 versus 3.62. This 0.28-unit difference is statistically significant at the level of p < 0.01 (see online appendix 6, table 13 for OLS regression results). The Cohen’s d effect size of 0.17 is small. This suggests that in addition to influencing racial assignment, religion plays a role in perceived skin pigmentation, with Muslims being perceived as darker, holding country of origin constant. This provides evidence to reject the null hypothesis and support hypothesis 4.2: the same image of the same person is perceived as darker when respondents think the hypothetical individual is Muslim.
Hypothesis 4.2
Next, I consider how both country of origin and religion interact and affect perceived skin color. Figure 7 shows that the Iranian Muslim is perceived to be the darkest of the images in all conditions (see online appendix 6, table 14 for OLS regression results). The differences between the perceived skin pigmentation of the Russian Muslim and the Iranian Christian are not statistically significant, as we saw with White assignment. Unlike the results with White assignment, the differences between the perceived skin pigmentation of the Russian Christian and the Iranian Christian are also not statistically significant. But the difference in means between the Russian Muslim (mean of 3.478) and the Iranian Muslim (mean of 3.77) is statistically significant at p < 0.05. This suggests that country of origin, when considered in conjunction with religion, influences perceived skin color. Those who are Russian, even if they are Muslim, are perceived as Whiter than their Iranian counterparts. Although hypothesis 4.1 was only partially accepted, this analysis shows that country of origin does play a role in perceived pigmentation if we consider Muslims from two different countries. Finally, the difference between the Iranian Muslim and the Iranian Christian is also statistically significant. Iranian Muslims are, on average, rated as 3.77 on the skin pigmentation scale, whereas Iranian Christians are rated, on average, as 3.385 on the skin pigmentation scale. This 0.39-unit decrease is statistically significant at a p < 0.01 level. The Cohen’s d effect size of 0.25 is small. When considering Iranians, the Christian Iranian is perceived as lighter than the Muslim counterpart.
Hypothesis 4.3
When comparing “White” characteristics—that is, of a Russian (who is Christian) with a Russian (who is Muslim) or of a Christian (from Russian) with a Christian (from Iran)—we see no statistically significant differences in perceived skin pigmentation. However, when comparing an Iranian (who is Christian) with an Iranian (who is Muslim) or a Muslim (from Russia) with a Muslim (from Iran), we see that religion and country of origin do play a role in altering perceived skin pigmentation. Russians are perceived as lighter than Iranians, even when both are Muslim; Christians are perceived as lighter than Muslims, even when both are Iranian. This partially explains why we see statistically significant differences between the perceived skin pigmentation of the Iranian Christian and Iranian Muslim, as well as between the Russian Muslim and the Iranian Muslim.
Thus, the empirical results offer strong support for hypotheses 1–3: religion and country of origin play distinct roles in racial assignment. The empirical results offer partial support for hypothesis 4. Hypothesis 4.1 is supported (with a directional significance test): country of origin plays a role in perceived skin pigmentation, all else constant, with Russians rated as having lighter pigmentation. Hypothesis 4.2 is supported: religion plays a role in perceived skin pigmentation, all else constant, with Christians rated as having lighter pigmentation. Finally, hypothesis 4.3 is partially confirmed, with Iranian Muslims rated as having the darkest skin pigmentation.
This study is not without its limitations, and there are certainly opportunities for future research. As mentioned, the experiment focuses on how Whites operationalize inclusion into Whiteness. But we do not know whether non-White respondents would make the same racialized judgments related to Whiteness. Moreover, just as this study problematizes who is White, future work should seek to understand how religion and country of origin influence Black inclusion. For example, Black MENA populations have been understudied. Research into how religion influences Blackness, coupled with the unique history of non-MENA Black Muslims, will enable a better understanding of Blackness and Islam in America. Future research could also vary the countries of origin. Just as with the court case in which Mohriez was ruled to be White, in part because of US foreign relations with Saudi Arabia, including other countries may give insight into the role of Whiteness relative to their geopolitical relationship with the United States. Other religions can also be studied. This study has focused on one of the most racialized and marginalized religions to make this case. Future work on other religions, such as Judaism, Buddhism, and Hinduism, will enhance our understanding of how religion and race shape racial identities, assignments, and classifications in the United States. Finally, future research should investigate the psychological mechanisms underlying the ethnoracial assignment decisions of White Americans.
Conclusion
This experiment supports research that religion, not just country of origin, plays a role in how Whites operationalize Whiteness. White Americans consider both characteristics when deciding who counts as White. Muslims, all else equal, are less likely to be assigned White than Christians, and Iranians are less likely to be assigned White than Russians. This contrasts with the state-sanctioned category of MENA individuals as White. Moreover, racial assignment decisions seem to have an additive effect, with religion and country of origin playing two distinctive—and not conflated—roles. Iranian and Russian Christians are more likely assigned White than their Muslim counterparts. Iranian Muslims are the least likely to be assigned White and are perceived to have the darkest skin pigmentation. This suggests that historical dynamics of the racialization of religious subjects have carried forward and remain in force today.
MENA and Muslims in the United States are often in the position of non-normative Whiteness. Although many are legally classified as White, the process of racialization—assigning social and political meaning to individuals based on reproduced traits—has pushed these groups to the margins of Whiteness. Both are classified among the most privileged ethnoracial groups while also being subjected to government surveillance, travel bans, and hate crimes. They are often discriminated against in schools (Atwal and Wang Reference Atwal and Wang2019; Pfaff et al. Reference Pfaff, Crabtree, Kern and Holbein2021) and even suffered disproportionately from COVID-19 relative to White Americans (Dallo et al. Reference Dallo, Kindratt, Seaton and Ruterbusch2022). Although recent research suggests that MENA individuals mostly do not self-identify as White (d’Urso Reference d’Urso2022; Maghbouleh, Schachter, and Flores Reference Maghbouleh, Schachter and Flores2022; Mathews et al. Reference Mathews, Phelan, Jones, Konya, Marks, Pratt, Coombs and Bentley2017) and many White converts to Islam question their White standing (Galonnier Reference Galonnier2015b), the question remains whether this is because MENA and Muslim individuals want to distinguish themselves from Whites or whether Whites also distinguish themselves from these groups. If Whites do not include these groups firmly as White, this would explain why these groups receive disparate treatment than other Whites. This article tests two traits—religion and country of origin—as the basis for disparate treatment and the marginalization of those on the cusp of Whiteness.
More than MENA and Muslim individuals, this article addresses the problematization of Whiteness. Rather than taking for granted who is White, it investigates how Whites understand Whiteness. This is important because Whites systemically benefit from Whiteness within the United States. Historically, Whites were the only ones who could own land, be naturalized, and vote, and today, Whites benefit politically and materially from their Whiteness. If we do not investigate constitutive traits of Whiteness, it becomes difficult to assess the marginalization of groups who may not be accepted within the White category. For example, the empirical findings may suggest that White converts to Islam will not be incorporated into Whiteness with the same regularity or certainty as White Christians. Perhaps the marginalization of MENA Christians is different from that of MENA Muslims because of the perception of MENA Christians as closer to being White. Although White converts, White Christians, MENA Christians, and MENA Muslims are all legally White, their positionality and ability to be included in American democracy may hinge on their religious characteristics. Ultimately, country of origin and religion matter much more in perceptions of Whiteness than the legal designation. This underscores ethnoracial classifications as social, not merely institutional, constructs.
Acknowledgments
Thank you to Beyza Buyuker, James Druckman, Alexandra Filindra, S. R. Gubitz, Aviral Pathak, Matthew Nelsen, Kumar Ramanathan, Jason Seawright, and Alvin Tillery, as well as the members of the Politics of Race, Immigration, and Ethnicity Consortium and the Druckman Research Lab at Northwestern University for their aid with this research.
Open access
