Making sense of race and racial classification

13
Race & Society 4 (2001) 235–247 Making sense of race and racial classification Angela James University of Southern California, 760 Westwd Plz Box 951759, Los Angeles, CA 90095-1759, USA Abstract As social scientists, race scholars, and demographers, how do we begin to make sense of recent changes in the Census Bureau’s system of racial classification, as well as of the popular response to those changes? This paper explores the lacuna between popular and scientific understandings of race. It reviews the theoretical understanding of race as a social construct, providing a brief history of racial classification in the United States. In addition, it examines the concepts of race mixing and racial ambiguity as a function of the peculiar and distinctive construction of race in the United States. Finally, the essay critically assesses how race is currently used in social research and how race might be more accurately represented and effectively employed in that research. © 2003 Elsevier Inc. All rights reserved. 1. Changing notions of race On March 7, 2001, the Washington Post published preliminary figures from Census 2000 and announced that the Hispanic population of the U.S. had “drawn even” with the African American population in size (Cohn & Fears, 2001). Whereas the Hispanic population consti- tuted 12.5% of the total U.S. population, the Black population was “only” 12.3%. Of course, that finding was true only if Blacks who responded as identifying with more than one race were not counted as Black, and if Blacks who also identified themselves as Hispanic origin were counted solely as Hispanic. The tone of this headline, however, presages the shape of emergent racial construction processes. Who “counts” in public policy is represented as being determined by which “minority” group can claim the largest population tallies. In this manner, marginalized groups are cast as competitors over scarce social resources. Few ask why the ethnic designation “Hispanic” is usually compared to racial designations. Would it be equally relevant to compare the number of people with “British” ethnic ancestry in New England with Tel.: +1-323-839-3515; fax: +1-310-794-6297. E-mail address: [email protected] (A. James). 1090-9524/$ – see front matter © 2003 Elsevier Inc. All rights reserved. doi:10.1016/S1090-9524(03)00012-3

Transcript of Making sense of race and racial classification

Race & Society 4 (2001) 235–247

Making sense of race and racial classification

Angela James∗

University of Southern California, 760 Westwd Plz Box 951759, Los Angeles, CA 90095-1759, USA

Abstract

As social scientists, race scholars, and demographers, how do we begin to make sense of recentchanges in the Census Bureau’s system of racial classification, as well as of the popular response tothose changes? This paper explores the lacuna between popular and scientific understandings of race.It reviews the theoretical understanding of race as a social construct, providing a brief history of racialclassification in the United States. In addition, it examines the concepts of race mixing and racialambiguity as a function of the peculiar and distinctive construction of race in the United States. Finally,the essay critically assesses how race is currently used in social research and how race might be moreaccurately represented and effectively employed in that research.© 2003 Elsevier Inc. All rights reserved.

1. Changing notions of race

On March 7, 2001, theWashington Postpublished preliminary figures from Census 2000and announced that the Hispanic population of the U.S. had “drawn even” with the AfricanAmerican population in size (Cohn & Fears, 2001). Whereas the Hispanic population consti-tuted 12.5% of the total U.S. population, the Black population was “only” 12.3%. Of course,that finding was true only if Blacks who responded as identifying with more than one racewere not counted as Black, and if Blacks who also identified themselves as Hispanic originwere counted solely as Hispanic. The tone of this headline, however, presages the shape ofemergent racial construction processes. Who “counts” in public policy is represented as beingdetermined by which “minority” group can claim the largest population tallies. In this manner,marginalized groups are cast as competitors over scarce social resources. Few ask why theethnic designation “Hispanic” is usually compared to racial designations. Would it be equallyrelevant to compare the number of people with “British” ethnic ancestry in New England with

∗ Tel.: +1-323-839-3515; fax:+1-310-794-6297.E-mail address:[email protected] (A. James).

1090-9524/$ – see front matter © 2003 Elsevier Inc. All rights reserved.doi:10.1016/S1090-9524(03)00012-3

236 A. James / Race & Society 4 (2001) 235–247

the number of Blacks in that area? Or does the comparison between Blacks and Hispanicsmerely reflect an emergent popular understanding of Hispanic/Latino as constituting a raciallydistinct group? The journalistic “fact” of Hispanics “overtaking” Blacks is as much a fabrica-tion of racial designation as it is of the demographic growth of the emerging group (Goldberg,1997).

As social scientists, race scholars, and demographers, how do we begin to make sense ofrecent changes in the Census Bureau’s system of racial classification, as well as of the popularresponse to those changes? This essay explores the lacuna between popular and contemporarytheoretical understandings of race. It starts by reviewing the idea of race as a social construct,as well as presenting a brief history of racial classification in the United States, then criticallyexamines the role of social research in perpetuating flawed, essentialist understandings of race.The essay ends with an assessment of the question of how race can be most accurately andeffectively used in social research. I argue that there are two major approaches to understandingrace and that these contrasting conceptions underlie the different ways social scientists approachthe examination of race in the social world. Specifically, race can be understood either as asocial construction or as a fixed characteristic. Although contemporary theoretical views ofrace flow from understanding it as social construction, this insight is lost in statistical analysisof race, which most often uses the construction as if it were a fixed characteristic. There aresignificant political and social costs associated with the continued use of race as a fixed qualityin statistical analyses. This essay highlights those costs, and offers suggestions for how raceshould be employed in statistical analysis.

2. Race as social construction

Race is an exceedingly slippery concept. Although it appears omnipresent and real in sociallife, the construct is hard to pin down in any objective sense. This is because the idea of race isriddled with apparent contradictions. For example, while race is adynamicphenomenon rootedin political struggle, it is commonly considered afixedcharacteristic of human populations;while it does not exist in terms of human biology, people routinely look to the human body forevidence about racial identity; while it is a biologicalfiction, it is nonetheless a socialfact.

2.1. The origin of race

The emergence of contemporary understandings of race coincided with European coloniza-tion and domination of native peoples around the world. When English voyagers first laid eyeson Africans some time after 1550, the stage was set for the race relations that have unfoldedin the United States ever since. As historian Winthrop Jordan has noted, “Englishmen actuallydescribed Negroes asBlack—an exaggerated term which in itself suggests that the Negro’s com-plexion had powerful impact upon their perceptions” (Jordan, 1974, p. 7, emphasis original).

In other words, the early English read the African body as “text” (Hall, 1988), as a signifierof important meanings about the nature of the souls (or lack thereof) contained within. Thelabeling of African bodies as “Black,” of course, became all the more meaningful becausethe label represented the perfect contrast to the ideal against which the English measured

A. James / Race & Society 4 (2001) 235–247 237

themselves. Adds Jordan, “It was important, if incalculably so, that English discovery of BlackAfricans came at a time when the accepted English standard of beauty was a fair complexionof rose and White. Negroes seemed the very picture of perverse negation” (1974, p. 7).

The role of this “perverse negation” in the subsequent history of race in America cannotbe overstated. It is a history rooted in “binary oppositions” (Hall, 1988), specifically betweenthe opposites “Black” and “White.” Because the meaning of an object is always established inrelation to another object, that is, through the differences between the objects, meaning is mosteffectively expressed through comparisons between binary opposites (e.g., day and night,male and female, Black and White). Eventually, the basic White (English)/Black (African)distinction would give birth to extended chains of related equivalences: English= civilized =Christian= superior= free, whereas African= uncivilized= heathen= inferior = un-free.The African thus became the foil against which the English defined who they wished to be.Indeed, the rationalized morality of the slavery system (“un-free” labor) depended upon thesedistinctions, which also worked to pacify poor Whites (“free” labor) with the relatively higherlevels of status it bestowed upon them. In a very real sense, America’s emerging racial orderwas predicated on this binary and continues to be organized by it today, despite the rise ofother-raced groups.

2.2. The nature of race

In our earliest conceptualizations of race, scientists and common-folk alike understood raceto be a matter of bloodlines. In other words, racial characteristics were considered to be “in theblood” and tied to an individual’s ancestry as well as appearance. Following the early Europeanvoyagers, observers generally accepted race as a fact of nature, as a biological, fixed essencethat could be used to objectively distinguish one human group from another.

Even though common-sense notions of race continue to echo this early view of racial differ-ences being rooted in biological distinctions between groups, most scholars today understandrace as a “social construction,” the ongoing, dynamic product of political contest (Omi &Winant, 1994). In practical terms, to say that race is “socially constructed” means that it isneither“mere illusion” and thus irrelevant,norsomething that should be viewed as “objective”and “fixed.” As a social phenomenon, race is rather unique in its marriage of fictitious andfactual attributes. Even though there are physical characteristics that one can point to in distin-guishing among various groups, most social scientists would agree that the physical differencesare much less significant than the social meanings attached to them (recall the chains of equiv-alences associated with the Black/White binary). In fact, researchers have shown that thereis often more phenotypic variationwithin so-called “racial” groups, than there isbetweenthegroups (Marks, 1994). This is why race as a social phenomenon is distinct, although not entirelydisconnected, from observable physical differences.

At the same time, the social construction of race refers to the processes by which people, bothindividually and collectively, negotiate their identities and related interests in an environmentwhere resources are assumed to be scarce. That is, race entails subjective attachments of theindividual to a racially defined group—either a dominant racial group (which uses its powerto subordinate other racial groups defined as inferior, i.e., practices “racism”), or subordinatedracial groups (whose members often join forces to challenge the racial status quo, i.e., enact

238 A. James / Race & Society 4 (2001) 235–247

“anti-racism”). In this sense, race acts as a proxy for a common set of historical experiences andcontemporary interests shared by individuals (Guinier, 1995). What it “means” to belong to arace at any given moment is created intertextually as part of an ongoing, ideological struggleover societal boundaries, inclusion, and equity (Hall, 1988; Hunt, 1997). These constructionsare the product of both the racial subject and the racial others. Each is intricately connected tothe racial structuring of society, as well as to ongoing interactions with each other.

In short, social events, both past and present, are key to the formation of any racializedcontext. Economic, political, and cultural factors continually shape the definitions of race andof racial groups in the United States (Omi & Winant, 1994). It is impossible to understandcontemporary social debates about race and racism without acknowledging how deeply theyare embedded in the specifics of the moment. At the same time, the social meaning of racetoday, although somewhat different from past understandings, cannot be entirely disconnectedfrom this history; it can be most fruitfully considered as part of a longstanding, dynamic processof racial construction (Omi & Winant, 1994). While the human body informs the roles peopleplay in this process, it does not determine the outcome.

Current research on human origins emphasizes common ancestors that branch off into sepa-rate subpopulations (Morning, 2000). This view contrasts sharply with the commonly employedconcept of four “racial groups” (Black, White, American Indian, and Asian), which have beenused in the United States. This concept of four races (putting aside for the moment the currentproject of racialization among Latinos) relies on the belief that separate, unmixed racial popula-tions exist. Yet, each group can only be viewed as constituting distinct “populations” by ignoringthe substantial and ongoing intermingling between groups. For example, in order to discussthe demise and/or disappearance of the Native American population from Mississippi, onehas to accept the premise that Blacks and Whites in Mississippi with known Native Americanancestry arenotNative Americans. Membership in a given category, especially if that categoryis “Black,” has traditionally meant exclusion from membership in any other racial category.

Fig. 1. Representing race: 2000 Census for United States. Source: Census 2000,Summary File 1(SF 1), 100% data.

A. James / Race & Society 4 (2001) 235–247 239

The following table, based on data from the 2000 Census (U.S. Bureau of Census, 2001),illustrates a more fluid understanding of race. In this Census, for the first time, respondentswere allowed to identify with more then one racial category (Fig. 1).

Here each racial group can be represented in terms of those who identify with only oneracial group, as well as those who identify with any other group(s). While this representationresults in population counts exceeding the number of “warm bodies” in the United States, it hasthe advantage of illuminating the relative permeability across groups. At the same time, thisrepresentation of all racial groups in nonexclusive terms underscores the socially constructednature of racial definitions.

3. The U.S. Census and its use of race for classification

Racial classification is made necessary by the political and social centrality of race inAmerican society. As the previous section on the history of race demonstrates, oppressionpredated the naming of differences among races; it is therefore naı̈ve to believe that simplyby eliminating classification, we could also eradicate racialized hierarchies. However, whileclassification may be necessary, it is also necessarily problematic. Because race is a sociallyconstructed concept with no fixed reality, all classification efforts are fraught with imprecision.Classification schemas always reflect the imperfect choices made in response to particularpolitical agendas and imperatives. Institutionalizing these choices over the long term alwayshas the effect of “naturalizing” constructed understandings of race. This reification explainswhy a great deal of struggle has generally been required to change Census classifications, andhighlights the necessity of interrogating change and continuities in categorization.

The federal government has always measured race in some form. Although the categorieshave changed over time, along with the rationales given for the collection of this informa-tion, Whites, Blacks, and Indians have been in various ways designated as “separate races”from the inception of the republic (Rodriguez, 2000). The Census has been entangled in thenation’s racial dilemmas, from the initial instruction of the “Founding Fathers” in the U.S.Constitution to count slaves as three-fifths of a person, to current debates over identity andrepresentation (Anderson & Fienberg, 1999; Skerry, 2000). Indians “not taxed” were not to beenumerated, while those who “renounced tribal rule” were to have their “color” described asIndian (Anderson & Fienberg, 1999, p. 12). In 1870, the Census Bureau initiated a new cate-gory for “Chinese,” which was followed legislatively by the Chinese Exclusion Act of 1882.Japanese, Filipino, Hindu, and Korean categories were all added in response to successiveimmigration waves. In fact, some have argued that the history of the Census Bureau’s attemptsto classify the populace can be read as an ongoing debate over exclusion of various parts ofthe population from full participation in the nation-state (Anderson & Fienberg, 1999).

Until 1960, the racial classification of individuals in the Census relied upon the reports andobservations of Census enumerators (Anderson & Fienberg, 1999). Race, during most of thisperiod was viewed as both self-evident and fixed. Significantly, where there was ambiguity,enumerators were instructed to look for other social signifiers (Rodriguez, 2000). For example,in the 1870 Census enumerators were instructed to use the socially dependent criteria ofresidential community to classify “half-breed” Indians. So, if such persons lived with Whites,

240 A. James / Race & Society 4 (2001) 235–247

and had the “habits of life” of Whites, they should be classified as Whites. However, if theylived among Indians they were to be classified as Indian (Rodriguez, 2000).1

For the last several decades, racial classification has reflected an individual’s self-identifica-tion with one in a fixed set of categories. This change reflected the recognition of race as sociallyconstructed, although the implication was still that race was fixed. Further, this technique forclassifying by race did not allow for any recognition of the possibility of mixed racial heritage.This, of course, reflected a social understanding that mixed-race individuals were not White, andthat other racial distinctions were less important. According to this schema, when a mixed-raceperson identified as White, they were “passing.”

In the most recent Census, taken in 2000, for the first time individuals were instructed toidentify themselves using as many racial categories as they wished. This recent change resultedfrom intense political pressure by multiracial advocates, in addition to the continuing increasein recognition of the fluid and subjective nature of race for any given individual, as well asacross groups. This change was opposed by advocates of several “single” race groups. InLos Angeles, for example, several local Black politicians and community leaders urged itsconstituents to simply “check the Black box,” particularly if their direct parentage was Black.This campaign was undertaken in recognition that there is strength in numbers. Debates overclassification schemes are not merely theoretical. The ideological and political significanceassigned to numeric representations ensures that any changes to classification systems willhave substantial ramifications.

Indeed, recently a local reporter called to ask me about the social and cultural consequencesof Black population decline in the city of Los Angeles. The reporter described a 75,000-persondecline in the number of Blacks in the city of Los Angeles between 1990 and 2000. After probingfor the source of the numbers with which he was working, I discovered that the numbers weredeflated by their exclusion of any Blacks who had reported more than one race. If all personslisting Black as a part of their racial identity were instead included in the “Black” category,the decline in the number of Blacks from the city of Los Angeles drops to a level easilyexplained by movement within the same metropolitan area. Although this particular articleon the impact of Black population decline in Los Angeles was never written, the reporter’smisunderstanding reflects a widely-held belief that Black influence in Los Angeles is waningand is being superseded by Latino influence. The results of this view range from support ofBlacks in the city for a White mayoral candidate over a Latino candidate with similar, if notmore liberal, political positions, to the uproar over the recent loss of the Black police chief.

3.1. Race versus ethnicity in the Census

The Census Bureau’s treatment of race directly contrasts its handling of ethnicity. Themanners in which race and ethnicity are treated by the Census, as well as by the popular media,tell us a great deal about the tortured history of race in American thought. Sections on ethnicityand ethnic ancestry have long incorporated questions about the respondent’s own country ofbirth along with the country of birth of each of their parents. For instance, in 1980 and 1990,the Census Bureau asked respondents to indicate the ancestryor ancestries with which theymost closely identified, notably not requiring the respondent to specify an exclusive ethnicidentity. As this demonstrates, the question of ethnic identity has never been considered rigid,

A. James / Race & Society 4 (2001) 235–247 241

while race has been assumed fixed across generations. Further, whereas the ancestry sectionsof the questionnaire are clearly labeled as such, respondents are not similarly alerted to thefact that they will be asked to identify their race. The respondent is asked to simply complete asentence, “This person is. . . ” that is followed by four racial categories—White, Black, NativeAmerican, and Asian/Pacific Islander—and one residual category, “other.” Hence, even theway the race question is displayed demonstrates the institutional discomfort with the projectof race classification. In fact, the questionnaire includes an overt denunciation of race as aphysical/biological, as opposed to social, fact. The description of race and the race question aspresented by the Office of Management and Budget:

The concept of race as used by the Census Bureau reflects self-identification; it does notdenote any clear-cut scientific definition of biological stock. The data for race representself-classification by people according to the race with which they most closely identify. Fur-thermore, it is recognized that the categories of the race item include both racial and nationalorigin or sociocultural groups. (U.S. Census Bureau, Census, 2000 Questionnaire)

This statement underscores the subjective and social-political nature of the data being col-lected. Census officials have explicitly distanced themselves from any understanding of racerooted in biology.

4. From ethnicity to race: contemporary racial construction and Hispanics

As the previous discussion highlights, despite the broadening conceptualization of race theCensus continues to be inextricably bound to issues of race and representation. As such, socio-cultural groups are continually redefined and “racialized.” For example, “Hispanic American”has been used in the U.S. to designate members of a group whose population could be of anyrace. However, in the last several decades, 50–60% of Hispanic Americans responded that theywere racially White, much smaller percentages respond that they were Black (3.6%), AmericanIndian (1%), or Asian (1%), and over 40% of Hispanic Americans marked “other” in responseto the racial question (Rodriguez, 2000).

The racially ambiguous status of Hispanics is becoming an increasingly important part of theracial self-portrait that is the U.S. Census. Alongside the social and political contestations ofpeople from “mixed” racial heritage, the difficulty of classifying Hispanics is transforming theboundaries of race in U.S. society. Historically, however, the political discourse about race inthe United States has been most fundamentally shaped by the definitions of who is “Black” andwho is “White” (Davis, 1991). The “one-drop rule” (Davis, 1991; Myrdal, 1944; Williamson,1984) has long placed everyone with “any known Black ancestry” in the Black category. An-thropologists use the term “hypo-descent rule” to describe this system whereby people withmixed ancestry are assigned the social status of the subordinate group (Harris, 1964). Althoughthere have at various times been attempts to describe intermediate racial statuses with termslike “mulatto, quadroon, or octoroon,” these terms have been merely descriptive of differingdegrees of Black ancestry and have never carried any legal or social significance. Other racialgroups in the United States have not been subjected to the one-drop rule. These groups rep-resent “intermediate” categories in the U.S. racial stratification schema (Hacker, 1992). Thedegree to which “racially-mixed” personswithoutBlack ancestry have been accorded the lower

242 A. James / Race & Society 4 (2001) 235–247

racialized status has been linked to both appearance and class. In general, these other groups ofracially-mixed people have been treated as assimilated into the dominant White group after thefirst generation of “miscegenation” (Davis, 1991). Whether Hispanic Americans will become a“race” or whether they, like other “ethnic” groups in the United States, will experience gradualincorporation and assimilation is an open historical question.

Racial classification is a dominant feature of all social interactions, both personal and in-stitutional, in the United States. Along with gender, knowledge of a person’s racial originsinfluences both the terms and the manner in which individuals are likely to interact. Racialclassification systems are the systemic and official accounting of salient social distinctions. Assuch, recent changes in racial classification schemas deserve attention insomuch as they reflectbroad understandings of what race means in the contemporary American context.

In the United States, racial categories reflect our own particular political and economic strug-gles. There are other racialized contexts with very different historical trajectories with regardsto race, most notably in Latin America, Brazil, and South Africa (Degler, 1986; Rodriguez,2000). In many of these contexts, race is defined in terms of appearance and class, rather thanancestry. While racial stratification with Whites at the top and Blacks at the bottom of thestatus hierarchy characterizes each of these other contexts, the boundaries separating races arenot as rigidly drawn. For example, it is not uncommon in many parts of Latin America to findwithin the same family members who are classified as “Moreno” (dark-skinned) and otherswho are considered “Jabao” (high-yellow) (Degler, 1986; Rodriguez, 2000). The categories arenot meaningless; they are attached to differential treatment and life circumstances (Rodriguez,2000). Categories also reflect the dominance and value of White “racial” characteristics overthose of Blacks and Indians. However, it is appearance rather then ancestry which determinesan individual’s experience of race and racism (Degler, 1986; Rodriguez, 2000).

Most research in this area has found that Hispanic Americans understand “race” to meansome combination of nationality, national origin, ethnicity, culture, and skin color (Rodriguez& Cordero-Guzman, 1992). Further, Hispanics tend to see race along a continuum, rather thanin dichotomous terms of “White” or “Black.”

Increasingly, Latin American conceptualizations of race are having an impact on the ongo-ing process of race construction in the United States. Only in one Federal Census (Anderson &Fienberg, 1999) have Hispanic Americans been classified as a racial group.2 This classificationwas resisted and has never been repeated. However, since 1970, Hispanic origin persons havehad two opportunities to identify themselves, once in the ancestry question included in thelong form, and once in the Hispanic origin question included in the short form. Thus, althoughHispanic origin is officially represented as an ethnicity, the political treatment suggests someambiguity as to whether and how Hispanic ethnicity is connected to overarching racial under-standings. As such, Census treatment of Hispanic groups in the United States exemplifies thesocial constructedness of race (Rodriguez, 2000). The vast majority of respondents marking“other” in response to the question of racial identity are of Hispanic origin, demonstrating theirsense that the racial taxonomies of the U.S. do not match their own identities.

In addition to the mismatch between U.S. and Hispanic American conceptualizations of race,the exclusionary nature of the concept of “Whiteness” should be acknowledged. Despite severalhistorical changes in racial categorization in the United States, there has been an overridingconcern with marking the truly White. Beginning with concern over how to classify mixed-race

A. James / Race & Society 4 (2001) 235–247 243

individuals at the turn of the twentieth century, keeping the “non-Whites,” with the notableexception of Indians, separate has been an ongoing interest. Contemporary race representationssuggest that this is still a concern. In March 30, 2001, CNN broadcast a news report that“non-Hispanic Whites are now a minority in California.” Many racial projections predict the“browning” of America. The apprehension with which the subject is broached indicates thatHispanic Whites are not really seen as White. The implied differences of non-Hispanic andHispanic Whites suggest that Hispanics may be an emerging racial group, and that this processof racial construction is related to both internal and external group pressures.

4.1. Mixed-race and racial stratification

Another important social challenge in the ongoing struggle over the social meaning of racehas been issued from “mixed-race” or “multiracial” individuals. Like Hispanics, this grouphas grown increasingly larger and has been able to successfully argue that it does not fit intothe traditionally rigid set of racial categories employed in the U.S. Census. In recent years,parents of mixed-race children and multiracial individuals themselves, claiming the right toself-identify, have argued that selecting one race forced individuals to deny the actual racialheritage of at least one parent. As a consequence, the one major change introduced in the 2000Census was the method of racial identification. For the first time, individuals were instructed toindicate all the racial categories with which they identified. As long as this classification systemstands, then, people responding with multiple racial categories should not be seen as revealingeither their actual parentage or their full ancestry. Rather, respondents’ choices merely indicatethe groups with which they social identify.

The historical progenitors of contemporary treatment by the Census of “mixed-race” individ-uals can be found in decisions and rules governing the enumeration of “mulattoes, quadroons,and octoroons,” or people with some African American ancestry, as well as the somewhatinversed “blood quantum” enumerations of “mixed blood” American Indians in the early partof the last century. People of mixed African and other heritage were counted for the first time in1850 (Williamson, 1984). Concern with the fluidity of this “mulatto” category was expressedin the instructions for enumerators in 1870, which advised them to be “particularly carefulin reporting the class Mulatto. The word is generic and includes quadroons, octoroons, andall persons having any discernable trace of African blood”. Similarly, in 1860, “half-breeds”were listed separately from Whites and Indians. In 1870, Census Bureau officials again de-bated this issue, questioning: “shall they be classified with respect to the superior or to theinferior blood?”(Rodriguez, 2000, p. 74). It was decided that the “one-drop” criteria appliedto the “former slave population” should not be applied to Indians. Ultimately, in what wasperhaps the first recognition of the socially-situated nature of racial meaning, the Census choseto classify “half-breeds” as White if they lived among Whites, and Indian if they lived amongIndians.

4.2. The strange history of race in social science research

Given the convoluted and contradictory understandings of race throughout society, the treat-ment of race in the social sciences is perhaps not so strange. Race and racial identity are both

244 A. James / Race & Society 4 (2001) 235–247

treated by social scientists variously as a firm and intrinsic characteristic, or as a fluid and situ-ational characteristic. Researchers on race in the social sciences can be divided into two majorgroupings: those who study race and racial dynamics, and those who routinely use the conceptof race in their studies. This division largely mirrors the division in the conceptualization of race.

Those who study race and ethnicity as social phenomena understand race as dynamic andsituational (Espiritu, 1992; Hunt, 1997; Omi & Winant, 1994; Waters, 1990). In this vein ofresearch, race is seen as a profoundly social characteristic. The dynamism and fluidity of raceis often used to better understand related social processes. For example, in his study of the 1992Los Angeles uprisings,Hunt (1997)focused on the process by which race is reconstructed inmedia as well as the ways in which media perceptions are shaped by race. Similarly, in herstudy of Asian Americans,Espiritu (1992)focuses on the process by which pan-ethnicity, orbroader ethnic bonds, is forged in response to both political threats and opportunities. In eachof these examples of contemporary race scholarship, race is seen as an ongoing social process.Racial delineations, according to this conceptualization, result from historical patterns of racialhierarchy imbedded in ongoing interactions, both spectacular (as in the case of a civil uprising)and mundane.

On the other hand, those whouserace or ethnicity in their research, as opposed to thosewhostudyrace, tend to treat it as a primordial, or fixed characteristic. (For an extended criticaldiscussion of this tendency in statistical research seeZuberi, 2001.) Despite the widespreadacademic understanding of race as a social construction, most studies using race in quantita-tive analysis treat race as a function of fixed differences between “populations.” In this way,race is most often conceptualized as acauseof a myriad of social processes and distinctions(Zuberi, 2001). From SAT scores to out-of-wedlock births, race is presented as part of a causalexplanation for various social distinctions.

Essentialized understandings of race are frequently made rather explicitly, as in the1994book by Herrnstein and Murray, which argued that racial differences in IQ reflected differ-ences in genetic endowment between groups, or as in a recent article inDemographythat foundenvironment strongly influenced racial differences in birth weight and then argued that an un-measured “genetic component” played a role in determining racial differences in birth weight(van den Oord & Rowe, 2000). Most commonly, however, the understanding of race as a fixedcharacteristic is expressed implicitly, as in an article in a recent issue of theAmerican Socio-logical Review, which presented an analysis in which a dummy variable distinguishing “Blackcouples from non-Black couples” was included as an “independent” or causal variable explain-ing marital dissolution (South, 2001). In another example from the same journal, researchersused race as a “control” variable to explain variation in children’s well-being. In all of theseexamples, the methods by which race affects social outcomes are not examined directly, andthus true causal relationships are not determined. In the first two examples, genetic differencesare identified as racially distributed and these genetic differences are considered the causefor the racial differences in the various outcomes under consideration. In other words, race isconceptualized as a fixed characteristic, rooted in biological or genetic differences betweeneasily distinguishable “groups.”

Explicit arguments about the presumed biological or genetic basis of race are not normallyarticulated. In this sense, the second set of examples is far more ubiquitous in the social sci-ences. In each of this second set of examples, race is used as an independent variable that

A. James / Race & Society 4 (2001) 235–247 245

helps explain some other social phenomenon. Often race is presented as a “demographic” or“control” variable, implying a theoretical neutrality not supported by the substance of the ar-guments or techniques used in the research. Through such usage, race has become, to use abad pun, the black hole of social scientific research. To illustrate this point with recent issuesof major research journals, in the Fall 2001 issues of theAmerican Sociological Review, De-mography, Social Forces, andAmerican Journal of Sociologyfully 67% of the articles thatmention race use race as a disembodied variable in a regression model. The use of race as acontrol variable flattens out the meanings of racial differences and replaces it with a genericnotion of difference. This technique represents a seemingly atheoretical and presumed neutralusageof race. However, using race as an independent variable without any contextualization orexplanation implies that the causes for the social differences under study exist within thepop-ulationsdefined by the various categories. This is most definitely not a “neutral” assumption.”When race is presumed tocausedifferences in family behavior, test-taking, and psychologicalwell-being—that is, without comment or argument about how or why the experience of racein U.S. society may result in different outcomes for individuals who face different racializedexperiences—the conceptual understanding of race as a fixed characteristic is being promoted.

5. Conclusions

Virtually everyone in America is of mixed “racial” origins. The long and continuing historyof race “mixing” does not imply that such mixing will solve the race problem. In fact, anextensive mulatto population was enumerated as early as the Census of 1850 (Williamson,1984). High proportions of Black Americans are thought to have some White ancestry, and anon-trivial proportion of White Americans likely have some Black ancestry. However, suchlongstanding genealogical legacies have not dismantled the Black–White divide in social, po-litical and economic life. Similarly, history suggests that simply changing the system of racialclassification will not reform the way people understand race. Census racial classifications havechanged with almost every decade without sparking a social movement, or even a perceptibleshift in common understanding. However, classifications and the struggle over racial classi-fication are not meaningless. There are important interrelationships between racial ideology,racial classification, and social scientific research using race (Fig. 2).

While I view racial ideology as logically preceding classification schemas and research,each continues to influence the others in whatOmi and Winant (1994)describe as the ongoingpolitical contestation over racial meanings.

This essay has focused on the use of race in studies and statistical models of individualcharacteristics and behavior. I argue that social scientists, in particular those using statisticalmodels that include race as an independent variable, have contributed to the conceptualizationof race as a fixed characteristic. There are several ways to incorporate a more socially andhistorically situated view of race into models of social behavior. First and foremost, the use ofrace as a “demographic, control” variable must be contested. The atheoretical use of race as ademographic “background” variable is misleading and promotes an essentialized understand-ing of race by hiding the social underpinnings of presumed racial differences. When racialdifferences are found in a model of social behaviors, or attitudes, I argue that it is incumbent

246 A. James / Race & Society 4 (2001) 235–247

Fig. 2. Conceptualizing the linkages: race, categorization, and statistics.

upon the researcher to critically examine possible causes for this difference. Instead of merely“controlling” for the difference of the aberrant “others,” racial differences should be assessedand grounded in the set of historical and social circumstances that give meaning to the raceconcept. To do otherwise is to suggest that differences are rooted in the characteristics or“nature” of the groups themselves.

The use of separate models and interaction terms may help establish the areas in whichbroader patterns of social processes are similar and different based on racial group iden-tity/membership. However, the most important step toward incorporating a more groundedconceptualization of race in social research is to adequately theorize race as a social and ahistorical fact in our thinking about racial effects.

Notes

1. Notably, by 1950 enumerators were given the following instructions: “A person of mixedIndian and Negro blood should be returned as a Negro, unless the Indian blood verydefinitely predominates and he is accepted in the community as an Indian” (U.S. Bureauof Census, 1953). This illustrates the peculiar marginalization of African Americans aswell as Indians.

2. In the 1930 Census, the Census Bureau listed “Mexican” as a possible answer to the“color or race” question (Anderson & Fienberg, 1999).

References

Anderson, M., & Fienberg, S. E. (1999).Who counts? The politics of census taking in contemporary America. NewYork: Russell Sage.

Cohn, D., & Fears, D. (2001, March 7). Hispanics draw even with blacks in the new census.Washington Post.Davis, F. J. (1991).Who is Black?: One nation’s definition. University Park, PA: The Pennsylvania State University

Press.Degler, C. (1986).Neither Black nor White: Slavery and race relations in Brazil and the United States. Madison,

WI: University of Wisconsin Press.

A. James / Race & Society 4 (2001) 235–247 247

Espiritu, Y. L. (1992).Asian American panethnicity: Bridging institutions and identities. Philadelphia, PA: TempleUniversity Press.

Goldberg, D. T. (1997).Racial subjects: Writing on race in America. New York: Routledge.Guinier, L. (1995). The representation of minority interests. In P. E. Peterson (Ed.),Classifying by race(pp. 21–49).

Princeton, NJ: Princeton University Press.Hacker, A. (1992).Two nations: Black and White, separate, hostile, unequal. New York: Scribner.Hall, S. (1988). New ethnicities. In K. Mercer (Ed.),Black film, British cinema. London: Institute of Contemporary

Arts.Harris, M. (1964).Patterns of race in the Americas. New York: W.W. Norton.Herrnstein, R. J., & Murray, C. (1994).The bell curve: Intelligence and class structure in American life. Touchstone

Books.Hunt, D. M. (1997).Screening the Los Angeles “riots”: Race, seeing, and resistance. Cambridge: Cambridge

University Press.Jordan, W. D. (1974).The White man’s burden: Historical origins of racism in the United States. London: Oxford

University Press.Marks, J. (1994). Black, White, other: Racial categories are cultural constructs masquerading as biology.Natural

History, 1, 32–35.Morning, A. (2000). Who is multiracial?Sociological Imagination, 27, 209–229.Myrdal, G. (1944).An American dilemma. New York: Harper and Bros.Omi, M., & Winant, H. (1994).Racial formation in the United States: From the 1960s to the 1990s. New York:

Routledge.Rodriguez, C. E. 2000.Changing race: Latinos, the census, and the history of ethnicity in the United States.New

York: New York University Press.Rodriguez, C., & Cordero-Guzman, H. (1992). Placing race in context.Ethnic and Racial Studies, 15, 523–542.Skerry, P. (2000).Counting on the Census?: Race, group identity, and the evasion of politics. Washington DC:

Brookings Institutional Press.South, S. J. (2001). Time dependent effects of wives employment on marital dissolution.American Sociological

Review, 66, 226–245.U.S. Bureau of the Census. (1953).Census of the population: 1950, population, characteristics of the population,

part I (Vol. 2). Washington DC: US Government Printing Office.U.S. Bureau of the Census. (2001).Summary file 2—California. Washington DC: US Government Printing Office.van den Oord, E. J. C. G., & Rowe, D. C. (2000). Racial differences in birth health risk: A quantitative genetic

approach.Demography, 37 (3), 285–298.Waters, M. C. (1990).Ethnic options: Choosing identities in America.Berkeley, CA: University of California Press.Williamson, J. (1984).New people: Miscegenation and mulattoes in the United States.New York: New York

University Press.Zuberi, T. (2001).Thicker than blood. Minneapolis: University of Minnesota Press.