Social Construction, Biology, and Race

How should one understand the connection between the genetics of ancestry and racial identity? This is a new version of the old debate about whether distinctions among races and the concept of “race” itself are biologically based, purely human inventions, or some combination.

Although the idea has been extant at least since the nineteenth century, social constructivism now predominates among social scientists. The American Anthropological Association’s (AAA) 1998 Statement on Race depicted it as “a worldview, a body of prejudgments that distorts our ideas about human differences and group behavior” (p. 713). The AAA argued that scholars should abjure its use (American Anthropological Association 1998).^{Footnote 2} Many geneticists and medical researchers concur; as Luca Cavalli-Sforza and his co-authors put it in the pathbreaking History and Geography of Human Genes: “Classification into races has proved to be a futile exercise. . . . The level at which we stop our classification is completely arbitrary. . . . We can identify ‘clusters’ of populations . . . [but] at no level can clusters be identified with races, since every level of clustering would determine a different partition and there is no biological reason to prefer a particular one” (Cavalli-Sforza et al., Reference Cavalli-Sforza, Menozzie and Piazza1994, p. 19).

Nevertheless, the historically stronger association between biology and race persists, in a new version. Researchers are exceedingly careful when they link race to genetic inheritance, but some nonetheless find it scientifically necessary to do so. One professor of medicine, for example, argues that “the debate . . . should not be over the existence of population differences, but how to describe those differences with more precision. . . . Railing against what some claim are misguided efforts to use racial, ethnic, or geographic distinctions does not make the differences disappear” (Cohn Reference Cohn2006, p. 553). Even Cavalli-Sforza and his co-authors provide color maps showing “very distinctly the [genetic] differences that we know exist among the continents: Africans (yellow), Caucasoids (green), Mongoloids, including American Indians (purple), and Australian Aborigines (red)" (Cavalli-Sforza et al., Reference Cavalli-Sforza, Menozzie and Piazza1994, p. 136).^{Footnote 3}

Firms selling DNA tests straddle this divide; they state explicitly that results will not identify the user’s race or ethnicity, but some of their language is more ambiguous. For example, geneticists Mark Shriver and Rick Kittles point to a “potential negative consequence” if data about ancestral categories are “misinterpreted as indications of ‘real’ racial divisions”—but they also assert that a personalized genetic history “has the potential to affect ethnic, religious, racial, and family identities” (Shriver and Kittles Reference Shriver, Kittles, Koenig, Lee and Richardson2008, p. 209). DNA Spectrum, an online seller of ancestry tests, advertises that its “propriety Full Spectrum Plus test [an autosomal DNA test—see below] has helped customers like James Shoemaker discover their true heritage” (PRWeb 2013). Thus a person choosing to do a DNA ancestry test, or confronted with questions about her views of such a test, is implicitly participating in a long-standing discussion that can be as subtle and as sophisticated as Shriver and Kittles’, or as blatant as eugenicists’ claims about ethnic bloodlines.

Singular and Multiple Ancestral Lineages

The decision to take a DNA ancestry test leads to another decision, about which type of test to take. In very quick compass, one has two basic alternatives (although, unlike Robert Frost’s traveler, one can take both roads^{Footnote 4} since the tests are not empirically or analytically incompatible).^{Footnote 5} Autosomal (non-sex-linked) tests seek to determine the proportions of a person’s ancestry from different regions of the world. Such a test might show one hundred percent ancestry from one population group, but given human migration and marital patterns, it is likely to show descent from more than one region of the world. Mitochondrial (also known as mt-DNA) tests and Y-DNA tests, in contrast, determine a particular ancestral lineage, often described as a tribe or indigenous population, from the paternal line for men or maternal line for women. This test does not show population group mixtures since it provides information about one line of forebears.

Thus a person taking a DNA test necessarily participates in a second old debate on the relative merits of attachment to a single group versus embrace of multiple ancestries or identities. Like willingness to grant legitimacy to DNA tests, this is a political and moral as well as personal and idiosyncratic choice. Endorsement of singular racial identity and clear boundaries between groups goes back at least to the racial nationalism of David Walker and Martin Delany, the inventors of blood quantum to determine Native American status, and (with a different political connotation) the White Citizens League. For Blacks, singular racial identity may be associated with commitment to the betterment of the group (Shelby Reference Shelby2005) and pride in one’s previously denigrated heritage (Joseph Reference Joseph2007). The slogans “Black is beautiful” and “Black power” arose at roughly the same time, often among the same people, and usually conveyed the same political message: “[F]rom now on we shall view ourselves as . . . Black people who are in fact energetic, determined, intelligent, beautiful, and peace-loving” (Ture and Hamilton, Reference Ture and Hamilton1992, p. 38).^{Footnote 6}

Support for multiracial complexity and the blurring or dissolution of boundaries has different though equally robust historical roots: de Crèvecoeur’s celebration of “the American, this new man”; Israel Zangwill’s melting pot; and José Vasconcelos’ raza cósmica. In the African American population, engagement with racial mixture has ranged from shame and outrage at slave owners’ exploitation to pride in “good” hair and light skin color to endorsement of the right to celebrate the breadth of one’s heritage, whatever it is. As Barack Obama put it, “I can no more disown [the Rev. Jeremiah Wright, Jr.] than I can disown my White grandmother;” he once observed that the kind of dogs he would most like to give his daughters would be “mutts like me.” Henry Louis Gates, Jr. makes the same point, more exuberantly: Now that “we know [through DNA testing] that Jane Gates’ children were fathered by an Irishman, I’m in pursuit of him twenty-four hours a day. He’s my great-grandfather. I want nothing more than to find his identity and travel to Europe and claim his people for my own” (Morton 2012).

Biology and Race

DNA ancestry testing draws one, willy nilly, into the issues of whether biology has any relationship to racial identity and whether one should focus on a discrete, singular racial identity or on a heterogeneous mixture of group identities. Consider the first issue first. We know of no systematic theories, never mind a developed literature, on why people might endorse, accept, or resist the idea of an association between race and biology, so we provide the initial outline of such a theory and its links to DNA ancestry testing and African Americans below.

If a person believes it to be an indisputable fact that race is purely a social construction, he or she is likely to reject DNA ancestry testing as misleading.^{Footnote 7} Similarly, if a person is a social constructivist out of ideological commitment or disgust at the uses to which purported racial science has been put, he or she is likely to reject DNA ancestry testing as dangerous. Conversely, a person who is willing to do a DNA ancestry test or who gives credence to its results can be assumed to accept the legitimacy of some sort of association between race and biology.

Given the uses to which racial science has been put, one might expect African Americans to reject any suggestion of such an association, and therefore to reject DNA ancestry testing. However, we theorize that Blacks might instead be drawn to DNA ancestry testing, perhaps even more than members of other groups are, and therefore willing to overcome any lingering fears of racial science. Our “broken chains” theory has two elements.

The first element is shared with other Americans—the profound urge to find one’s roots. As Businessweek puts it, “today, genealogy ranks second only to porn as the most searched topic online” (Falconer Reference Falconer2012, p. 1). Forbes magazine reports that the genealogy industry is worth $2 billion a year (Anderson Reference Anderson2014), with the website Ancestry.com reporting over two million paying subscribers as of 2012 (Farnham Reference Farnham2012). Groups vary in the geographic stability and therefore traceability of their ancestry, with American Blacks at the end of the continuum of difficulty. Unlike in other groups, the vast majority of Blacks whose ancestors came to America before the twentieth century are almost always unable to use conventional records to trace their antebellum ancestry. Slave ships counted but did not name their cargo. The United States census recorded only the number of slaves held by a (named) White owner through 1840, and only occasionally included (first) names in the 1850 and 1860 censuses. People changed their names or their names were changed regardless of their preferences. Few Blacks owned land or were on the tax rolls, many Black marriages and births were never officially recorded, few attended formal schools, and so on—written records are scant and incomplete at best. So African Americans typically find popular genealogical sites such as Ancestry.com to be useless beyond a few generations.

The second element in the broken chains theory is distinctive, perhaps unique, to Black descendants of enslaved Africans. Finding one’s ancestry can represent a type of victory over subjugation by Europeans:

Thus one woman cried for joy when discovering her Kenyan ancestry. “She carried a message from her ancestors in every cell [sic] of her body, in her own DNA. It was a secret document, a talisman of her African roots. . . . It had slid unseen past slave owners who had strived to erase any sense of individual identity from her ancestors” (Sykes Reference Sykes2012, p. 135). By this logic, the deeper the erasure of history and family through enslavement, the more precious is the knowledge that can be gained only from DNA ancestry testing. “For 400 years, to have no clue as to your name, culture, heritage, where you came from—how could you not search? You need to find out who you are, just to get that fulfillment, even if it is just for a day, a second” (Ataiyero Reference Ataiyero2007, p. A5).^{Footnote 8}

The broken chains theory generates our first hypothesis:

Singular or Multiple Heritage

As with the issue of links between race and biology, the extensive literature on racial mixture has not produced a systematic theory about when or why people will prefer singular or plural group identities, or how that preference connects to DNA ancestry testing. So we begin with the broken chains theory and develop it in a new direction in order to generate a testable hypothesis about views on racial mixture.

We expect two sets of people to be especially committed to a singular, or pure, group identity. The first set includes those whose social superiority depends in large part on belonging to a disproportionately favored group. Examples include Brahmins in the caste system of sub-continental India, the Cohanim among observant Jews, and Anglo-Saxon Whites through much of American history. The second set includes those reacting with anger and solidarity against the social inferiority predicated on belonging to a disproportionately disfavored group. Examples include homosexuals embracing the pejorative appellation of “gay” or “queer,” the Roma across much of Europe, and Black Americans. In contrast to both sets, members of groups with mixed, mid-range, or highly variable social status are less likely to place a great deal of stock in singular or pure ancestry.

Following this logic, we expect African Americans who are interested in DNA ancestry testing to be especially focused on and responsive to tests promising knowledge of singular African ancestry. This could either be an autosomal test that shows mostly African descent, or a mitochondrial or Y-DNA test that explicitly traces an African lineage.

A strong commitment to singular racial identity need not conflict, either logically or empirically, with acceptance of or enthusiasm about racial mixture. Analysts who embrace multiracialism argue that it permits a fluid racial identity, such that one’s self-definition can vary with context or preferences.^{Footnote 9} That perspective seems less plausible for those whose social superiority depends on racial purity than for those who resist social inferiority; the former will lose status if they admit to multiple heritages, while the latter may gain status, or at least not lose it. Thus, African Americans who are interested in DNA ancestry testing can take pride in their African heritage without having to reject the possibility of non-African ancestors. Hence our second hypothesis, which emerges from an extension of the broken chains theory:

Expanding Awareness

If only a few Americans are aware of the revolution in genomic science, or of genetics more generally, DNA ancestry testing will be of little interest except to a handful of specialists. But that is not the case.

Three initial analyses show that public attention to biology and genetics is growing or reasonably high. The first uses published books. A Google Books Ngram (Figure 1) shows the diffusion of the words “genetic,” “DNA,” “biological,” “gene[s],” and “genomic” in books published between 1940 and 2008.^{Footnote 10} As Figure 1 shows, the proportion of books with all of these terms, except the specialized “genomic,” increased markedly over those seventy years.

Fig. 1.

Google Ngrams for “DNA,” “Genetic[s],” “Biological,” “Gene[s],” and “Genomic[s],” 1940-2008

Note: The y-axis represents the percentage of Ngrams that contained the above words out of the universe of Ngrams in English-language books.

The second piece of evidence about public attention to biology and genetics uses newspapers rather than books; it focuses specifically on genomic science and race or ethnicity. We conducted an automated content analysis of what we believe to be all newspaper articles published in the United States on the topic of genomics and race or ethnicity through 2009.^{Footnote 11} We used the broadest available database on the grounds that it would best show how the print media are reporting this topic. For the same reason, we retained duplicates; an article’s appearance in more than one newspaper is directly relevant to the question of public attention to biology and genetics. We erred on the side of being over-inclusive in order not to bias our corpus of texts toward or against our expectation of increasing attention, or toward or against our second hypothesis on the singular or multiple framing of DNA ancestry.

Specifically, we searched Lexis-Nexis Academic for texts using a genetics-related term (“DNA,” “genetic,” or “genomics”)^{Footnote 12} within the same sentence as a word associated with race, ethnicity, or heredity (“race,” “racial,” “ethnicity,” “genealogy,” “descendant,” “ancestor,” or “lineage”). We added keywords to exclude articles clearly outside our inquiry, such as those on rape or murder forensics, genetic food modification, television schedules, and death notices. We avoided words that would slant the search toward any particular group (e.g., “Black,” “White,” “African,” “Celtic”). The final search yielded 5580 newspaper articles from 1969 through 2009; their yearly frequency is shown in Figure 2.

Fig. 2.

Newspaper Articles per Year Addressing Race or Ancestry and DNA, 1988-2009^{Footnote 38}

The curve in Figure 2 resembles that of Figure 1; the annual number of newspaper articles linking DNA and race rose from fewer than 50 in 1990 to as many as 500 by the late 2000s. The number of relevant articles peaks around 2006, coinciding with Gates’ first public television documentary on prominent African Americans’ use of genetic testing. Journalistic interest remained strong through 2009.^{Footnote 13}

Perhaps the best way to determine if someone is paying attention to an issue is to ask him or her directly; that is our third measure of awareness of genetics. By that criterion, 13% of respondents to the 2010 GSS have heard “a great deal” about genetic testing, and an additional 66% have heard “something” about it.^{Footnote 14} Similarly, in 2007, 49% of survey respondents claimed to have thought “a great deal” or “a small amount” about “issues to do with genes or genetics” over the previous few months (VCU Center for Public Policy 2007). We have no way of knowing if these figures are accurate; they are best understood as an indicator that half (in 2007) to four-fifths (in 2010) of Americans think they should be paying attention to genetics.

Media Presentations of Singular and Multiple Heritage

Given that Americans are learning about genetics—or at least books and newspapers are providing more information and Americans think that they should be learning about genetics—we can move to the more focused question of what they are learning. The database of 5580 newspaper articles provides a lens through which to view the initial stages of public learning about links between the new genomic science and the construction or discovery of racial and ethnic categories.^{Footnote 15} In particular, the database enables a determination of whether the distinction between group singularity and multiplicity, on which H2 is based, is broadly meaningful.

We sorted the media reports of links between DNA ancestry tests and race or ethnicity into two bins: (1) those that emphasize multiplicity by focusing on group mixture or the softening of distinctions among conventionally defined races or ethnicities, and (2) those that emphasize singularity by focusing on reinforcement of one ancestral line or single-group identity, or on the hardening of distinctions among conventionally defined races or ethnicities (we also created two other bins for articles that do both, and articles that do neither). Since we could not code 5600 articles by hand, we used ReadMe (Hopkins and King, Reference Hopkins and King2010), an automated content analysis method for the R statistical computing language. Two research assistants and the authors hand coded over 600 articles selected across the full period, dividing them into four mutually exclusive categories. This was the “training set.” We ran the content analysis program on the training set, thereby allowing the program to use the proportion of documents observed with each pattern of word profiles to make predictions. We then used the same classification method to analyze all 5580 articles (the “test set”). The program estimated the proportion of articles in each substantive category with good accuracy (Hopkins and King, Reference Hopkins and King2010).

The four categories and the coder instructions were:

For example, a description of recent research that included analysis of a newly discovered gene shared among humans from all regions of the world would be coded as “Blend,” since it conveys the idea that DNA points to a shared human origin. In contrast, an article discussing the relatively high prevalence of breast cancer among Jewish women and speculating that genetic factors unique to Jews are linked to the increased risk would be coded as “Strengthen,” since it implies that genomic science reveals biological distinctions among racial or ethnic groups.

Table 1 shows the estimated proportion of articles in each category, as generated by the automated content analysis. About three-tenths of the articles framed DNA ancestry testing in terms of singular racial identifications, while a fifth focused on multiplicity across racial groups. Another tenth addressed both concepts.^{Footnote 16} Methodologically, these results show that the dichotomous framework of plurality and singularity is meaningful and useful. We can effectively hand code, and then automatically code, articles according to this distinction, and the coding yields interpretable and internally consistent patterns. Substantively, the results show that since genomics research has come into public view, both singularity and plurality of heritages have been included in public discourse. However, Americans have been more exposed to the idea that genomic science distinguishes races genetically or that people “belong” to only one group than to the idea that genomic science erodes group boundaries or shows people to have mixed biogeographical origins. This emphasis on racial singularity could have political impact—but whether it reifies harmful and artificial barriers between groups or encourages solidarity and group pride remains to be seen.

Table 1.

Proportion of Newspaper Articles Focused on Plurality, Singularity, or Both, 1969-2009

Figure 3 provides a more dynamic portrait of what newspapers are conveying to Americans about genomics and race. It shows the proportion of articles in each of the three substantive coding categories separately by year from 1988 through 2009. (For ease of interpretation, we removed the fourth category, “Neither,” from the denominator in calculating these proportions.)

Fig. 3.

Proportion of Newspaper Articles on Plurality, Singularity, or Both, 1988-2009

The proportion of articles discussing both multiplicity and singularity (the dashed line) is consistent but low. More interestingly, after the instability due to a small number of articles in the early 1990s settles down, in almost every year the articles focus as much or more on a single racial group (dark line) as on mixture or multiple groups (dotted line). The first exception is 2001, which is probably due to publicity about the new mixed-race option on the U.S. census, and the public unveiling of the Human Genome Project. J. Craig Venter’s statement in his White House speech that geneticists have shown that “the concept of race has no genetic or scientific basis” (Venter Reference Venter2007, p. 315) was widely noted in that year, but Figure 3 shows that its message did not last in the media. The only other exception to the dominant message about racial singularity is 2009, which is probably due to publicity about President Obama’s racially mixed ancestry. ^{Footnote 17}

H1: Views of DNA Ancestry Testing

The Ngrams, newspaper content analysis, and survey results show that the media are providing Americans with information about genetics and the links between genetics and race. They also show that a considerable portion of Americans is aware of this new endeavor. The distinction between singular and multiple ancestries in newspaper stories reveals a meaningful structure in ongoing public discussions about genomics and race or ethnicity. That set of findings provides a sufficient empirical foundation to enable us to turn directly to the hypotheses. We focus first on H1: African Americans are interested in, persuaded by, and influenced by DNA ancestry tests—and are more interested, persuaded, and influenced than are members of other American races and ethnicities.

Initial support for H1 comes from our 2010 GSS module. So far as we can tell, this is the first nationally representative survey to ask views on the use of DNA to research one’s ancestry. According to the survey, 63% of the respondents were “strongly” or “somewhat” favorable to the use of DNA for this purpose and only 9% were unfavorable. Blacks were more likely than Whites and Hispanics to be “strongly” favorable (proportions were 42%, 29%, and 39%, respectively), as well as more likely to “somewhat” favor it. Overall, three-fourths of Blacks, compared with just over three-fifths of Whites and Hispanics, approved of DNA ancestry testing.

The GSS result is clear, but thin. For a richer understanding of just what people approve and why, we turn to a new vignette survey, first reported here. It included a random sample of 1095 American adults, divided among 242 non-Hispanic Whites, 201 non-Hispanic Blacks, 233 non-Hispanic Asian Americans, 205 Hispanics, and 214 non-Hispanic multiracials.^{Footnote 18} All respondents identified their own race or ethnicity. The survey was conducted over the Internet by Knowledge Networks (now part of GfK) in August 2010, through the peer-reviewed Time-sharing Experiments for the Social Sciences (TESS) program.

In order to examine difficult concepts with which few people have direct experience, we developed vignettes (Finch Reference Finch1987) featuring imaginary people who have just received the results of a DNA ancestry test. Each participant responded to four vignettes. The respondent’s gender always matched that of the imagined subject. Two vignettes (one each for racial plurality and racial singularity) asked the respondent to imagine that he or she was the vignette individual; in these cases, the respondent’s pre-test race (including Hispanic as a separate category) matched the pre-test race of each imaginary subject. The other two vignettes (again, one each for racial plurality and racial singularity), asked the respondent to imagine how the vignette individual would feel after receiving test results. Here the respondent's pre-test race differed randomly from the pre-test race of each imaginary subject.

These variations allow for an analysis of the internal validity of responses to the vignettes. That is, given the United States’ fraught racial history, we anticipate that meaningful answers will vary in intelligible ways depending on whether the respondent is considering a vignette subject with the same or a different race from his or her own. Furthermore, people may consider their own experiences differently than they think about the experiences of others. For example, a Black respondent might convey more enthusiasm when asked to identify with a Black vignette character than when asked to assess the likely level of enthusiasm held by someone of a different race.

Each vignette had the same three sets of response categories, addressing: (1) emotional reactions to the ancestry test on a seven-point scale from “pleased” to “displeased”; (2) cognitive responses on a seven-point scale from “believable” to “unbelievable”; and (3) the test’s impact with two alternatives, “would the results matter a lot to your identity” or “not matter at all to your identity.”^{Footnote 19} For each vignette, the screen showed a map of the world with the relevant continent(s) highlighted, as in Figure 4.

Fig. 4.

Map Showing Highlighted Continents, Vignette Survey, 2010

Table 2 provides examples of how the vignettes appeared to a respondent.^{Footnote 20}Table 3 presents results for the emotional response of “pleased,” after collapsing the first three points on the seven-point scale.^{Footnote 21} H1, which posits that Blacks are especially favorable to DNA ancestry testing, receives strong support. The simplest measure makes the point most clearly. Averaging across the four vignettes, 51.5% of both Blacks and Hispanics would be pleased with either result or would expect others to be pleased with either result. This compares with 43.8% of multiracials, and only 38% of Asians and 37% of non-Hispanic Whites. Using simple Chi-square tests, overall differences in response categories between Blacks and Whites are significant at the 0.01 level (with a test statistic of 51.5); the differences between Whites and Hispanics are also significant at the 0.01 level (with a test statistic of 52.1).

Table 2.

Illustrative Items in the Vignette Survey, 2010

Table 3.

Emotional Reactions to Vignette DNA Test Results, 2010 (% of Respondents in Each Cell)

Dividing the results in Table 3 between vignettes that present the respondents’ views of how they would feel about a DNA ancestry test (row 1) and those that present respondents’ expectation of how others (from a randomly selected group) might feel (row 2) provides a bit of nuance while reinforcing H1. With regard to their own view, averaging across the two conditions of singular and plural test results, a higher proportion of Blacks (53%) express pleasure than is the case in any other group. Hispanics are close behind at 49%, but the proportions decline to a third among Whites.

For the question prompts looking only at the multiple-ancestry narrative, the differences between Blacks and Whites, and Hispanics and Whites, are all significant at conventional levels of statistical significance, as determined by a series of weighted Chi-squared tests. For the White-Black comparison under the multiple-ancestry prompt, the weighted Chi-square statistic was 8.06 (p-value = 0.018) and the White-Latino comparable Chi-square statistic was 8.60 (p-value = 0.013). For the question prompts looking only at the singular-ancestry narrative, the differences between Whites and Blacks, and between Whites and Hispanics, are only narrowly non-significant.

With regard to expectations of others’ reactions, a slightly smaller proportion of Blacks, and slightly more than half of Latinos (50% and 54%, respectively), anticipate that others would feel as they do. Some African Americans, that is, perceive their own group’s exceptionalism.^{Footnote 22} The share of multiracials, Whites, and Asians who anticipate others’ pleasure with DNA ancestry test results range downward to 38% among Asian Americans. However, these differences are not significant at the standard levels.

Some Blacks are indifferent to the results posited in the four vignettes, but only an average of 11% would be displeased with the outcomes of the enterprise (that is the same or marginally fewer than in any other group; results not shown).^{Footnote 23} A majority of African Americans surveyed, in sum, do not harbor mistrust toward DNA ancestry testing—instead, they hold positive views of it. They are not associating DNA ancestry testing with the old and dangerous tropes of eugenics or racial science; it seems reasonable to infer that they perceive it to be a way of restoring the links in the broken chain of genealogy.

We turn next to the second set of responses to the vignettes, asking whether respondents would meet DNA test results with skepticism or acceptance. The results do not contradict H1 but they support it only marginally (data not shown). There is no substantial difference across the groups in agreement on whether the DNA test results would be believable if the respondent were the person in the vignette;^{Footnote 24} the average proportion saying “yes” varies only from 59.5% for Asian Americans to 62.5% for multiracials and Blacks. On the vignettes about whether a target person of another race would find the test results believable, the proportion saying yes ranges from 58% for multiracials to 64.5% for Whites; Blacks are in the middle, at 61.5% saying yes. At the other end of the scale, Blacks and multiracials are less likely to say that the results are not believable than are the other three groups (14% and 13% for Blacks and multiracials, respectively, and 17% to 19% for the other three groups). Nonetheless, a series of Chi-square tests show no meaningful differences between Blacks and Whites, Hispanics and Whites, and also across all of the groups in levels of belief. In sum, a solid majority of respondents in all groups trust the outcome of a hypothetical DNA ancestry test—a result reassuring in terms of external validity of the survey but not strongly supportive of H1 about Black exceptionalism.^{Footnote 25}

We turn finally to the third set of responses to the vignette, which queried respondents’ views about the impact of the DNA ancestry test result on their self-understanding or identity. Our goal for this item was to operationalize the rather abstract question of whether Americans see race or ethnicity as purely a social construction or whether, in their view, it has a meaningful biological or “factual” component. Since this is the most cognitively difficult item, we offered only two stark choices rather than a 7-point range: would the test results “matter a lot/not matter at all to your [or to his/her] identity?” Table 4 shows the results.

Table 4.

Impact on Identity of Vignette DNA Test Results, 2010 (% of Respondents in Each Cell)

The results support H1. Averaging across the four vignettes asked of each respondent, 45% of African Americans agree with the strong statement that the test results would “matter a lot” to identity. That compares with 41.5% of Latinos, 36.5% of Asian Americans, 32.5% of multiracials, and only 24.5% of Whites. These differences are highly statistically significant. For example, for the multiplicity prompt asking respondents to identify with the vignette character, the differences between Blacks and Whites (Chi-square statistic = 18.9, p-value = 0.00), Hispanics and Whites (Chi-square statistic = 12.7, p-value = 0.00), and across all of the groups (Chi-square = 28.9, p-value = 0.00) are significant at conventional levels. The desire not to be associated with the old racial science probably leads some Whites to deny that genetic inheritance affects identity—but social desirability seems unlikely to explain why almost twice as many Blacks as Whites agree that knowing one’s genetic heritage would strongly affect one’s self-understanding. Other responses to this survey have reassured us about its external validity, so the relatively and absolutely high level of Black agreement that DNA tests matter to identity seems worth taking seriously.

As before, looking at rows 1 and 2 separately helps to refine the general finding of support for H1 that we have seen so far in Figure 4. Members of all groups except African Americans are more likely to claim that DNA test results would matter to other people than to admit that they would matter to themselves. Put the other way around, Blacks are the only group among the five to agree that they would be as much influenced by ancestry tests as they would expect others to be. The consistency of this pattern is reassuring from the perspective of internal validity, since it implies that respondents were meaningfully distinguishing between projecting themselves into the vignette and predicting how others might react. Why both White and non-White Americans (except Blacks) think DNA ancestry tests will matter to others but not themselves remains one of the many topics for further exploration that this research has opened up. The central point is clear, however: As our theory proposed, DNA ancestry tests are especially if not uniquely meaningful to Black Americans, many of whom want to find links that can repair the broken chain caused by the Middle Passage and enslavement.

H2: Singular and Multiple Ancestries

Given evidence of Blacks’ unusually strong interest in DNA ancestry testing, we can turn to H2: African Americans are especially favorable toward tests that reveal a single ancestral line, and are more favorable to racial singularity than are all other groups except White Americans. Unlike White Americans, however, African Americans are also favorable toward DNA ancestry tests that reveal multiple ancestral lines.

As with H1, broadly framed survey questions provide initial support for H2. African Americans have long expressed a strong sense of group solidarity: across ten rounds of the GSS from 1996 through 2014, with a total sample size of 14,000, on average three in five Blacks report feeling “very close” to other Blacks (9 on a 9-point scale). About a third of Whites on average feel equivalently close to other Whites. One must discount the low White results to some unknown degree because of social desirability effects; the only conclusion one can draw with confidence is that of Black solidarity.^{Footnote 26} Similarly, a 2004 GSS item asked which, on a list of ten options, is “most important to you in describing who you are.” Respondents could give three answers; combining respondents’ first, second, and third choices, two-fifths of Blacks and “others” (mostly Hispanic), but only a tenth of Whites, chose their race or ethnicity.

But Blacks’ racial consciousness and solidarity need not imply rejection of multiracialism or of an identity as mixed race. From 1982 forward, every relevant GSS survey has shown stronger support among Black than White respondents for interracial marriage, by twenty percentage points or more in the first decade. Blacks’ support eventually became almost unanimous. In contrast to the opposition of almost all members of the Congressional Black Caucus on this issue, as early as 1995 half of Black, compared with just over a third of White, respondents to a Newsweek poll agreed that “the U.S. Census should add a ‘multiracial’ category to population surveys so some people aren’t forced to deny part of a family member’s heritage by having to choose a single racial category” (authors’ analysis of Newsweek (1995)).

These data underlie the second hypothesis, predicting both Black endorsement of racial singularity and Black acceptance of racial mixture or multiple heritages in DNA ancestry tests. Given the weight of social desirability, we do not anticipate strong support for singular ancestry from Whites despite our theoretical prediction. All of those expectations are borne out in the vignette survey. Consider again Table 3. Averaging across both “self” and “other” vignettes, African Americans are more likely than most other groups to express pleasure at single-continent-of-origin outcomes. Just over three-fifths (61.5%) do so, compared with the same proportion of Hispanics (60%), but only 48% of multiracials and 45.5% of both Whites and Asian Americans.^{Footnote 27} The difference between Blacks and Whites, as we noted above, is highly significant.

However, and crucially for H2, Blacks are also more likely than some other groups to express pleasure at multiple-continent-of-origin outcomes. The proportion is smaller, 41.5%, but it remains higher than among Whites (28.5%) or Asian Americans (30.5%).^{Footnote 28} Support for H2 is reinforced by looking at the other end of the scale, at the proportions expressing displeasure at the purported DNA test results (results not shown). As our theory would suggest, more Whites (17%) and Asian Americans (20%) than Blacks (13%) were displeased with tests showing a plural ancestry for themselves as the vignette subject.^{Footnote 29}Table 3, row 2, shows that African Americans expect members of other races to be displeased with plural results—again suggesting that Blacks see themselves as distinctive, in this case distinctively open to multiple ancestries.^{Footnote 30}

With regard to the question of whether the respondent trusts the DNA ancestry test results, we have already reported very little difference across groups for H1. A similar pattern holds with regard to H2 (results not shown). As we reported above, overall three-fifths of the respondents would find the DNA test trustworthy. There are only minor differences across the five groups with regard to whether respondents would find test results showing either singular or multiple ancestry compelling. Blacks are in the middle of the range.

The most interesting feature of this question is its reinforcement of the pattern found in the newspapers’ portrayal of DNA and race. As Table 1 and Figure 2 showed, the media consistently report more single group than multiple group results in articles on genetics and race. We cannot claim causality here, but we note that across the board, respondents trust single group results even more than they trust plural group results. Overall, 54% find the latter believable compared with 71% who find the former believable; the difference is highly statistically significant. Furthermore, as Table 3 shows, nine of the ten within-group comparisons show that respondents would be more “pleased” with singular than with plural test results in the vignettes. Thus, in addition to revealing systematic differences between Blacks and other groups of Americans, the vignette study shows that DNA ancestry tests, like newspapers, do more to reinforce than to dissolve racial categories in the understanding of the American public.

Returning directly to H2, the results in Table 4, with regard to identity or self-understanding, again show Blacks’ distinctive commitment to singular racial ancestry in combination with their acceptance of multiple heritages. Averaging across the two conditions (how would you be affected, how would another person be affected), Blacks are much more likely than any other group to say that discovery of a singular ancestry would affect their identity “a lot.” Agreement ranged from 47.5% among African Americans down to 24.5% among Whites.^{Footnote 31} The difference in means in responses between Black and White respondents on this dimension is highly significant (t-statistic = 6.94, p-value = 0.00). Crucially for H2, the same result holds for the discovery of a plural ancestry; Blacks anticipate more impact regardless of what the DNA test reveals. Agreement that a multiple heritage would affect identity ranges from 42.5% among African Americans down to 24% among Whites, with a difference-in-means t-statistic of 5.83 (p-value = 0.00).^{Footnote 32}

As our theory predicted, the most historically subordinated group held a stronger commitment to its denigrated racial label than other groups did with regard to their own labels, and it held a stronger commitment to its denigrated racial label than to the proffered alternative. But, arguably due to their desire to mend the broken chain of ancestral connections, Blacks’ positive response to racial singularity was almost matched by their positive response to racial multiplicity.