The Confederate Battle Emblem as Symbol

Before analyzing contemporary opinion about the flag, we turn to the historical record to assess the intended and constructed meaning of the flag’s display in Southern capitals. The battle over Confederate symbolism can be fairly characterized by two diametrically opposed perspectives. The first is that the Confederacy was fundamentally racist, and thus that honoring the Confederacy is racist. This view holds that notions of the honor of the Confederacy are both historically and morally wrong and deserve to be repudiated, and that Southern resistance to repudiation is indicative of the perseverance of the Southern orthodoxy and a continued disregard for African American opinion (Thornton Reference Thornton1996). The second perspective is that the Southern heritage of bravery and idealism—the fight for liberty—is real, and that demands for the repudiation of Confederate symbols needlessly vilify Southern Whites and Southern identity generally and are concomitant to a demand for a new Southern identity in which apology is the central feature. ^{Footnote 1} For proponents of the “heritage” position, denunciation of the Confederacy is concomitant to admission that ancestors of Southerners are “uniquely guilty,” that they are “the Nazis of the American past” (Thornton Reference Thornton1996, p. 234). Each perspective is on display in the current debates over the public display of Confederate symbols—not just the rebel flag, but also of statues and memorials to Confederate soldiers and the like. It is important to note that both sides accuse the other of being willfully ignorant of the relevant historical facts, which nearly dooms any hope for finding a middle ground from the outset.

The most recent Confederate flag controversy has been over the display of the flag itself, but its incorporation into state flag designs in the South has also been controversial. After the Civil War, the Confederate flag was rarely displayed and absent from official state displays until its reappearance in the 1950s, when it was adopted as an element of several redesigned state flags (and also reappeared itself in prominent state-sponsored locations, as until recently in South Carolina). Georgia and South Carolina have had some of the most visible confederate flag controversies. Many claim that the incorporation of the Confederate battle emblem was motivated by racism—most explicitly triggered by the Supreme Court’s 1954 decision in Brown v. Board of Education, which barred racial segregation in schools as an unconstitutional violation of the Constitution’s promise of equal protection under the law. ^{Footnote 2} Others claim that the move was not racially motivated, but rather represents a more benign official recognition of the state’s heritage and history. In this section, we trace the modern roots of the flag’s symbolic deployment and the ideas behind its inclusion and display.

From Appomattox through the 1930s, the Confederate banner was rarely displayed, appearing mostly at memorial tributes for Confederate veterans (Davis Reference Davis1998). During this time, it was apparently not widely produced or available, as even Senator Coleman Livingston Blease (D-SC) had to make a general appeal and have one especially made for him by the Daughters of South Carolina in 1930 (Congressional Record 1930). During World War II it again appeared briefly, this time as a symbol of regional pride among servicemen. Throughout this time, the flag remained little used and largely apolitical (Davis Reference Davis1998; Martinez Reference Martinez2008).

The flag emerged as a potent political symbol in the wake of President Harry Truman’s modest civil rights initiatives in the late 1940s. Many Southern Democrats viewed these measures as an infringement upon their state rights, which ultimately led to the Dixiecrat revolt of 1948. The Dixiecrats enjoyed little electoral success, but they adopted the Confederate battle emblem as their symbol, and their use of the Confederate symbol generated popular interest in the emblem, and forged a strong link between the flag and racism (Davis Reference Davis1998). Shortly thereafter, the battle flag’s increasing salience was “dismissed as a fad” by some, despite an early 1950s “rage for displaying the Confederate battle flag on neckties, windshields, and wherever else a place may be found” (Atlanta Journal and Constitution 1952). In 1951, Rep. John Rankin (D-MS), perhaps the most outspoken racist and segregationist in Congress at the time, proudly observed that he had “never seen as many Confederate flags in all my life as I have observed floating here in Washington during the last few months.” Rankin himself wore a Confederate flag necktie to help drive home his opposition to “indignities” of “Communist” fair employment and the “beastly” integration of schools, playgrounds, and the military (Congressional Record 1951). ^{Footnote 3}

The Confederate Emblem in the Georgia State Flag

The flag appeared much more frequently, and gained official recognitions in the South after the Supreme Court’s decision in Brown v. Board, along with other more concrete forms of resistance, such as approval of a measure to divert state money into private schools in order to avoid integration (see generally Klarman Reference Klarman1994, Reference Klarman2004). The push to change Georgia’s state flag to include the battle emblem began in 1955, when the Association of County Commissioners (ACC) suggested a new state flag design that incorporated the battle emblem. The ACC was a conclave for a number of Georgia’s county commissioners—121 of them, which represented less than one-third of the state’s population, but a majority of the county-unit votes, and as such, was a force in Democratic politics in the state. The flag containing the emblem was designed by the ACC’s attorney, John Sammons Bell, who also happened to be the state chairman of the Democratic party and one of the nation’s staunchest and most outspoken pro-segregationists (Martinez Reference Martinez2008). The ACC adopted and disseminated a series of resolutions suggesting that the Confederate emblem should be adopted because it is “a symbol of loyalty and devotion of a people to [Georgia’s] government,” and is “symbolic of the traditions it represents,” “regardless of what the Federal government or any division thereof says or does” (quoted in Davis Reference Davis1998, p. 317). The ACC’s resolutions stressed that Brown was “an affront and challenge to the traditions of our people...[and] this Association and its members...pledge to the Governor and all public officials of the State, full support in each and every way or means required...to protect and maintain the segregation of the races in our schools” (quoted in Davis Reference Davis1998, p. 317). Notably, the ACC’s resolutions never mention any intention of memorializing Confederate soldiers. The terms “tradition” and “heritage” appeared frequently in that publication, but did so primarily in reference to maintenance of segregation and the county-unit system. The ACC resolution did argue the flag’s design represented both Georgia’s role in the original United States and the “Confederacy under which [their] forefathers fought so bravely and valiantly,” but the document’s emphasis was the usefulness of the symbol, “which has meaning for Georgians,” in the South’s new battle against integration, and soon thereafter, reapportionment (Davis Reference Davis1998). ^{Footnote 4}

The flag’s popular re-emergence and inclusion in official designs was controversial even among Southern traditionalists: some supported it as a symbol of Southern heritage, others opposed the seeming defiance represented by its re-emergence. Indeed, flying the flag was illegal in Washington, DC, and the United Daughters of the Confederacy supported the ban, contending that the “sacred banner was being trivialized and dishonored” by frivolous displays and opposed inclusion in state flags on the basis that the flag “belongs to no one state of the Confederacy and placing it on our Georgia flag…will cause strife” (quoted in Davis Reference Davis1998, p. 313). The Georgia Division of the Children of the Confederacy adopted a similar resolution (Atlanta Constitution 1955). In Georgia, there were also many concerns over the cost of replacing flags at all public buildings, and the editors of the Atlanta Constitution (1955) called for a committee to study the matter, rather than endorse the change. ^{Footnote 5}

There was also recognition of the indignity Black Georgians would suffer in response to the new flag, particularly as state law required it to fly at all schools (including Black schools). The Georgia Federation of Womens’ Clubs resolved that the flag change would be a “backward step toward prejudice and sectionalism.” The editors of the Atlanta Daily World (Atlanta’s Black daily) agreed, noting that Denmark Groover’s “all good Southerners” seemed to leave out “a certain segment of citizens” that would find it difficult to “salute or insist devotion in something that stood for the enslavement of its people and which revives and enlivens [the] cause” of White supremacy. They identified the flag resolution as a “bitter retaliation against…the activation of a real democracy on our home front” (Atlanta Daily World 1956).

The state senate voted on February 1, 1956, to add the Confederate battle emblem to the flag by a vote of 41–3. In that chamber, the bill’s champion was Sen. Jefferson Lee Davis, who spoke of Georgia’s role in the “War Between the States.” The vote in the house was closer: 107 supported the change, 32 opposed, and 66 abstained (Azarian and Fesshazion, Reference Azarian and Fesshazion2000). No legislator (so far as Davis could find to report) stated any intention of memorializing or paying tribute to Confederate veterans. Rather, House leader Denmark Groover argued that the new design had “deep meaning in the hearts of all true Southerners” adding that “anything we…can do to preserve the memory of the Confederacy is a step forward” (Davis Reference Davis1998, p. 325).

Davis argues that an “overwhelming amount of evidence indicates that those who introduced the flag change (and those legislators who voted for it) were not motivated by a desire to offer a memorial to Confederate soldiers, but were influenced by the Supreme Court’s desegregation rulings and by the fear that the Court would find Georgia’s county-unit system unconstitutional” (1998, p. 307–8). ^{Footnote 6} Indeed, Governor Marvin Griffin, who signed the new flag into law, had campaigned on the promise to maintain “Georgia’s two greatest traditions—segregation and the county-unit system” (quoted in Davis Reference Davis1998, p. 307), and Groover, who had guided the change through the state Assembly, later admitted that “the Confederate symbol was added mostly out of defiance to federal integration orders” (Firestone Reference Firestone2001).

Even so, when Governor Zell Miller proposed in 1993 to remove the Confederate emblem from the state flag because of its racist connotations, the flag’s defenders rallied to the familiar cry “Heritage! not Hate!” Polls revealed that most Georgians viewed a change as unnecessary (Leib Reference Leib1995), however, these state-wide polls masked racial variations. Three-quarters of Whites reported that they viewed the flag as a symbol of Southern pride, while more than half of Blacks viewed it as a racist symbol (Reingold and Wike, Reference Reingold and Wike1998). Still, Miller’s effort to replace the flag failed. Indeed, the Georgia legislature responded to Miller’s call by proposing a bill that would discontinue funding for any municipalities that did not display the state flag on state property, and another that would have made it a crime to deface Confederate monuments, punishable by up to seven years in prison (Martinez Reference Martinez2008).

In 2001, Governor Roy Barnes took up the issue of the flag, again proposing to remove the Confederate emblem and replace it with the state seal (the new design also included each of the states’ previous flags in miniature in a banner beneath the seal). Barnes’ supporters revealed the bill on January 21, 2001, and his allies in the legislature called for a vote the same afternoon. The Confederate emblem’s supporters had no time to respond or to marshal public opinion, allowing Barnes’ proposal to narrowly pass (94–82). The measure was approved by the state Senate on January 30 (34–22). Barnes’ flag design flew over the state capitol from 2001 to 2003. The abrupt adoption of the Barnes’ flag caused a political controversy, and was likely a contributing factor in Barnes’ loss to Sonny Perdue in the ensuing gubernatorial election. Making good on a campaign promise, Perdue proposed a referendum which would allow voters to choose between the Barnes’ flag and the pre-1956 state flag—strategically omitting the choice of the flag which featured the Confederate battle emblem. ^{Footnote 7} In the March 2, 2003 referendum, Georgia voters approved the pre-1956 flag over the Barnes flag three-to-one. This flag remains the official flag of Georgia today.

The Confederate Emblem in South Carolina

The Confederate battle emblem was raised over the South Carolina capitol in 1962. ^{Footnote 8} White political leaders universally claimed that the flag’s placement was commemorative, and not motivated by racial malice. The state legislature debated removing the emblem in 1993, amidst concerns that the flag was a deterrent to international investment (Leib Reference Leib1995; Woliver et al., Reference Woliver, Ledford and Dolan2001). These debates over the emblem led the legislature to propose a non-binding referendum on the flag in 1994, in which 76% of respondents voted to keep the battle emblem flying (oddly, the referendum was not voted on in an at-large election, but was tied to the Republican party primary (see Webster and Leib, Reference Webster and Leib2001)). In 1996, Governor David Beasley responded to a wave of racially-motivated violence (especially the burning of several Black churches) by calling for the removal of the flag. In a public address, he stated that it was time to “compromise on the Confederate flag, and teach our children that we can live together” (quoted in Webster and Leib, Reference Webster and Leib2001, p. 278). Beasley ultimately proposed a compromise bill that would remove the flag from the capitol dome and move it to a Confederate monument on the capitol grounds. A statewide poll taken in December 1996 found modest support for Beasley’s proposal (52% to 43%). Interest groups coalesced on both sides of the issue: the Council of Conservative Citizens, the League of the South, and the Ku Klux Klan opposed the removal of the battle emblem (it is important to note that the CCC and the League of the South did not welcome the Klan’s involvement), while dozens of groups led by the National Association for the Advancement of Colored People and the South Carolina Chamber of Commerce agitated for its removal (Woliver et al., Reference Woliver, Ledford and Dolan2001). After a series of votes on proposals facially similar to Beasley’s (the legislature declined to directly vote on the Governor’s plan), the legislature rejected any plan to remove the battle emblem from the capitol dome.

Gerald Webster and Jonathan Leib (Reference Webster and Leib2001) analyze the legislature’s votes on these proposals. They find that partisanship and the racial composition of a legislator’s district strongly influenced vote choice. Republicans were highly supportive of the emblem, as were representatives from overwhelmingly White districts. Representatives from majority Black districts, in contrast, overwhelmingly opposed the measure. Notably, Webster and Leib (Reference Webster and Leib2001) also find that many Democrats were absent for these votes, suggesting that many Democrats did not want to officially announce a position on the issue. Further, they find that similar patterns (i.e., support broke along partisan and racial lines) hold for a subsequent vote on (optional) incorporation of the Confederate emblem into state license plates.

After the NAACP called for a tourism boycott of South Carolina for as long as it chose to fly the Confederate flag in 1999, debates began anew about its removal. In 2000, Governor Jim Hodges called for the flag to be removed from the capitol dome. Later that year, the state Senate voted to remove the flag from the dome and place it on a flagpole at the Confederate memorial in front of the capitol building. This proposal gained broad bipartisan support from both Black and White senators. (Webster and Leib, Reference Webster and Leib2001). The state House narrowly passed the measure (63–56) after several amendments aimed to make the flag more prominent, which cost the measure the support of many Democrats and all but three of its Black members (Webster and Leib, Reference Webster and Leib2001).

2004 Georgia Survey

The Survey Research Laboratory in the Andrew Young School of Policy Studies, Georgia State University, conducted the panel survey before and after the March 2, 2004 referendum described above. The first survey was conducted by telephone using random-digit dialing in the period January 20 through February 29; a follow-up survey of the same respondents was conducted during the period May 24 through June 28. The sampling procedure was stratified to ensure representative sampling of residents in the five-county Atlanta metropolitan area and in the remainder of the state. The interviewing staff included both African American and White interviewers. We analyze responses from 341 Whites who completed both waves.

This survey suits our purposes admirably because it has a number of questions that reflect the distinct sides of the “heritage or hatred” debate. For example, we reason that if some Whites are motivated to support the flag because it represents a unique Southern historical legacy, we might expect to see that knowledge about Southern history, along with feelings of attachment to the South, are greater among White supporters of the Confederate battle emblem than among White opponents. On the other hand, if racial prejudice leads Whites to support the Confederate battle emblem, we should see that anti-Black attitudes are more widely held among White supporters of the Confederate battle emblem than among White opponents. ^{Footnote 9}

To assess these competing perspectives, we regress support for the Confederate flag on knowledge about Southern history, feelings of attachment to the South, and a number of racial attitude measures: racial resentment (Kalmoe and Piston, Reference Kalmoe and Piston2013), opposition to interracial dating, and denial that Blacks are worse off than Whites when it comes to jobs and education. Control variables include ideology (measured as support for limited government), party identification, and select demographics (sex, age, education, birthplace, urbanicity, and home ownership). We estimate a logistic regression due to the structure of the dependent variable: respondents were coded “1” if they preferred the version of the state flag that prominently featured the Confederate battle emblem, and “0” otherwise.

Knowledge about Southern history was measured through two questions: 1) whether the respondent can correctly identify the famous Union general, William Tecumseh Sherman, and 2) the number of Civil War battles the respondent can name (in our analyses, no credit was given for additional battles named after the first two). We reason that this measure captures the latent concept of pride in Southern heritage, in that those with high levels of Southern pride should also exhibit higher levels of knowledge of Southern history than those who have no such pride (after controlling for education). While imperfect, this measure is a significant improvement over previous works that relied upon residence in the South to proxy “heritage” or “pride.”

Feelings of attachment to the South were measured with a question asking how close the respondent felt to Southerners. The racial resentment measure was a battery of three standard questions (e.g., Tesler and Sears, Reference Tesler and Sears2010; the additional standard question about whether Blacks are getting less than they deserve was not included). “Old-fashioned” racism was captured by a question asking whether the respondent would object if their child dated someone of a different race. The final independent variable of interest, perceptions of White advantage, was a battery of two questions. The first asked whether the respondent thought that the average Black person was worse off (or better off, or about the same) than the average White person with respect to income, and the second asked the same except with respect to education. Coefficient estimates from the regression are presented in Table 1. ^{Footnote 10}

Table 1. Predictors of Support for Confederate Flag 2004 Georgia Survey Data

***p < 0.001; **p < 0.01; *p < 0.05; cell entries are logistic regression coefficients (standard errors in parentheses). All variables are coded from 0 to 1. White respondents only.

We begin with those variables intended to reflect the “heritage” side of the debate. Contrary to the argument that respect for Southern heritage drives White support for the Confederate flag, we find that those Whites with more knowledge about Civil War history are actually less supportive of the state flag prominently featuring the Confederate battle emblem. White supporters of the Confederate battle emblem are distinguished not by their knowledge of Southern history, but rather their ignorance of it.

Of course, it is possible that one could feel an affiliation for the South without knowing much about the Civil War. We therefore also examine whether those Whites who say they “feel close to Southerners” are more likely to support the Confederate flag. But after holding other factors constant, the results indicate no meaningful relationship between feelings of closeness to Southerners and support for the Confederate battle emblem.

Turning now to the racial prejudice side of the debate, we see a positive and statistically significant relationship between racial resentment and support for the Confederate battle emblem. Since the racial resentment scale is controversial—some have argued it is contaminated with (though not reducible to) non-racial ideology (e.g., Feldman and Huddy, Reference Feldman and Huddy2005)—it is important to note that the results hold when we drop the two controversial questions from the battery. Furthermore, while opposition to interracial dating is not statistically associated with opinion about the Confederate flag, this changes if racial resentment is not included in the model. Finally, even after controlling for racial resentment and opposition to interracial marriage, denial that Blacks are worse off than Whites when it comes to income and education is positively associated with White support for the Confederate battle emblem. The magnitude of these associations is significant, as a predicted probability plot (control variables set to their means) in Figure 1 indicates. At the low end of the racial resentment scale, the predicted probability of preferring the Confederate battle emblem is 0.34; at the high end, the predicted probability more than doubles to 0.75. Conversely, Figure 1 shows also that individuals at the low end of knowledge of Southern history are predicted to support the Confederate flag at 0.59, while those with the highest knowledge are predicted to support it at only 0.39, net of controls. That is, to the extent that knowledge of Southern Civil War history captures “Southern pride” in the Confederacy, such pride (or heritage) is negatively associated with support for the rebel banner. In sum, among our sample of White Georgians the heritage argument fares poorly—in fact, the coefficient is statistically significant but in the opposite direction as proponents of the heritage argument would suggest—but the racial prejudice argument fares well as racial attitudes are tightly bound up with support for the Confederate battle emblem. ^{Footnote 11}

Fig. 1. Predictors of Confederate Flag Support 2004 Georgia Survey Data

2014 Winthrop Survey

The Winthrop Poll at Winthrop University (South Carolina) interviewed 852 adult South Carolinians between November 9 and November 16, 2014. The survey was conducted using random digit dialing (RDD), supplemented with wireless phone number sampling. Calls were made on weekday evenings, and on weekends, in order to avoid oversampling non-working populations or systematically excluding groups working second or third shifts. Similarly, wireless sampling helps ensure that there are not geographic or age disparities in the sample. This survey included individuals of multiple races; we analyze only responses of the 581 White respondents.

While the Winthrop survey is more recent than the Georgia survey, it does have a more limited range of measures of Southern pride and racial attitudes. In fact, there is only one question about attachment to the South that asks whether the respondent considers him/herself to be a native Southerner, a converted Southerner, or a non-Southerner. There is also only one question measuring racial attitudes. This question is a slightly modified version of a standard question in the racial resentment battery, asking whether generations of slavery and discrimination make it difficult for Blacks to work their way out of the lower class.

Here too, in order to assess competing explanations, we regress support for the Confederate battle emblem on Southern identity and racial attitudes. We have two separate questions that we use as dependent variables, each of which measures support for the Confederate battle emblem. The first asks whether the respondent feels that the Confederate flag should be flown on the grounds of the South Carolina State House. This was followed up with a branching question asking whether the respondent felt this way strongly or not strongly. The second question asks whether the respondent feels very positive, positive, neutral, negative, or very negative about the Confederate flag. We conduct two separate ordered logistic regressions, one for each of these dependent variables. In addition to the independent variables described above, we include controls for partisanship and demographics: gender, age, education, and income. Coefficient estimates are presented in Table 2. ^{Footnote 12}

Table 2. Predictors of Support for Confederacy 2014 Winthrop Survey Data

*** p < 0.001; **p < 0.01; *p < 0.05; cell entries are ordered logistic regression coefficients (standard errors in parentheses); cut points suppressed. The omitted variable for Southern identity is “non-Southerner.” All variables are coded from 0 to 1. White respondents only.

This analysis uncovers more support for the heritage thesis than did the analysis of the Georgia survey data. Both those who identify as “converted Southerners” and those who identify as “native Southerners” are more likely than those who identify as “non-Southerners” to believe that the Confederate flag should fly outside the South Carolina State House and to report that they feel positive toward the flag. That said, the results here support the racial prejudice thesis as well. Those Whites who deny that generations of slavery and discrimination have made it difficult for Blacks to work their way out of the lower class are also more likely than their counterparts to claim that the Confederate flag should fly outside the South Carolina State House and to report that they feel positive toward the flag.

Furthermore, the magnitude of this effect is large. Predicted probability plots are shown in Figure 2: all control variables are set to their means. At the low end of the racial resentment scale, the predicted probability of reporting that the respondent strongly feels the flag should be flown is 0.23; at the high end of the racial resentment scale, this probability more than doubles to 0.63. While fewer Whites report feeling “very positive” toward the flag (the modal response is neutral), the pattern here is similar. At the low end of the racial resentment scale, the predicted probability of feeling very positive is 0.07, while at the high end, this probability more than triples to 0.24. As in the case of the Georgia survey, the South Carolina survey indicates that racial attitudes and White attitudes toward the Confederate flag are closely associated.

Fig. 2. Racial Attitudes and Confederate Flag Support 2014 Winthrop Survey Data

2008 ANES Survey

Of course, the foregoing analyses grant Southern nostalgists the benefit of the doubt as to whether the remembrance and recognition of the concept of “Southern heritage,” when embraced by White flag supporters, is analytically separable from racial prejudice or White supremacy. When the pride and prejudice hypotheses are evaluated simultaneously, the data cast considerable doubt on this move. The first thing to note is that although the flag may plausibly be a part of Southern culture, it is definitely not a symbol admired by all southern groups. African American Southerners overwhelmingly disapprove of the flag. This is why we do not include them in our analyses—not because they are not a part of Southern culture, but because this would make the case too easy. Thus the Confederate flag is at best a racially polarizing symbol of White Southern culture, which some Southern Whites claim to be a non-racial symbol. But while not all flag supporters express negative attitudes toward Blacks, the two sentiments do vary together empirically.

We now build on our analyses of surveys of statewide samples that directly examine support for Southern symbols to analyze the relationship between prejudice and its supposedly innocent alternative—pride or attachment to the South. To do so, we turn to the 2008 American National Election Studies time series survey. The response rate of the ANES is high, and the interviews are conducted face-to-face in order to produce high-quality data. Consistent with other ANES studies conducted during years of presidential elections, interviews were conducted in two waves. The pre-election wave was conducted during the two months preceding the November election, and the post-election wave was conducted during the two months following the election.

While the ANES does not have measures of Confederate flag support, it does have a measure of global warmth toward Southerners: a 101–point feeling thermometer (rescaled from 0 to 1). This allows us to examine whether racial attitudes are bound up with Southern symbols beyond the Confederate flag. Furthermore, the nationally representative sample allows us to examine associations between racial attitudes and warmth toward Southerners among those Whites who live both inside and outside the South (defined as residence in one of the states that seceded during the Civil War).

To do so we conduct a series of ordinary least squares regressions, broken out by region of respondent. In all cases the dependent variable is warmth toward Southerners. The independent variable of interest is either racial resentment (the standard four-item battery) or a separate racial attitude measure: a differential between warmth toward Whites and warmth toward Blacks. This approach (Hutchings Reference Hutchings2009) allows us to account for respondent tendencies to rate any racial group positively (or negatively). Control variables, consistent with previous analyses, include limited government, partisanship, and demographics: age, gender, education, income, and participation in the labor force. Coefficient estimates are presented in Table 3. ^{Footnote 13}

Table 3. Racial Prejudice Predicts White Warmth Toward Southerners 2008 ANES Survey

*** p < 0.001; **p < 0.01; *p < 0.05; cell entries are ordinary least squares regression coefficients (standard errors in parentheses). All variables are coded from 0 to 1. White respondents only.

Among White respondents who live in the South, we see a strong association between racial attitudes and warmth toward Southerners, net of controls. This association is statistically significant in both cases and large in magnitude; movement from the low end to the high end of the White/Black thermometer differential is associated with movement covering 40% of the Southern feeling thermometer, while movement from the low end to the high end of the racial resentment battery is associated with movement covering 16% of the Southern feeling thermometer. For non-Southern Whites, in contrast, the coefficient on racial resentment is statistically significant, while the coefficient on White/Black feeling thermometer is not. In sum, especially among Southerners, racial prejudice is tied up with not only support for the Confederate flag in particular (as we see in the Georgia and South Carolina polls) but also warmth toward the South more generally. While measures of support for specific Confederate symbols in the national sample would have bolstered the specific case above, this finding—that White supremacy and affection for the South itself are twinned, both in the survey data and historically—means that the empirical conclusions from the state-level analyses are conservative; it is also possible that White respondents who express affection for “the South” or southern heritage are subtly (perhaps not so subtly) invoking affection for White supremacy.