On Acorns and Oaks: Revisiting the Partisan Realignment of the White South 1 David O. Sears University of California, Los Angeles Nicholas A. Valentino The University of Michigan April 4, 2005 (34) 1 Abstract The mass realignment of Southern whites to the Republican party since the 1960's has sharply changed the partisan balance of the nation as a whole. We explore the convergence of (1) historically high levels of racial animosity among white Southerners, (2) the particular historic salience of race in the South, and (3) the conservative values at the heart of the surging religious fundamentalism in the South, hypothesizing that they are responsible for much of the shift of white Southerners to the Republicans. One set of analyses shows strong continuity between the state-by-state presidential vote results between the present era and the two previous periods in which the parties were sharply divided on racial policy, the antebellum and civil rights eras. The current party alignment does not resemble those in the much longer periods in which the parties did not differ on racial issues. A second set of analyses, using NES data, show that (1) symbolic racism has become increasingly closely associated with Republican party identification over the last thirty years in the old Confederacy, but not in the North and West; (2) Southern white evangelicals are considerably more racially conservative than are other white Southerners, Northern white evangelicals, or people of other faiths; (3) white evangelicals are also more conservative than other whites in non-racial domains, but without systematic regional differences, presumably because the political distinctiveness of the white South is mainly limited to racial issues; (4) evangelical Protestantism is associated strongly with Republican party identification in both South and North, but that association is explained by racial conservatism only in the old Confederacy; and (5) evangelical Protestants dominate the Southern white electorate, but are a much smaller minority elsewhere in the nation. These findings suggest the continuing power of specifically racial conservatism, especially as associated with evangelical Protestantism, in producing the mass partisan realignment of the white Southern electorate. 2 The purpose of this paper is to reexamine the partisan realignment of the American voting public that has been occurring since the 1960's. Realignments normally are understood as involving shifts in the party coalitions over time (Key, 1959; Burnham, 1965; Bartels, 1998; Mayhew, 2002; Petrocik, 1984). Since 1960, region has become associated less closely with partisanship, most notably as the once solidly Democratic South has converged toward the rest of the nation. As shown in Figure 1, the proportion of white Southerners who are Republican has surged, whereas the proportion who are Democrats has dropped precipitously, according to the National Election Studies’ surveys conducted over the past half century. [Insert Figure 1 about here] Region is of course not the only demographic dimension that has changed in relation to partisanship. Race, religious observance, gender, and socio-economic status have become associated more closely with partisanship, as whites, especially the more religiously observant, males, and those with higher socio-economic status have moved toward the Republicans (Abramowitz & Saunders, 1998; Manza and Brooks, 1999; Miller & Shanks, 1996; Petrocik, 1987; Kauffmann and Petrocik, 1998). These changes have changed the balance of political power in the United States. A nation with a consistent Democratic majority at mid-century has evolved into one that is closely divided, if not tilting toward the Republicans. The Democrats have not captured a majority of the presidential vote in a quarter of a century. In 1994 the Republicans took full control of the Congress for the first time in nearly half a century, and seem to be solidifying their position. In state politics, the once healthy Democratic majority of the governorships has switched to a strong Republican majority, and Republicans have come at least to parity in the state legislatures. In terms of party identification, Republicans have overcome the stable majority once held by Democrats (Abramson et al., 2002; Flanigan and Zingale, 2003; Miller & Shanks, 1996). A phenomenon of this importance always generates extensive scholarship. Any brief summary of its results risks doing violence to its inevitable nuances. We advance a somewhat 3 different account than the current conventional wisdom, also a risky proposition, and it will ultimately require a lengthier exposition than possible here. We hope that its contours can be sketched out here with sufficient evidence to make a plausible case. Conventional Explanations of the Realignment Race as the initial trigger The literature on the contemporary party system would seem to reflect a consensus about the events triggering at least the regional component of this realignment. Perhaps the most complete analysis, by Carmines and Stimson (1989), concludes that “[t]he struggle over race, at its peak the dominant issue of American political life for only some three years in the mid1960’s…, permanently rearranged the American party system” (p. xiii). They argue that realignment was triggered by the ideological polarization of the two party caucuses in Congress after the election of 1958, shifting the balance of power in each party away from the center. The exogenous pressures of the civil rights movement forced these more polarized parties to take more distinctive positions on civil rights issues, departing from their similar levels of support for the racial status quo in the years previous. The Republicans’ “Goldwater gamble” in 1964 campaigned on a states-rights justification for opposing federal intervention to protect blacks’ civil rights. It triggered a heightened emphasis on race in the two parties’ platforms, increased polarization of party activists on racial issues, and heightened perceptions within the mass public of that party polarization on desegregation. Later in the decade the Nixon “Southern strategy” accelerated the realignment of Southern whites into the Republican camp. As careful and persuasive as the Carmines and Stimson analysis is, for our purposes it is incomplete in two key respects. First, it deals only with the period from the end of World War II to 1980. The realignment has continued into the present, we believe. Indeed the post-1980 Reagan era is often cited as central to the Republican surge (e.g., Black and Black, 2002). Second, they do not explicitly address the regional basis for change. In our view the most 4 important cause of the change in the partisan balance of power is the conversion of the once-solid Democratic South to a region that at least tilts Republican (Black and Black, 2002). A dwindling role for race? The conventional wisdom about this continuing regional realignment no longer focuses on race. Rather, it emphasizes a regularization of the relationship between ideology and party identification in the post civil rights era: “....a secular realignment of party loyalties along ideological lines” (Abramowitz & Saunders, 1998, p. 634; also see Rabinowitz, Gurian, & Macdonald, 1984). Regularization of the link between party and ideology has clearly occurred. But ideological conservatism can be based in a number of values and attitudes. The general consensus seems to be that the key issues attracting Southern whites to the Republican party have been support for lower taxes, a less intrusive, New Deal-style big federal government, a reduction in the welfare state, a strong national defense, freedom to bear arms, religious conservatism and associated moral issues such as abortion and family values, as well as an increasingly class-based party cleavage in the South (Abramowitz, 1994; Sniderman and Carmines, 1997; Petrocik, 1987; Abramowitz and Saunders, 1998; Carmines and Stanley, 1990; Green et al., 1998; Kellstedt, 1990; Steed et al., 1990). Most writing on the subject have concluded that even if the realignment of white Southerners may have been triggered initially by racial issues, today their support for Republicans is due primarily to other factors. As a result, race and racial issues have played a minor role in most of these accounts of white Southerners’ continuing political realignment. This mirrors a common line of thinking about American politics more generally, that much of the once-considerable force of racial prejudice has been spent: “A quarter century ago, what counted was who a policy would benefit, blacks or whites” (Sniderman & Piazza, 1993, pp. 4-5), while “the contemporary debate over racial policy is driven primarily by conflict over what the government should try to do, and only secondarily over what it should try to do for blacks” [emphasis in the original] (Sniderman & Carmines, 1997, p. 4). As a result, today “prejudice is 5 very far from a dominating factor in the contemporary politics of race” (Sniderman & Carmines, 1997, p. 73). Three contrary indications We are less sure. As a starting point, there would seem to be three presumptive reasons to take a closer look at race. One is the fact that the major shift to the Republican Party has occurred in the South, a historic bastion of racial conservatism.2 The “Solid South” voted reliably more for Democratic presidential candidates than for Republicans from the “Compromise of 1877" to the Dixiecrat rebellion of 1948, years when most Southern blacks were excluded from voting. Thereafter the South began slipping away from the Democrats, and in 2000 was lost entirely. The Democrats held the majority of Southern Senate and House seats well into the 1980's, as well as a wide advantage in Southerners’ party identifications. Democratic support in the South has dropped sharply in all these respects, though at different rates depending on the indicator. They have lost, in succession, their once-hegemonic lead in the U. S. Senate, the House, gubernatorial seats, and state legislatures. This partisan realignment of Southern whites has been widely noted, though there is some disagreement about the timing of the realignment, and about whether a transitional period of dealignment occurred (e.g., Beck, 1977; Black and Black, 1992, 2002; Campbell, 1977a, b; Carmines & Stanley, 1990; Converse, 1966; Jewett, 1999; Miller and Shanks, 1996; Petrocik, 1987; Stanley, 1988; Sundquist, 1983; a most detailed, state-by-state account is provided by Bullock and Rozell, 2003). Second, the Southern movement from Democrats to Republicans has occurred almost exclusively among white voters. Southern blacks mobilized greatly in this period, and overwhelmingly supported Democratic candidates (Abramowitz & Saunders, 1998; Black & Black, 1992; Burnham, 1996; Miller & Shanks, 1996). As the Blacks’ (2002) careful analysis documents, the partisan solidarity among Southern African Americans, as well as their considerable numbers, has at least given the Southern Democratic party a thoroughly racially integrated face, and often perhaps a predominantly black one. 6 Third, the Southern white realignment has occurred predominantly, though not exclusively, among whites native to the South. There is good evidence that migration from the North was not solely responsible; many native Southerners themselves changed parties, especially in the Deep South and to a surprising degree, among older cohorts (Beck, 1977; Campbell, 1977a, b; Carmines and Layman, 1998; Carmines and Stanley, 1990; Glaser and Gilens, 1997; Petrocik, 1987; Stanley, 1988). Finally, some of the emphasis on putatively non-racial issues may be suspect. A number of issues that, while not explicitly racial in nature, have become racialized in recent decades; for example, some have opined that “when the official subject is presidential politics, taxes, welfare, crime, rights, or values . . . the real subject is race” (Edsall and Edsall, 1991). At an empirical level, issues such as welfare, crime, and taxes have been shown to have a racial basis (Gilens, 1999; Mendelberg, 2001; Sears and Citrin, 1985). Theoretical Conditions for A Continuing Role of Race These doubts thus lead us to reappraise the possible role of racial conservatism as a continuing factor in the realignment of Southern whites. If so, four conditions need to hold. The first is that long-term continuities in Southern culture may have led to a continuing pattern of unusual racial conservatism in the white South, both at the aggregate and at the individual level. The two-caste system was deeply embedded in Southern society, and indeed central to its social and political organization, for centuries. Half a century ago, when analyzing Southern politics, one had first and foremost to focus on race, as V.O. Key, Jr. (1949) so clearly documented. On both cultural and psychological grounds we are skeptical that it has disappeared as a primary focal point of Southern politics in what is the short historical moment since the 1960's. The long history of racial animosity and exclusion in the white South might therefore raise suspicions about an important continuing role of racial issues in its political realignment. The two-caste system was deeply embedded in Southern society, and indeed central to its social and political organization, for centuries. Half a century ago, when analyzing Southern politics, one 7 had first and foremost to focus on race, as V.O. Key, Jr. (1949) so clearly documented. On both historical and psychological grounds we are skeptical that it has disappeared as a primary focal point of Southern politics in what is the short historical moment since the 1960's. Historically, continuity of contention over racial issues is one of the hallmarks of American political history. Race was a most conflictual issue at each of the young nation’s three early pivotal moments: the debates over the language of the Declaration of Independence and the Constitution, and in the politics that triggered the Civil War and dealt with its aftermath. Much later it was the nation’s most conflictual issue in the 1960’s during the wrenching abandonment of the Jim Crow system. In each case the conflict over race was largely sectional. The core documents agreed to at the founding of the nation essentially avoided the issue of slavery. Many historians today believe that tabling the issue of slavery was the price of bringing Southern elites to support independence, first, and then the Union.3 They did so not because it was a minor issue, because the number of slaves was large, the economic importance of slavery was great, and political contention over slavery was vigorous. Rather, slavery was a “deal-breaker.” However the historical record is unequivocal with regard to the Civil War in the 1860's, the Civil Rights Act of 1964, and the Voting Rights Act of 1965. They were tangibly and on the record cleavages between North and South. They revolved around race, and specifically around Southern white opposition to equality for African Americans. Second, at the individual psychological level, the South has been the home of substantially greater anti-black prejudice and other negative racial attitudes than has the rest of the country for as long as it has been measured systematically (Schuman et al., 1997). Moreover, racial attitudes have been shown to be among the most stable of Americans’ political attitudes at the intra-individual level (Converse and Markus, 1979; Kinder and Sanders, 1996; Sears, 1983). Presumably long-term stability of political attitudes depends on a number of factors, such as the recurrence of the attitude object on the political agenda, the volume of communication about it, the cognitive simplicity of the attitude object, and so on (Sears, 1983). All of those conditions 8 would seem to be met by racial attitudes. Finally, continuity at the aggregate level derives from the intergenerational transmission of racial conservatism among native white Southerners. Given this pattern of continuing greater racial conservatism among Southern adults, it would be quite surprising if young Southern whites have not been socialized to substantially greater racial animosity than young whites elsewhere in the country (e.g., Steeh and Schuman, 1991). As a result, the inertial power of such central cultural and psychological states should not be underestimated. The third condition is the easiest: a clear differentiation between the two parties in their racial conservatism. Few would doubt that the Democrats have emerged since the 1960's as the more liberal party on race. If one needed concrete indicators, the fact that the percent of blacks identifying as Republicans (including leaning independents) in the NES has ranged between 3% and only 12% since 1964 (Abramson et al., 2002, pp. 174-5). Or, the congressional Black Caucus has been almost unanimously Democratic since its inception. The parties clearly differ on race. The fourth step is that race would have to be particularly salient in the South, and therefore more psychologically accessible and easily primed, than elsewhere in the country. Again we appeal to the historic circumstances of life in the South. Race has from the beginning been a more central element of society as a whole in the South than elsewhere. Blacks have always represented a considerably higher percentage of the population in the South. They have always been more visible in the South, because they have been distributed more broadly in the population than in the North and West, with their tradition of black segregation in large metropolitan ghettoes. And as the Blacks (2002) have noted, in politics blacks are more salient as well, representing a considerably higher proportion of Democratic candidates and voters than is true elsewhere. For these reasons, race is likely to be more chronically accessible and more frequently primed in electoral situations in the South than elsewhere. If so, racial attitudes should be more politically powerful among white Southerners, everything else being equal. 9 Continuities in the Sectional Conflict: Aggregate Level Our first hypothesis is that the relatively recent shift of the South to the Republicans obscures an important long-term continuity in American politics. If we are correct, and racial issues are central ingredients in this shift, then accounts of contemporary party realignment are incomplete without examining the sectional division between North and South as the oldest basis for divisions over race in American life. The realignment of the past four decades has been primarily sectional in nature, as has been widely noted. But we suggest that the realignment has served to restore the two parties’ support bases to those sections of the country that were most politically divided during, and only during, the two previous eras in which the two parties were divided most centrally on racial issues: the antebellum period, and the civil rights era. It has swung the states of the old Confederacy into the Republican base, and the states that most supported the Union and civil rights into the Democratic column. Presidential voting has been the leading indicator, but other political indices have swung in the same direction with some lag. In fact, American political history yields only two prior periods in which the parties have been sharply polarized over issues of race. The first was the antebellum period from 1848 through 1860, beginning with the defeat of the Wilmot Proviso in 1847, the founding of the Free Soil Party in 1848, the passage of the Kansas-Nebraska Act in 1854, the Dred Scott decision in 1857, and ending with Lincoln’s victory in 1860 that triggered the secession of the Confederacy. The second was the civil rights era, beginning in 1963, when the Kennedy administration first began to take a strong position on civil rights, through the passage of the Civil Rights and Voting Rights Acts in 1964 and 1965 (though both, in the end, with substantial Republican support) and the clash of the Johnson and Goldwater campaigns over civil rights legislation, through the “Southern strategy” of Richard M. Nixon, the George Wallace campaign in 1968, and the rise of black militancy. The two parties had, of course, almost exactly reversed positions on racial issues in the meantime. Lincoln, the Republican who signed the Emancipation Proclamation, was succeeded by Goldwater, who voted against the Civil Rights Act. Douglas, the Democrat who 10 promoted the Kansas-Nebraska Act, was succeeded by Lyndon Johnson, who strongly supported civil rights legislation. We carry out a quasi-experiment to test this. We compare presidential voting in the present era with that in these two past eras, using the full periods in question (1848 to 1860 and 1964 to 1972). We also use an individual focal year for each prior period: 1856, because it was the first presidential election in which the Republican party provided a candidate, and immediately preceded 1857, which Stampp (1993) argues was the pivotal year for the ultimate breakdown of sectional harmony, and 1964, in line with most recent scholarship (e.g., Carmines & Stimson, 1989). Continuities across eras of party polarization over race Indeed, the partisan alignment of today shows a striking level of continuity with its alignment in those two most racialized periods in the history of American political parties, in several respects. First of all, the geographical pattern of the presidential vote is strikingly similar across these two periods, tracking the almost complete inversion of the two parties’ positions on racial issues in the interim. The Pearson correlation of the state-by-state Democratic presidential vote across the two focal years, 1856 and 1964 was r = -.60. Considering those eras more broadly, the average correlation was r = -.56 between the antebellum and civil-rights-era elections, taken a pair at a time. Second, we find striking continuities between the presidential votes cast in both eras with those cast in 2000. The states taken by the Democrat James Buchanan in the focal year of 1856 were much the same as the “red states” taken by George W. Bush in 2000, although the two elections were separated by 144 years, and the states taken by the Republican John C. Fremont were much the same as the “blue states” taken by Al Gore. Figure 2 shows this continuity in detail. Texas, Alabama, and Mississippi stand out as the strongest supporters of both Buchanan and Bush, while Massachusetts, Rhode Island, and New York stand out as strong supporters of both Fremont and Gore. The correlation between the state-by-state vote tallies in the two years is 11 r = -.68, and the average correlation of the results in the elections in the full antebellum period with the 2000 results is r = -.59, as shown in Table 1. Much the same continuity holds from the second racialized period, the civil rights era, to the present, though that is less surprising. The correlation of the focal year of 1964 with the year 2000 is r = .55, and the average correlation between the results from the full civil rights era and the 2000 results is r = .65, [Insert Figure 2 and Table 1 about here] We have used the year 2000 as our focal year for the present era, but it is not at all unusual. We could have selected virtually any presidential election in the past two decades to make the same point. Table 2 demonstrates the same continuity for all the presidential elections since Reagan-Mondale in 1984. The antebellum elections correlate on average r = - .50 with elections over that two-decade period, and the civil rights era, an average of r = .66. [Insert Table 2 about here] Third, if our hypothesis is correct, the present should show more continuity with those two earlier eras in which the parties were racially polarized than with the far more numerous elections in which both parties accepted the racial status quo to an equivalent degree. Indeed the correlations of the results for all elections since 1828 with the 2000 results emphasizes how uniquely those two past racialized eras foreshadow the present. As mentioned above, the antebellum elections and the civil rights era elections correlate r = -.68 and r = .55 with the 2000 result, respectively. But the correlations are much smaller for the elections before and between these two brief racialized eras, elections when race was neither a major national political issue nor a point of great difference between the parties. Table 1 shows those correlations grouped into different eras, but all show the same low absolute level: for example, the average correlation of the results for the 1872-1956 elections, inclusive, with the year 2000 is r = -.18. It exceeded plus or minus .30 only once in those 20 elections, in 1896. 12 Sectional conflict Finally, what the two past racialized eras and the present have in common is the continuity of specifically sectional conflict. This is revealed in two ways. It is apparent from Figure 1 that the states of the old Confederacy were strongest for Buchanan and for Bush, and the states of the Northeast were strongest for Fremont and Gore. The slave states that remained in the Union were intermediate. If we examine differences among individual elections, the centrality of sectional conflict is also apparent. There was a substantial aggregate-level switch of Southern states away from the Democrats to the Republicans after World War II, anticipated in the Dixiecrat revolt in 1948 and in the 1960 election, and then solidified in the 1964 election. As shown in Table 2, to a lesser extent the nomination of Bill Clinton in 1992 weakened continuity with the civil rights era, but by 1996 it is clear that Southern whites’ affinity for their native sons had largely eroded, and almost completely disappeared in 2000. Indeed the Southerner Al Gore’s vote reflected the antebellum alignment just as strongly as did the Massachusetts liberal Michael Dukakis. In other words, the partisan alignment has, for the past decades, resembled, more than any other in American history, the sectional cleavages of the antebellum and civil rights eras. James Buchanan, George H. W. Bush, and George W. Bush all drew their voting strength from about the same areas. In addition, the contemporary party alignment is quite dissimilar from the one that held in the long period between the Civil War and the civil rights era, when race was essentially off the national political agenda. These aggregate-level results are of course not conclusive evidence about individual-level processes; they are merely circumstantial. But they are consistent with the hypothesis that divisions over race have been central to the sectional cleavage in all three eras. We next move to public opinion data to examine their role in the contemporary era at the individual level. 13 Continuing Regional Differences in Racial Conservatism Our second hypothesis is that Southern whites have continued to hold substantially more negative racial attitudes than do whites living elsewhere, net of other factors. To begin with, we need to specify what we mean by racial attitudes. If one sought to find a basis for white Southerners’ partisan realignment in the old-fashioned (or “Jim Crow”) racism of half a century ago, based on support for segregation, formal discrimination, and a belief in inherent black inferiority, one would be disappointed. That form of racism has little support or political influence (Sears et al., 1997; Kluegel, 1990; Schuman et al., 1997; Kinder & Sanders, 1996). We also have some doubts about the continuing value of such traditional measures of racial animosity as trait stereotypes and group thermometers, because of their blunt and blatant form in an era of strengthened norms about suppressing the expression of racial prejudice (Mendelberg, 2001). Rather, the nature of white Americans’ racial attitudes has changed in important ways. A new form of racism seems to have largely replaced the traditional forms of racial prejudice just described, variously been described as “symbolic racism” (Kinder and Sears, 1981; Sears et al., 1997), “modern racism” (McConahay, 1986), or “racial resentment” (Kinder and Sanders, 1996; Mendelberg, 2001). The basic tenets of symbolic racism are widely accepted by whites today, tending to split whites about evenly (Henry and Sears, 2002; Tarman and Sears, 2004). Moreover, it has been shown to have more political influence than do other racial attitudes (Sears et al., 1997; Kinder and Sanders, 1996; Hughes, 1997; Bobo, 2000; Sears and Henry, 2004; Sears et al., 1997). If the historically greater levels of racial animus in the white South have persisted, as we expect, symbolic racism should have been consistently higher in the South than elsewhere throughout the last several decades, despite the decline of Jim Crow racism. Trends in racial conservatism Mean levels of Jim Crow racism did indeed sharply decline throughout the country between the 1970’s and the 1990’s, as shown in the top panel of Figure 3. A multivariate analysis (see Valentino and Sears, 2004) yields a large and highly significant drop for the entire sample, 14 from .39 to .14 on a 0-1 scale (p < .001). But both the Deep and Outer South continued to house substantially higher levels of Jim Crow racism than the North throughout the last three decades, as reflected in the large coefficients for the two regional dummies (b = .31 and .21, both p < .001). The size of the regional difference does diminish slightly over time, as reflected in the two small but significant region*year interactions, but mostly due to a floor effect for the scale. Thus small regional differences in Jim Crow racism have persisted over time, as it has been disappearing in the North and barely hanging on in the South. [Insert Figure 3 about here] The story for symbolic racism is much different than that for Jim Crow racism. First mean levels of symbolic racism have actually increased slightly over time, rather than declining (see the second panel of Figure 3), as reflected in a significant main effect for year in the regression analysis. Second, the regional difference in symbolic racism is very large, with considerably more symbolic racism in the South than the North and West at all time points. This is shown in Figure 3, and revealed statistically in significant region main effects. But third, the South did not converge to the rest of the country in symbolic racism between the 1970’s and the 1990’s; neither region*year interaction on symbolic racism is statistically distinguishable from zero (Valentino and Sears, 2004). Unlike Jim Crow racism, then, there is no declining trend in symbolic racism in any region. Do we find Southern whites in the contemporary period to be considerably more racially conservative than whites living elsewhere, above and beyond their more general political conservatism, on all dimensions of racial animosity, or are they specific to Jim Crow and symbolic racism? To test the generality of current regional differences, we examine the period beginning in the 1990’s, and broaden the range of racial attitudes by pooling available surveys across years (all NES surveys since 1992 with relevant measures, and all the GSS surveys since 1990). We again examine Jim Crow and symbolic racism, but add conventional measures of 15 traditional prejudice, stereotyping of blacks, and the feeling thermometers. To isolate the racial element from other factors, we control for ideology and relevant demographics. On each of these four measures, the South, and especially the Deep South, includes a disproportionate share of the highly racially conservative (Valentino and Sears, 2004). For example, 55% of the whites in the Deep South, but only 32% of those in the North and West, fall in the top third of the national distribution of symbolic racism. If so, are such differences merely a spurious effect of Southerners’ more general conservatism or demographic distinctiveness? The impact of region is highly statistically significant for each measure in analyses of variance that include controls for education, age, gender, and ideology. White Southerners are today more racially conservative than whites living elsewhere on all conventional dimensions of racial attitudes. These regional differences in racial conservatism are not explained by regional differences in more general political conservatism or demographics. Trends in political ideology Is the Southern realignment driven by an increased conservatism among whites that goes beyond race? In fact the Southern white electorate has not become relatively more ideologically conservative. The mean political ideology among whites has been quite stable over the available period (1972 to 2000), shifting only from 4.23 to 4.42 on the one to seven liberal-conservative scale. There appear to be no large regional differences in this small conservatizing trend. As shown in Figure 4, the Deep and Outer South were slightly but consistently more conservative than the rest of the country throughout this time period. The slightly expanded regional difference in 1998 and 2000 comes too late to explain the long realigning trends across three decades. The multivariate analyses (see Valentino and Sears, 2004) show that the Deep South*year interaction is non-significant, while the Outer South*year interaction shows a slightly increasing convergence to the rest of the country. In other words, the ideological complexion of the three regions has been relatively stable over time, so the massive shifts in Southern whites’ party 16 identifications cannot be explained as resulting from comparable conservatizing shifts in political ideology in the South.4 [Insert Figure 4 about here] In sum, party identification has swung sharply to the Republican party among white Southerners since the Civil Rights era, but not elsewhere in the country. Jim Crow or “oldfashioned” racism has diminished drastically, but the South has consistently retained more of it than the rest of the country. In contrast, symbolic racism has, if anything, been increasing over time, but again the South has been consistently higher than the rest of the country. We do find that white Southerners have been consistently more conservative in terms of political ideology than other whites, but there has been no change in the size of that difference. Changes in regional differences in racial attitudes and/or political ideology are unlikely candidates for explaining the white Southern realignment over the past three decades, therefore. Roots of Southern Realignment in Racial Conservatism? Presumably racial issues did not strongly differentiate the two parties in the pre-1963 South. But as Democratic and Republican elites more and more clearly polarized on race after 1964, and conservative white Southern elites moved to the Republican party, the mass public’s party identifications should have gradually fallen in line. Racial conservatism should have become increasingly with associated with Republican rather than Democratic partisanship in the South. Moreover, if we are correct that race has historically been more salient in both Southern politics and its society more generally than in the North, the strength of that linkage between racial conservatism and partisanship should have become greater in the South. As a result, our third hypothesis is that the distinctive realignment of Southern whites to the Republican party is associated significantly with their racial conservatism, even considering sources of political conservatism with no manifest racial content. Specifically racial conservatism continues to be a stronger force in attracting Southern whites to the Republican party than is true for whites elsewhere in the nation. 17 Indeed when we look at a time series analysis of the contribution of symbolic racism to Republican partisanship, we find a sharp regional difference. Figure 5 shows the results of regressing party identification on symbolic racism in the South and in the North.5 Here we are mindful of the critique leveled by some in the past that symbolic racism does not really reflect racism at all but non-racial conservatism instead (e.g., Sniderman and Tetlock, 1986; Tetlock, 1994; Roth, 1994). There is much recent evidence that symbolic racism has a strong and distinctive racial component that cannot be reduced to general political conservatism, though it does take some contribution from the latter (Henry and Sears, 2002; Sears and Henry, 2003, 2004; Tarman and Sears, 2004). Moreover, it seems clear that political conservatism has itself absorbed a distinctly racial component over recent decades. Nevertheless, because of such concerns, we also control on ideology in those analyses (even though this could plausibly be argued to “over-control” for ideology).6 [Insert Figure 5 about here] As can be seen, symbolic racism is strongly associated with Republican party identification in the South today, even with ideology controlled. The association is present but considerably more modest outside the South. White Southerners also show a considerable rise in the association of symbolic racism with Republican party identification since the Reagan years, as the most racially conservative whites left the Democratic fold, whether by choice or by death, and/or joined the electorate as Republicans. No such historic change has occurred in the North, where presumably race is a less central, though not trivial, partisan issue. Here, then, is evidence that the sharp surge in support for Republicans in the South during the past 15 years is closely associated with whites’ racial conservatism. Alternative explanations for these effects of symbolic racism might focus on general political ideology, both because it is thought by some to explain Southern realignment and because some have speculated that symbolic racism taps non-racial conservatism as well or better than it does racial resentment (e.g., Sniderman and Tetlock 1986). However, the greater impact 18 of symbolic racism in the South is not merely an artifact of stronger general political conservatism. Table 3 shows a regression analysis for both presidential vote and party identification in the contemporary era. The fourth row shows that ideology is, by itself, a strong and significant predictor of both the vote and party identification outside the South. But the two non-significant ideology*region interaction terms in the fifth row indicate that ideology has no greater impact on partisanship in the South than elsewhere. Moreover the interaction term for the vote actually falls in the opposite direction from the interaction for party identification, reducing the likelihood of falsely accepting the null hypothesis. In other words, in the contemporary period ideology has no greater partisan impact in the South than elsewhere, once the especially strong impact of racial conservatism in the South has been taken into account. [Insert Table 3 about here] To compare the magnitudes of symbolic racism and ideology effects across regions, we have converted the logistic regression coefficients in the vote preference model into probabilities. These results are presented in Figure 6. The height of each column represents the change in the probability of voting for the Republican presidential candidate associated with moving from the lowest to the highest value of either the symbolic racism or the liberal-conservative ideology scales. The bars on the left show that moving across the full symbolic racism scale leads to a 60 point increase in the probability of voting for the Republican candidate in the South, but only a 20 point increase elsewhere. The bars on the right show that the impact of ideology is larger than that for symbolic racism, but there is no regional difference in the size of that effect. The regional difference in the effects of symbolic racism, and the lack of any regional difference in the effects of ideology, suggests that specifically racial attitudes have played a fundamental role in the Southern-based partisan realignment of the past four decades. [Insert Figure 6 about here] We have relied primarily on symbolic racism to make the case that racial attitudes are central to Southern realignment. To isolate its specifically racial component we have controlled 19 on standard political ideology. But it is worth making further effort on this front, since the nonracial hypothesis is a prominent alternative among both academics and the lay public. Another approach is to compare the impact of racial, as opposed to non-racial, policy preferences. The latter are particularly prominent in explanations of white Southerners’ presidential votes for Republicans. To do so, we compared the effects of racial policy issues (government aid to minorities and affirmative action) with those of legalized abortion and national defense, two issues often invoked to explain Southern conservatism but deliberately chosen to have little spillover from racial issues. The dependent variable is dichotomous Republican vote choice with all other candidates in the excluded category, as before. We test the main effects of each policy attitude and of region on the vote (with demographics controlled; we do not control for ideology because the racial vs. non-racial contrast is carried out by the comparison across issues). The key term is the interaction of policy attitudes and region, which we hypothesize will be significantly positive for racial issues but close to zero for non-racial issue preferences. The model for the impact of racial issues is shown Table 4. The first row shows that opposition to assisting blacks strongly boosts the probability of Republican presidential voting outside the South, since that is the excluded group in the dummy variable for region. The second and third rows indicate that opinions about abortion and defense also have a large impact on Republican vote (again, in the North). The fourth row shows that once racial policy attitudes have been controlled, the remaining regional difference in the vote is small and non-significant. Then, most important, the fifth row shows that the racial policy attitudes*region interaction is positive and statistically significant, indicating that racial policy attitudes are significantly stronger predictors of the vote in the South than they are in the North.7 These logistic coefficients can be transformed into probabilities. Holding demographic controls constant at either their mean values or at “male” for gender, the impact of moving from the lowest to the highest levels of opposition to racial policies boosts the likelihood of Republican voting in the South by 63 percentage points, from 10% to 73%. Outside the South, it boosts the likelihood of voting Republican by only 38 20 percentage points, from 14% to 52%. In other words, opposition to special assistance to blacks has a significantly larger influence on support for Republican presidential voting in the South than it does in the North and West. [Insert Table 4 about here] In contrast, there are no significant regional differences in the linkages of non-racial policy attitudes to vote preference. Not surprisingly, opposition to abortion and support for defense spending both significantly boost support for Republican presidential candidates in the North and West, as shown in the second and third rows of Table 4. However, the sixth and seventh rows show that there is no regional difference in the impact of these opinions, and in fact the small interaction terms actually fall in the reverse direction. In short, racial issue opinions are more strongly linked to vote choice in the South than in the North and West, while non-racial issues operate almost identically across region. These findings contradict what we earlier described as the conventional wisdom about contemporary party realignment, then. They do so by examining both regional differences and racial issues more directly than has previous research. The white South is particularly sensitive to racial issues, and the racial conservatism of Republican elites, and their divergence from Democratic elites, has given racial issues an important role in realignment. The New Politics of Religion As mentioned at the outset, religion has been a second important domain of partisan realignment. The old denominationally-based model emphasized in the early voting studies focused on cleavages between the Democratic-leaning Catholics and Jews and the Republican leanings of white Protestants (e.g., Lazarsfeld, Berelson, and Gaudet, 1948). More recent research focusing specifically on religion, has carefully classified the church denominations and religious beliefs of ordinary voters, grassroots political activists, and active church members (e.g., Jelen, 1993; Layman, 1997; Kellstedt et al., 1999; Leege et al., 2002; Bolce and DeMaio, 1999a; Wuthnow, 1988).8 21 The Christian Right It has found that those global denominational differences are now considerably less politically important than cleavages between people with more orthodox and those with more liberal religious beliefs. The most significant cleavage is between white Protestants described as “evangelical,” “fundamentalist,” or “born-again” Christians,who tend to believe in the literal truth of the Bible, and members of “mainline” Protestant denominations and/or those with more liberal religious beliefs.9 Fundamentalist white Protestants, especially those who are particularly observant, are the strongest supporters of the Republican party, while liberal Protestants seem to have shifted toward the Democrats. For example, exit polls in 1988 and 1992 showed that fundamentalist religious belief was one of the strongest, if not the strongest, predictors of Republican presidential voting in virtually all the larger states (Jackson and Carsey, 1999; Manza and Brooks, 1999; also see Bolce and DeMaio, 1999b; Burnham, 1996; Busch, 2000; Green et al., 1998). The political cleavage driven by Christian fundamentalism has considerable generality: it seems to have had equally profound effects across presidential voting, congressional voting, and underlying party identification (Kellstedt et al., 1998). In other words, fundamentalists are becoming more politically homogeneous and distinctive from other religious factions within and outside Christianity.10 At first blush, the political activation of evangelical Protestants would seem to be mostly catalyzed by moral issues such as abortion, and traditional family values concerning parental rights, prayer in the schools, religious training in the upbringing of children, traditional gender roles, and opposition to gay and lesbian rights (e.g. Jelen, 1993; Kellstedt et al., 1998; Kohut et al, 2000; Wilcox, 1996). Traditional party differences on economic issues and the size of government do not drive the distinctive political preferences of fundamentalists to the same degree. Race and the Christian Right 22 For our purposes it is noteworthy that political cleavages between the Christian Right and their opponents are rarely ascribed to differences of opinion about racial issues.11 Nor do blacks themselves have much visibility in this literature, even though they are among the strongest opponents of the Christian Right’s broad conservative political agenda (e.g., Bolce and DeMaio, 1999a; Jelen, 1993; Kellstedt et al., 1998; Layman, 1997; Pastor et al., 1999; Wilcox, 1996).12 However, we are less sure that race is incidental to the surging role of evangelicals in the Southern Republican party. This argument, we anticipate, will be the most difficult and controversial for many readers. We come to it for both historical and psychological reasons. Our general theoretical orientation ascribes great inertial power to continuities with the historical past. Most white Southerners today were born and bred in the South, and are products of Southern culture, however much certain aspects of that have changed. Historically, churches were the centerpieces of Southern life, for whites and for blacks, to a greater extent than has been true elsewhere. Those churches were segregated racially, and remain so today, with some exceptions. Prior to the civil rights era, the most devout and regular churchgoers among white Southerners were also those who most supported traditional Jim Crow beliefs (Pettigrew, 1959). When the Johnson administration, in 1965, began to put pressure on the South to implement the Brown v. Board of Education desegregation decision, many white Southerners joined together to develop a parallel system of private Christian academies that, being religiously-based organizations, were beyond the reach of federal desegregation orders. But they provided a central refuge for the racial status quo, housing opposition to the ending of Jim Crow within religious belief and institutions. According to one account, the Christian Day School movement generated literally thousands of such private schools, and by the early 1970's were multiplying at a rate of nearly one a day (Martin, 1996, p. 169). In 1975, the IRS revoked tax-exempt status for Bob Jones University, which had maintained strict racial segregation until 1971, and continued to ban interracial dating. During the Carter administration, the IRS proposed in 1978 to revoke tax-exempt status for the Christian 23 academies, an event that mobilized an enormous and ultimately successful protest among evangelical Protestants. The primary issue at stake was the maintenance of de facto segregation, according to Frank Weyrich, one of its central organizers, although that conclusion must be speculative because the rhetoric of the protest did not emphasize race (Martin, 1996, p. 173). Political organizing among Christian fundamentalists then began in earnest. The next year, Jerry Falwell formed the “Moral Majority” to create a distinctive Christian vote in politics. Most date the mass mobilization of conservative Christians from that era, though later accelerated by the candidacy of Pat Robertson in 1988, and later the organization of the Christian Crusade. Christian fundamentalists today, unlike the more other-worldly pietists of years gone by, increasingly believe in the appropriateness, and even desirability, of expressing and implementing their religious convictions in the political arena (Kohut et al., 2000; Kellstedt et al., 1998). As a result, they have become much more politicized, and are now much more active politically (Layman, 1997; Manza and Brooks, 1999). Our psychological reasons for expecting a link between religious and racial conservatism revolves around the nature of symbolic racism, the most politically prevalent and powerful form of contemporary racism. It has long been theorized to blend anti-black prejudices with conservative, traditional American moral values that themselves have no manifest racial content. It was thought to originate partly in a conservative “moral code” (Sears and Kinder, 1971, p. 66; also see Kinder and Sears, 1981) or “....the traditional religious and value socialization of secular American civil Protestantism...” (McConahay and Hough, 1976, pp. 38-9). Later descriptions cited a number of specific Protestant virtues, such as “...hard work, individualism, thrift, punctuality, sexual repression, and delay of gratification, as opposed to laziness, seeking of favoritism and handouts, impulsivity, and so on....” (Sears, 1988, p. 72).13 At the most general level, symbolic racism does in fact draw upon both antiblack affect and non-racial conservatism (Henry and Sears, 2002; Sears and Henry, 2003). In terms of more specific traditional values, 24 recent evidence points to significant contributions of individualism and morality, especially the perception that blacks tend to violate them (Sears and Henry, 2003; Sears et al., 1997). Such traditional American moral values are especially important to evangelical Protestants. They favor “playing by the rules,” working hard, obeying the law, insuring that children obey their parents, living a moral life according to traditional interpretations of the Bible, and so on. Many feel so strongly about instilling these conservative values in their children that they fight for more religious training in the public schools, such as prayer in school or the visible display of the 10 Commandments, or even home school their children to insure adequate exposure to traditional religious values. Finally, Christian fundamentalism, like traditional racial conservatism, must be understood, at least to start with, in terms of its sectional base. Evangelical Protestants dominate the white South, but are a much smaller minority elsewhere in the country. According to a 1996 survey reported by Kohut and his colleagues (2000, p. 130), 52% of the “committed white evangelical Protestants” lived in the South, and only 10% in the Northeast, 13% in the West, and 25% in the Midwest (also see Green et al., 1998). Of course it would be incorrect to say that the white evangelical Protestant phenomenon is exclusively a Southern one. But it is difficult to assess its political effects without placing its Southern sectional base front and center.14 Fundamentalism and Political Conservatism Our fourth hypothesis has several parts, then. Southern white evangelicals are more racially conservative than either white evangelicals elsewhere or other white Southerners. White evangelicals everywhere are indeed more conservative than non-evangelicals on non-racial political issues, but region by itself does not provide any additional distinctive contribution to that difference. Finally, we speculated that such distinctive racial attitudes of Southern white evangelicals might be especially noteworthy in terms of the symbolic racism belief system. To test this we pooled the National Election Studies surveys from 1986 to 2000 that contained the same four symbolic racism items. We divided the white subsample into three 25 religious groups: evangelical Protestants, mainstream Protestants, and all others.15 The mean levels of symbolic racism are shown for each religious group in the South, and outside the South, at the top of Figure 7. In the old Confederacy, evangelical Protestants are indeed far higher in symbolic racism than are non-Protestants, with mainline Protestants falling in between. Indeed, of all white Americans, Southern evangelical Protestants stand out as by far the highest in such racial conservatism. By contrast, religion is not associated closely with symbolic racism in the North. The interaction of region and religion is statistically significant (p < .001). [Insert Figure 7 about here] This analysis is useful for our purposes because it shows the great strength of the difference between Southern evangelicals and others. The reliability of the finding is enhanced as a result of its being based on six NES surveys. But it includes no controls on other relevant factors. As a result we repeated the same comparison but using regression analysis, and controls on relevant demographic factors (Sears and Valentino, 2001). Evangelicals continued to be higher in symbolic racism than non-Protestants in the old Confederacy, despite controls on other relevant demographics. But there is no such difference in the North and West (the evangelical x region interaction is significant, at p < .001, as is the mainstream x region interaction, at p < .05). This again testifies to the unique level of negative racial attitudes among Southern white evangelicals. But perhaps this close link between evangelical Protestantism and racism in the white South is limited to symbolic racism. As indicated earlier, the concept of symbolic racism involves both racial animosity and traditional morality, so it might be especially appealing to Southern evangelicals. But symbolic racism has also been charged with being little more than a measure of general conservatism, rather than racism. How do we know that these findings reflect unusual levels of negative racial attitudes among Southern white evangelicals? We have two ways of showing that. 26 One is that other measures of racial animosity show the same thing. Figure 7 also shows comparable results for both the stereotype that blacks are “unintelligent” and opposition to federal assistance for blacks. In each case, the old sectional conflict over African Americans emerges: on all three measures, Southern whites are significantly more anti-black than Northern whites. More important for the present argument, Southern evangelicals stand out as having the strongest racial animus in each case. Each region X religion interaction is statistically significant. One must conclude that there is something unique about the combination of being a white Southerner and identifying with evangelical religion that predisposes one to negative attitudes about blacks; none of those elements by itself seems to be sufficient. Second, Southern evangelicals look very much like evangelicals outside the South on non-racial measures of conservatism do not yield uniquely high levels among Southern evangelicals. Figure 8 repeats the same analyses for four measures of conservatism cited widely in the literatures on realignment and on the Christian Right: opposition to abortion, anti-gay affect, support for defense spending, and support for traditional roles for women. In each case evangelicals are quite clearly more conservative than are either of the other religious groups, with highly significant religion main effects. But on these non-racial issues, the South is not very distinctive once regional differences in religion have been taken into account. To be sure, white Southerners of whatever religious persuasion are unusually supportive of a strong defense, but there are no other significant regional main effects. None of the special conservatism of Southern evangelicals that yielded region x religion interactions on racial attitudes yield such interactions outside the domain of race (the region x religion interaction is significant for abortion, but it is in the opposite direction from the expected: evangelicals outside the South are actually slightly more opposed to abortion than Southern evangelicals, and the effect is a small one). [Insert Figure 8 about here] Put another way, outside of racial issues, there is nothing distinctive about Southern evangelicals. The combination of being a white Southerner and identifying with evangelical 27 religion seems to produce the most negative attitudes toward blacks, but has no distinctive effects outside the historically troubled domain of race. Race and the Partisan Realignment of Southern White Evangelicals Now we return to the question of partisan realignment. Our fifth hypothesis was that symbolic racism would have unusual political force among Southern white evangelicals, and indeed would explain much of their attraction to the modern Republican party. To test this, we use the 1992 and 2000 NES datasets (the other NES surveys in the 1990's did not have the requisite measures). We first replicated the finding familiar from other studies that Southern white evangelicals do indeed support the Republican party more than do other Southern whites (e.g., Kohut et al., 2000; Kellstedt et al., 1998). The first column of Table 5 presents the effects of the two Protestant dummy variables on party identification for white residents of the old Confederacy, with demographic controls, It shows that being an evangelical Protestant is strongly associated with Republican party identification, while being a mainline Protestant is not. This is a substantively large effect: being Evangelical moves white Southerners about half a point on the seven-point party identification scale. In the North, both produce significant associations (column 4). Both are familiar findings: that evangelical Protestants are especially Republican throughout the nation has been reported numerous times. So has the finding that Catholics, Jews, and the non-religious are more Democratic than even mainline Protestants in the North. [Insert Table 5 about here] The next question is what explains this affinity of evangelicals for the Republicans? We want a relatively conservative test of our hypothesis that specifically racial conservatism is at work among the residents of the old Confederate states. So we first test a model that includes only measures of non-racial conservatism: woman’s role, defense spending, and opposition to abortion (anti-gay affect was not included in both datasets). There it can be seen that, not at all surprisingly, non-racial conservatism has a strong effect in producing support for the Republican 28 Party among both Southern and Northern whites (columns 2 and 5). The Republican party is popular all over the country among those who are conservative on non-racial issues. And including non-racial conservatism substantially reduces the coefficient for evangelical Protestantism, especially in the South. So white Southern evangelicals like the Republicans partly for their conservative stands on non-racial issues.16 Does racial conservatism also explain Southern evangelicals’ support for the Republicans? The answer would appear to be a definite “yes.” The third column shows that once racial attitudes are included in the equation, symbolic racism has a strong effect in promoting Republican party identification in the South. Moreover, the evangelical Protestantism term dwindles to non-significance. So symbolic racism appears to explain virtually all of the extra support given by Southern evangelicals to the Republicans in the South that has not already been explained by other conservative attitudes. And the increment in variance explained is not trivial, given these simple models. In the North, by contrast, symbolic racism also has a significant effect, but it does not come close to erasing the particular affinity of evangelicals to the Republican party. In other words, racial conservatism is not as central in explaining Republican partisanship in the North. The old Confederacy continues to be unique in generating great support for a conservative political party based distinctively on a racial conservatism -- just as it did in the years before the Civil War, and during the period of aggressive civil rights action in the 1960's. The realignment of Southern whites to the Republican party, then, may have been led by evangelical Protestants, because of their greater numbers, their greater racial conservatism, and because racial conservatism continues to have more political power in the white South than elsewhere.17 Racism and evangelical Protestantism have, in this view, converged as a political force that undergirds the movement of white Southerners to the Republican party, and to support of its candidates at all levels. 29 Timing of the Racially-Based Realignment The final part of our story concerns the timing of this long sequence of realigning events. This will ultimately require a longer exposition, but we can sketch out two preliminary findings here (also see Sears and Valentino, 2003; Sears, Citrin, and Kosterman, 1987). The late 1960's: Fundamentalists and racial animus We earlier indicated that the Johnson administration began to place strong pressure upon segregated school districts in the South to desegregate their schools. That pressure precipitated a major effort to organize local private school systems under religious auspices -- the so-called Christian academies. One possible consequence might have been an acceleration of Southern evangelicals’ distinctly racial conservatism. In other words, backlash against the Johnson administration’s racial liberalism, and the rise of Christian academies as alternative havens for the preservation of segregation, might have caused Southern fundamentalists to begin to separate themselves from other white Southerners in terms of racial animosity. Figure 9 shows the time series in the only measure of racial prejudice available in the NES surveys through this period: the difference between liking for whites and liking for blacks, on the feeling thermometers. As can be seen in Figure 9A, in 1964 the evangelicals and mainline Protestants in the South had almost exactly equal levels of prejudice. By 1968 evangelicals were considerably more prejudiced, and that gap did not narrow at any point in the period up to 1980. Figure 9B shows that the differences between evangelicals and mainline Protestants in that 1968 to 1980 period were small and inconsistent. We see in the mid-1960's, then, the emergence of a distinctive racial conservatism among Southern white evangelicals. The Republicans did not exploit this opening until slightly later. The Nixon “Southern Strategy” began to be implemented following the 1968 campaign, as the Nixon administration began to observe the success of George Wallace as a third-party candidate. In Table 6 we present logistic regression equations in which we predict to the presidential vote from a measure of racial policy opposition available through the 1950's and a measure of church attendance among 30 Protestants. As can be seen, opposition to racial policies begins to turn decisively against the Democrats in the South in the 1968 election, and religious attendance, in 1972. In contrast, racism already began to split the parties in the North in 1964, when many Southern racial conservatives were still Democrats. And religion began to have diminishing effects in the North after the 1960 campaign of John F. Kennedy, who was the target of anti-Catholic religious passions among observant Protestants (also see Converse et al., 1963). [Insert Table 6 about here] 1984, the Jackson campaign, and white flight In 1984, after the rise of the Moral Majority, and Ronald Reagan’s first term, Jesse Jackson ran for president. He had some success in the early primaries, especially in the South. Our hypothesis is that his success accelerated the flight of nervous racially conservative white Southern Democrats to the Republican party. Perhaps they also were attracted by his appeals on other non-racial issues we have discussed above, but the data indicate a timing of the white flight that is a bit sudden for that. We are fortunate that in 1984, the “rolling thunder” sample of the NES interviewed small numbers of cases every day from January through the end of the campaign. Table 17 shows an approximate parity of North and South through the beginning of October. But from that point through the beginning of December, movement out of the Democratic party into the Republican party accelerated. And in the South in became a stampede. By early December a slight Democratic lead among Southern Democrats had become an overwhelming Republican advantage of 58% to 34%. [Insert Table 7 about here] Our hypothesis is that this panicky exodus was related to the specter of Jesse Jackson’s popularity among blacks. In Table 8 we present the correlates of party identification in each of these phases of the campaign, by region. As can be seen there is a sharp increase in the correlation of evaluation of Jesse Jackson with party identification among white Southerners 31 toward the end of the campaign, compared to either earlier in the campaign or in the North. We believe along with other commentators (e.g., Black and Black, 2002; Dalleck, 2004) that the Reagan presidency was a critical element in the realignment of the white South. But we suspect it had more to do with racial conservatism than those accounts suggest. [Insert Table 8 about here] Summary and Conclusions Over the past several decades, partisanship in the American mass public has undergone a profound sectional realignment, with Republicans coming to dominate the South, and Democrats, the Northeast, the upper Midwest, and the Pacific Coast. Our general hypothesis is that this realignment is associated intimately with the longstanding and still unresolved racial conflicts in American society, especially in the new South. We offered three main propositions about how white evangelical Southerners differ from whites living elsewhere in the country in terms of their racial attitudes and religious convictions, then. First, the South is distinctive in two ways. White Southerners are more racially conservative than whites elsewhere in the country, and race is likely to be more salient to white Southerners than, on average, whites elsewhere, and therefore more psychologically accessible and easily primed. Second, evangelical Protestants are likely to be distinctive as well, in being likely to hold more conservative moral values than whites of other faiths. But there would be no reason to expect a regional difference in this respect, above and beyond the greater concentration of evangelicals in the South, since what we are arguing is unique about the South is its history of racial exclusion and animosity. Third, we would expect that these two traditions, of strong racial prejudice among white Southerners and the dedication of evangelicals everywhere to conservative moral and religious values, would come together in special force among Southern white evangelicals, generating especially high levels of racial conservatism. The historic salience of race in the South might also lead that racial conservatism to be especially powerful in drawing them toward a conservative Southern Republicanism. 32 Our findings suggest the continuing power of specifically racial conservatism in producing the mass partisan realignment of the white Southern electorate. As the numerically and culturally dominant group in the white Southern population, white evangelicals have much to do with the partisan realignment of the South in the aggregate. The linkages between evangelical religion and racial animosity seem to be fueling much of the shift toward conservative Republicanism among white Southerners. The high level of racial animosity among Southern white evangelicals is particularly noteworthy given their numerical dominance within the Southern white electorate. As shown in Table 9, evangelical Protestants constitute almost 50% of the whites in the states of the old Confederacy, but only 19% of whites living in the North, according to National Election Studies data collected over the past 15 years.18 In a sense the political culture of Southern whites has become one in which all others -- mainline Protestants, Catholics, Jews -- are, like blacks, outnumbered minorities. [Insert Table 9 about here] Finally, how central is symbolic racism to our story, relative to other forms of racial animosity? The theory of symbolic racism holds that it is based on a blend of moral traditionalism with primitive racial animosities (Sears, 1988). The former would seem to be stronger among evangelicals than among those of other religious persuasions, ceteris paribus, and the latter should be stronger among white Southerners than among other Americans, again everything else being equal. But Figure 3 does show similar effects of religion and region on all three standard measures of racial prejudice. We have no ready explanation for this parallelism. We can say two things, however. One is that the origins of symbolic racism in a blend of individualism and anti-black affect have been investigated fairly carefully, and verified (Sears and Henry, 2002). However less is known about its hypothesized roots in other, more moralistic, forms of traditional values. Second, symbolic racism consistently shows stronger political effects than other measures of racial animosity (e.g., Sears et al., 1997). So even if white Southern 33 evangelicals show the strongest racial animosity on all standard measures of it, that distinctiveness is likely to have more political consequence through the vehicle of symbolic racism than through older, more traditional forms of racial prejudice. 34 Bibliography (partial) Abramowitz, Alan I. (1994). Issue evolution reconsidered: Racial attitudes and partisanship in the U.S. electorate. American Journal of Political Science. 39:1-24. Abramowitz, Alan I. and Saunders, Kyle L. (1998). Ideological realignment in the U. S. electorate. Journal of Politics. 60:634-652. Beck, Paul Allen (1977). Partisan dealignment in the postwar South. American Political Science Review. 71:477-498. Bermanzohn, Sally Avery (1998). The legacy of the Ku Klux Klan in far right terrorism today. Presented at the 1998 annual meeting of the American Political Science Association, Boston. Black, Earl, and Black, Merle (1992). The vital South: How presidents are elected. Cambridge: Harvard University Press. Bobo, Lawrence (2000). Race and beliefs about affirmative action, In D.O. Sears, J. Sidanius and L. Bobo (Eds.), Racialized politics: The debate about racism in America. Chicago: University of Chicago Press. Bolce, Louis, & De Maio, Gerald (1999a). Religious outlook, culture war politics, and antipathy toward Christian fundamentalists. Public Opinion Quarterly, 63, 29-61. Bolce, Louis, & De Maio, Gerald (1999b). The anti-Christian fundamentalist factor in contemporary politics. Public Opinion Quarterly. 63: 508-542. Bullock, Charles S., & Rozell, Mark J. (Eds). (1998). The new politics of the old South. Lanham, MD.: Rowman and Littlefield. Burnham, Walter Dean (1996). Realignment lives: The 1994 earthquake and its implications. In Bert Rockman (Ed.). Busch, Beverly G. (2000). A retreat of the Christian Right? The evangelical vote in midterm congressional elections 1986-1998. Presented at the annual meeting of the American Political Science Association, Washington, D.C. Campbell, Bruce A. (1977a). Change in the Southern Electorate. American Journal of Political Science, Vol. 21, No. 1. (Feb., 1977), pp. 37-64. Campbell, Bruce A. (1977b). Patterns of change in the partisan loyalties of native Southerners. Journal of Politics, 39:730-761. Carmines, E. G., & Stimson, James A. (1989). Issue evolution: Race and the transformation of American politics. Princeton: Princeton University Press. Carmines, Edward C. and Layman, Geoffrey C. (1998) “When prejudice matters: the impact of racial stereotypes on the racial policy preferences of Democrats and Republicans" In Hurwitz, Jon, and Mark Peffley, Editors. (1998) Perception and prejudice. New Haven: Yale University Press. 35 Carmines, Edward C. and Stanley, Harold. (1990) . In Robert P. Steed, Laurence W. Moreland, & Tod A. Baker (Eds.). The disappearing South?: Studies in regional change and continuity. Tuscaloosa, AL: University of Alabama Press. Carmines, Edward G., & Merriman, W.R., Jr. (1993). The changing American dilemma: Liberal values and racial policies. In P.M. Sniderman, P.E. Tetlock, & E.G. Carmines (Eds.), Prejudice, politics, and the American dilemma, 237-255. Stanford, CA: Stanford University Press. Converse, P. E. (1966). Religion and politics: the 1960 campaign. In A.Campbell, P. E. Converse, W. E. Miller, & D. E. Stokes (Eds.), Elections and the political order (pp. 96124). New York: Wiley. D’Souza, Dinesh (1995). The end of racism. New York: The Free Press. Edsall, T. B., & Edsall, M. D. (1991, May). When the official subject is presidential politics, taxes, welfare, crime, rights, or values . . . the real subject is race. The Atlantic Monthly, pp. 53-86. Firebaugh, G., & Davis, K. E. (1988). Trends in antiblack prejudice, 1972-1984: Region and cohort effects. American Journal of Sociology, 94, 251-272. Glaser, James M., & Martin Gilens. 1997. Interregional Migration and Political Resocialization: A Study of Racial Attitudes Under Pressure Public Opinion Quarterly, 61:72-86. Green, J. C., Kellstedt, L. A., Smidt, C. E., and Guth, J. L. (1998). In Bullock, Charles S. III and Rozell, M. J. (Eds.). The new politics of the old South: An introduction to Southern Politics. (pp. 261-276). Lanham, MD: Rowman and Littlefield. Green, John C., Guth, James L., Smidt, Corwin E., & Kellstedt, Lyman A., Eds. (1996). Religion and the culture wars: Dispatches from the front. Lanham, MD.: Rowman and Littlefield. Green, John C., Rozell, Mark J., & Wilcox, Clyde (Eds.). (2000). Prayers in the precincts: The Christian Right in the 1998 elections. Washington: Georgetown University Press. Hughes, Michael (1997). Symbolic racism, old-fashioned racism, and whites’ opposition to affirmative action. In S.A. Tuch & J.K. Martin (eds.), Racial attitudes in the 1990s: Continuity and change, Westport, CT: Praeger. Hunter, James Davison (1991). Culture wars: The struggle to define America. New York: Basic Books. Jackson, Robert A., & Carsey, Thomas M. (1999). Group components of U. S. presidential voting across the states. Political Behavior, 21, 123-151. Jelen, Ted G. (1991). The political mobilization of religious beliefs. New York: Praeger. Jelen, Ted G. (1993). The political consequences of religious group attitudes. Journal of Politics. 55: 178-190. 36 Jennings, M. Kent, and Laura Stoker. (1999). The Persistence of the Past: The Class of 1965 Turns Fifty. Prepared for presentation at the Midwest Political Science Association Convention, Chicago, IL., April 1999. Jewett, Aubrey (1999). Examining two state realignment in the Southern states. Presented at the annual meeting of the American Political Science Association, Atlanta. Kellstedt, L. A. (1990). Evangelical Religion and Support for Social Issue Policies: An Examination of Regional Variation. In Robert P. Steed, Laurence W, Moreland, & Tod A. Baker (Eds.). The Disappearing South? Studies in Regional Change and Continuity. Tuscaloosa, AL: University of Alabama Press. Kellstedt, Lyman A., Green, John C., Smidt, Corwin E., & Guth, James L. (1999). Faith and the vote: The role of religion in political alignments. Presented at the annual meeting of the American Political Science Association, Atlanta. Key, V. O., Jr. (1949). Southern politics in state and nation. New York: Vintage Books. Kinder, D. R., & Sanders, Lynn. M. (1996). Divided by color: Racial politics and Democratic ideals. Chicago, IL: University of Chicago Press. Kinder, D. R., & Sears, D. O. (1981). Prejudice and politics: Symbolic racism versus racial threats to the good life. Journal of Personality and Social Psychology, 40, 414-431. Kluegel, J. R. (1990). Trends in whites' explanations of the black-white gap in socioeconomic status, 1977-1989. American Sociological Review, 55, 512-525. Kohut, Andrew, Green, John C., Keeter, Scott, & Toth, Robert C. (2000). The diminishing divide: Religion’s changing role in American politics. Washington: Brookings Institution. Kuklinski, James H., Michael D. Cobb, & Martin Gilens. 1997. Racial attitudes and the "new south." Journal of Politics, May, 59:323-349. Layman, Geoffrey C. (1997). Religion and political behavior in the United States: The impact of beliefs, affiliations, and commitment from 1980 to 1994. Public Opinion Quarterly. 61: 288-316. Lazarsfeld, Paul F., Berelson, Bernard, & Gaudet, Hazel. (1948). The people's choice (2nd ed). New York: Columbia University Press. Leege, David C., Lyman A. Kellstedt, and Kenneth D. Wald. 1990. Religion and Politics: A Report on Measures of Religiosity in the 1989 NES Pilot Study. Report to the NES Board of Overseers. McConahay, John B. (1986). Modern racism, ambivalence, and the modern racism scale. In J. F. Dovidio & S. L. Gaertner (Eds.), Prejudice, discrimination, and racism (pp. 91-126). New York: Academic Press. Miller, Warren E., & J. Merrill Shanks. 1996. The New American Voter. Cambridge, MA: Harvard University Press. 37 Pastor, Gregory S., Stone, Walter J., & Rapaport, Ronald B. (1999). Candidate-centered sources of party change: The case of Pat Robertson, 1988. The Journal of Politics. 61: 423-444. Petrocik, J.R. (1987). Realignment: New party coalitions and the nationalization of the South. Journal of Politics, 49, 347-375. Petrocik, John R. (1998). Reformulating the party coalitions: The “Christian Democratic” Republicans. Presented at the annual meeting of the American Political Science Association, Boston. Pettigrew, Thomas F. & Meertens, Roel W. (1995). Subtle and blatant prejudice in Western Europe. European Journal of Social Psychology, 25, 57-75. Rabinowitz, George, Gurian, Paul-Henri, & MacDonald, Stuart Elaine. (1984). The structure of presidential elections and the process of realignment, 1944-1980. American Journal of Political Science. 28:611-635. Roth, Byron M. (1994). Prescription for failure: Race relations in the age of social science. New Brunswick, N.J.:Transaction. Schuman, Howard, Charlotte Steeh, Lawrence Bobo, & Maria Krysan. (1997). Racial Attitudes in America: Trends and Interpretations. Revised Edition. Cambridge, MA: Harvard University Press. Sears, D. O., Citrin, J., & Kosterman, Richard. (1987). Jesse Jackson and the Southern white electorate in 1984. In Robert P. Steed, Laurence W. Moreland, & Tod A. Baker (Eds.), Blacks in Southern politics (pp. 209-225). New York: Praeger. Sears, D. O., P. J. Henry, & R. Kosterman. (2000). Egalitarian Values and Contemporary Racial Politics. In D. O. Sears, J. Sidanius, & L. Bobo (Eds.), Racialized Politics: The Debate About Racism in America. Chicago, IL: University of Chicago Press. Sears, D.O., van Laar, C., Carrillo, M., & Kosterman, R. (1997). Is it really racism? The origins of white Americans opposition to race-targeted policies. Public Opinion Quarterly, 61, 16-53. Sears, David O., & Valentino, Nicholas A. (1999). Long-Term Continuities in the Politics of Race. Prepared for presentation at the Annual Meeting of the American Political Science Association, Atlanta, Georgia, September 3. Sears, David O., Hetts, John J., Sidanius, Jim, and Bobo, Lawrence (2000). Race in American politics: Framing the debates, In D.O. Sears, J. Sidanius & L. Bobo (Eds.), Racialized politics: The debate about racism in America. Chicago: University of Chicago Press. Sheatsley, Paul B. (1966). White Attitudes Toward the Negro. Daedalus, The Negro American. Winter. Sidanius, J. & Pratto, F. (1999). Social dominance: An intergroup theory of social hierarchy and oppression. NY: Cambridge University Press. 38 Sniderman, P. M., & Carmines, E. G. (1997). Reaching beyond race. Cambridge, MA.: Harvard University Press. Sniderman, Paul M., & Piazza, Thomas. (1993). The scar of race. Cambridge, MA.: Harvard University Press. Stanley, Harold (1988). . Southern partisan changes: Dealignment, realignment or both? The Journal of Politics, Vol. 50, No. 1. (Feb.), pp. 64-88. Steed, Robert P., Laurence W. Moreland, and Tod A Baker. (1990). Searching for the Mind of the South in the Second Reconstruction. In Robert P. Steed, Laurence W. Moreland & Tod A. Baker (Eds.). The Disappearing South? Studies in Regional Change and Continuity. Tuscaloosa, AL: University of Alabama Press. Steeh, C., & Schuman, H. (1991). Changes in racial attitudes among young white adults, 1984-1990. American Journal of Sociology, 96, 340-367. Sundquist, James L. (1983). Dynamics of the party system (rev. ed.). Washington: Brookings Institution. Tetlock, P. E. (1994). Political psychology or politicized psychology: Is the road to scientific hell paved with good moral intentions? Political Psychology, 15, 509-529. Tuch, S.A., & Hughes, M. (1996). Whites’ racial policy attitudes. Social Science Quarterly, 77(4): 723-745. Tuch, S.A., and J.K. Martin. (1997). Regional differences in whites’ racial policy attitudes. In S.A. Tuch and J.K. Martin (Eds.), Racial Attitudes in the 1990's: Continuity and Change. Westport, CT: Praeger. Valentino, Nicholas A., & David O. Sears. (2004). Race, sectional conflict, and partisan realignment in the contemporary era. Submitted for publication. Wilcox, Clyde (1996). Onward Christian soldiers: The religious right in American politics. Boulder: Westview Press. Wright, G.C., Jr. (1977). Contextual models of electoral behavior: The Southern Wallace vote. The American Political Science Review, 71, 497-508. Wuthnow, Robert (1988). The restructuring of American religion: Society and faith since World War II. Princeton: Princeton University Press. 39 Figure 1. Democratic Vote: 1856 by 2000 80 Democratic Vote Percent in 1856 70 AR TX AL MS VA GA NC 60 KY FL MO LA TN DE 50 MD 40 ME RI NY 30 MA VT Region 20 Confederacy North 10 30 40 50 60 Democratic Vote Percent in 2000 40 70 Table 1. Average Pearson’s Correlation between Democratic Presidential Vote in 2000 and Selected Prior Periods. Period Years r # of Elections Storm Clouds (1828-1844) -0.41 5 Antebellum (1848-1860) -0.59 4 Transition (1972-1876) -0.13 2 Black Invisibility (1880-1956) -0.18 20 Republican Dominance (1880-1928) -0.23 13 New Deal (1932-1944) -0.19 4 Postwar Era (1948-1956) 0.05 3 (1960) 0.48 1 (1964-1972) 0.65 3 Transition Civil Rights Era Note: Entries are the correlations of the presidential vote for the Democratic candidate in each state in the 2000 election by that vote in earlier presidential elections, averaged over the prior years as specified. For example the entry for the Postwar Era is the average correlation of 2000 X 1948, 2000 X 1952, and 2000 X 1956: r = .05. The number of states equals the number of states that cast votes in the presidential election in both 2000 and the comparison year, ranging from 22 states in 1828 to 51 in 2000 (D.C.’s vote is entered separately). 41 Table 2 Pearson’s Correlations of Democratic Presidential Vote in Recent Elections by Previous Periods of Partisan Polarization over Race. Years 1856 1976 Antebellum Era (1848-1860) .28 1964 .38 Civil Rights Era (1964-1972) .30 1980 .18 .29 .28 .15 1984 -.50 -.56 .67 .49 1988 -.53 -.65 .76 .62 1992 -.28 -.25 .56 .42 1996 -.60 -.70 .68 .55 2000 -.59 -.68 .65 .54 Note: In first and third columns, the entries for the two eras are mean correlations for all years in the period. In the second and fourth columns, the entries are the correlations for the years 1856 and 1964, respectively, by Democratic vote in each recent election. 42 .17 Figure 2. The Impact of Symbolic Racism on Party Identification, by Year and Region, controlling for Ideology. 2.5 B 2 B The association between Symbolic Racism on Party Identification 1.5 B B B 1 0.5 J J J J J J 0 -0.5 -1 B 1986 1988 1990 B 1992 South 1994 J 1996 1998 2000 North Entries are unstandardized B's from OLS regression analyses with 7-point party identification scale as the dependent variable. Controls in each model are: Education, age, gender, and ideology. Symbolic racism is coded 0-1, so coefficients represent the change in the 7-point party identification scale that would result from moving from the lowest to the highest level of symbolic racism. 43 44 Table 3. The Impact of Religious Identification on Symbolic Racism among Whites, across Region. Former Confederacy B Nort h+ Wes t Beta B Beta Constant .74 .77 Evangelical Protestant .06*** .18 .01 .03 Mainline Protestant .03** .08 .01 .02 Age -.0001 -.01 .0003 .02 Gender -.003 -.01 -.02*** -.06 Education -.02*** -.22 -.03*** -.27 Income .02*** .09 .03*** .05 R Squared .07 .07 N 1722 4525 Note: Excluded category, captured by the constant, consists of Catholic, non-religious, Jews, and Non traditional orthodox, and non-Jewish Non-Christian. Symbolic racism is scored on a 0 to 1 scale, with higher values representing higher levels of racism. * p<.05; ** p<.01; *** p<.001, two-tailed. Source: NES 1986, 1988, 1990, 1992, 1994, 2000. 45 Figure 4. Mean Non-Racial Attitudes across Religious Identification and Region. National Election Studies 1986-2000. 0.68 4A. Opposition to Abortion 0.64 Mean Opposition to Abortion (0-1) 0.6 J B 0.56 0.6 0.56 0.52 B J 0.48 B J 0.44 0.4 Evangelical 0.68 0.64 B J B J JB North + West J B Evangelical Mainline Anova results Region Religious Id Region * Religious Id F .49 345.62*** 7.57* Other F .25 38.66*** 1.22 4D. Support for Traditional Woman's Role 0.56 0.52 B J Former Confederacy 0.32 Other Former Confederacy Woman's role is 0.48 in the home (0-1) 0.44 0.4 B J 0.52 0.48 North + West 4C. Support for Spending on Defense 0.6 0.56 0.36 Mainline Anova results Region Religious Id Region * Religious Id Mean Support for Defense Spending (0-1) 0.52 Mean Feeling 0.48 Thermometer toward 0.44 Homosexuals (0-1) 0.4 Former Confederacy 4B. Affect toward Homosexuals North + West North + West JB 0.36 0.44 0.32 0.4 J B JB Mainline Other Former Confederacy 0.28 Evangelical Anova results Region Religious Id Mainline - Region * Religious Id Other F 100.27*** 37.27*** .15 46 Evangelical Anova results Region Religious Id Region * Religious Id F .15 84.29*** 1.18 Table 4. The Impact of Racial Attitudes and Religion on Partisanship, Across Region. NES 1992, 2000. Dependent variable is Party Identification (1-7; hi=Rep) Evangelical Prot. (0,1) Mainline Prot. (0,1) Former Confederacy Demographic + Non-racial model attitudes .64*** .36* .22 .30 .18 .20 North + West Demographic + Non-racial model attitudes .86*** .64*** .73*** .79*** .57*** .81*** Age (years) Female (0,1) Education (1-5) Income (1-3) -.012** -.36* .17*** .13 -.018*** -.21 .25*** .10 -.017*** -.15 .29*** .16 -.006* -.30*** .03 .24*** -.008** -.18* .09*** .25*** -.010*** -.21* .12*** .21*** Woman’s role (0-1) Defense Spending (0-1) Opp. to Abortion (0-1) ---- 1.27*** 1.73*** .97*** 1.36*** 1.51*** .93*** ---- 1.22*** .84*** .99*** 1.10*** .64** .95*** Symbolic Racism (0-1) Black stereotype (0-1) --- --- 2.03*** .48 --- --- 1.57*** .12 + Racial attitudes Constant 3.16 1.22 -.66 2.94 1.50 R Squared .07 .15 .19 .05 .12 N 844 785 681 2224 2258 Entries are unstandardized B’s from an OLS regression analysis. The dependent variable is the standard 7 point party identification scale, with higher values corresponding to Republican attachment. Excluded group is “other,” including Catholics, Jews, non-religious, non-traditional orthodox, and non-Jewish/non-Protestants. 47 + Racial attitudes .62 .13 1798 Table 5. Percent of Population belonging to Various Religious Groups. Former Confederacy North + West Difference All States Evangelical Protestant 49% (1445) 19% (1416) 30% 27% (2861) Mainline Protestant 23% (691) 31% (2353) -8% 29% (3044) Catholic 15% (456) 31% (2392) -16% 27% (2848) Other 13% (388) 19% (1477) -6% 18% (1865) Regional Total 100% (2980) 100% (7638) Note: Entries are percentage falling above the population median in symbolic racism. Cell N’s in parentheses. Columns do not add to 100% due to rounding. “Other” consists of Jews, non-religious, nonChristian non-Jewish, and non-traditional Orthodox. Source: Pooled National Election Studies: 1986, 1988,1990, 1992, 1994, 2000. 48 101% (10618) 49 Endnotes 1. We would like to express our appreciation to Sharmaine V. Cheleden and Christopher Tarman for their contributions to various aspects of this project. We would also like to recognize the contribution of David Leege in developing new measures of religiosity for the National Election Studies. This project would not have been possible had we not been able to use those measures. Chris Achen, Martin Gilens, Vincent Hutchings, and John R. Petrocik provided valuable feedback at various points in the theoretical development and analysis. Pat Luevano helped us understand the process whereby David Leege’s research was operationalized in the NES Cumulative Study religion codes. 50 2. Throughout this paper, we will refer to the South interchangeably with the eleven former Confederate states: Texas, Louisiana, Mississippi, Alabama, Georgia, Florida, South Carolina, North Carolina, Virginia, Arkansas, and Tennessee. 3. This is not to claim that either North or South was unanimous on the subject of slavery at that time; much support for it existed in the North, and it was not fully supported in the South. But those who advocated abolition, however gradual, were quickly dispatched to the sidelines at the writing of both the Declaration of Independence and Constitution (see, for example, Ellis, 1999; Klincknor & Smith, 1999). 4 Our primary goal is to explain aggregate regional differences over time, not develop models explaining the maximum variance in vote choice or opinion. Hence the usual concerns about low R-squareds do not pertain here. Multicollinearity is also of little concern here, as region (South vs. other) and time are uncorrelated (Pearson’s r = .04). 5. The symbolic racism scale is composed of four items..... 6. It might be objected that this control is insufficient because a single-item ideology item is less reliable than the multi-item symbolic racism scale, and measurement error therefore renders the ideology control incomplete. This argument is plausible a priori but in practice unlikely to explain the effects we report. Symbolic racism normally has far stronger effects than ideology on racial policy preferences than does ideology when both are included in the same equation (e.g., Sears et al., 1997). But when both are measured with multi-item scales, symbolic racism is actually somewhat less reliable (e.g., Sears et al., 1997, found an average Cronbach’s alpha of .72 for symbolic racism across four surveys with an average of five items each, and an average alpha of .83 for three-item ideology and party identification scales). The gross disparity in effects is not likely to be explained by this small difference in reliability. 7 We combined the two racial policy items into a single scale items to save space, and on the basis of extensive evidence that they work in parallel (e.g., Sears et al. 1997). When analyzed separately, both produce significant and nearly identical interaction effects. 8. A few of the more important book-length products in this extensive area of research include Kohut, Green, Ketter, & Toth, 2000; Green, Rozell, & Wilcox, 2000; Rozell & Wilcox, 1995; Wilcox, 1996; Leege & Kellstedt, 1993; Green et al., 1996; Jelen, 1991). 9. The databases we rely on do not have sufficient numbers of cases to allow us to make sharp distinctions among these terms, so we use them more or less interchangeably. 10. Similar cleavages have been found among Catholics and Jews, as well as among black Protestants, but their political impact is less (e.g., Kohut et al., 2000). This line of research has also touched on the rise of intense anti-fundamentalist attitudes, but that is beyond the scope of this paper (Bolce & DeMaio, 1999a, 1999b). 51 11. Although see Bolce and DeMaio (1999b), who explore the role of symbolic racism in whites’ attitudes toward fundamentalists. 12. Kohut and his colleagues (2000), however, report a complex pattern of similarities and differences between white and black evangelicals. For example, they find that they differ sharply on issues related to race. At the same time, blacks do not oppose the narrower moral agendas of fundamentalist Christians as strenuously (see Kohut et al., 2000). However, the complexities of racial differences in the political impact of religion go beyond the scope of this paper. 13. The importance of values in the theory is illustrated by the fact that Kinder and Sanders (1996) switched to the term, “racial resentment” on the grounds that “symbolic racism” did not convey the central role of the violation of values in the theory, leaving the theory open to the misinterpretation that symbolic racism was just racism and that values merely provided an epiphenomenal justification for racial animosity. 14. For an important exception, that focuses specifically on the realignment to the Republicans of committed Southern evangelicals, see Green, Kellstedt, Smidt, and Guth (1998). 15. We have borrowed our religious classification from that developed for the National Election Studies by David Leege. We have done so deliberately in order to insure that the religious classification, which is central to our data analysis, is neutral with respect to our own hypotheses. The refinement most useful for our purposes is what Leege et al. refer to as “religious tradition” in their pilot report, which establishes the distinction between “mainline” and “evangelical” Protestants in each of the National Election Studies beginning in 1990. Through a series of branching questions, Leege et al.’s new measure is able to identify those respondents who identify with evangelical or fundamentalist denominations within the larger Protestant families such as Lutherans, Pentacostal, Baptist, Episcopal, etc. Leege et al. test validity of the distinction between evangelicals and mainline Protestants extensively in their pilot study report. For example, 82% of those classified as evangelical Protestant based on denomination alone identified themselves as Born Again, compared to only 27% of non-evangelical Protestants. In addition, 69% of evangelical Protestants claimed that the bible is the literal word of God, while only 31% of nonevangelicals expressed this belief. Finally, evangelicals were far more likely to self-identify as evangelical than were mainline Protestants (67% compared to 34%). We believe this classification is the most valid one available, but any who are interested in the finer workings of the measure are encouraged to review it in the Pilot Study Report of 1990. 16. We are taking at face value the non-racial character of these issue attitudes. 17. At the extreme right wing, Christian identity is closely linked to white identity, support for the KKK, and violent antagonism toward Jews and peoples of color is manifested in support for “white bastions” in the Northwest. Indeed the Christian Identity movement is explicitly racebased (Bermanzohn, 1998). We would not claim that this reflects the views of the mainstream in any section of the country, but it does provide models for this linkage. 52
© Copyright 2026 Paperzz