Shirley Harvell and other voters filed suit in 1989, challenging the election procedure for school board members in the Blytheville, Arkansas, school district as violative of Section 2 of the Voting Rights Act, 42 U.S.C. § 1973 (“VRA”). The district court dismissed the complaint, finding that Harvell failed to set forth a prima facie violation of Section 2. Harvell v. Ladd, 759 F.Supp. 525, 529-30 (E.D.Ark.1991). We reversed and remanded to the district court for further findings concerning the factors set forth in Thornburg v. Gingles, 478 U.S. 30, 36-37, 106 S.Ct. 2752, 2758-59, 92 L.Ed.2d 25 (1986). Harvell v. Ladd, 958 F.2d 226, 230 (8th Cir.1992). On remand, the district court again dismissed the complaint because it found that Harvell failed to establish that the unsuccessful black candidates were the minority’s preferred candidates. We again reversed in part the district court’s subsequent decision following remand, but affirmed its denial of sanctions against Harvell. Harvell v. Blytheville Sch. Dist. # 5, 33 F.3d 910 (8th Cir.1994). We subsequently agreed to rehear the case en bane and vacated the panel opinion. We now reverse the decision of the district court in part and affirm the denial of sanctions.
I.
The underlying facts of this dispute are set out in detail in our earlier opinions. The voting age population of the Blytheville school district is 14,500, of which 70% is white and 29% black.1 Each member of the eight member at-large school board serves a four-year term, and two terms expire each year. Historically, school board elections were determined by a plurality vote. In 1987, however, the Arkansas legislature altered the voting scheme for school board elections from one of plurality to one of majority, which would occasion the need for a run-off election between the voter-preferred candidates in the event that no candidate was able to garnish a majority of the voters in the initial election. Ark.Code. Ann. § 6-14-121. AH winning candidates since 1987 have been elected by a majority of votes cast in the first round. Following two elections under this scheme Harvell filed suit. None of six separate black candidates has defeated a white candidate in eight attempts following the 1987 election, although Dr. Helen Nunn was reseated without opposition in 1990.2
II.
To mount a successful challenge to multi-member districts under Section 2,3 a plaintiff must initially satisfy the three preconditions delineated in Gingles. These preconditions are 1) that the minority group is large enough and geographically compact enough that it would be a majority in a single-member district; 2) that the minority group is politically cohesive; and 3) that the white majority votes sufficiently as a bloc to enable it, in the absence of special circumstances, usually to defeat the minority’s preferred candidate. 478 U.S. at 50-51, 106 S.Ct. at 2766-67. Gingles also set forth nine non-exclusive factors mentioned in the Senate report accompanying the VRA to assist in determining whether, under the totality of the circumstances, a challenged electoral scheme dilutes the minority vote. These include (1) the history of voting-related discrimination in the state or political subdivision; (2) the extent to which voting in the state or subdivision is racially polarized; (3) the extent to which the state or subdivision has used voting practices or procedures that tend to enhance opportunities for discrimination against the minority group; (4) whether minority candidates have been denied access *1386to any candidate-slating process; (5) the extent to which minorities have borne the effects of past discrimination in relation to education, employment, and health; (6) whether local political campaigns have used overt or subtle racial appeals; (7) the extent to which minority group members have been elected to public office in the jurisdiction; (8) whether there is a significant lack of responsiveness on the part of elected officials to the particularized needs of members of the minority group; and (9) whether the policy underlying the use of voting qualifications is tenuous. 478 U.S. at 36-37, 106 S.Ct. at 2768-59.
We must analyze the elements of a Section 2 case in context, according deference to the district court where necessary and applying legal constructs where appropriate. The district court’s findings regarding the factual context giving rise to the claim are reviewed for clear error. See Gingles, 478 U.S. at 78-79, 106 S.Ct. at 2780-81. But the legal conclusions it employs, “including those that may infect a so-called mixed finding of law and fact, or a finding of fact that is predicated on a misunderstanding of the governing rule of law,” are subject to plenary review. Id. at 79, 106 S.Ct. at 2781 (quoting Bose Corp. v. Consumers Union of U.S., Inc., 466 U.S. 485, 104 S.Ct. 1949, 80 L.Ed.2d 502 (1984)).
The district court found that Harvell established the first two Gingles preconditions at trial. The school district does not contest these findings of geographic compactness and political cohesiveness. We therefore accept them as established. The district court also found that voting in the school board elections is racially polarized. Har-vell, 759 F.Supp. at 527-28. This undisputed finding is borne out by the record and weighs heavily in favor of finding the third Gingles precondition established. Jeffers v. Clinton, 730 F.Supp. 196, 205 (E.D.Ark.1989) (three-judge court), aff'd mem., 498 U.S. 1019, 111 S.Ct. 662, 112 L.Ed.2d 656 (1991); Smith v. Clinton, 687 F.Supp. 1310, 1314-15 (E.D.Ark.1988) (three-judge court) see Collins v. City of Norfolk, 816 F.2d 932, 935 (4th Cir.1987) (existence of polarization establishes the power of white bloc voting). The district court found, however, and the school district contends, that the low voter turnout sufficiently indicates that the minority candidates in the elections from 1988 to 1992 were not preferred by the minority voters and precluded Harvell from satisfying the third Gin-gles precondition. Harvell asserts, and we agree, that the district court misapprehended the definition of who is eligible for “minority-preferred candidate” status and that the evidence establishes that the candidates in those elections were in fact the minority-preferred candidates.
We do not categorically state that a candidate is the minority-preferred candidate simply because that candidate is a member of the minority. Such stereotyping runs afoul of the principles embodied in the Equal Protection Clause. See Miller, — U.S. at -, 115 S.Ct. at 2486 (state assignment of voters on basis of racial assumptions is “offensive and demeaning”). But see Jenkins v. Red Clay Consol. Sch. Dist. Bd. of Educ., 4 F.3d 1103, 1126 (3d Cir.1993) (stating that “practical experience leads to the inference that any particular minority candidate is minority-preferred”), cert. denied, — U.S. -, 114 S.Ct. 2779, 129 L.Ed.2d 891 (1994). Accordingly, such an inference is insufficient to establish that any particular candidate is minority-preferred. The preferences of the minority voters must be established on an election-specific basis, viewing all the relevant circumstances. See Jenkins, 4 F.3d at 1126.
The record makes clear in this case that the race of a candidate is by far the determinative factor in minority voting patterns in Blytheville school board elections. Harvell’s regression analysis strongly supports the conclusion that those black candidates who did run in the elections from 1988-1992 were in fact the minority-preferred candidates.4 See id. (inference derived from candidate’s race combined with statistical evidence of voting patterns is sufficient to establish minority preference); Clarke v. City of Cincinnati, 40 F.3d 807, 810 n. 1 (6th Cir.1994) (“[Cjourts generally have understood blacks’ preferred candidates simply to be those can*1387didates who receive the greatest support from black voters.”)) cert. denied, — U.S. -, 115 S.Ct. 1960, 131 L.Ed.2d 851 (1995); see also Jeffers, 730 F.Supp. at 208 (noting race-conscious voting behavior of Arkansans). The school district does not contest this statistical evidence and offers no other evidence to contradict the statistical preference. The consistency of the data over time lends additional support to this conclusion. Gingles, 478 U.S. at 57, 106 S.Ct. at 2769-70.
There may be situations in which voter apathy may be linked to disapproval of a particular candidate, but there is no indication that such is the case here. The silence of the minority voters is not so deafening as to warrant a finding that they disapproved of six different minority candidates in light of the uncontroverted statistical evidence that supports a finding of overwhelming support from those blacks who did vote. Nor are the numbers so low as to reduce their statistical significance to a nullity. Speculation regarding reasons for low minority turnout is inappropriate. Gomez v. City of Watsonville, 863 F.2d 1407, 1415-16 (9th Cir.1988) (looking at actual voting patterns to determine political cohesion rather than turnout rates), cert. denied, 489 U.S. 1080, 109 S.Ct. 1534, 103 L.Ed.2d 839 (1989). The school district’s interpretation of the low turnout underestimates the legal significance of the years of polarized voting evident in the Blytheville school board elections. See Gingles, 478 U.S. at 57, 106 S.Ct. at 2769-70; see also Clark v. Calhoun County, 21 F.3d 92, 96 (5th Cir.1994) (ability of majority to defeat the minority-preferred candidate is “ordinarily established through evidence of racially polarized voting”). Similarly, the related natures of cohesiveness, polarization, and bloc voting demonstrate the incongruity of any reliance on low voter turnout in this ease. See Clark, 21 F.3d at 96; Jenkins, 4 F.3d at 1133 n. 32 (discussing relation between political eohe-siveness and the minority-preferred candidate); Collins v. City of Norfolk, 883 F.2d 1232, 1237 (4th Cir.1989) (polarization relevant to determination of eohesiveness and whether white bloc voting defeats minority-preferred candidates); Collins, 816 F.2d at 935 (“[Tjhe existence of racially polarized voting ... establishes both cohesiveness of the minority group and the power of white bloc voting to defeat the minority’s candidates.”) (emphasis omitted); see also Gingles, 478 U.S. at 52 n. 18, 106 S.Ct. at 2767 n. 18 (noting interchangeable use of polarization and bloc voting); United States v. Marengo County Comm’n, 731 F.2d 1546, 1567 (11th Cir.) (“The surest indication of race-conscious politics is a pattern of racially polarized voting.”), cert. denied, 469 U.S. 976, 105 S.Ct. 375, 83 L.Ed.2d 311 (1984). The fact that black candidates have lost universally in contested post-1987 elections cannot be explained away simply on the basis of the black electorate’s dissatisfaction with its candidates. In the face of the finding of consistent polarization, a legally significant white cross-over vote does not exist under the current election scheme. Cf. Cane v. Worcester County, 35 F.3d 921, 926 (4th Cir.1994) (finding average 19% white crossover vote insufficient to salvage at-large, one-on-one election scheme under the facts of that case), cert. denied, — U.S. -, 115 S.Ct. 1097, 130 L.Ed.2d 1065 (1995).
That some minority-preferred candidates did achieve electoral success under the old system does not necessitate a finding that those minority candidates who do not now succeed are not the minority’s preferred candidates, particularly when no candidates in these elections received greater minority voter support. Moreover, we should not require as a condition to a finding of minority-preferred status heroic, mythic qualities of candidacy better suited to a romantic view of the electoral process than to the rough-and-tumble world of contested elections. Much as the relevant population is the population able to vote, see African Am. Voting Rights Legal Defense Fund, Inc. v. Villa, 54 F.3d 1345, 1352 (8th Cir.1995), the relevant candidate should generally be one able to receive those votes, not some idealized figure whose absence from the ballot keeps a disappointed electorate at home. Under the school district’s proffered view, the black voters of Blytheville have gone five consecutive years, six candidates, and eight campaigns without stating a preference, a proposition we cannot accept.
*1388There are a number of factors in addition to the statistical data that indicate that the district court’s reliance on low turnout is too speculative. Low voter turnout can be explained by any number of socio-economic factors, or may be explained by unabashed voter apathy. It does not, however, lend itself so easily to an inference of a sub silentio predisposition against any particular minority candidate. Not only were the minority turnouts low, but voter turnout as a whole for these elections has been historically low. Additionally, we note that following the change in the law the average vote totals declined for both white and minority candidates, although much more precipitously for black candidates. In light of the fact that low voter turnout has often been considered the result of the minority’s inability to effectively participate in the political process, see Gomez, 863 F.2d at 1416 n. 4; United States v. Dallas County Comm’n, 739 F.2d 1529, 1538 (11th Cir.1984), it stands to reason that when an external stimulus dampens the white turnout it may impact even more greatly on a group that has faced historic disadvantages.5
Similarly, black voters need have only looked at their plurality successes in 1974 and 1975 to realize that they faced a much lower possibility of success under the present scheme. This may also account for the lower turnout. Finally, varying election dates, the number of seats up for election, and the presence or absence of other ballot issues that may draw the electorate to the polls also have an impact on turnout. Such considerations have no bearing on whether the candidates who ran were or were not the minority-preferred candidates. Indeed, there are so many possibilities for explaining the low turnouts that selecting any one in this ease is purely impermissible speculation.
Nor can we accept the school district’s attempts to characterize the unsuccessful candidates as militant fringe candidates who disenfranchised themselves from the black community. There is no evidence in the record that the minority community viewed any of the six different candidates as in any way inadequate representatives of its interests. In the absence of any such evidence, the countervailing voting patterns conclusively rebut these denigrations.
The school district relies heavily on the historical fact of proportional or near proportional representation of the black population on the school board as evidence of the election scheme’s validity. At the outset we note that proportional representation is an important factor to consider in evaluating the validity of an electoral process. It is not, however, the statutory touchstone, and does not provide an absolute safe harbor in which a defendant can seek refuge from the totality of the circumstances. See Johnson v. De Grandy, — U.S. -, -, 114 S.Ct. 2647, 2660-61, 129 L.Ed.2d 775 (1994);6 Barnett v. Daley, 32 F.3d 1196, 1202 (7th Cir.1994); see also Zimmer v. McKeithen, 485 F.2d 1297, 1307 (5th Cir.1973) (listing possible explanation for success at the polls despite vote dilution), aff'd sub nom., East Carroll Parish Sch. Bd. v. Marshall, 424 U.S. 636, 96 S.Ct. 1083, 47 L.Ed.2d 296 (1976). Just as proportional representation is not mandated under Section 2, it also does not preclude finding a violation, because racial reference points do not necessarily reflect political realities. See De Grandy, — U.S. at - & n. 17, 114 S.Ct. at 2661-62 & n. 17 (probative value of proportionality *1389varies with the facts). Similarly, the white majority has no right under Section 2 to ensure that a minority group has absolutely no opportunity to achieve greater than proportional representation in any given race.
The proportional representation on which the school district relies rests on infirm ground. The electoral success that black candidates achieved under the plurality system is no longer present. No black candidate has won a contested election since the change in the law. The district contends that none of those candidates would have won under a plurality system either, thus proving that the majority vote requirement is not to blame, and further requiring us to weigh heavily the historic successes of black candidates in rejecting Harvell’s complaints. This sort of back-to-the-futurism on the part of the school district cannot withstand analysis, for if we control the results of the elections prior to 1988 with the majority vote requirement, it becomes plain that no proportional representation would have occurred in the past. Both the 1974 and 1975 elections would have gone to white candidates under this scenario; thus, Norvell Moore would have never won initially to enable him to serve multiple terms as a mostly unopposed incumbent. Dr. Nunn’s 1982 victory would have provided the only successful black candidate in Blytheville history had the current scheme been in place then. Such success does not indicate substantial minority voting strength, nor does it indicate the presence of a consistent, legally significant white cross-over vote.
Further, even accepting the historical effects of the plurality element, most of the elections won by black candidates were done so as incumbents in the face of no opposition. This is precisely the type of special circumstance recognized in Gingles as not vitiating any element of the claim. 478 U.S. at 51,106 S.Ct. at 2766-67. Even in an extreme case of total vote dilution a candidate running in the face of no opposition is ensured success. The two instances in which a black candidate beat a white candidate head-to-head involved Moore’s 23-vote victory as a three-term incumbent7 against a candidate who was not even the white majority’s preference as a previous candidate, and Dr. Nunn’s election in 1982. So in the final analysis, when those campaigns involving readily apparent special circumstances are removed, black candidates still have won only one election in thirteen attempts since 1969, fitting precisely in the Gingles test as to whether the white majority does indeed vote “sufficiently as a bloc to enable it ... usually to defeat the minority’s preferred candidate.” Gingles, 478 U.S. at 51, 106 S.Ct. at 2766-67.
We emphasize that we are not assessing the validity of the law that existed prior to 1988. It no longer exists, and its legitimacy is inconsequential. The election results under that scheme and the type of representation they yielded, however, are relevant to our inquiry to the extent that the past political reality sheds light on the totality of the circumstances of the present scheme. Proportional representation may well have sufficed to protect the former scheme from a Section 2 claim, but the truth of the current system is that minority candidates face substantial impediments to election8 that are too stark for us to dismiss their attendant results as mere happenstance. It is a system that allows only for victory of majority-preferred minority candidates. In essence, any minority serving on the board does so at the sufferance of the majority — a hollow proportional representation indeed when the minority must rely on majority benevolence to ensure the adequacy of its representation. A system that works for minorities only in the *1390absence of white opposition is a system that fails to operate in accord with the law.
Satisfaction of the necessary9 Gingles preconditions carries a plaintiff a long way towards showing a Section 2 violation, Jenkins, 4 F.3d at 1116 n. 6, 1135; see Clark, 21 F.3d at 97, but in the final analysis Harvell must still show that the challenged electoral scheme provides minority voters “less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice.” 42 U.S.C. § 1973(b). The totality of the circumstances on the record before us establishes that dilution. See Gomez, 863 F.2d at 1411 (remand not necessary when the record “permits only one resolution of the factual issue” (quoting Pullman-Standard v. Swint, 456 U.S. 273, 287 [sic; 292], 102 S.Ct. 1781, 1789 [sic; 1792], 72 L.Ed.2d 66 (1982))).
The two primary factors considered in our totality analysis are the extent to which voting is racially polarized and the extent to which minorities have been elected under the challenged scheme. Gingles, 478 U.S. at 48-49 n. 15, 106 S.Ct. at 2765-66 n. 15. As we have discussed, voting practices in the school board elections are highly polarized. There also has been only minimal electoral success under the present scheme, with Dr. Nunn’s unopposed reseating as an incumbent in 1990 representing the only minority victory in nine attempts.
The remaining totality factors, although not essential to Harvell’s claim, see id., support our conclusion that Harvell has identified a Section 2 violation. No one party to the litigation denies the long history of racial discrimination in the electoral process in Arkansas. See Perkins v. City of West Helena, 675 F.2d 201, 211 (8th Cir.), aff'd mem., 459 U.S. 801, 103 S.Ct. 33, 74 L.Ed.2d 47 (1982). It is true, as noted by the district court, that strides have been made since the dawn of the civil rights movement, but Arkansas remains by no means idyllic for black voters. Jeffers, 730 F.Supp. at 204. We conclude that the district court did not accord sufficient weight to the vestiges of that history. See Westwego Citizens for Better Gov’t v. City of Westwego, 872 F.2d 1201, 1211-12 (5th Cir.1989). So too, the recognized historic effects of discrimination in the areas of health, employment, and education impact negatively on minority political participation. The majority vote requirement, staggered terms, and at-large structure also tend to suppress minority voters’ influence. See De Grandy, — U.S. at -, 114 S.Ct. at 2660-61; Collins, 883 F.2d at 1236 (at-large system and staggered terms susceptible of diluting minority votes). The manner in which those black candidates who were elected managed to obtain and retain their seats attenuates the district’s reliance on proportional representation, and similarly heightens our concern over the new election scheme, which while retaining the at-large structure,10 removes the potential for winning on the basis of a split white vote.
No evidence of a candidate-slating process was adduced, nor was any evidence of racial appeals presented, but the presence of such insidious accoutrements is unnecessary in situations where the process controls the result, and their absence does not preclude .finding a Section 2 violation. Further, the district court’s finding that the school board was not unresponsive to the minority community was predicated on its belief that because those black candidates who did serve on the board were the minority-preferred candidates, those who did not win election to the board must have been rejected by black voters as potentially less responsive. Our earlier discussion of the minority-preferred *1391candidate status counters that proposition. Even accepting the finding of responsiveness as not clearly erroneous, however, it is similarly insufficient to counter the other factors that censure this scheme. See Westwego Citizens for Better Gov’t v. City of Westwego, 946 F.2d 1109, 1123 (5th Cir.1991).
The totality of the circumstances on this record leaves us with the firm conviction that the Blytheville school board electoral scheme is institutionally restrictive, thereby creating “an inequality in the opportunities enjoyed by black and white voters to elect their preferred representatives.” Gingles, 478 U.S. at 47, 106 S.Ct. at 2764. Racial polarization effectively cripples minority attempts to effect representation on the school board in the absence of the fortuity of white voter acquiescence. The fact that the current school board election system is less accessible to minority participation than that of ten or twenty years ago runs contrary to the intent and standard of Section 2.
The Supreme Court’s recent redistricting decision in Miller does not alter our analysis of the Gingles factors or our ultimate decision in this appeal. Miller analyzed the equal protection problems involved in drawing voting districts along race-based lines, but did not purport to alter our inquiry into the vote-dilution claim. See — U.S. at -, 115 S.Ct. at 2485-86. We do, however, sound a cautionary note to the district court on remand to steer clear of the type of racial gerrymandering proscribed in Miller, while keeping in mind the need to vindicate the rights of the minority voters.
Our holding that plaintiffs are entitled to relief on their claim perforce forecloses defendant’s application for sanctions pursuant to Fed.R.Civ.P. 11 and attorneys’ fees pursuant to 42 U.S.C. § 1988.
The judgment is affirmed in part and reversed in part, and the ease is remanded to the district court for the entry of an appropriate remedial decree.
. These are the 1980 census figures that were used at trial. According to the 1990 census, the Blytheville school district has a total population of 23,057, of which approximately 37% are black and 62% white, but we do not have the relevant voting age populations before us.
. Appendix I contains the vote breakdown by candidate for each election in which a black candidate has run.
.This vote dilution claim is "analytically distinct” from a challenge to voting districts on equal protection grounds. Miller v. Johnson, U.S. - , -, 115 S.Ct. 2475, 2485, 132 L.Ed.2d 762 (1995) (quoting Shaw v. Reno, - U.S. -, -, 113 S.Ct. 2816, 2830, 125 L.Ed.2d 511 (1993)).
. Appendix II contains the results of Harvell’s expert statistical evidence.
. We further note that in 1990 the polling place of a precinct known as a historic black voter stronghold was moved to a predominantly white area. Although there is no evidence that this move was predicated on an attempt to suppress the black vote, our inquiry looks to the effect of voting practices on equal political opportunity rather than to their intent, see Chisom v. Roemer, 501 U.S. 380, 383-84, 111 S.Ct. 2354, 2357-58, 115 L.Ed.2d 348 (1991); Jeffers, 730 F.Supp. at 203-04, and the 1990 election produced a record low black voter turnout for that particular precinct.
. De Grandy resolved a claim involving “proportionality," which "links the number of majority-minority voting districts to minority members’ share of the relevant population.” - U.S. at - n. 11, 114 S.Ct. at 2658 n. 11. Here, because we address a claim involving a single at-large district, the analyses between proportionality and proportional representation are essentially the same. Cf. Villa, 54 F.3d at 1352-53 n. 10 (distinguishing proportional representation in case involving proportionality claim and single-member districts).
. We agree that incumbency is the least “special" of the special circumstances, compare Clarke, 40 F.3d at 813-14 (de-emphasizing incumbency element) with Collins, 883 F.2d at 1243 (emphasizing the importance of incumbency), but accept its potential as such and view this election as a special circumstance because of the multiple incumbency, board composition, and opposition involved.
. In its petition for rehearing the district pointed out that another black person has been appointed to the school board, once again bringing representation into proportion with the black population. This appointment is irrelevant to the ability of black voters to elect the representatives of their choice. Nor does his unopposed reseat-ing in 1993 convince us of the soundness of the election scheme.
. The Supreme Court hedged slightly on this notion in' De Grandy, at one point calling the preconditions “generally necessary.” -U.S. at -, 114 S.Ct. at 2657. The preeminence of these preconditions in the totality analysis has consistently been recognized in recent opinions by the Supreme Court, however, see Voinovich v. Quilter, 507 U.S. 146, 156-60, 113 S.Ct. 1149, 1157-58, 122 L.Ed.2d 500 (1993); Growe v. Emison, 507 U.S. 25, 37-42, 113 S.Ct. 1075, 1083-85, 122 L.Ed.2d 388 (1993), and we continue to accord them their deserved primacy.
. Following our decision in the first appeal in this case, the Arkansas legislature amended its school board election law to require seven single-member districts, or five single-member districts and two at-large representatives in all school districts having a minority population of 10% or more. If a school district has been found to not be in violation of the VRA, it is excepted from this provision. Ark.Code Ann. § 6-13-631.