We consider here a challenge under Section 2 of the Voting Rights Act of 1965, as amended, 42 U.S.C. § 1973, to a municipality’s at-large method of electing members to its governing body. In 1985, the City of Niagara Falls, New York (“City” or “Niagara Falls”) approved by referendum the formation of a seven-member City Council, the members of which are elected at large. The first election under this system took place in the fall of 1987. Plaintiffs filed suit in September 1989, contending that the at-large system operates to “dilute” black voting strength in violation of § 2. They seek to replace it with a single-member-district method of electing members to the City Council, including a requirement that one district have a majority black population.
We review the judgment of the United States District Court for the Western District of New York (William M. Skretny, Judge), denying the plaintiffs declaratory and injunctive relief after a bench trial held in October and November 1993. The district court held that the plaintiffs failed to show that the white majority in Niagara Falls votes sufficiently as a bloc to enable it, in the absence of special circumstances, usually to defeat blacks’ preferred candidate — the third prong of the so-called Gingles test. See Thornburg v. Gingles, 478 U.S. 30, 51, 106 S.Ct. 2752, 2766-67, 92 L.Ed.2d 25 (1986). We first consider whether, in reaching this conclusion, the district court gave too little weight to elections that took place before the City adopted the contested governing scheme. We then consider whether the court gave too much weight to the electoral *1005successes of white candidates supported by black voters. Finally, we review the district court’s ultimate conclusion that, under the totality of the circumstances, the plaintiffs failed to prove vote dilution. We hold that the district court erred in finding that the plaintiffs failed to satisfy the third Gingles prong. But we hold that its ultimate conclusion that, under the totality of the circumstances, the plaintiffs failed to prove vote dilution under § 2, is not clearly erroneous. On this basis, we affirm.
I
The Electoral System
From 1915 until January 1, 1988, the City of Niagara Falls was governed by a mayor and a separate Council of four members, all of whom were elected at large for four-year terms. Terms were staggered, with two Council members elected every two years. A City Charter Revision Commission studied this system and recommended four options for change, on which a referendum was held in November 1985. The voters of Niagara Falls chose between two basic forms of municipal government, and for each basic type there was a choice between an at-large election system and an election system based on single-member districts. The majority of voters chose the so-called Mayor-Council form of government, with the mayor as the chief executive officer, and a separate Council of seven members, all to be elected at large. The voters rejected the option of having the seven Council members elected from seven separate, single-member districts. Under the new system, put into place January 1,1988, the mayor and Council members are elected to four-year terms; the elections of Council members are staggered and held every two years; and voters are able to vote for as many candidates as there are seats available. There are no district residency requirements. Elections have been conducted under the new scheme since 1987.
The Racial Composition of Niagara Falls
In 1970, blacks constituted 10.06% of the voting-age population in the City of Niagara Falls and 9.3% of the total population. In 1980, the black population grew to 12.94% of the total population, but its percentage of the voting-age population remained roughly the same. By 1990, blacks constituted 15.58% of the total population and 13.02% of the City’s voting-age population.1
The Plaintiffs’ Case
The plaintiffs include the National Association for the Advancement of Colored People (“NAACP”), Renae Kimble (its political action chairperson in Niagara Falls), and eighteen other registered black voters of Niagara Falls. When the plaintiffs filed this suit in September 1989 against the City, its Mayor, the members of the City Council, and the City Clerk, no black had ever been elected to the City Council. At that time, only one election to fill seats on the newly enlarged City Council had been held. In 1991, Andrew Walker, a black candidate, won the Democratic Party primary for City Council, and he received the most votes in the general election. He was chairman of the Niagara Falls City Council at the time of the trial and is thus a defendant in this case.
Plaintiffs’ proof consisted in large part of expert analyses of various elections dating back to 1969 that involved black and white candidates. One of their experts, Professor Michael McDonald, analyzed elections using simple regression analysis2 and, to a lesser *1006extent, extreme case analysis.3 The defendants’ expert, Professor Harold Stanley, analyzed many more elections, including those that involved only white candidates, and he primarily applied double regression analysis.4 The parties agreed, and the district court found, that the results obtained by the different methodologies were not materially different. The court relied on the testimony of both experts in entering its findings of fact.
McDonald testified that there were no significant differences in voter turnout by race in the City Council primary elections.5 He testified that black voter turnout trailed that of whites in the general elections, but the district court found that the “plaintiffs did not undertake to analyze comprehensively the significance of these turnout differences.” Decision and Order at 15. The plaintiffs do not challenge this finding on appeal.
II
In 1982 Congress amended § 2 of the Voting Rights Act in large part to eliminate the requirement, declared by a plurality of the Supreme Court in City of Mobile v. Bolden, 446 U.S. 55, 100 S.Ct. 1490, 64 L.Ed.2d 47 (1980), that plaintiffs in § 2 eases prove that the challenged electoral system was created or maintained for the purpose of discriminating against minorities. Gingles, 478 U.S. at 35, 106 S.Ct. at 2758. The amendments to § 2 of the Act, together with a Senate Judiciary Committee Report that accompanied the bill amending § 2, S.Rep. No. 417, 97th Cong., 2d Sess. (1982) (“S.Rep.” or “Senate Report”), reprinted in 1982 U.S.C.C.A.N. 177, made clear that proving such discriminatory intent was not a statutory requirement. Instead, § 2 plaintiffs could prevail by proving that minorities were denied an equal opportunity to participate in the political process and elect candidates of their choice as a result of an electoral practice. See Gingles, 478 U.S. at 43-44 & n. 8, 106 S.Ct. at 2762-63 & n. 8 (discussing how amended § 2 reinstated the “results test” applied by courts before the 1980 Bolden decision).
Section 2, as amended, reads as follows:
(a) No voting qualification or prerequisite to voting or standard, practice, or procedure shall be imposed or applied by any State or political subdivision in a manner which results in a denial or abridgement of the right of any citizen of the United States to vote on account of race or col- or....
(b) A violation of subsection (a) is established if, based on the totality of circumstances, it is shown that the political processes leading to nomination or election in the State or political subdivision are not equally open to participation by members of a class of citizens protected by subsection (a) in that its members have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice. The extent to which members of a protected class have been elected to office in the State or political subdivision is one circumstance which may be considered: Provided, That nothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion in the population.
42 U.S.C. § 1973 (all emphasis supplied except “Provided,”).
The Supreme Court interpreted this amended statute for the first time in a chal*1007lenge to a multimember at-large districting scheme enacted by the North Carolina General Assembly. Gingles, 478 U.S. at 34, 106 S.Ct. at 2758. The Gingles Court held that, in order to establish unlawful dilution of minority voting strength, or “vote dilution,” a plaintiff must show that (1) the minority group “is sufficiently large and geographically compact to constitute a majority in a [hypothetical] single-member district”; (2) the minority group is “politically cohesive”; and (3) “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.” Gingles, 478 U.S. at 50-51, 106 S.Ct. at 2766. Regarding the third prong, Gingles instructed that “in general, a white bloc vote that normally will defeat the combined strength of minority support plus white ‘crossover’ votes rises to the level of legally significant white bloc voting.” Id. at 56, 106 S.Ct. at 2769.
Proof of these three factors is not enough to demonstrate vote dilution. Even if these factors are satisfied, a court must consider whether, under the totality of the circumstances, the challenged standard, practice or procedure “impair[s] the ability of ... [minority] voters to participate equally in the political process and to elect candidates of their choice.” Id. at 80, 106 S.Ct. at 2781. Gingles explicitly treated as relevant to this inquiry the factors set forth in the Senate Judiciary Committee Report (the “Senate Report Factors”), see id. at 36-37, 44-45, 106 5.Ct. at 2758-59, 2762-64, which, according to the Senate Report, were based on the Supreme Court’s approach in White v. Regester, 412 U.S. 755, 765-70, 93 S.Ct. 2332, 2339-41, 37 L.Ed.2d 314 (1973), as articulated in Zimmer v. McKeithen, 485 F.2d 1297, 1305 (5th Cir.1973) (en banc), 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). S.Rep. at 28 n. 113, 1982 U.S.C.C.A.N. at 206 n. 113. The Senate Report lists the relevant factors as follows:
1.the extent of any history of official discrimination in the state or political subdivision that touched the right of the members of the minority group to register, to vote, or otherwise to participate in the democratic process;
2. the extent to which voting in the elections of the state or political subdivision is racially polarized;
3. the extent to which the state or political subdivision has used unusually large election districts, majority vote requirements, anti-single-shot provisions, or other voting practices or procedures that may enhance the opportunity for discrimination against the minority group;
4. if there is a candidate slating process, whether the members of the minority group have been denied access to that process;
5. the extent to which members of the minority group in the state or political subdivision bear the effects of discrimination in such areas as education, employment and health, which hinder their ability to participate effectively in the political process;
6. whether political campaigns have been characterized by overt or subtle racial appeals;
7. the extent to which members of the minority group have been elected to public office in the jurisdiction.
S.Rep. at 28-29, 1982 U.S.C.C.A.N. 206-07 (footnotes omitted), cited in Gingles, 478 U.S. at 36-37, 44-45, 106 S.Ct. at 2758-59, 2762-64.
As Gingles notes, the Senate Report lists two additional factors that may have probative value, 478 U.S. at 45, 106 S.Ct. at 2763-64:
• whether there is a significant lack of responsiveness on the part of elected officials to the particularized needs of the members of the minority group[;]
• whether the policy underlying the state or political subdivision’s use of [the challenged] voting qualification, prerequisite to voting, or standard, practice or procedure is tenuous.6
*1008S.Rep. at 29, 1982 U.S.C.C.A.N. at 207 (footnote omitted).
Gingles points out that the Senate Report reposes broad discretion in the courts. The list of factors “is neither comprehensive nor exclusive.” 478 U.S. at 45,106 S.Ct. at 2763. Rather, in deciding whether § 2 has been violated, courts are to engage in a “searching practical evaluation of the ‘past and present reality.’” Id. at 45, 106 S.Ct. at 2764. (quoting S.Rep. at 30, 1982 U.S.C.C.A.N. at 208).
The Supreme Court recently emphasized that the three Gingles factors are “generally necessary” to prove a § 2 violation, but not sufficient. Courts are required to look explicitly at the totality of the circumstances, because “ultimate conclusions about equality or inequality of opportunity were intended by Congress to be judgments resting on comprehensive, not limited, canvassing of relevant facts.” Johnson v. De Grandy, — U.S. -, -, 114 S.Ct. 2647, 2657, 129 L.Ed.2d 775 (1994).
Finally, Justice Brennan, writing for the Gingles majority,7 asserted that “[mjulti-member districts and at-large election schemes ... are not per se violative of minority voters’ rights.” Gingles, 478 U.S. at 48, 106 S.Ct. at 2765.8
Ill
The district court found no dispute regarding whether the plaintiffs had satisfied the first and second prongs of the Gingles test— namely, that the minority group “is sufficiently large and geographically compact to constitute a majority in a [hypothetical] single-member district,” and that the minority group is “politically cohesive.” Gingles, 478 U.S. at 50-51, 106 S.Ct. at 2766-67; see Decision and Order at 4, 39, 41. These findings remain uncontested on appeal. The district court found, however, that the plaintiffs failed to satisfy the third prong of Gingles— that is, they had not demonstrated that the white majority in Niagara Falls votes sufficiently as a bloc usually to defeat black voters’ candidates of choice. Gingles, 478 U.S. at 51, 106 S.Ct. at 2766-67; Decision and Order at 41. We first consider whether the district court erred in concluding that the plaintiffs failed to satisfy this third prong of the Gingles test.
The question of whether “legally significant white bloc voting” has occurred is a question of fact reviewed under the “clearly erroneous” standard. Rangel v. Morales, 8 F.3d 242, 245 (5th Cir.1993). However, the district court’s application of legal standards in reaching its findings is subject to de novo review. See id. The district court’s ultimate finding on the question of vote dilution is subject to the “clearly erroneous” standard of review, as set forth in Rule 52(a) of the Federal Rules of Civil Procedure.9 Gingles, 478 U.S. at 79, 106 S.Ct. at 2781.
The plaintiffs argue that the district court’s finding that they failed to satisfy the third prong of Gingles majority bloc voting— derived from two main errors of law. They maintain that the district court erred in giving little weight to elections that occurred before the challenged system was first operative, in 1987. They further contend that the court erred in affording what the plaintiffs call “equal weight” to elections involving only white candidates (so-called “white-white” elections, as opposed to “black-white” elections). The plaintiffs claim, and the district court found, that in black-white elections (several of which occurred before 1987) there was evidence of a “high degree of racial polarization.” Appellants’ Br. at 35; Decision and Order at 53. According to Gingles, racial bloc voting exists when there is a correlation between the “race of the voter and the way in which the voter votes,” or, *1009said another way, it occurs when “black voters and white voters vote differently.” 478 U.S. at 53 n. 21, 106 S.Ct. at 2768 n. 21 (internal quotation marks and citation omitted). The plaintiffs contend that in light of this evidence of racial bloc voting in black-white elections, the court erred in giving what they call “equal weight” to white-white elections in deciding whether the white majority votes sufficiently as a bloc usually to defeat the preferred candidates of minorities. Appellants’ Br. at 26. The plaintiffs state: “If black voter[s] cannot be expected to elect a black candidate of choice under the current at-large system — then having the opportunity to vote on occasions for winning white candidates doesn’t amount to equal opportunity.” Appellants’ Br. at 43. For the reasons described below, we conclude that the district court erred in finding that the plaintiffs failed to satisfy the third Gingles prong.
At the outset, we pause to clarify what the record actually reflects regarding the success of black candidates. When the plaintiffs suggest that black candidates in Niagara Falls are consistently defeated, they brush over the fact that in 1991, a black candidate, defendant Andrew Walker, won the City Council Democratic Party primary, with support from 38.6% of white voters and 79.4% of black voters, and that he was the leading vote-getter in the general election, with support from 42.9% of white voters and 84.8% of black voters.10 This record, then, is not one in which black voters have been unable to elect a black candidate. Cf. Citizens for a Better Gretna v. City of Gretna, 834 F.2d 496, 504 (5th Cir.1987) (in affirming § 2 violation, noting that no black had ever been elected to challenged municipal office), cert. denied, 492 U.S. 905, 109 S.Ct. 3213, 106 L.Ed.2d 564 (1989); Smith v. Clinton, 687 F.Supp. 1310, 1312 (E.D.Ark.1988) (three-judge panel) (no black ever elected state representative in challenged district). Nevertheless, the election of one black candidate does not preclude the plaintiffs from establishing vote dilution. See Gingles, 478 U.S. at 75, 106 S.Ct. at 2779. And until Walker’s election, no black candidate had ever been elected to the City Council of Niagara Falls. The district court nevertheless concluded that the plaintiffs failed to show that the white majority votes sufficiently as a bloc usually to defeat minorities’ preferred candidates. Decision and Order at 85. We disagree.
The plaintiffs presented evidence on nine black-white elections for the City Council, five of which occurred before the challenged electoral scheme was adopted and four of which occurred afterwards. As the district court found, these elections overall reflected “a high degree of racial polarization.” Decision and Order at 53.
a. Pre-Reform Elections11
1. 1969 City Council General Election
Willie Abrams, who is black, received support from 23.4% of black voters and 1.4% of white voters in a five-person contest for two seats. He lost.
2. 1971 City Council Democratic Party Primary
Arthur Ray, who is black, received support from 92.1% of black voters and 26.3% of *1010white voters in a four-person contest for two seats. He lost.
3. 1971 City Council General Election
Arthur Ray, who ran on a minor party line in a five-person contest for two seats, received support from 78.2% of black voters and 9.8% of white voters. He lost.
1. 1977 City Council Democratic Party Primary
Two black candidates were among six who ran for two seats. One black candidate, Joseph Profit, received support from 71.3% (72.7%) of black voters and 7.2% (8.3%) of white voters. The other black candidate, George Matthews, received support from 65.8% (68.3%) of black voters and 4.4% (5.2%) of white voters. Both lost the primary election. Of the six candidates, Profit and Matthews received the lowest percentage of white-voter support.
Black Support White Support
Martino (W) virtually none 44.0%
Matthews (B) 65.8% 4.4%
O’Dea (W) virtually none 32.3%
Profit (B) 71.3% 7.2%
Tangent (W) 13.1% 47.5%
Walsh ON) 12.0% 40.1%
5. 1979 City Council Democratic Party Primary
Matthews was one of six candidates — and the only black — running for two seats. He received support from 97.2% (88.0%) of black voters and 6.9% (7.3%) of white voters. The candidate who received the second-highest black-voter support, Guy Tom Sottile, who is white, won a seat. He received support from 11.2% of black voters. Of the six candidates, Matthews received the lowest percentage of white-voter support.
Black Support White Support
Martel (W) virtually none 41.0%
Martino ON) 3.8% 34.3%
Matthews (B) 97.2% 6.9%
O’Donnell ON) 7.8% 30.5%
Sottile (W) U.2% 35.1%
White (W) 7.8% 31.0%
b. Post-Reform Elections
1. 1987 City Council Democratic Party Primary
Renae Kimble, who is black and a plaintiff in this case, was the only black competitor among nine candidates for five seats. She received support from 97.1% (92.2%) of black voters and 27.5% (29%) of white voters. She placed sixth, losing by six votes out of the 45,505 votes cast. Four of the five successful candidates, who were white, received the highest percentage of black-voter support after Kimble, as follows: Agnello, 29.7%; Rendina, 28.7%; Aversa, 20.7%; and Henry Buchalski, 13.8%.12 Of the nine candidates, Kimble received the lowest percentage of white-voter support.
Black Support White Support
Agnello (W) 29.7% 41.4%
Aversa (W) 20.7% 41.8%
Buchalski (W) 13.8% 43.5%
Giove (W) 3.5% 44.0%
Kimble (B) 97.1% 27.5%
Pasquantino (W) 11.1% 33.0%
Rendina (W) 28.7% 49.3%
Sottile ON) 11.7% 52.7%
Zito (W) 2.1% 37.2%
2. 1987 City Council General Election
Kimble, who ran on the Right-to-Life and Liberal party lines, received support from 77.0% (57.0%) of black voters and 8.3% (7.9%) of white voters in an eleven-candidate contest for five seats; she was the only black candidate. She lost. Three of the five successful candidates, all of whom were white, received the third-, fourth-, and fifth-highest percentages of black-voter support, as follows: Sottile, 37.9%; Rendina, 23.5%; and Buchalski, 17.2%. Of the eleven candidates, Kimble received the lowest percentage of white-voter support.
Black Support White Support
Agnello (W) 38.3% 39.4%
Aversa ON) 13.7% 42.7%
Buchalski ON) 17.2% 45.4%
Caso (W) virtually none 44.3%
Darby ON) virtually none 18.7%
Gawel ON) virtually none 43.9%
Geracitano ON) virtually none 47.8%
Kimble (B) 77.0% 8.3%
Mayes (W) 1.7% 20.3%
Rendina (W) 23.5% 47.1%
Sottile (W) 37.9% 51.7%
*1011 3. 1991 City Council Democratic Party Primary
Andrew Walker, who is black, received 85.7% (79.4%) of black-voter support and 37.6% (38.6%) of white-voter support in a five-person contest for three seats in which he was the only black competitor. He won. Sottile, who is white, received the third-highest percentage of black-voter support. He won with 9.1% of black-voter support and 58.5% of white-voter support. Of the five candidates, Walker received the lowest percentage of white-voter support.
Black Support White Support
Aversa (W) 4.6% 43.6%
Morello (W) 0.6% 50.9%
Rendina (W) 14.9% 41.8%
Sottile (W) 9.1% 58.5%
Walker (B) 85.7% 37.6%
Jf. 1991 City Council General Election Walker received support from virtually all black voters (according to the defendants’ statistics) (84.8% according to the plaintiffs’) and 43.2% (42.9%) of white voters, in a nine-candidate race. Sottile, who received the second-highest percentage of black-voter support, placed second in the race for three seats, winning with support from 45.7% of black voters and 45.1% of white voters. Of the nine candidates, Walker received the third-highest percentage of support from white voters.
Black Support White Support
Aversa (W) virtually none 7.5%
Geracitano (W) virtually none 45.3%
Joseph (W) virtually none 32.6%
LaRocca (W) virtually none 37.4%
Morello (W) 35.4% 38.3%
Rendina (W) virtually none 4.9%
Scattering (W) virtually none virtually none
Sottile (W) 45.7% 45.1%
Walker (B) virtually unanimous 43.2%
Based on this statistical record, the district court concluded that “[i]in general, elections in which at least one African American candidate contested were marked by a high degree of racial polarization.” Decision and Order at 53. But the district court found that the elections preceding 1987 held “diminished” probative value, either because they occurred when the minority population was too small to satisfy the first prong of Gingles, or because they occurred before the challenged, reformed governing body had been put in place, or both. Decision and Order at 39-46. A close study of the district court’s approach suggests that, for purposes of the third Gingles prong inquiry, the court actually gave these elections almost no weight at all, but focused entirely on the post-1987 elections. Decision and Order at 41-48.
The district court found that the 1971 elections were not probative because they occurred at a time when blacks made up only 9.3% of the city’s population and 10.06% of the voting population, and that the minority population was not sufficiently large then to form a hypothetical single-member district— the first Gingles prong. This reasoning erroneously conflates the various functions of the first and third Gingles prongs.
The first Gingles prong — whether the minority population is “sufficiently large and geographically compact to constitute a majority in a single-member district,” Gingles, 478 U.S. at 50, 106 S.Ct. at 2766 — concerns the existence of legally cognizable injury. As Justice Brennan said, “[ujnless minority voters possess the potential to elect representatives in the absence of the challenged structure or practice, they cannot claim to have been injured by that structure or practice.” Id. at 50 n. 17, 106 S.Ct. at 2766 n. 17. No matter how severe the racial polarization, if black voters could not constitute a majority in a hypothetical single-member district, then they cannot claim to be worse off by virtue of the multimember scheme.
This issue, however, is distinct from whether the voting patterns indicate that the white majority votes sufficiently as a bloc to enable it usually to defeat minority preferences under a multimember scheme — the third Gingles prong. Id. at 51, 106 S.Ct. at 2766-67. While a substantial change in the character of the population arguably could affect white voting patterns and the ability of white voters to defeat the minority’s preferred candidates, this kind of change was not the basis for the district court’s decision to disregard these early elections. Rather, it said these elections were less significant because they occurred at a time when the first Gingles factor could not be satisfied. Decision and Order at 39-40, 45-46. That conclusion was error.
*1012The question of discounting elections on the grounds that they occurred under a now-discarded electoral scheme is more difficult. Starting with the 1987 elections, Niagara Falls changed the size of its City Council from four to seven members. This suit was filed in 1989. According to Justice Brennan, “[t]he essence of a § 2 claim is that a certain electoral law, practice, or structure interacts with social and historical conditions to cause 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 2765. Elections held under a significantly different electoral structure, then, may be less probative as to whether whites are voting as a bloc usually to defeat the minority’s preferred candidate for a seat on the challenged electoral body. Yet, Gin-gles instincts that “a pattern of racial bloc voting that extends over a period of time is more probative of a claim that a district experiences legally significant polarization than are the results of a single election.” Id. at 57, 106 S.Ct. at 2769.
In this case, the district court found, and we agree, that racial polarization in elections in which a black candidate ran has been less pronounced since the City Council was expanded. While black candidates received less than ten percent of white-voter support in most of the elections in the 1970s, for example, that percentage increased after the Council expanded. In the 1987 Democratic Party primary, Kimble received 27.5% of white-voter support. In 1991, Walker received 37.6% white-voter support in the primary, and 43.2% in the general election.
Nevertheless, this shift is not so pronounced as to justify what was, in essence, the district court’s wholesale disregard of the pre-reform elections for purposes of the third Gingles prong. The proper focus for the second and third Gingles factors is “the probative value of the proffered evidence as an indicator of the voting behavior of the relevant polity.” Westwego Citizens for Better Gov’t v. City of Westwego, 872 F.2d 1201, 1209 (5th Cir.1989). Here, the same polity was voting under both the four-member system and the seven-member system. Moreover, if the relevant evidence from City Council elections were limited to those elections occurring after 1985, the court would have been left with only two primary and three general elections. Where a district court is faced with so slim a record, it should be especially hesitant to disregard the results of earlier elections for the same office and by the same electorate. Cf. Citizens for a Better Gretna, 834 F.2d at 502-03 (holding that district court properly considered evidence of voting patterns in presidential and statewide elections in challenge to aldermanic body on grounds that Gingles “suggests flexibility in the face of sparse data”). In sum, the pre-1987 elections provide important evidence of racial voting patterns in City Council elections and should not have been effectively disregarded.
In so concluding, we note that differences between the four- and seven-member electoral systems are relevant to the strength of the plaintiffs’ vote dilution claim. The Supreme Court has recently made clear that the totality-of-the-circumstances inquiry is important precisely because “bloc voting [is] a matter of degree, with a variable legal significance depending on other facts.” De Grandy, — U.S. at -, 114 S.Ct. at 2657 (citing Gingles, 478 U.S. at 55-58, 106 S.Ct. at 2768-70). Thus, the expansion of the Council from four to seven members is a development that deserves full consideration in the totality-of-the-circumstances inquiry.
For purposes of the third Gingles prong, however, a review of the black-white City Council Democratic Party primary and general elections suggests that white bloc voting usually has resulted in the defeat of the candidates supported by minority voters. As the plaintiffs point out, in every City Council Democratic Party primary and general election since 1975 in which a black candidate has run, the black candidate has received the greatest support from black voters and the least support from white voters, with the exception of the 1991 general election in which Walker was victorious. Walker is the only minority-preferred black candidate who has ever been elected to the Niagara Falls City Council. City elections results for other *1013offices in which a black candidate has run generally follow this pattern.13
Taken alone, all of this evidence would suggest that the plaintiffs have satisfied the third Gingles prong. However, the district court also considered elections analyzed by the defendants in which no black candidate ran, so-called white-white elections. The results of these elections, as represented by the defendants’ statistics, were as follows:
1. 1975 City Council Democratic Party Primary
Seven candidates ran for two seats. Bat-taglia and Gallagher won.
Black Support White Support
Battaglia 16.1% 31.9%
Dillon 14.2% 13.4%
Gallagher 31.3% 42.7%
Gingino 19.4% 18.9%
Mains 21.3% 27.3%
Narkiewiecz virtually none 26.6%
Sinclair 33.4% 24.6%
2. 1975 City Council General Election
Seven candidates ran for two seats. Bat-taglia and Gallagher won.
Black Support White Support
Battaglia 52.5% 42.3%
Gallagher 58.7% 50.2%
Martino 8.1% 10.0%
Murty virtually none 3.8%
Narkiewiecz virtually none 12.2%
Szabo 4.1% 26.7%
Warner 12.1% 37.5%
S. 1977 City Council General Election
Five candidates ran for two seats. Smith and Tangent won.
Black Support White Support
Martino 79.5% 39.4%
Pitarresi virtually none 42.6%
Smith 9.4% 45.4%
Tangent 71.6% 42.3%
Walsh virtually none 10.7%
1. 1979 City Council General Election
Five candidates ran for two seats. Bazzo-ni and Martel won.
Black Support White Support
Bazzoni virtually none 57.9%
Martel 25.6%. 47.8%
Martino virtually none 6.0%
Pitarresi virtually none 36.1%
Sottile virtually unanimous 31.9%
5. 1981 City Council Democratic Party Primary
Eight candidates ran for two seats. O’Farrell and Rendina won.
Black Support White Support
Agnello 9.2% 12.9%
Fama 0.8% 20.4%
Garabedian 3.0% 12.1%
O’Farrell 30.6% 36.7%
Redding 33.4% 25.3%
Rendina 2.6% 39.0%
Tangent 33.3% 24.4%
Whit 11.9% 11.2%
6. 1981 City Council General Election
Seven candidates ran for two seats. Smith and Cook won.
Black Support White Support
Cook virtually none 45.9%
Fama virtually none 4.4%
O’Farrell 73.3% 35.6%
Redding 4.3% 2.0%
Rendina 37.3% 42.3%
Smith virtually none 51.3%
Tangent 4.1% 3.4%
*1014 7. 1983 City Council Democratic Party Primary
Three candidates ran for two seats. Red-ding and Martel won.
Black Support White Support
Martel 35.1% 60.3%
Redding 56.3% 61.1%
Rendina 27.6% 48.5%
8. 1983 City Council General Election
Four candidates ran for two seats. Red-ding and Martel won.
Black Support White Support
Bazzoni virtually none 44.8%
Martel 79.5% 42.1%
Redding 90.6% 52.3%
Geracitano virtually none 42.0%
9. 1985 City Council Democratic Party Primary
Six candidates ran for two seats. Soda and Quaranto won.
Black Support White Support
Cook 50.3% 14.6%
Giove 0.5% 26.7%
Kostoff 13.5% 6.7%
Quaranto 21.9% 51.2%
Simon 18.8% 42.2%
Soda 35.2% 44.4%
10. 1985 City Council General Election
Seven candidates ran for two seats. Soda and Quaranto won.
Black Support White Support
Badorian 8.9% 23.4%
Cook ' 3.9% 4.5%
Geracitano 8.8% 44.2%
Quaranto 50.5% 56.3%
Simon virtually none 2.7%
Smith virtually none 2.5%
Soda 46.3% 46.3%
11. 1989 City Council General Election14
Eleven candidates ran for four seats. Bu-chalski, Gawel, Palillo, and Quaranto won.
Black Support White Support
Albond 58.5% 26.2%
Aversa 40.1% 33.6%
Buchalski 40.8% 40.0%
Carl virtually none virtually none
Caso virtually none virtually none
Gawel virtually none 49.8%
LaRocca virtually none 42.8%
Martel virtually none 36.7%
Palillo virtually none 48.4%
Quaranto 47.4% 56.1%
Roffle 1.4% 2.3%
The defendants also analyzed seven mayoral elections in which no black candidate ran.15
The plaintiffs argue that, inasmuch as the black-white elections showed a strong pattern of white bloc voting resulting in the *1015defeat of black candidates, the district court’s finding that whites do not vote sufficiently as a bloc usually to defeat black voters’ preferred candidates is based — in error — on City Council elections involving only white candidates, as well as the electoral successes of white candidates supported by some black voters in elections involving a black candidate. White-white elections, they contend, are an improper evidentiary source for the third Gingles prong in light of the other evidence in this case of racially polarized voting patterns.16
Courts for long have grappled with the appropriate weight to afford white-white elections in § 2 cases. Indeed, the question is part of a broader debate about the extent to which plaintiffs must prove racial bias to prevail under § 2, or, similarly, whether they must disprove that it is partisan politics — not racial bias — that accounts for the alleged deprivation of minorities’ rights to enjoy equal opportunities in political participation. See League of United Latin American Citizens (LULAC) v. Clements, 999 F.2d 881, 861 (5th Cir.1993) (en banc) (where evidence shows that divergent voting patterns are best explained by “partisan affiliation,” vote dilution not established), cert. denied, — U.S. -, 114 S.Ct. 878, 127 L.Ed.2d 74 (1994); Baird v. Consolidated City of Indianapolis, 976 F.2d 357, 361 (7th Cir.1992) (“The Voting Rights Act does not guarantee that nominees of the Democratic Party will be elected, even if black voters are likely to favor that party’s candidates.”), cert. denied, — U.S. -, 113 S.Ct. 2334, 124 L.Ed.2d 246 (1993). By looking beyond the statistical voting patterns of whites and blacks to see who is actually running — whites or blacks — courts attempt to determine whether minorities are voting for certain candidates because they are “truly” minorities’ representatives of choice, and whether the majority is voting against candidates for reasons of race.
The Voting Rights Act does not explicitly address the question of how much weight, if any, courts should afford white-white elections. The Act states that the election of minority candidates (as opposed to the election of candidates supported by minorities) is one factor to be considered in a § 2 claim. 42 U.S.C. § 1973(b). Yet, the law is designed to ensure that members of a protected class have an equal opportunity “to elect representatives of their choice, ” id. (emphasis added), not necessarily members of their class. Often the representatives of choice of a protected class will be members of that class. See Gingles, 478 U.S. at 68, 106 S.Ct. at 2775 (Justice Brennan, for plurality, noting that “it will frequently be the case that a black candidate is the choice of blacks, while a white candidate is the choice of whites”). And the Act seeks to ensure that members of a class have the right to an equal opportunity to obtain representation by members of their class. Butts v. City of New York, 779 F.2d 141, 148 (2d Cir.1985), cert. denied, 478 U.S. 1021, 106 S.Ct. 3335, 92 L.Ed.2d 740 (1986). But the Act does not delegitimate elections in which a substantial portion of a protected class expresses a preference for a candidate who is not a member of that class. The Voting Rights Act does not provide that race alone, to the exclusion of all other factors political and social, is the Rosetta Stone of an election. The statute does not require this species of racialism, and basic American constitutional principles make us recoil from any such idea. Indeed, the Act explicitly provides: “[Njothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion in the population.” 42 U.S.C. § 1973(b).
The Supreme Court has not defined the term “representatives of ... choice” or the “minority’s preferred candidate.” Gingles, 478 U.S. at 51, 106 S.Ct. at 2766-67. In Gingles, Justice Brennan, writing for a plu*1016rality,17 stated that “the race of the candidate per se is irrelevant to racial bloc voting analysis.” Id. at 67, 106 S.Ct. at 2774-75 (emphasis in original). Rather, “it is the status of the candidate as the chosen representative of a particular racial group, not the race of the candidate, that is important.” Id. at 68, 106 S.Ct. at 2775 (emphasis in original). Justice Brennan maintained that the “central inquiry” of § 2 was whether voting patterns were “correlated with the race of the voter,” not whether the patterns were “caused by race.” Id. at 63, 106 S.Ct. at 2772-73 (emphasis in original). He contended that a requirement that minority voters prove racial hostility on the part of white voters would be at odds with the congressional intent behind the 1982 amendments to § 2. Gingles, Id. at 64-67, 106 S.Ct. at 2773-75; see also LU-LAC, 999 F.2d at 855-59 (describing Justice Brennan’s rationale for treating race of candidate as irrelevant).
Five other justices declined to join in this portion of the opinion. Writing only for himself, Justice White suggested that if blacks and whites were voting differently simply because they were voting along partisan lines, there would be no unlawful racial polarization, because “interest group politics,” not “racial discrimination,” would explain the outcomes. Gingles, 478 U.S. at 83, 106 S.Ct. at 2783 (White, J., concurring). Justice O’Connor, in whose opinion three other justices joined, wrote that “[t]he overall vote dilution inquiry neither requires nor permits an arbitrary rule against consideration of all evidence concerning voting preferences other than statistical evidence of racial voting patterns.” Id. at 101, 106 S.Ct. at 2792 (O’Con-nor, J., concurring in the judgment). Both Justices White and O’Connor said the Gin-gles plurality did not have to reach the question of the import of the candidate’s race versus the voters’ race, and Justice O’Connor stated that the idea that the candidate’s race was irrelevant conflicted with the Supreme Court case law on which the Senate Report was based. Gingles, 478 U.S. at 83, 101, 106 S.Ct. at 2783, 2792-93.
Most Courts of Appeals have rejected Justice Brennan’s position on the importance— or lack of importance — of a candidate’s race, a position that, as described above, did not enjoy the support of a majority of the Court. Several courts instead have held, in various formulations, that elections in which a black candidate runs are more probative on the question of white bloc voting than are white-white elections. See, e.g., Clarke v. City of Cincinnati, 40 F.3d 807, 812 (6th Cir.1994) (finding that “a candidate’s race can be relevant to a § 2 inquiry”), cert. denied, — U.S. -, 115 S.Ct. 1960, 131 L.Ed.2d 851 (1995); Jenkins v. Red Clay Consol. Sch. Dist. Bd. of Educ., 4 F.3d 1103, 1128 (3d Cir.1993) (“As a general matter ... elections involving white candidates only are much less probative of racially polarized voting than elections involving both black and white candidates.”), cert. denied, — U.S. -, 114 S.Ct. 2779, 129 L.Ed.2d 891 (1994); cf. Sanchez v. Bond, 875 F.2d 1488, 1495 (10th Cir.1989) (“[T]he district court has a right to consider [elections with only ‘Anglo’ candidates] and to give them such weight as the circumstances warrant.”), cert. denied, 498 U.S. 937, 111 S.Ct. 340, 112 L.Ed.2d 305 (1990).
We decline to adopt an approach precluding the possibility that a white candidate can be the actual and legitimate choice of minority voters. Such an approach would project a bleak, if not hopeless, view of our society — a view inconsistent with our people’s aspirations for a multiracial and integrated constitutional democracy. No legal rule should presuppose the inevitability of electoral apartheid — least of all a rule interpreting a statute designed to implement the Fourteenth and Fifteenth Amendments to the Constitution. Indeed, any such rule would be in tension with the Voting Rights Act’s directive that members of a protected class have no right to proportional representation. 42 U.S.C. § 1973(b).
The trial court thus properly attempted to analyze white-white elections, as well as to evaluate the success of white candidates in black-white elections, to determine whether *1017whites voted as a bloc to defeat blacks’ preferred candidates. We nevertheless conclude that the district court erred for two primary reasons in finding that legally significant bloc voting did not exist in City Council elections in Niagara Falls. First, the district court apparently gave equal weight to the results of general and primary elections, even when a candidate strongly preferred by minority voters lost the primary election. Second, in both primary and general elections, the district court treated successful candidates with some black support as preferred candidates, despite the fact that other unsuccessful candidates had received significantly more support.
The Voting Rights Act asks whether the political processes leading to nomination or election in the State or political subdivision are not equally open to participation by members of a class of [protected] citizens ... in that its members have 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) (emphasis supplied). As Chief Judge Myron Thompson of the Middle District of Alabama has observed, a focus on general elections, without regard to primary results, can fail to “present a full picture of the extent to which white bloc voting is preventing minority-preferred candidates from electoral success.” White v. Alabama, 867 F.Supp. 1619, 1654 (M.D.Ala.1994). While recognizing that the court was “treading on dangerous ground by attempting to ascertain who is really the black community’s candidate of choice,” Judge Thompson noted that potential black candidates often do not reach the general election because they lose in the primaries. Id.
Judge Thompson’s insight is evident in this case. For example, in examining the 1977 City Council general election, the trial court emphasized that blacks’ second-most preferred candidate was successful in a race for two seats, despite the fact that the candidate was not one of the two candidates most preferred by whites. This analysis ignores the fact that the candidate the court treats as blacks’ second-most preferred candidate re-eeived only 18.1% of black votes in the primary, while two losing candidates received 65.8% and 71.3% of black support in the primary.. Similarly, the court uses the 1985 general election as evidence of a lack of polarization, because the two candidates most preferred by blacks, also most preferred by whites, won. However, a candidate with greater black support lost the primary. In light of the language of the Voting Rights Act, we believe that general elections are less probative determinants of the success of the minority’s preferred candidate in these circumstances.
In analyzing which candidates were preferred by minority voters, the trial court also failed to account for great differentials in black voter support in a single election. In effect, the district court treated as “minority-preferred” successful candidates who had significantly less black support than their unsuccessful opponents. In its analysis of the 1987 City Council Democratic Party primary, for example, the court concluded that blacks “were able to elect four out of five of their most preferred candidates.” Decision and Order at 45. Given the large differential in support — virtually unanimous black support for the losing candidate and between 13.8% and 29.7% black support for the winning candidates — the court clearly erred in treating the four winning candidates as minority-preferred. Similarly, the court noted as evidence of a lack of polarization the fact that blacks were able to elect three of their five preferred candidates in the 1987 City Council general election. Again, the court ignored a large differential in black voter support of the unsuccessful and successful candidates; indeed, the court treated as minority-preferred a candidate who had 17.2% black voter support, despite the fact that a candidate with 77.0% black voter support lost.
In light of this evidence, we conclude that the trial court’s finding of no legally significant bloc voting was clearly erroneous. Black-white City Council elections, both post-reform and pre-reform, show a high degree of racial polarization. Of the white-white elections, the court treated the 1979 and 1981 City Council general elections as racially polarized. Against this evidence of polariza*1018tion, the trial court apparently balanced the remaining white-white elections, including some that it concluded were somewhat polarized. As noted above, success of a minority-preferred candidate in a general election is entitled to less weight when a candidate with far greater support was defeated in the primary; the 1977 and 1985 general elections therefore have less probative value than the court accorded them. Only the 1975 general election, the 1983 Democratic Party primary election, and the 1983 general election can properly be characterized as elections that were not polarized. In those elections, candidates with significant black voter support (and, incidentally, significant white voter support) were successful. And although the candidate with the greatest black support in the 1975 primary election lost, support for that candidate was not dramatically higher than support for one of the successful candidates; there is therefore no reason to discount the 1975 general election.
Given the stark evidence of racial polarization in black-white races in City Council elections in Niagara Falls, evidence of black voter support for candidates in white-white elections or for white candidates in black-white elections is insufficient to alter the conclusion that “the white majority votes sufficiently as a bloc to enable it ... usually to defeat the minority’s preferred candidate.”18
In concluding that the district court too broadly categorized candidates as “minority-preferred,” we recognize that the district court was forced to give meaning to the concept of a minority-preferred candidate without the benefit of our guidance. Faced with similar concerns, other circuits have attempted to provide guidance for determining whether a white candidate may properly be considered minority voters’ representative of choice. The Third Circuit has said that, to determine whether a white candidate is, “as a realistic matter, the minority voters’ representative of choice,” courts may look at whether “the minority community can be said to have sponsored the candidate.” Jenkins, 4 F.3d at 1129. This inquiry entails scrutinizing, among other things, the level of minority involvement in advancing the candidate and in conducting or financing the campaign, as well as asking whether the candidate “campaigned in predominately minority areas or addressed predominately minority crowds and interests.” The Fourth Circuit has said that, in elections in multimember systems, where successful candidates receive support from more than 50% of minority voters, but an unsuccessful candidate has received more minority votes, a court must “review[ ]” the situation “individually” to determine whether successful candidates “can be fairly considered as representatives of the minority community.” Collins v. City of Norfolk, 816 F.2d 932, 937 (4th Cir.1987).19
We sympathize with these Circuits in their efforts to grapple with the often-conflicting requirements of the Voting Rights Act, but we believe that evaluating whether a person is, “as a realistic matter,” minority-preferred—based on subjective indicators such as “anecdotal testimonial evidence”—is a dubious judicial task, and one that can degenerate into racial stereotyping of a high order. Questions such as whether a candidate, in a campaign, “addressed predominately minority crowds and interests” suggest the existence of a racial political orthodoxy that courts should not legitimate, much less profess or promote.
Rather than require district judges to assess candidates’ authenticity in matters racial—an unavoidably malleable, highly subjective inquiry—we instead adopt a bright-*1019line rule: For purposes of analyzing the third prong of Gingles, in § 2 cases in which the plaintiffs seek to replace an at-large, multimember electoral system with a series of single-member districts of which one or more would be a so-called majority-minority district, a candidate cannot be “minority-preferred” if that candidate receives support from fewer than 50% of minority voters. When a candidate receives support from 50% or more of minority voters in a general election, a court need not treat the candidate as minority-preferred when another candidate receiving greater support in the primary faded to reach the general election. Finally, even if a candidate receives 50% or more of the minority vote, a court need not treat the candidate as minority-preferred if another candidate receives significantly higher support. This approach arguably may create results over- or underinclusive at the margin. But we think that the ballot box provides the best and most objective proxy for determining who constitutes a representative of choice.20 This rule means that, in any given election in a multimember scheme — where more than one contestant is elected to office in one election — a court may treat more than one candidate as the “minority’s preferred candidate,” or none at all.
Such a rule brings into focus the concerns raised above. Its application to identify minority-preferred candidates would foreclose the use of tepid black voter support for a successful candidate in a primary or general election — or tepid black voter support for a successful candidate in a general election, when a strongly supported candidate has lost the primary — to counterbalance evidence of racially polarized voting. Taking as an example the eleven white-white general and primary City Council elections analyzed by the defendants in this case, only two would provide evidence of the success of minority-preferred candidates in support of the defendants’ position: the 1983 Democratic Party primary and general elections. In 1983, a minority-preferred candidate, Redding, was successful in both the primary and general elections. In all other elections, a minority-preferred candidate lost (1985 primary; 1979, 1981, and 1989 general elections); a minority-preferred candidate won a general election, but another candidate received more support in the primary (1975, 1977, and 1985 general elections); or no candidate could be classified as minority-preferred (1975 and 1981 primaries). Indeed, looking at both black-white and white-white elections as analyzed by the defendants, Redding and Walker were the only candidates who received more than 50% of the black vote in the primary and went on to win the general election with more than 50% of the black vote. The rule we adopt today, then, reinforces the conclusion that legally significant bloc voting occurred in City Council elections in Niagara Falls.
IV
After its effort to apply the third Gingles prong, the district court found that under the “totality of the circumstances” presented on this record, the plaintiffs faded to show that the challenged voting structure impairs the plaintiffs’ rights to enjoy an equal opportunity to participate in the political process and elect candidates of their choice. This inquiry was proper, as the Supreme Court has recently emphasized that while the three Gingles factors are important and necessary, they are not sufficient to prove a § 2 violation. De Grandy, — U.S. at -, 114 S.Ct. at 2657. The requirement of this broader inquiry follows from the Act itself, which states that a violation of the Act “is established if, based on the totality of circumstances, it is shown” that the political process is not equally open to participation.21 *102042 U.S.C. § 1973(b). Reviewing the district court’s ultimate conclusion under the clearly erroneous standard, we affirm the district court’s judgment.
The district court carefully considered each of the factors enumerated in the Senate Report and set forth in Gingles, 478 U.S. at 36-37, 44-45, 106 S.Ct. at 2758-59, 2762-64. It concluded that, while there was substantial evidence of racially polarized voting dating back to elections in 1969, many of the other Senate Report factors weighed against a finding that the at-large system “operates to minimize or cancel out [blacks’] ability to elect their preferred candidates.” Gingles, 478 U.S. at 48, 106 S.Ct. at 2765.
The district court found that the plaintiffs provided no significant evidence of a history of official discrimination that has “touched the right of the members of the minority group to register, to vote, or otherwise to participate in the democratic process.” S.Rep. at 29,1982 U.S.C.C.AN. at 206. The plaintiffs introduced the testimony of an expert historian, Professor Lillian Williams. The district court, however, gave little weight to Professor Williams’s testimony, on several grounds. First, her analysis of Niagara Falls stopped at approximately 1980. Second, Williams relied heavily on the testimony of one witness at a state commission hearing, and in cross-examination Williams admitted that she knew almost nothing about that witness. Decision and Order at 50. Third, many of the personal interviews on which Williams relied were interviews of the plaintiffs or Williams’s family and friends, and Williams did not keep notes of these interviews. Fourth, many of Williams’s sources did not concern Niagara Falls in particular. Finally, the court reviewed Williams’s testimony in light of her stated vehement opposition to at-large districts under any circumstances. Id. at 51. The district court was entitled to find unpersuasive Williams’s testimony for these reasons. The plaintiffs also proffered evidence of official discrimination in public housing in the 1950s and 1960s. The district court found that such discrimination did exist, but it also found that more recently the City has made “great strides in integrating those facilities.” Id. at 63-64.
In attacking the district court’s conclusion on this factor, the plaintiffs direct our attention to evidence that blacks are socioeconomically disadvantaged, but they point to no relevant evidence of a history of official discrimination that has touched on their rights to participate in the political process. Therefore, the district court’s findings on this factor are not clearly erroneous.
The district court next found that the City of Niagara Falls “does not employ [electoral] practices or procedures that would enhance the possibility of discrimination against African Americans.” Id. at 55. The plaintiffs testified that they have not faced obstacles in “voting, registering to vote, affiliating with a particular party, forwarding themselves as candidates, or collecting petitions to be included on the ballot.” Id. at 56. There is no majority vote requirement, candidates do not run for numbered posts or specific seats, and there is no full-slate rule requiring that voters cast all of their votes in a multi-candidate election. Id. at 55.
The district court did identify staggered elections as one possible exception to these findings. The plaintiffs offered testimony that, in general, staggered elections dilute minority voting strength by reducing the number of seats open at any given election. Under the current system, with staggered elections, all seven City Council seats are never open at once. The initial 1987 election involved five open seats; subsequent elections have involved three or four open seats. The district court noted that the plaintiffs did not demonstrate the effects of this principle in Niagara Falls elections. In any event, as Gingles noted, “there is no requirement that any particular number of factors be proved, or that a majority of them point one way or the other.” 478 U.S. at 45 (quoting S.Rep. at 29, 1982 U.S.C.C.A.N. at 207).
*1021The district court next found that there is no candidate slating process in Niagara Falls. The court noted further that there were no impediments to blacks’ becoming political candidates in Niagara Falls. Decision and Order at 56-58. The plaintiffs do not challenge this finding.
The district court then found that blacks in Niagara Falls are “significantly disadvantaged” socioeconomically compared to whites, id. at 58, but it found that this circumstance did not “hinder their ability to participate effectively in the political process,” see S.Rep. at 29, 1982 U.S.C.C.A.N. at 206. In so concluding, the court noted that the plaintiffs “introduced no evidence of significant racial differences in voter registration rates.” Decision and Order at 65. Moreover, while turnout of black voters was lower than that of whites in City Council general elections, there was no evidence of a difference in primary elections. Id.
Plaintiffs argue that the district court misapplied the Senate Report factors by requiring the plaintiffs to prove a causal nexus between blacks’ low socioeconomic status, which the court acknowledged, and blacks’ “depressed political participation in terms of significant[ly] lower rates of voter turnout in City Council general elections and political candidacies.” Appellants’ Br. at 43. According to the Senate Report, voting rights plaintiffs need not establish such a nexus where both disparate socioeconomic conditions and depressed political participation are shown to exist. S.Rep. at 29 & n. 114; 1982 U.S.C.C.A.N. at 206-07.
It is true that the district court noted that the plaintiffs did not explain the significance of the lower black voter turnout rates in City Council general elections. Decision and Order at 65. But other evidence indicated active political participation on the part of blacks in Niagara Falls. While there have certainly been many more white than black candidates for the City Council, there were several black candidates in the 1970s, and recent years have seen competitive campaigns by Kimble and Walker. Moreover, as the district court noted elsewhere in its opinion, blacks have held several elected positions on the Board of Education, they have been appointed to other local boards and commissions, and they have held positions on the Niagara Falls Democratic Committee. Id. at 57, 61, 66-67. Based on these findings, we cannot deem clearly erroneous the district court’s finding that socioeconomic disadvantages have not hindered blacks’ ability to participate in the political process.
The district court next found that there was “no credible evidence of either overt or subtle racial appeals in any political campaigns in the City of Niagara Falls.” Id. at 66. The plaintiffs do not challenge this finding.
Regarding the extent to which minorities have been elected to public office in Niagara Falls, the district court noted that Walker was elected to the City Council in 1991. The district court explicitly found that the plaintiffs had failed to show that the 1991 primary and general elections, which occurred after this lawsuit was filed, reflected an attempt by white voters to thwart the success of this litigation. Id. at 43, 46;22 cf. Gingles, 478 U.S. at 75-76, 106 S.Ct. at 2779-80 (holding that district court is permitted, on the basis of all relevant circumstances, to “view[ ] with some caution” black candidates’ success in election held after filing of lawsuit). The court also noted that several blacks have been elected to positions on the Board of Education. Decision and Order at 66.
The district court’s findings regarding this factor enjoy further support from some of the evidence regarding the increase in the size of the Council from four to seven members — evidence that the district court considered in its analysis regarding the third Gin-gles prong. As Justice Brennan wrote for the court in Gingles: “The essence of a § 2 claim is that a certain electoral law, practice, *1022or structure interacts with social and historical conditions to cause 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. Thus, we must examine the challenged structure — the seven-member City Council, which took effect with the 1987 elections. As the district court correctly recognized, minority voters have had some success under this new system. Decision and Order at 43.
As noted earlier, Walker was successful in the 1991 Democratic Party primary and in the general election. While Democratic candidates historically have enjoyed more success in Niagara Falls than Republicans, there is no suggestion in the record or in the submissions of counsel that successful Democratic Party primary contenders easily or automatically prevail in the general election. Testimony by the defendants’ expert indicated that from 1975 to 1991, sixteen Democrats and eight Republicans were elected to the City Council. When this case was tried before the district court, Walker had been elected by the members of the Council as its chairman.
The district court also noted that Kimble failed by only six votes to win one of the available nominations in the 1987 Democratic Party primary for the City Council. To be sure, Kimble only placed sixth in a contest of nine, and she received the least white-voter support of all the candidates. Nevertheless, she received more white-voter support than any other black candidate had before, and the crossover vote was nearly enough to secure a seat.
Had these elections occurred under the discarded, four-member system, we might have regarded them as aberrational. But both the plaintiffs’ and defendants’ experts testified that a larger number of seats can enhance the opportunities for minorities to elect representatives of their choice. The plaintiffs’ expert, Professor James W. Loewen (an expert in race relations and political sociology with a particular expertise in “issues associated with voting rights litigation”), testified that the more seats that are open, the better the opportunity for members of the black community to elect representatives of their choice. When more seats are open, more candidates enter the field. And single-shot voting — in which minority voters focus their support on one candidate — is more effective the more candidates there are in the field, because the majority vote may then splinter among the many candidates. Under the reformed system in Niagara Falls, candidates now vie for three or four seats, rather than two, and, in turn, more candidates have entered the fields.23
As Gingles emphasized, the “extent to which minority group members have been elected to public office in the jurisdiction” is one of the most important Senate Report factors, along with racial polarization. 478 U.S. at 48 n. 15, 106 S.Ct. at 2765-66 n. 15. Under the newly expanded at-large system, the minority’s preferred candidates have enjoyed a proven equal opportunity to achieve — and one candidate has in fact *1023achieved — electoral success. We recognize the admonition in Zimmer — one of the cases from which the Senate Report factors were derived, S.Rep. at 28 n. 113, 1982 U.S.C.C.A.N. at 206 n. 113 — that the success of a black candidate (or, more generally, a minority-preferred candidate) is not disposi-tive. But based on all of the evidence adduced above, particularly the election results under the newly formed Council, we cannot conclude that these results were achieved “despite the relative political backwardness of black residents in the electoral district.” Zimmer, 486 F.2d at 1307.
As for the remaining two factors — elected officials’ responsiveness to the particularized needs of the minority group and whether the contested practice or structure is tenuous— the district court’s findings support its ultimate conclusion. We cannot accept the plaintiffs’ argument that only unresponsiveness — and not responsiveness — is relevant to a § 2 inquiry. The district court cited numerous efforts made by the City to address the needs of members of the minority community. The Niagara Human Rights Commission (established in the 1960s) began an Affirmative Action Task Force to increase the number of minority residents who apply for and qualify for positions in the Niagara Falls police and fire departments. The City has sought grants for increased “community policing” in certain black neighborhoods. In 1970 the Niagara Falls Board of Education established a school integration program that has achieved substantial success. In 1978 the City Council adopted a Fair Housing Law. In 1992 the Council established a Minority Business Loan fund to provide loans to black business owners. Although these efforts have not all been wholly successful— the Human Rights Commission, for example, has been criticized as ineffective — they lend support to the district court’s conclusion that elected officials in Niagara Falls have been responsive to some of the minority community’s needs.24 Decision and Order at 67. Finally, the plaintiffs do not challenge the district court’s conclusion that the seven-member at-large system is not a “tenuous” policy. Decision and Order at 69. Rather, they only point out that, according to the Senate Report, they are not required to prove otherwise to show vote dilution. S.Rep. at 29 n. 117, 1982 U.S.C.C.A.N. at 207 n. 117.
The district court concluded, in essence, that this is not a record where “a histoiy of persistent discrimination reflected in the larger society and its bloc-voting behavior portend[s] any dilutive effect” from the seven-member at-large council. De Grandy, — U.S. -, 114 S.Ct. at 2668. Put another way, Niagara Falls, New York is not what the Senate Report termed “one of those communities in our Nation where racial politics do dominate the electoral process.” S.Rep. at 33, 1982 U.S.C.C.A.N. at 211. In reviewing the district court’s conclusion, we are aware that the multi-factor, totality-of-the-eircumstances inquiry has been criticized as failing to provide a “rule for deciding a vote dilution claim.” See Holder v. Hall, — U.S. -, -, 114 S.Ct. 2581, 2615 (1994) (Thomas, J., concurring in the judgment). But this is the test followed by the Supreme Court in Whitcomb v. Chavis, 403 U.S. 124, 91 S.Ct. 1858, 29 L.Ed.2d 363 (1971) and White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973), set forth in the Senate Report and in the Act itself, and reaffirmed recently by the Supreme Court in the De Grandy case. Unless and until we are directed otherwise, we must determine whether the district court’s ultimate conclusion on the question of vote dilution was *1024clearly erroneous. See Gingles, 478 U.S. at 79, 106 S.Ct. at 2781 (“the clearly erroneous test ... is the appropriate standard for appellate review of a finding of vote dilution”).
CONCLUSION
Keeping in mind that the district court enjoys “the benefit of ... [a] particular familiarity with the indigenous political reality,” Gingles, 478 U.S. at 79, 106 S.Ct. at 2781, we have reviewed the record upon which the district court entered its findings. We cannot find that the court’s conclusion that the plaintiffs failed to prove a § 2 violation was clearly erroneous. The judgment of the district court is affirmed.
. The district court called the "majority” population of Niagara Falls “white” and the minority population "black” or "African-American.” The court found that the number of non-African-Americans who were not Caucasian was statistically insignificant for purposes of the litigation. The parties do not challenge this approach, and we will adhere to it. For purposes of this opinion, then, the terms "majority” and "white” are synonymous, and the terms "minority,” "black,” and “African-American” are synonymous.
. The district court explained that simple regression analysis
measures a candidate’s share of the votes received in a particular election district as a percentage of the number of voters at the polls in that district. That percentage is correlated with the racial composition of the district, measured in terms of percentage of the voting age population in that district.... A "correlation coefficient” is generated, demonstrating how consistently voter support for a candidate or group of candidates varies with the racial composition of the election districts- Simple regression does not allow for the effects of racial differences in voter turnout; it assumes *1006that turnout rates between racial groups are the same.
Decision and Order at 13-14; see also Jenkins v. Red Clay Consol. Sch. Dist. Bd. of Educ., 4 F.3d 1103, 1119 n. 10 (3d Cir.1993), cert. denied, - U.S. -, 114 S.Ct. 2779, 129 L.Ed.2d 891 (1994).
.This statistical method, also called homogenous precinct analysis, examines the actual voting percentages received by candidates in the most heavily white and most heavily black districts. Jenkins, 4 F.3d at 1120 n. 12.
. Double regression analysis, as opposed to simple regression analysis, see supra note 2, takes into account racial differences in voter turnout rates. Decision and Order at 14.
. The parties only analyzed Democratic Party primaries, not Republican Party or any other party primaries, because between 90 to 97 percent of the registered black voters in the City of Niagara Falls are Democrats. Decision and Order at 12. Unless otherwise indicated, all subsequent references to primary elections are to Democratic Party primaries.
. Regarding the idea of a "tenuous” practice, the Senate Report states that a "marked[] depart[ure] from past practices or from practices elsewhere in the jurisdiction” would "bear[] on the fairness of [a practice's] impact.” S.Rep. at 29 n. 117, 1982 U.S.C.C.A.N. at 207 n. 117.
. Certain sections of Justice Brennan’s opinion in Gingles did not enjoy the support of a majority of the court. See infra note 17.
. Although Gingles concerned a § 2 challenge to multimember districts, § 2 challenges may be brought against an electoral system consisting of single-member districts as well. See De Grandy, - U.S. at -, 114 S.Ct. at 2658.
.Rule 52(a) provides in pertinent part: "Findings of fact, whether based on oral or documentary evidence, shall not be set aside unless clearly erroneous, and due regard shall be given to the opportunity of the trial court to judge of the credibility of the witnesses.”
. These statistics were provided by the plaintiffs. According to the defendants’ double regression analysis, Walker received support from 37.6% of white voters and 85.7% of black voters in the Democratic Party primary; he received support from 43.2% of white voters and virtually unanimous support from black voters in the general election.
. Only the plaintiffs analyzed the 1969 general election, the 1971 Democratic Party primary, and the 1971 general election. For all but these elections, the statistics in the text are drawn from the double regression analysis of the defendants’ experts; the plaintiffs' statistics, where available, are provided parenthetically. The defendants’ statistics indicate not only voter support for black candidates, but also voter support for white candidates.
All tables in this opinion indicating relative black and white voter support reflect the defendants’ double regression analysis. As the defendants' expert explained, the percentages generated by this ahalysis represent an estimate of the percentage of voters at the polls (as opposed to a percentage of the voters in the voting population generally) who1 cast one of their available votes for a particular candidate. In the 1977 Democratic Party primary, for example, the estimated percentage of black voters at the polls who cast one of their two votes for Joseph Profit was 71.3%.
. The record did not indicate the first names of several of the candidates.
. Results of these elections were as follows:
1. 1975 Mayoral Democratic Party Primary Profit, who is black, lost a three-candidate race against two white candidates. O'Laughlin won.
Black Support White Support
O’Laughlin (W) 22.9% 65.3%
Profit (B) 50.5% 1.7%
Sottile (W) 19.3% 28.1%
2. 1991 Mayoral Democratic Party Primary Kimble, who is black, lost a three-candidate race against two white candidates. Quaranto won.
Black Support White Support
Kimble (B) virtually unanimous 6.1%
Massaro (W) virtually none 31.6%
Quaranto (W) virtually none 61.0%
3. 1982 New York State Lieutenant Governor Democratic Party Primary
H. Carl McCall, who is black, received support from 96.7% of black voters, and 36.8% of white voters. The plaintiffs only analyzed the voter support for McCall, and the defendants did not analyze this election. McCall lost.
4. 1988 Presidential Democratic Party Primary Jesse Jackson, who is black, received support from virtually all black voters of Niagara Falls, and 10.9% of white voters. Jackson lost.
. In 1989 the City Council Democratic Party primary election was uncontested; four candidates ran for four available seats.
. The results of these elections were as follows:
1. 1975 Mayoral General Election
Three candidates ran. O’Laughlin won.
Black Support White Support
Chille 0.5% 37.5%
O’Laughlin 78.4% 56.5%
Sottile 5.4% 1.8%
2. 1979 Mayoral Democratic Party Primary
Two candidates ran. O'Laughlin won.
Black Support White Support
O’Laughlin 54.1% 56.2%
Sinclair 17.6% 41.8%
3.1979 Mayoral General Election
Two candidates ran. O'Laughlin won.
Black Support White Support
Cook virtually none 44.5%
O’Laughlin 95.2% 50.5%
4.1983 Mayoral General Election Two candidates ran. O'Laughlin won.
Black Support White Support
Cook virtually none 42.2%
O’Laughlin 89.7% 54.0%
5.1987 Mayoral Democratic Party Primary Four candidates ran. O'Laughlin won.
Black Support White Support
Massaro 33.1% 31.6%
Martel 23.1% 16.8%
O’Laughlin 28.6% 35.5%
Redding 10.6% 14.4%
6.1987 Mayoral General Election Four candidates ran. O'Laughlin won.
Black Support White Support
O’Laughlin 66.9% 50.3%
Palillo 5.8% 44.4%
Redding 0.7% 1.7%
Scattering virtually none 0.1%
*10157. 1991 Mayoral General Election Four candidates ran. Palillo won.
Black Support White Support
Massaro 0.6% 1.7%
Palillo virtually none 52.3%
Quaranto 89.9% 44.4%
Scattering virtually none virtually none
. The parties, as well as the district court, focus on the City Council elections in their analysis. This approach makes sense, since it is the City Council election scheme that is being challenged. See Carrollton Branch of NAACP v. Stallings, 829 F.2d 1547, 1558 (11th Cir.1987) (in Gingles inquiry, elections for seats on challenged governmental body more probative than those of other *1016elections in region), cert. denied, 485 U.S. 936, 108 S.Ct 1111, 99 L.Ed.2d 272 (1988). Moreover, our review of the other elections indicates that they do not cast doubt on conclusions yielded by the City Council elections.
. Only three Justices joined in this portion of Justice Brennan's opinion. Justice White, who *1017agreed with all other portions of Justice Brennan’s opinion, expressed his disagreement with this portion in a separate concurrence.
. There was no dispute that the black candidates in the elections analyzed in this case were black voters' candidates of choice. In every City Council election in which a black candidate ran, except in the 1969 general election, the black candidate received more than 50% of the minority vote—usually much more.
. In addition, Chief Judge Tjoflat of the Eleventh Circuit has suggested that a white candidate may be considered black voters' candidate of choice if the candidate is proven to be "strongly preferred" by black voters, or, put another way, if voters were "energized to support a particular white candidate.” Nipper v. Smith, 39 F.3d 1494, 1540 (11th Cir.1994) (Tjoflat, C.J., joined by Anderson, J., on rehearing en banc). Such support can be proved by "anecdotal testimonial evidence, polling data, a review of the appeals made during the campaign, and turnout information." Id.
. To paraphrase Winston Churchill’s oft-quoted comment on democracy, this rule may be the worst possible rule, except for all the others. See Winston Churchill, Remarks in the House of Commons (Nov. 11, 1947), in The Oxford Dictionary of Quotations 150 (3d ed. 1980) ("Many forms of government have been tried, and will be tried in this world of sin and woe. No one pretends that democracy is perfect or all-wise. Indeed, it has been said that democracy is the worst form of Government except all those other forms that have been tried from time to time.”).
. We agree, however, with the Third Circuit in noting that
it will be only the very unusual case in which the plaintiffs can establish the existence of the three Gingles factors but still have failed to *1020establish a violation of § 2 under the totality of circumstances. In such cases, the district court must explain with particularity why it has concluded, under the particular facts of that case, that an electoral system that routinely results in white voters voting as a bloc to defeat the candidate of choice of a politically cohesive minority group is not violative of § 2 of the Voting Rights Act.
Jenkins, 4 F.3d at 1135.
. The district court’s opinion states that the plaintiffs portrayed Walker's election as a " ‘special circumstance’ because of Walker’s outstanding credentials and popularity.” Decision and Order at 66. The plaintiffs do not press such arguments on appeal, nor do they draw our attention to any evidence in the record to indicate that the district court erred in rejecting their contentions. It borders on the absurd, of course, to suggest that an individual’s success in politics should be discounted as aberrational because the person is qualified and popular.
. Our focus on these aspects of the seven-member system is not at odds with the Supreme Court's recent plurality opinion in Holder v. Hall, — U.S. - 114 S.Ct. 2581, 129 L.Ed.2d 687 (1994). In Holder, a plurality of the court held that a vote dilution challenge could not be maintained under § 2 where the plaintiff was challenging the size of the governing body. The plurality reasoned that because the size of a governing body could vary greatly, the choice was “inherently standardless” and no “reasonable benchmark” was available by which to determine whether the challenged system was dilutive. Id. at -, 114 S.Ct. at 2586, 2588 (internal quotation marks and citations omitted).
Under this reasoning, the plaintiffs in this case likely could not sustain a § 2 challenge on the ground that a change from a four- to seven-member body or vice versa diluted their voting strength. But the size of the body is not being challenged here. Rather, we have to decide whether the seven-member, at-large system— compared to a system of single-member districts that the plaintiffs advocate — results in vote dilution. See Holder, - U.S. at -, 114 S.Ct. at 2589 (O'Connor, J., concurring in part and concurring in the judgment) (“In a challenge to a multimember at-large system ... a court may compare it to a system of multiple single-member districts.”). For the reasons described above, the elections conducted under the seven-member system properly enjoy added weight in the determination of whether that system dilutes minority voting strength. See Gingles, 478 U.S. at 56, 106 S.Ct. at 2769 (noting relevance of number of seats open and number of candidates in field).
. We pursue this inquiry with some reluctance, as it entails our deciphering what policy steps qualify as responses to the "needs of members of the minority community.” S.Rep. at 29, 1982 U.S.C.C.A.N. at 207. As the Senate Report itself recognizes, this factor is less “objective” than other factors that the Report sets forth. Id. at 29 n. 116, 207 n. 116. Nevertheless, we are somewhat more comfortable engaging in this inquiry, which the Report and Gingles direct us to undertake, than with the prospect of looking at anecdotal evidence to determine whether a particular candidate is "minority preferred.” See supra part III. The “responsiveness” inquiry here involves review of tangible efforts of elected officials and the impact of these efforts on particular members of the community. In contrast, a review of anecdotal evidence to determine who in a campaign is minority preferred would more likely involve the proffer of testimony by individuals in the community purportedly representing the minority's "voice” with regard to whom the minority community “really” prefers.