League of United Latin American Citizens, Council No. 4434 v. Clements

*734I.BACKGROUND..........................................................739

A. Texas’ Method of Electing District Court Judges.....................739

B. Procedural History.................................................. 740

II. THE ACCEPTED FRAMEWORK FOR ANALYZING SECTION 2 VOTE DILUTION CLAIMS ................................................... 741

A. The Threshold Inquiry: The Gingles Factors......................... 742

1. Size and Geographical Compactness of the Minority Group.........743

2. Political Cohesiveness of the Minority Group.......................743

3. Legally Significant White Bloc Voting............................. 744

B. The Broader Inquiry: The Totality of the Circumstances............747

1. The Senate Report Factors........................................747

a. History of discrimination touching the rights of minorities to participate in the political process......................747

b. Extent of racially polarized voting........................... 747

c. Use of voting practices that enhance the opportunity for discrimination____1........................................749

d. Minority access to the slating process...................'.....750

e. Lingering socioeconomic effects of discrimination............750

f. Use of racial appeals in campaigns..........................750

g. Extent to which minority candidates have been elected to public office ...............................................750

h. Responsiveness of elected officials to particular needs of the minority group............................................752

i. Tenuousness of the policy underlying the challenged practice ........................................................752

2. Other Relevant Factors, Including Racial Animus in the Electorate----753

C. The Ultimate Inquiry: Unequal Opportunity to Participate on Account of Race or Color ...........................................754

III. THE PROPOSED BALANCING FRAMEWORK FOR ANALYZING SECTION 2 VOTE DILUTION CLAIMS....................................755

A. The Accepted Role of State Interests in Section 2 Analysis..........756

B. The Proposed Role for State Interests in Section 2 Analysis.........756

C. Problems with the Proposed Balancing Framework..................757

1. The Legal Problem...............................................757

a. Congressional intent......................................... 757

b. Federalism principles........................................758

c. The: Supreme Court’s decision in Houston Lawyers’ Association........................................................760

2. The Practical Problem............................................ 763

3. Summation....................................................... 764

D. Applying the Proposed Balancing Framework in this Case: Evaluating Texas’ Asserted Interests......................................764

1. Identifying the Threatened State Interests.........................764

2. Scrutinizing the Threatened State Interests........................ 765

a. Texas’ interest in preserving the administrative advantages of the current at-large system..............................' 766

b. Texas’ interest in allowing judges to specialize...............766

c. Texas’ linkage interest.......................................767

d. Texas’ interest in preserving the function of district court judges as sole decision-makers.....'.........................769

3. Assigning a Weight to the Threatened State Interests.............772

IV. REVIEW OF THE DISTRICT COURT’S SECTION 2 LIABILITY FINDINGS................................... 772

A. Standard of Appellate Review....................................... 773

B. Review of the District Court’s Vote Dilution Findings Under the Accepted Section 2 Framework....................................774

*7351. Statistical Methodology....................................... 774

2. Review of District Court's Vote Dilution Findings ............. 776

a. Bexar County............................................ 777

(i) Gingles factors....................................... 777

(ii) Totality of circumstances factors..................... 778

(iii) Ultimate vote dilution finding....................... 780

b. Dallas County............................................ 780

(i) Gingles factors....................................... 780

(ii) Totality of circumstances factors..................... 781

(iii) Ultimate vote dilution finding ....................... 785

c. Ector County............................................. 785

(i) Gingles factors....................................... 786

(ii) Totality of circumstances factors..................... 786

(iii) Ultimate vote dilution finding....................... 788

d. Harris County............................................ 788

(i) Gingles factors....................................... 788

(ii) Totality of circumstances factors..................... 789

(iii) Ultimate vote dilution finding ....................... 791

e. Jefferson County......................................... 792

(i) Gingles factors....................................... 792

(ii) Totality of circumstances factors..................... 793

(iii) Ultimate vote dilution finding....................... 794

f. Lubbock County.......................................... 794

(i) Gingles factors....................................... 794

(ii) Totality of circumstances factors..................... 795

(iii) Ultimate vote dilution finding....................... 796

g. Midland County.......................................... 797

(i) Gingles factors....................................... 797

(ii) Totality of circumstances factors..................... 798

(iii) Ultimate vote dilution finding....................... 799

h. Tarrant County........................................... 799

(i) Gingles factors....................................... 799

(ii) Totality of circumstances factors..................... 800

(iii) Ultimate vote dilution finding....................... 801

i. Travis County............................................ 801

3. Effect of District Court’s Refusal to Consider Partisan Voting Evidence................................................... 803

a. The Partisanship Evidence................................ 803

(i) History of partisan politics in Texas................. 803

(ii) How partisan politics operate in Texas district court elections ............................................. 803

(iii) The limitations of the partisanship evidence.......... 805

(iv) Summation of partisanship evidence................. 805

b. The District Court’s Treatment of the Evidence............ 806

c. The District Court’s Error................................ 807

d. The Effect of the Error; Whitcomb Considered............ 807

4. Summation................................................... 813

C. Review of the District Court’s Vote Dilution Findings Under the Proposed Balancing Framework............................... 813

V. REMEDY............................................................ 814

VI. CONCLUSION....................................................... 815

DISSENT OF PATRICK E. HIGGINBOTHAM, Circuit Judge 819

APPENDIX A TO JUDGE HIGGINBOTHAM’S DISSENT ... 842

*736Before KING, JOHNSON and HIGGINBOTHAM, Circuit Judges. KING, Circuit Judge:

This case is before us on remand from the Supreme Court's decision in Houston Lawyers' Association v. Attorney General of Texas, — U.S. —, 111 S.Ct. 2376, 115 L.Ed.2d 379 (1991), in which the Court held that Section 2 of the Voting Rights Act, 42 U.S.C. § 1973, applies to all Texas judicial elections. We must now address a question that is undoubtedly easier to frame than to answer. In particular, we must decide whether the district court erred in concluding that the method by which Texas elects district court judges — as that method operates in nine counties — violates Section 2 of the Voting Rights Act. After a careful review of the record, we hold (1) that the district court correctly concluded, in eight of the counties at issue, that Texas' method of electing district court judges violates Section 2, but (2) that the district court erred in finding a Section 2 violation in Travis County. We therefore affirm the district court's decision in part, reverse the decision in part, and remand the case to the district court for consideration and imposition of an appropriate remedy.

In view of the length of this opinion,1 a summary of the pertinent facts and major legal conclusions may be helpful to the reader:

As with all cases under the Voting Rights Act, this one is driven by the facts. In this case, certain key facts are best summarized by the following table which sets forth, with respect to each of the nine Texas counties at issue, the population of the relevant minority group, the total number of district court judges elected in the county, the number of judges who are members of the relevant minority group and the percentage of the total number of district judges who are members of the relevant minority group.

[[Image here]]

*737The table portrays graphically what is inescapable from the record in this case— that in Texas district court elections, minority voters have less opportunity than white voters to participate in the political process and to elect representatives of their choice. The powerlessness of Texas minority voters in state district court elections stands in marked contrast to the increasing ability of those voters to participate in the political process and to elect representatives of their choice in the context of federal and state legislative elections and in local city council and school board elections. The strides that minority voters have made in the latter elections are due in no small part to the Voting Rights Act. In the years following the passage of the Act, minority plaintiffs throughout the state mounted successful challenges under Section 2 to at-large election schemes for the federal and state legislative branches and for local city councils and school boards. Thus, the face of this state’s legislative branch and of local government more and more reflects the face of this state’s people. By contrast, the face of the judicial branch in Texas continues to be—as it has always been—overwhelmingly white.

The state of Texas and the other defendants argue that the Voting Rights Act, so helpful to minorities in these other contexts, affords minorities no relief in the context of the election of state district judges. This is so, we are told, because the state has a compelling interest in the maintenance of the present electoral system— an interest which outweighs the interest of minorities in having an opportunity equal to that of the state’s white citizens to elect district judges of their choice. Specifically, we are told by the state that electing district judges from an area no smaller than a county is necessary to ensure that district judges are independent or accountable to all litigants equally and that no particular group will have undue influence over the decisions that must be made alone by the district judge. We are told further that the mechanism by which that interest is implemented is the state’s venue rules which operate to ensure the accountability of a state district judge to all the citizens of the county in which he is elected. Upon close inspection, we find, not surprisingly, that the venue rules in Texas, like the venue rules in the federal courts and in the courts of other states, are predicated on considerations of convenience for the litigants and the witnesses; in general, they afford a defendant some protection from being forced to defend an action in a district court remote from his residence in this vast state, or remote from the place where the events underlying the controversy occurred and the place where evidence is most likely available. In short, here as elsewhere, the venue rules were not designed to ensure judicial accountability.

What does ensure that state district judges are independent and accountable to all litigants equally and that no particular group will have undue influence over the decisions that must be made are—first, the Texas Code of Judicial Conduct and second, the integrity of the individual judges. The Texas Code of Judicial Conduct charges judges with applying the law and states that an honorable judiciary separated from the influence of others is “indispensable to justice in our society.” The Code also stresses that “[a] judge should be unswayed by partisan interest, public clamor, or fear of criticism.” There is no evidence in this record that judges elected with the support of minorities will be any less obedient to the commands of the Texas Code of Judicial Conduct than are state district judges who are currently elected with the support of many other interest groups, *738such as the personal injury bar and the defense bar. Nor is there any support in this record for the proposition that persons elected with the support of minorities would somehow be lacking in the same high level of integrity that has in the past characterized the Texas bench.

In summary, the compelling interest proffered by the state and the other defendants for the maintenance of the current system is, at best, little more than tenuous. Arrayed against it is the Texas Constitution, which was amended by the Texas legislature and by the voters of this state in 1985 to permit the election of state district judges from areas smaller than a county. The voters of this state have made provision for one of the remedies available to the district court in this case—namely, the remedy of subdistricting. Surely, the legitimate state interests in this case can permissibly be defined, in part, by this provision of its Constitution.

We turn finally to a brief summary of certain other important legal issues involved in the decision of this case. In particular, we have been asked by the state of Texas and several state district judges to decide what Congress meant in the Voting Rights Act when it prohibited voting practices that result in a denial or abridgement of the right to vote “on account of race or color.”

With respect to the significance of Congress' use of the phrase "on account of race" in Section 2, our holdings may be summarized as follows: First, we hold that, under the plain language of Section 2, minority plaintiffs must demonstrate that the challenged election practice, under the totality of the circumstances, results in the denial or abridgement of the right to vote "on account of," or based on, race or color. See Chisom v. Roemer, — U.S. —, —, 111 S.Ct. 2354, 2363, 115 L.Ed.2d 348 (1991). The phrase "on account of race or color" has been broadly defined by the Supreme Court and by Congress. Specifically, the totality of circumstances factors listed by Congress as relevant to a determination of Section 2 liability, as well as the threshold factors for proving vote dilution set forth in Thornburg v. Gingles, 478 U.S. 30, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986), point to minority voters' unequal opportunity to participate in the political process on the basis of race or color. Thus, minority plaintiffs seeking to satisfy the "on account of race or color" requirement — which is indisputably their burden — may present evidence of the existence of racially polarized voting, racial campaign appeals, a history of official discrimination, the lingering socioeconomic effects of discrimination, and other features of the current or past racial climate of the community in question. Proof of some or all of these factors raises an inference that racial discrimination is responsible for minority plaintiffs' unequal opportunity to participate in the political process.

Therefore, we reject the argument that “on account of race or color” is a narrowly-defined phrase. In particular, we reject the contention that Congress has defined the phrase “on account of race or color” to require minority plaintiffs to present proof that their unequal opportunity to participate in the electoral process and elect representatives of their choice is “caused by racial animus in the electorate.” We also reject the related argument that, to prove legally significant white bloc voting in the context of partisan elections, as well as racially polarized voting in such elections, minority plaintiffs must prove that such voting patterns are caused by racial animus in the electorate. While proof of the presence of racial animus in the electorate would be a significant factor indicating that race or color is in some way responsible for minority plaintiffs’ unequal opportunity to participate in the electoral process, the absence of proof of racial animus in the electorate simply will not, by itself, preclude a finding that minority plaintiffs have an unequal opportunity to participate in the electoral process on account of race or color.

We have also been asked to decide whether evidence of a strong statistical correlation between the electoral success and the party affiliation of a candidate can override or negate other factors which Congress has indicated point towards vote dilution. With respect to the evidence in sever*739al of the counties at issue showing a strong statistical correlation between a candidate’s electoral success and his or her party affiliation, we hold that the district court erred in refusing to consider the evidence. The evidence is unquestionably relevant as an important feature of the political landscape. And, to the extent the evidence purported to demonstrate the absence of current racial animus in the electorate, it was also relevant to the question of whether race or color is responsible for the minority plaintiffs’ unequal opportunity to participate in the political process.

We further hold, however, that the district court’s failure to consider the particular evidence adduced in this case amounts to harmless error. The so-called “partisanship evidence” in this case was offered to demonstrate only that voters in elections in large Texas counties, not knowing the race or names of district court candidates, vote without specific racial animus toward those candidates. The evidence did not purport to explain why voters voted the way they did, but simply purported to rule out specific racial animus towards candidates. Moreover, to the extent that minority candidates find it harder, because of their lack of financial resources, to mount a county-wide campaign on a scale necessary to gain name recognition, the partisanship evidence of straight-ticket voting reinforces rather than negates minorities’ unequal opportunity to participate in the political process. Accordingly, we hold that, although it should not have been excluded from consideration as legally irrelevant, the partisanship evidence adduced here does not undercut the district court’s ultimate conclusion that minorities have an unequal opportunity to participate in the political process, an unequal opportunity that is tied to race or color.

I. BACKGROUND

This lawsuit, which is before this panel for a second time, encompasses nine different voting rights cases. It concerns the method by which Texas elects its district court judges in Bexar, Dallas, Ector, Harris, Jefferson, Lubbock, Midland, Tarrant, and Travis counties (“target counties”). Before we recount the procedural history of this lawsuit, it will be helpful to outline Texas’ method of electing district court judges and to describe how this method operates in the larger counties of the state.

A. Texas’ Method of Electing District Court Judges

Texas elects its 386 state district judges in partisan elections, which are conducted at the same time and in a substantially similar fashion as other state partisan races. Political parties nominate judicial candidates in general primaries and runoffs, and a candidate must receive a majority of the vote to qualify as the party’s nominee. See Tex.Elec.Code Ann. § 172.-003 (Vernon 1986). At the general election, judicial candidates must run for a specifically numbered district court, and their party affiliation is indicated on the ballot. To win the general election, a judicial candidate needs only a plurality of votes. Id. § 2.001.

One feature of Texas’ method of electing state district judges is unique—the size of the various judicial districts from which such judges are elected. Since 1985, the Texas Constitution has provided that judicial districts may not be smaller than a county unless a majority of the voters of the county authorizes smaller districts, see Tex. Const. art. V, § 7a(i), and to date, no district smaller than one county has been authorized by voters from any Texas county.2 In addition, although the Texas Constitution permits the creation of more than one judgeship per judicial district, see Tex. Const. art. V, § 7, the legislature has seldom invoked this provision. Consequently, as a general rule only one district judge is elected per judicial district, and the judicial districts, although no smaller than a county, vary in size from one county to six counties and in population from approximately 13,000 to 2.5 million. See generally The American Bench 2138-54 (6th ed. 1991) (breaking down judicial districts in Texas according to the counties they cover).

*740Thus, Texas’ method of electing district court judges operates as an at-large system in the larger counties—including the nine target counties—but not in the smaller counties. For example, in Harris County, which according to the 1980 Census has a population of some 2,409,544, there are fifty-nine (59) overlapping, single-judge, county-wide judicial districts. The persons running for those 59 positions may be voted on by all 1,685,024 registered voters in Harris County. By contrast, in Milam County, which according to the 1980 Census has a population of 22,732, there is only one judicial district. The registered voters in that county cast their ballots for only one district judge. See id.

B. Procedural History

On July 11, 1988, ten individuals, along with local and state chapters of the League of United Latin American Citizens (collectively, “Plaintiffs”), filed suit in federal district court, seeking declaratory and injunctive relief. Plaintiffs asserted that Texas’ method of electing district court judges, as that method operates in larger counties, (1) dilutes minority voting strength in violation of Section 2 of the Voting Rights Act, and (2) violates the Fourteenth and Fifteenth Amendments of the United States Constitution. The named defendants in the original lawsuit were: William P. Clements, the Governor of Texas; Jim Mattox, the Attorney General of Texas, who is charged with enforcing the laws of the state; George Bayoud, the Secretary of State of Texas, who is charged with administering the elections laws of the state; and members of the Judicial Districts Board of Texas, which is charged with reapportioning the districts from which Texas district court judges are elected (collectively, “State Defendants”).

Several months later, in March 1989, the district court permitted several parties to intervene in the lawsuit. The Houston Lawyers Association and the Texas Black Legislative Caucus intervened on behalf of Plaintiffs, as did certain individuals residing in Dallas County. Two Texas district court judges—Sharolyn Wood, 127th District Court in Harris County, and Harold Entz, 194th District Court in Dallas County—intervened in their personal capacities on behalf of the State Defendants.

Plaintiffs originally challenged Texas’ election method in forty-four counties. By the time of trial, however, they had narrowed their challenged to the nine target counties. Plaintiffs proceeded on behalf of Black voters in Harris, Dallas, Tarrant, and Jefferson counties, on behalf of Hispanic voters in Bexar and Travis counties, and on behalf of Black and Hispanic voters combined in Lubbock, Ector, and Midland counties.

The lawsuit was tried to the district court the week of September 18, 1989. After considering all the evidence, including much expert testimony, the district court rendered its liability decision and made findings of fact and conclusions of law. In its 94-page memorandum opinion of November 8, 1989, the district court rejected the Plaintiffs’ constitutional claims but found their statutory claim meritorious. The district court concluded that the Plaintiffs, on behalf of specified minority voters in each of the nine target counties, demonstrated a violation of Section 2 of the Voting Rights Act. The district court based its conclusion on its finding that Texas’ current method of electing district court judges in the nine target counties “interacts with social and historical conditions” to cause the minority voters in each of the nine counties to have less opportunity than white voters “to elect their preferred candidates.”

On appeal, a divided panel of this court held that the district court erred in concluding that Texas' method of electing district court judges violates Section 2 of the Voting Rights Act in the nine target counties. See League of United Latin American *741Citizens v. Clements, 902 F.2d 293 (5th Cir.), opinion on rehearing en banc, 914 F.2d 620 (1990), rev'd and remanded, — U.S. —, 111 S.Ct. 2376, 115 L.Ed.2d 379 (1991). We reasoned that, although Section 2 applies to some state judicial elections, it does not apply to elections of state district judges. Id. at 308. In reaching this decision, we focused specifically on the legislative history of the Voting Rights Act, as amended in 1982, and the nature of Texas district court judgeships. We noted that, unlike a judge of a multi-member body, "the district judge in Texas does his judging alone." Id. Because we concluded that "there can be no share of such a single member office," we held that county-wide elections of district judges "[do] not violate the Voting Rights Act." Id. Accordingly, we reversed the judgment of the district court.

On its own initiative, the members of this court decided to rehear the case en banc. Over Judge Johnson's dissent, a majority of this court concluded that Section 2 of the Voting Rights Act does not apply to any judicial elections. See League of United Latin American Citizens, 914 F.2d 620, 631 (5th Cir.1990) (en banc), rev'd and remanded, — U.S. —, 111 S.Ct. 2376, 115 L.Ed.2d 379 (1991). In reaching this conclusion, the majority relied primarily on Congress' use of the word "representative" in Section 2. It reasoned that Congress did not intend the term "representative" to include state judges. The majority stated: "Should Congress seek to install [the results] test for judicial elections, it must say so plainly. Instead, it has thus far plainly said the contrary." Id.

The Supreme Court granted certiorari for the limited purpose of considering the scope of Section 2's coverage. See Houston Lawyers' Ass'n v. Attorney General of Texas, — U.S. —, —, 111 S.Ct. 2376, 2380, 115 L.Ed.2d 379 (1991). Relying on its decision in Chisom v. Roemer, — U.S. —, 111 S.Ct. 2354, 115 L.Ed.2d 348 (1991), which was issued the same day, the Court concluded that Section 2 of the Voting Rights Act applies to Texas' method of electing district court judges. Id. — U.S. at —, 111 S.Ct. at 2381. The Court specifically held that "[i]f a State decides to elect its trial judges, as Texas did in 1861, those elections must be conducted in compliance with the Voting Rights Act." Id. at —, 111 S.Ct. at 2380. Accordingly, the Court reversed this circuit's en banc decision and remanded the case for further proceedings.

II. THE ACCEPTED FRAMEWORK FOR ANALYZING SECTION 2 VOTE DILUTION CLAIMS

When Congress amended Section 2 of the Voting Rights Act in 1982, it sought “to clearly establish the standards ... for proving a violation of that section.” S.Rep. No. 417, 97th Cong., 2d Sess., at 2 (1982), reprinted in 1982 U.S.C.C.A.N. 177, 178 [hereinafter S.Rep.].3 Congress specifically intended to “make clear that proof of discriminatory intent is not required to establish a violation of Section 2.” S.Rep. at 2, 1982 U.S.C.C.A.N. at 178. Congress also intended to restore “the legal standards under the results test by codifying” the vote dilution framework embraced in White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973). Id. at 2, 1982 U.S.C.C.A.N. at 179.

As amended, Section 2 establishes the basic framework for analyzing vote dilution claims. It provides:

(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 *742the United States to vote on account of race or color, or in contravention of the guarantees set forth in section 1973b(f)(2) of this title, as provided in subsection (b) of this section.
(b) A violation of subsection (a) of this section is established if, based on the totality of the 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) of this section 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. Subsection (a) makes clear that Congress intended to codify a results test, and subsection (b) provides guidance about how the results test is to be applied. See Chisom, — U.S. at —, 111 S.Ct. at 2364. Under this framework, then, a plaintiff can prevail “by showing that a challenged election law or procedure, in the context of the total circumstances of the local electoral process, ha[s] the result of denying a racial or language minority an equal chance to participate in the electoral process.” S.Rep. at 16, 1982 U.S.C.C.A.N. at 193.

The basic framework set forth in the language of Section 2, although it appears straightforward, has become increasingly complex. Both the Supreme Court and Congress have elaborated on this basic structure, and the product of their elaborations is a two-part framework for analyzing vote dilution claims brought by minority voters. Under the first part of this accepted Section 2 framework, a court must determine whether the minority voters have satisfied a threshold test enunciated by the Supreme Court. If minority voters satisfy the threshold test and demonstrate that their ability to elect representatives of their choice is being hindered by the challenged practice, a court must then ask whether the minority voters have demonstrated certain factors that Congress deemed relevant to the Section 2 inquiry—factors that raise an inference of vote dilution. Ultimately, a court must inquire whether minority voters have demonstrated an unequal opportunity to participate in the political process on account of race or color.

A. The Threshold Inquiry: The Gingles Factors

The Supreme Court first elaborated on the basic analytical framework established by Section 2 in Thornburg v. Gingles, 478 U.S. 30, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986), a case in which Black voters successfully challenged North Carolina's multimember legislative redistricting plan. Although the case produced a "complicated web of dissents, concurrences, and plurality opinions,"4 a majority of the Court agreed that at-large election procedures do not automatically violate Section 2. Id. at 46, 106 S.Ct. at 2764. A majority of the Court also agreed with the formulation of a three-part, threshold test for analyzing claims that an at-large election scheme dilutes minority voting strength. Under this threshold test, the minority group challenging an at-large election scheme must demonstrate that (1) it is sufficiently large and geographically compact to constitute a majority in a single member district, (2) it is politically cohesive, and (3) the white major*743ity votes sufficiently as a bloc to enable it — in the absence of special circumstances — usually to defeat the minority's preferred candidate. Id. at 48-51, 106 S.Ct. at 2765-2767. Failure to establish any one of the Gingles factors precludes a finding of vote dilution, because "[t]hese circumstances are necessary preconditions for multimember districts to operate to impair minority voters' ability to elect representatives of their choice...." Id. at 50, 106 S.Ct. at 2766; see also Overton v. City of Austin, 871 F.2d 529, 538 (5th Cir.1989) (failure to establish any one of the Gingles factors is fatal to plaintiffs' vote dilution claim).

1. Size and Geographical Compactness of the Minority Group

As stated above, to pass the Gingles threshold inquiry, the minority group challenging an at-large election scheme must first demonstrate that it is sufficiently large and geographically compact to constitute a majority in a single-member district. To satisfy this first Gingles factor, the minority group must ordinarily be able to draw a single member district in which a majority of the voting age population is minority. See Westwego Citizens for Better Government v. City of Westwego, 946 F.2d 1109, 1117 n. 7 (5th Cir.1991) (Westwego III); Overton, 871 F.2d at 536; see also Westwego Citizens for Better Gov't v. Westwego, 872 F.2d 1201, 1205 n. 4 (5th Cir.1989) (Westwego I) (noting that evidence of size of "voting age" population is critical to a vote dilution claim); Romero v. City of Pomona, 883 F.2d 1418, 1426 (9th Cir.1989) ("eligible minority voter population, rather than total minority population, is the appropriate measure of geographical compactness"). This requirement, according to the majority in Gingles, ensures that, in the absence of the multimember district and at-large voting scheme, "minority voters possess the potential to elect representatives...." Gingles, 478 U.S. at 50 n. 17, 106 S.Ct. at 2767 n. 17 (emphasis in original). In other words, if the minority group is not sufficiently large and geographically compact, "the multimember form of the district cannot be responsible for [minority] voters' inability to elect [their] candidates." Id. at 50, 106 S.Ct. at 2766 (emphasis in original) (footnote omitted).

2. Political Cohesiveness of the Minority Group

The minority group must also demonstrate that it is "politically cohesive" to pass the Gingles threshold inquiry. Id. at 51, 106 S.Ct. at 2766. One way of proving political cohesiveness, according to the Gingles majority, is to show "that a significant number of minority group members usually vote for the same candidates...." Id. at 56, 106 S.Ct. at 2769; see also Citizens for a Better Gretna v. City of Gretna, 834 F.2d 496, 501-02 (5th Cir.1987), cert. denied, 492 U.S. 905, 109 S.Ct. 3213, 106 L.Ed.2d 564 (1989). If the minority group is not politically cohesive, or does not engage in significant bloc voting, "it cannot be said that the selection of a multimember structure thwarts distinctive minority group interests." Gingles, 478 U.S. at 51, 106 S.Ct. at 2766 (citation omitted).

Although statistical evidence of racially polarized voting, see infra Part II.B.l.b., is frequently employed to demonstrate a minority group’s political cohesiveness, other evidence may also establish this second Gingles factor. We have on several occasions indicated that Gingles allows minority voters to prove their political cohesiveness even in the absence of statistical evidence of racial polarization. See Westwego III, 946 F.2d at 1118 n. 12; Brewer v. Ham, 876 F.2d 448, 453 (5th Cir.1989). In particular, political cohesiveness also may be demonstrated by testimony from persons familiar with the community in question, provided that such testimony is reinforced with other evidence or is not otherwise *744rebutted. See Brewer, 876 F.2d at 453-54; see also Overton, 871 F.2d at 536.

Political cohesiveness is related to, but distinct from, the concept of racially polarized voting. The notion of political cohesiveness contemplates that a specified group of voters shares common beliefs, ideals, principles, agendas, concerns, and the like such that they generally unite behind or coalesce around particular candidates and issues. See Monroe v. City of Woodville, 881 F.2d 1327, 1331 (5th Cir. 1989), modified, 897 F.2d 763, cert. denied, 498 U.S. 822, 111 S.Ct. 71, 112 L.Ed.2d 45 (1990). The term racially polarized voting, on the other hand, describes an electorate in which white voters favor and vote for certain candidates or propositions, and minority voters vote for other candidates or propositions. See Gingles, 478 U.S. at 53 n. 21, 106 S.Ct. at 2768 n. 21. (racial polarization exists where there is a consistent relationship between the race of the voter and the way in which the voter votes or where minority voters and white voters vote differently). Thus, while a showing of racially polarized voting will frequently demonstrate that minority voters are politically cohesive,5 a showing that minority voters are politically cohesive will not, by itself, establish racially polarized voting.

Finally, according to a majority of the Justices in Gingles, to satisfy the second threshold factor minority voters need not prove that the reason they vote for the same candidates is linked to race or color. Justice Brennan, writing for three other Justices, would have held that “the reasons [minority] and white voters vote differently have no relevance to the central inquiry of § 2.” Gingles, 478 U.S. at 63, 106 S.Ct. at 2773 (Brennan, J., joined by Marshall, Blackmun, and Stevens, JJ.). Although not entirely agreeing with Justice Brennan, Justice O’Connor, writing on behalf of three other Justices, agreed that defendants cannot rebut statistical evidence of a minority group’s political cohesiveness by “offering evidence that the divergent racial voting patterns may be explained in part by causes other than race.” Gingles, 478 U.S. at 100, 106 S.Ct. at 2792 (O’Connor, J., joined by Burger, C.J., Powell and Rehnquist, JJ., concurring in the judgment). Justice O’Connor specifically stated that statistical evidence of divergent racial voting offered solely to establish the minority group’s political cohesiveness may not be rebutted by evidence indicating that there is “an underlying divergence in the interests of minority and white voters.” Id.

3. Legally Significant White Bloc Voting

Under the third Gingles factor, the minority group “must be able to demonstrate that the white majority votes sufficiently as a bloc to enable it—in the absence of special circumstances, such as the minority candidate running unopposed—usually to defeat the minority’s preferred candidate.” Gingles, 478 U.S. at 51, 106 S.Ct. at 2766-2767 (internal references omitted). “In establishing this last circumstance, the minority group demonstrates that submergence in a white multimember district impedes its ability to elect its chosen representatives.” Id. As with demonstrating political cohesiveness, a minority group can demonstrate white bloc voting by introducing statistical evidence of racially polarized voting. See Westwego III, 946 F.2d at 1118. Moreover, where statistical evidence of racially polarized voting is used to establish white bloc voting, the elections that will usually be most probative are those in which the minority-preferred candidate is a member of the minority group. See Westwego I, 872 F.2d at 1208 n. 7.

The determinative question for this third Gingles factor “is not whether whites *745generally vote as a bloc, but rather, whether such bloc voting is legally significant.” Monroe, 881 F.2d at 1332. Legally significant white bloc voting means white bloc voting that usually defeats the minority group’s preferred candidate. Id. at 1333. As the Court recognized in Gingles, “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.” Gingles, 478 U.S. at 56, 106 S.Ct. at 2769. It is “the usual predictability of the majority’s success,” the Court continued, that “distinguishes structural dilution from the mere loss of an occasional election.” Id. at 51, 106 S.Ct. at 2767.

The amount of white bloc voting that rises to the level of legal significance will vary from location to location. That is, in determining whether white voting strength can generally “minimize or cancel” minority voters’ ability to elect representatives of their choice, courts must look at a number of factors. Id. at 56,106 S.Ct. at 2769 (quoting S.Rep. at 28). Among the factors affecting the level of legally significant white bloc voting are: (1) the nature of the allegedly dilutive electoral mechanism; (2) the presence or absence of other potentially dilutive electoral devices, such as majority vote requirements, designated posts, and prohibitions against bullet voting; (3) the percentage of registered voters in the district who are members of the minority group; (4) the size of the district; and (5) in multimember districts, the number of seats open and the number of candidates in the field. Gingles, 478 U.S. at 56, 106 S.Ct. at 2769.

Contrary to the arguments advanced by the State Defendants, Judge Wood, and Judge Entz, a minority group is not required to demonstrate that racial animus is responsible for a white bloc voting pattern. A requirement of demonstrating racial animus would, in the words of the Senate Report, be “unnecessarily divisive because it [would] involve charges of racism on the part of ... entire communities.” S.Rep. at 36, 1982 U.S.C.C.A.N. at 214. And, one of the reasons Congress decided to amend Section 2 was to obviate the need for such charges of racism. Id.6

In sum, a minority group can satisfy the third Gingles factor by showing that a substantial majority of white voters consistently vote against the candidates preferred by minority voters, such that the *746candidates preferred by the minority group usually lose. It has never been the case, however, that for white bloc voting to be legally significant, it must be shown to be motivated by racial animus. See e.g., Westwego III, 946 F.2d at 1118-20 (concluding, without inquiring as to the cause of voting patterns, that the record “unmistakably” demonstrated legally significant white bloc voting). We decline to hold so now.

*747B. The Broader Inquiry: The Totality of the Circumstances

Satisfaction of the Gingles factors, at least in this circuit, does not by itself establish a violation of Section 2. See Westwego III, 946 F.2d at 1116. The minority group must further demonstrate that, under the totality of the circumstances, “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). Unlike the Gingles threshold inquiry, the totality of circumstances inquiry is broad. It requires courts to make a “searching practical evaluation of the ‘past and present reality’ ” of the community and to take a “ ‘functional’ view of the political process.” Gingles, 478 U.S. at 45, 106 S.Ct. at 2764. Although courts typically should be guided by the factors mentioned in the Senate Report accompanying the 1982 amendments to Section 2, other factors may also be relevant. See Westwego III, 946 F.2d at 1120.

1. The Senate Report Factors

The Senate Report accompanying the 1982 amendments to Section 2 identifies seven "typical" factors and two "additional" factors that may be relevant to an analysis of the totality of the circumstances. These factors, which were derived from the analytical framework used by the Supreme Court in White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973), as articulated by this court in Zimmer v. McKeithen, 485 F.2d 1297, 1305 (5th Cir.1973), aff'd sub nom. East Carroll Parish School Bd. v. Marshall, 424 U.S. 636, 96 S.Ct. 1083, 47 L.Ed.2d 296 (1976), are meant to aid a court in assessing whether "the challenged system or practice, in the context of all the circumstances in the jurisdiction in question, results in minorities being denied equal access to the political process." S.REP. at 27, 1982 U.S.C.C.A.N. at 205. According to the Senate Report, a court must "assess the impact of the challenged structure or practice on the basis of objective factors, rather than making a determination about the motivations which lay behind its adoption or maintenance." Id. at 27, 1982 U.S.C.C.A.N. at 205 (emphasis added).

a. History of discrimination touching the rights of minorities to participate in the political process. The first factor mentioned in the Senate Report is “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.” S.Rep. at 28, 1982 U.S.C.C.A.N. at 206. In listing this factor, Congress made clear that it “was concerned not only with present discrimination, but with the vestiges of discrimination which may interact with present political structures to perpetuate a historical lack of access to the political system.” Westwego I, 872 F.2d at 1211-12. Evidence that is relevant under this factor includes the use of poll taxes and literacy tests, as well as the existence of racially segregated schools and public facilities. See White, 412 U.S. at 768, 93 S.Ct. at 2340; Zimmer, 485 F.2d at 1306; see also Andrew P. Miller and Mark A. Packman, Amended Section 2 of the Voting Rights Act: What is the Intent of the Results Test?, 36 Emory L.J. 1, 28 (1987).

b. Extent of racially polarized voting. The second factor that Congress deemed relevant to an analysis of the totality of the circumstances is “the extent to which voting in the elections of the state or political subdivision is racially polarized.” S.Rep. at 29, 1982 U.S.C.C.A.N. at 206. This factor is one of the two most important in the inquiry into the past and present reality of the challenged electoral structure. Westwego III, 946 F.2d at 1120. It is, we have stated, “the linchpin -of a § 2 vote dilution claim____” Gretna, 834 F.2d at 499; see also Samuel Issacharoff, Polarized Voting and the Political Process: The Transformation of Voting Rights Jurisprudence, 90 Mich.L.Rev. 1833, 1848-49 (1990) (opin*748ing that the inclusion of the racially polarized voting inquiry in the Senate Report marked “a major shift in voting rights law” and pointed to an “emerging doctrinal coherence for claims of minority vote dilution”).

As previously discussed, see supra Part II.A.2, racially polarized voting “exists where there is a ‘consistent relationship between [the] race of the voter and the way in which the voter votes,’ ... or to put it differently, where ‘[minority] voters and white voters vote differently.’ ” Gingles, 478 U.S. at 53 n. 21, 106 S.Ct. at 2768 n. 21. Racially polarized voting is not, as Judge Wood, Judge Entz, and the State Defendants suggest, racially motivated voting, or voting caused by racial animus.7 See Collins v. City of Norfolk, 816 F.2d 932, 935 (4th Cir.1987) (“The legal standard for the existence of racially polarized voting looks only to the difference between how majority votes and minority votes were cast; it does not ask why those votes were cast the way they were____”). Nor is racially polarized voting simply “the tendency of citizens to vote for candidates of their own race.” Miller & Packman, supra, at 16. Rather, racially polarized voting is an objective factor, which is established by evidence demonstrating that “voters vote along racial lines.” Westwego III, 946 F.2d at 1122.

To say that racial polarization does not simply measure the tendency of citizens to vote for candidates of their own race, however, does not mean that evidence of such tendencies are unimportant to the racial polarization inquiry. We recognized in Gretna that, in determining the existence of racial bloc voting, “the race of the candidate is in general of less significance than the race of the voter—but only within the context of an election that offers voters the choice of supporting a viable minority candidate.” 834 F.2d at 503. Furthermore, we have noted that “the evidence most probative of racially polarized voting must be drawn from elections including both [minority] and white candidates.” Westwego I, 872 F.2d at 1208 n. 7.8 Accordingly, in ascertaining whether a community’s elections are characterized by racially polarized voting, a court may properly give more weight to elections in which the minority-preferred candidate is a member of the minority group.

The Senate Report recognizes that racially polarized voting is a relative concept by instructing courts to assess not *749only the existence of racially polarized voting, but the extent of such voting. Indeed, the existence of racially polarized voting will frequently, if not always, be established by satisfying of the second and third Gingles factors. See supra Parts H.A.2., II.A.3; see also Romero v. City of Pomona, 883 F.2d 1418, 1423 (9th Cir.1989) (suggesting that the second and third Gingles factors, political cohesiveness and white bloc voting, are the “component parts” of racially polarized voting). Under the totality of circumstances inquiry, the focus is on the degree of racially polarized voting. When the pattern of racially polarized voting is severe—that is, when minority and white voters rarely engage in cross-over voting—an at-large election scheme is more likely to dilute minority voting strength in violation of Section 2. Conversely, when the pattern of racially polarized voting is slight, and there is substantial cross-over voting between minority and with voters, an at-large election scheme is less likely to dilute minority voting strength in violation of Section 2.

c. Use of voting practices that enhance the opportunity for discrimination. The Senate Report also directs courts to consider other voting practices that may interact with the challenged election scheme to dilute minority voting strength. In particular, courts are to consider “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.” S.Rep. at 29, 1982 U.S.C.C.A.N. at 206. Courts may also consider, at least in the context of at-large elections, whether the state or political subdivision employs what amounts to a numbered-post system and whether the system lacks a district residency requirement. See White v. Regester, 412 U.S. 755, 766, 93 S.Ct. 2332, 2339, 37 L.Ed.2d 314 (1973) (discussing Texas’ “place” rule). The potential of such features to dilute minority voting strength is well-documented. See generally Chandler Davidson, Overview, in Minority Vote Dilution 5-8 (Chandler Davidson ed. 1984).

Unusually large election districts may enhance vote dilution in a least two ways: First, such districts may make it more difficult for minorities to successfully campaign for office. See Rogers v. Lodge, 458 U.S. 613, 627, 102 S.Ct. 3272, 3280, 73 L.Ed.2d 1012 (1982); see also Whitcomb v. Chavis, 403 U.S. 124, 143, 91 S.Ct. 1858, 1869, 29 L.Ed.2d 363 (1971) (recognizing that, when large, multimember districts have an enhanced tendency to minimize or cancel out minority voting strength). Such districts may also create the problem of overly long ballots, making “an intelligent choice among candidates ... quite difficult.” Lucas v. Colorado Gen. Assembly, 377 U.S. 713, 731, 84 S.Ct. 1459, 1471, 12 L.Ed.2d 632 (1964).

The other features discussed above similarly enhance the possibility of vote dilution. Majority vote requirements can obstruct the election of minority candidates by giving white voting majorities a "second shot" at minority candidates who have only mustered a plurality of the votes in the first election. See Hearings Before the Subcomm. on Constitutional Rights of the Senate Comm. on the Judiciary, 94th Cong., 1st Sess. 491 (1975) (testimony of Dr. Charles Cotrell); see also Jones v. City of Lubbock, 727 F.2d 364, 383 (5th Cir. 1984); Major v. Treen, 574 F.Supp. 325, 351 n. 32 (E.D.La.1983) (three-judge court). Anti-single shot provisions prohibit voters from casting fewer than all of their votes and effectively force minority voters to cast votes for white candidates who may not be favored by minority voters. Westwego III, 946 F.2d at 1113 n. 3. Under a numbered post system, candidates are required to run for a designated seat. See White, 412 U.S. at 766, 93 S.Ct. at 2339. This requirement, the Supreme Court has recognized, "enhances [the minority group's] lack of access because it prevents a cohesive political group from concentrating on a single candidate." Rogers, 458 U.S. at 627, 102 S.Ct. at 3280. Finally, where there is no requirement that candidates reside in subdistricts of a multimember district, "all candidates may be selected from outside [a minority group's] residential *750area." White, 412 U.S. at 766 n. 10, 93 S.Ct. at 2340 n. 10.

d. Minority access to the slating process. The fourth totality factor listed in the Senate Report is “whether members of the minority group have been denied- access” to any candidate slating process. S.Rep. at 29, 1982 U.S.C.C.A.N. at 206. Slating has been defined as “the creation of a package or slate of candidates, before filing for office, by an organization with sufficient strength to make the election merely a stamp of approval of the preordained candidate group.” Overton, 871 F.2d at 534. The core inquiry as to slating, we have stated, “is the ability of [minorities] to get on the ballot.” Hendrix v. Joseph, 559 F.2d 1265, 1268 (5th Cir.1977).

Where a slating organization exists, "the ability of minorities to participate in that slating organization and to receive its endorsement may be of paramount importance." United States v. Marengo County Comm'n, 731 F.2d 1546, 1569 (11th Cir.) (Wisdom, J., sitting by designation), cert. denied, 469 U.S. 976, 105 S.Ct. 375, 83 L.Ed.2d 311 (1984). We have indicated, however, that a slating organization's endorsement of a few minority candidates will not preclude a finding that the minority group lacks access to the slating process, at least where those candidates are not preferred by the minority group. See Velasquez v. City of Abilene, 725 F.2d 1017, 1022 & n. 1 (5th Cir.1984) (referring to the Senate Report). Moreover, we have recognized that the absence of a slating organization will not mitigate evidence of an unequal opportunity to participate in the political process. See McMillan v. Escambia County, 638 F.2d 1239, 1245 (5th Cir.), cert. dismissed, 453 U.S. 946, 102 S.Ct. 17, 69 L.Ed.2d 1033 (1981).

e. Lingering socioeconomic effects of discrimination. Congress also has directed courts to consider “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____” S.Rep. at 29, 1982 U.S.C.C.A.N. at 206. In doing so, Congress recognized that the lingering socioeconomic effects of discrimination may hinder a minority group’s “ability to participate effectively in the political process.” Id. Where the minority group presents evidence that its members are socioeconomically disadvantaged and that their level of participation in politics is depressed, the group “need not prove any further causal nexus between [its members’] disparate socioeconomic status and the depressed level of political participation.” Id. at 29 n. 114, 1982 U.S.C.C.A.N. at 207 n. 114.

f.Use of racial appeals in campaigns. The Senate Report further directs courts, when looking at the present reality of the community’s political process, to consider “whether political campaigns have been characterized by overt or subtle racial appeals.” S.Rep. at 29, 1982 U.S.C.C.A.N. at 206. The presence of this factor, Congress undoubtedly recognized, could provide evidence that racial politics dominate the political process. See id. at 33, 1982 U.S.C.C.A.N. at 210-211. The absence of racial appeals, however, is not fatal to a minority group’s vote dilution claim under Section 2. As Judge Wisdom recognized in Marengo County, overt political racism has decreased over time. 731 F.2d at 1571. Therefore, while “[e]vidence of racism can be very significant if it is present[,]____its absence should not weigh heavily against a plaintiff proceeding under the results test of [S]ection 2.” Id.

g. Extent to which minority candidates have been elected to public office. The “extent to which members of the minority group have been elected to public office in the jurisdiction,” S.Rep. at 29, 1982 U.S.C.C.A.N. at 207, along with the extent of racially polarized voting, are the two most important factors that the district court must consider its analysis of the totality of the circumstances. See Gingles, 478 U.S. at 48 n. 15, 106 S.Ct. at 2767 n. 15; Westwego III, 946 F.2d at 1120. Indeed, minority candidate success is the only factor that is mentioned in the language of Section 2 itself.9 Subsection (b), *751which sets forth the totality of the circumstances standard, expressly provides: “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(b).

In the Senate Report, Congress indicated that the analysis of minority candidate success should be a cautious, relative inquiry. On the one hand, Congress recognized that, where “no members of a minority group have been elected to office over an extended period of time,” the lack of success “is probative.” S.Rep. at 29 n. 115, 1982 U.S.C.C.A.N. at 207 n. 115. Congress cautioned, however, against giving too much weight to the isolated instances of minority candidate success. Citing our decision in Zimmer, Congress noted that “the election of a few minority candidates does not ‘necessarily foreclose the possibility of dilution of the [minority] vote’ in violation of this section.” S.Rep. at 29 n. 115, 1982 U.S.C.C.A.N. at 207 n. 115. If such isolated successes were allowed to foreclose a Section 2 claim, Congress continued, “the possibility exists that the majority citizens might evade the section[, for example,] by manipulating the election of a ‘safe’ minority candidate.” Id.10

*752Congress did not, however, clearly establish a reference point for measuring the success of minority candidates. One possibility is to measure the success of minority candidates by reference to the proportion of minorities in the local population. Another way in which to measure the success of minority candidates is to focus on their success rate by comparing the number of minority candidates who have won with the number of minority candidates who have lost. And finally, as a third way of gauging minority success, a court could compare the number of minorities elected with the “qualified applicant pool,” as is currently done with employment discrimination claims brought under Title VII of the Civil Rights Act of 1964.

In our view, both of the first two methods discussed above are relevant to the inquiry of minority candidate success. Measuring the success of minority candidates by referring to the proportion of minorities in the local population is supported by the language of Section 2. See 42 U.S.C. § 1973(b). The method was also embraced by a majority of the Justices in Gingles. See Gingles, 478 U.S. at 77, 106 S.Ct. at 2780 (Brennan J., joined by White, J.) (finding that “persistent proportional representation” of minority voters in a multimember district precluded their vote dilution claim); id. at 104-05, 106 S.Ct. at 2794 (O’Connor, J., joined by Burger, C.J., Powell and Rehnquist, JJ., concurring in the judgment) (stating that “consistent and virtually proportional minority electoral success” is entitled to great weight in totality of circumstances analysis). The Senate Report provides support for relying on the success rate of minority candidates in determining the extent to which minorities have been elected. It specifically notes the probative value of “[t]he fact that no members of a minority group have been elected to office over an extended period of time____” S.Rep. at 29 n. 115, 1982 U.S.C.C.A.N. at 207 n. 115 (emphasis added).

We have serious reservations about the relevance of the third, “qualified applicant pool” approach. Such an approach seems inconsistent with our “refusal to preclude vote dilution claims where few or no [minority] candidates have sought offices in the challenged electoral system.” Westwego I, 872 F.2d at 1209 n. 9. It would, in our view, “allow voting rights cases to be defeated at the outset by the very barriers to political participation that Congress has sought to remove.” Id. Indeed, it seems perverse to measure the success of minority candidates by referring to the pool of qualified minority candidates, when the paucity of qualified minority candidates “is a likely result of a racially discriminatory system.” McMillan v. Escambia County, 748 F.2d 1037, 1045 (11th Cir.1984). But see Southern Christian Leadership Conference of Alabama v. Evans, 785 F.Supp. 1469, 1476-77 (M.D.Ala.1992) (measuring success of minority candidates under “qualified applicant pool” standard).

h. Responsiveness of elected officials to particular needs of the minority group. An “additional factor” that in some cases has probative value in the totality of circumstances inquiry is “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.” S.Rep. at 29, 1982 U.S.C.C.A.N. at 207. The Senate Report makes clear, however, that “[u]nresponsiveness is not an essential part” of a vote dilution claim. Id. at 29 n. 116, 1982 U.S.C.C.A.N. at 207 n. 116. “Therefore, defendants’ proof of some responsiveness [does] not negate [a minority group’s] showing by other, more objective factors ... that minority voters nevertheless [are] shut out of equal access to the political process.” Id.; see Campos v. City of Bay-town, 840 F.2d 1240, 1250 (5th Cir.1988).

i. Tenuousness of the policy underlying the challenged practice. A second “additional factor” that is sometimes relevant to totality of circumstances analysis is “whether the policy underlying the state or political subdivision’s use of [the challenged] voting qualification, prerequisite to voting, or standard, practice or *753procedure is tenuous. S.Rep. at 29, 1982 U.S.C.C.A.N. at 207. The tenuousness of the policy underlying the challenged practice, although less important under the “results test” embodied in Section 2, is nonetheless relevant. See Marengo County, 731 F.2d at 1571. Evidence of a tenuous underlying state policy, which is circumstantial evidence that the practice or procedure was motivated by an intent to discriminate, may also indicate that the practice or procedure produces a discriminatory result. See id. Moreover, “the tenuousness of the justification for a state policy may indicate that the policy is unfair.” Id. (citation omitted); see also S.Rep. at 29 n. 117, 1982 U.S.C.C.A.N. at 207 n. 117 (“If the [challenged] procedure markedly departs from past practice or from practice elsewhere in the jurisdiction, that bears on the fairness of its impact.”). The existence of a legitimate state policy underlying the challenged practice or procedure does not, however, preclude a finding of vote dilution. As noted in the Senate Report: “[E]ven a consistently applied practice premised on a racially neutral policy [does] not negate a plaintiff’s showing through other factors that the challenged practice denies minorities fair access to the process.” S.Rep. at 29 n. 117, 1982 U.S.C.C.A.N. at 207 n. 117.

2. Other Relevant Factors, Including Racial Animus in the Electorate

The list of relevant factors in the Senate Report is not exclusive. The Senate Report specifically states that, while the nine “enumerated factors will often be the most relevant ones, in some cases other factors will be indicative of the alleged dilution.” S.Rep. at 29, 1982 U.S.C.C.A.N. at 207. The report also instructs courts to consider “the challenged election law or procedure, in the context of the total circumstances of the local electoral process.” See id. at 16, 1982 U.S.C.C.A.N. at 193 (emphasis added). As we stated in Westwego III, “other factors may be relevant” under the totality of circumstances inquiry. 946 F.2d at 1120.

In determining whether a factor not expressly mentioned in the Senate Report is relevant for purposes of the totality of circumstances inquiry, courts should be mindful of Justice O’Connor’s admonition in Gingles to avoid arbitrary rules that prevent “consideration of all evidence concerning voting preferences other than statistical evidence of racial voting patterns.” 478 U.S. at 101, 106 S.Ct. at 2792 (O’Connor, J., concurring). Courts should also remember that the totality of circumstances inquiry is expansive and requires a “searching practical evaluation of the ‘past and present reality.’ ” S.Rep. at 30, 1982 U.S.C.C.A.N. at 208 (quoting White v. Reges ter ) (emphasis added). Finally, courts should determine whether the particular factor being advanced as relevant under the totality of circumstances inquiry is probative of the ultimate vote dilution inquiry. See infra Part II.C.

The State Defendants, Judge Wood, and Judge Entz argue that, under these principles, the presence or absence of racial animus in the electorate must be relevant to the totality of circumstances inquiry. We agree and accordingly hold that racial animus in the electorate is an “other factor” that is relevant to the totality of circumstances inquiry. After all, as Justice O’Connor recognized in her Gingles concurrence, the Senate Report accompanying the amendment to Section 2 repeatedly emphasizes the- relevance of “racial politics” and “racial political considerations” to the vote dilution inquiry. See Gingles, 478 U.S. at 101, 106 S.Ct. at 2792 (O’Connor, J., concurring) (citing S.Rep. at 33-34, 1982 U.S.C.C.A.N. at 210-212). Moreover, by listing the presence of racial campaign appeals as a relevant factor in the totality of circumstances inquiry, Congress has already instructed courts to consider, at least in a limited fashion, whether racial politics are a part of the local political landscape.

The presence of racial animus in the electorate may be a signal of vote dilution. Indeed, like the presence of racial campaign appeals, racial animus in the electorate “can be very significant if it is present.” Marengo, 731 F.2d at 1571. Where white voters are motivated by racial animus, minority voters are not only likely to be prevented from electing representatives *754of their choice, but they are also likely to be unable to influence the representatives who are elected. See Gingles, 478 U.S. at 100-01, 106 S.Ct. at 2792 (O’Connor, J., concurring). In addition, the presence of racial animus in the electorate makes it substantially more likely that an election practice with the potential to dilute minority voting strength is in fact operating to dilute minority voting strength.

The absence of racial animus in the electorate, however, like the absence of racial campaign appeals, “should not weigh heavily against a plaintiff proceeding under the results test of [S]ection 2.” Marengo, 731 F.2d at 1571. The limited relevance of the absence of racial animus is mandated by two considerations. The first is the nature of racial animus as a subjective rather than an objective factor pointing toward vote dilution. In the Senate Report accompanying the amended Section 2, Congress expressed a preference for objective rather than subjective indicators of vote dilution. See, e.g., S.Rep. at 27, 1982 U.S.C.C.A.N. at 204. Placing great weight on the absence of racial animus in the electorate would, in our view, be contrary to congressional intent. The second consideration mandating that little weight be accorded to the absence of this factor is the problem of proving the existence of racial animus in the electorate. As Judge Wisdom has recognized, because overt political racism has decreased over time, racial animus in the electorate may be difficult, if not impossible, to detect. See Marengo, 731 F.2d at 1571. Anecdotal testimony to the effect that voters are pure of heart, by its very nature, will not bear much weight. Moreover, as the experts in this case admitted, see infra Part IV.B.3.a.(iii), there are inordinate difficulties in using statistical analyses to explain why voters vote for or against certain candidates. One expert, on whom the State Defendants placed heavy reliance, testified that it would be impossible to use a multivariate analysis to account for the various factors that might influence voters’ decisions. See id.

Thus, racial animus in the electorate, like other factors in the totality of circumstances inquiry, is neither necessary nor dispositive. We reiterate that a plaintiff proceeding under Section 2 need not establish racial animus in the electorate to demonstrate minority political cohesiveness, legally significant white bloc voting, or racially polarized voting. Moreover, because racial animus in the electorate is not an essential element of a vote dilution claim, evidence suggesting the absence of racial animus cannot, by itself, defeat a vote dilution claim.11

C. The Ultimate Inquiry: Unequal Opportunity to Participate on Account of Race or Color

In the end, the Gingles threshold inquiry and the expansive, totality of circumstances inquiry provide only guidance for answering the ultimate Section 2 inquiry: whether it has been 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 [racial or language minority] 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). As the Senate Report accompanying the amended Section 2 cautions, courts should not become bogged down in “mechanical ‘point counting.’ ” S.Rep. at 29 n. 118, 1982 U.S.C.C.A.N. at 207 n. 118. No one of the factors is dispositive, and the minority group need not prove a majority of the factors mentioned in the Senate Report. Westwego III, 946 F.2d at 1120. Rather, based on “the totality of circumstances and guided by those relevant factors in the particular case,” a court is required to make an overall judgment of “whether the voting strength of minority voters is ... ‘minimized or can*755celed out.’ ” S.Rep. at 29 n. 118, 1982 U.S.C.C.A.N. at 207 n. 118; see also Gingles, 478 U.S. at 47, 106 S.Ct. at 2764 (“The 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 [minority] and white voters to elect their preferred representatives.”).

The focus of the ultimate Section 2 inquiry is on minority voters’ unequal opportunity to participate in the electoral process on account of race or color. The factors that must be considered in both the Gingles and the totality of circumstances inquiries point to minority voters’ unequal opportunity to participate in the political process. Those factors also point to an unequal opportunity to participate on the basis of race or color. That is, to prevail under Section 2, a minority group has the burden to demonstrate (1) that, because of the challenged scheme or practice, its members are unable to fully participate in the political process and (2) that their inability to participate is tied to race or color. The framework set forth by the Supreme Court in Gingles and by Congress in the Senate Report is the only means through which a minority group meets this burden.

By framing the ultimate vote dilution inquiry in terms of whether the minority group has demonstrated an unequal opportunity to participate in the political process that is tied to race or color, we give effect to Congress' intent in amending Section 2. The language of Section 2, as well as the Senate Report accompanying the amendment of that section, indicate that Congress was concerned with minority groups having an unequal opportunity to participate in the political process. See 42 U.S.C. § 1973(b) (violation established if minority group has "less opportunity" to participate in the political process under the totality of circumstances); S.REP. at 27, 1982 U.S.C.C.A.N. at 205 (violation established if minority group demonstrates under totality of circumstances that, because of the challenged practice, its members are being "denied equal access to the political process"); S.REP. at 30, 1982 U.S.C.C.A.N. at 207 (ultimate question is whether the political processes are "equally open" to minority candidates). The language of Section 2 also indicates that Congress was concerned with voting practices resulting in an unequal opportunity to participate that is linked to race or color. By prohibiting only election procedures that result "in a denial or abridgment of the right ... to vote on account of race or color...." 42 U.S.C. § 1973(a) (emphasis added), Congress expressed its intent that the minority group's unequal opportunity be tied to race or color. See also Chisom v. Roemer, — U.S. —, —, 111 S.Ct. 2354, 2363, 115 L.Ed.2d 348 (1991) ("Now plaintiffs can prevail under § 2 by demonstrating that a challenged election practice has resulted in the denial or abridgement of the right to vote based on color or race.") (emphasis added).

To show that an at-large election scheme results in an unequal opportunity to participate in the political process on account of race or color, however, a minority group need not demonstrate current racial animus in the electorate. See S.Rep. at 28 n. 109, 1982 U.S.C.C.A.N. at 205 n. 109 (“[I]t is patently clear that Congress has used the words ‘on account of race or color’ in the Act to mean ‘with respect to’ race or color, and not to connote any required purpose of racial discrimination.”). Rather, the minority group may, by demonstrating some mix of factors under the totality of the circumstances—such as the existence of racially polarized voting, racial campaign appeals, a history of official discrimination, the lingering socioeconomic effects of discrimination, and other features of the current or past racial climate—raise an inference that their unequal opportunity to participate is linked to race or color. In short, members of a minority group seeking relief may satisfy the ultimate vote dilution inquiry by offering circumstantial evidence that racial discrimination is responsible for their unequal opportunity to participate in the political process.

III. THE PROPOSED BALANCING FRAMEWORK FOR ANALYZING SECTION 2 VOTE DILUTION CLAIMS

The State Defendants, as well as Judge Wood and Judge Entz, argue that the ana*756lytical framework set forth in Part II of this opinion is incomplete. Specifically, they argue that the Gingles inquiry, the totality of circumstances inquiry, and the ultimate vote dilution inquiry fail to adequately take into account Texas’ asserted interests in maintaining its current method of electing district court judges. Thus, they propose a balancing framework for analyzing Section 2 vote dilution claims—a framework in which a court would be required to weigh the extent of proven vote dilution, as established by the factors set forth in Part II of this opinion, against the magnitude of the state’s asserted interests in the election scheme or practice.

A. The Accepted Role of State Interests in Section 2 Analysis

All parties concede that the state interests underlying an allegedly dilutive election practice are relevant to Section 2 analysis in two ways. First, as discussed above, see supra Part II.B.1.i., a state's interests in the challenged election practice may be relevant in the totality of circumstances inquiry. The district court may consider as an "additional factor" whether the state policies underlying the challenged practice are tenuous. See S.REP. at 29, 1982 U.S.C.C.A.N. at 206. Second, a state's interests in the challenged practice must be taken into account during the remedial phase of a vote dilution lawsuit. In Whitcomb v. Chavis, 403 U.S. 124, 160-61, 91 S.Ct. 1858, 1878, 29 L.Ed.2d 363 (1971), the Supreme Court held that the district court may not "broadly brush[] aside state apportionment policy" when remedying an allegedly dilutive electoral mechanism. Instead, when choosing among remedial plans or fashioning one of its own, the district court must honor state policies and intrude on those policies only to the extent necessary. See Upham v. Seamon, 456 U.S. 37, 40-42, 102 S.Ct. 1518, 1520-1521, 71 L.Ed.2d 725 (1982); White v. Weiser, 412 U.S. 783, 795, 93 S.Ct. 2348, 2355, 37 L.Ed.2d 335 (1973); see also United States v. Dallas County Comm'n, 850 F.2d 1430, 1432 (11th Cir.1988) ("[W]hen devising election plans, federal courts should defer to state legislative policy and modify the state's plan only to the extent necessary to cure statutory or constitutional defects."), cert. denied, 490 U.S. 1030, 109 S.Ct. 1768, 104 L.Ed.2d 203 (1989).

B. The Proposed Role for State Interests in Section 2 Analysis

The State Defendants, Judge Wood, and Judge Entz propose a new role for state interests in Section 2 analysis. Although the details for their proposals are different, all the proposals have one thing in common. Under each of the proposals, state interests in an underlying election scheme would have to be weighed against proven vote dilution before a Section 2 violation could be found.

Judge Wood proposes a three-step balancing framework. Under her framework a court would first be required, “[i]n each vote dilution case, [to] separately determine the existence and severity of vote dilution.” Judge Wood’s Brief on Remand at 29. The court would then be required to separately “assess the state interests that would be affected.” Id. Finally, the court would have to “balance the severity of the dilution against the intrusion on state interests necessary to cure it to determine if a remediable violation of § 2 exists under the totality of the circumstances.” Id. If the state interests outweighed the severity of the dilution, then there would be no violation of Section 2.

The balancing framework proposed by Judge Entz is somewhat more complex. Under his proposed framework, the minority group would first have to satisfy the Gingles threshold inquiry. Once the minority group did so, the court would then have to consider whether the state interests are legally compelling—that is, (1) whether they relate to a central aspect of state sovereignty, and (2) whether tampering with them would result in undue federal entanglement with the operations of state government. If the state interests were compelling, they would “prevent a section 2 violation.” Judge Entz’s Brief on Remand at 9. If the state interests were not legally compelling, the district court would still be required to weigh the state *757interests in the challenged practice, “along with all other factors ... to determine whether the [minority group] proved dilution by clear and convincing evidence.” Id.

Although similar to the framework proposed by Judge Entz, the framework proposed by the State Defendants is yet a third version of the balancing framework. Under this proposal, the district court would first, before inquiring into any alleged vote dilution, ascertain whether the challenged election procedure is necessary to achieve a compelling state interest. If this Fourteenth Amendment strict scrutiny standard were satisfied, the State Defendants argue, the state interest would be “sufficiently strong to outweigh as a matter of law the other elements in the totality of the circumstances____” State Defendants’ Brief on Remand at 21. If the strict scrutiny standard were not satisfied, then, under the State Defendants’ proposal, the district court would still be required to engage in some kind of weighing: The plaintiffs would have to set forth facts to satisfy the Gingles inquiry. The burden would then “shift to the state to produce evidence of its interests in the maintenance of the system and the non-discriminatory reasons for retaining the system.” Id. at 24. And finally, if the state met its shifting burden, the plaintiffs would have to prove their vote dilution case by clear and convincing evidence. Id. at 24-25.

C. Problems with the Proposed Balancing Framework

In evaluating the appropriateness of balancing state interests against proven vote dilution under Section 2 analysis, we discover two problems. The first problem is that the proposed balancing framework may not be grounded in law. The second is that the proposed balancing framework appears to unnecessarily complicate an already complex analysis. Although both of these problems call into question the appropriateness of the proposed balancing framework, we ultimately need not decide whether such a framework is required in this case.

1. The Legal Problem

The proposed balancing framework may be grounded in law in one of several ways. First, in amending Section 2 and providing for a totality of circumstances analysis, Congress may have expressly intended for courts to balance the “total circumstances” pointing toward vote dilution against the state’s asserted interests in the challenged election practice. Second, even if Congress did not expressly intend to provide for a balancing framework, constitutionally-rooted federalism principles may nonetheless require such a framework. Or finally, the Supreme Court may have erected a balancing framework when it remanded this case to us in Houston Lawyers’ Association v. Attorney General of Texas, — U.S. —, 111 S.Ct. 2376, 115 L.Ed.2d 379 (1991).

a. Congressional intent. Before addressing any constitutional concerns, we first ask whether Congress demonstrably intended for state interests to be balanced against proven vote dilution before finding a violation of Section 2. If Congress expressed an intent that courts should use a balancing framework in analyzing vote dilution claims, we are undoubtedly bound by that intent. Of course, in discerning congressional intent, we apply the traditional tools of statutory construction, looking first to the language of Section 2 and then to the legislative history if the language of the statute is ambiguous. See Blum v. Stenson, 465 U.S. 886, 896, 104 S.Ct. 1541, 1548, 79 L.Ed.2d 891 (1984).

Our inquiry necessarily begins with the language of Section 2. On its face, Section 2 does not allow a state to defeat proof of vote dilution by showing that it has a good reason for maintaining the challenged electoral practice. It provides, in absolute terms, that a violation is established if, “based on the totality of the 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 racial or language minority]____” 42 U.S.C. § 1973(b). It makes no exceptions.

The Senate Report accompanying the amendment of Section 2 speaks in similarly *758absolute terms. The report states: “If as a result of the challenged practice or structure plaintiffs do not have an equal opportunity to participate in the political processes and to elect candidates of their choice, there is a violation cf. this section.” S.Rep. at 28, 1982 U.S.C.C.A.N. at 206 (emphasis added). It does not make an exception for election practices or structures that are justified by strong or compelling state interests. And, although the Senate Report allows courts to consider whether the state interests underlying the challenged procedure are tenuous, this consideration is qualitatively different from balancing state interests against proven vote dilution. Moreover, Congress emphasized that “even a consistently applied practice premised on a racially neutral policy would not negate a plaintiff’s showing through other factors that the challenged practice denies minorities fair access to the process.” Id. at 29 n. 117, 1982 U.S.C.C.A.N. at 207 n. 117. Thus, the Senate Report, like the language of Section 2 itself, appears inconsistent with the balancing framework proposed in this case.

b. Federalism principles. The State Defendants, Judge Wood, and Judge Entz argue that, even if Congress did not expressly intend for courts to employ a balancing framework under Section 2, federalism principles nonetheless require us to construe Section 2 to allow for such an approach. Relying primarily on the Supreme Court’s decision in Gregory v. Ashcroft, — U.S. —, 111 S.Ct. 2395, 115 L.Ed.2d 410 (1991), they argue that we should employ a plain statement rule of statutory construction. In particular, they argue that we should construe Section 2 to require a balancing framework unless Congress has unmistakably indicated its intent to upset the usual constitutional balance of federal and state powers.

In Gregory, the Supreme Court granted certiorari to decide whether the mandatory retirement provision for state judges in the Missouri Constitution violated the federal Age Discrimination in Employment Act of 1967 (ADEA), 29 U.S.C. §§ 621-634. See — U.S. at —, 111 S.Ct. at 2398. The ADEA prohibits any employer, including state employers, from discharging an employee who is at least 40 years old because of such individual’s age. 29 U.S.C. § 630(b)(2). The definition of “employee” under the ADEA, however, excludes all elected and most high-ranking government officials. See 29 U.S.C. § 630(f). The crucial question in Gregory thus became whether state judges, who are appointed by. the Governor of Missouri and thereafter subject to retention elections, were “employees” within the meaning of the ADEA. — U.S. —, 111 S.Ct. at 2403.

In order to avoid potential constitutional problems, the Court in Gregory applied a “plain statement” rule of statutory construction, requiring Congress to unmistakably express its intention to alter the usual constitutional balance between the states and the federal government. Because Congress did not plainly express its intent to include appointed state judges within the terms of the ADEA, the Court held that such judges were not covered. — U.S. at —, 111 S.Ct. at 2404. The Court reasoned that it “would not read the ADEA to cover state judges unless Congress has made it clear that judges are included.” Id. (emphasis in original).

The Court’s application of the plain statement rule in Gregory rested on two concerns. The first was the nature of the state interest involved. The Court emphasized that the Missouri constitutional provision was not just “an area traditionally regulated by the States, [but] a decision of the most fundamental sort for a sovereign entity.” — U.S. at —, 111 S.Ct. at 2400. According to the Court, “the authority of the people of the States to determine the qualifications of their most important government officials ... lies at ‘the heart of representative government.’ ” Id. at —, 111 S.Ct. at 2402. The second concern was the nature of the power being exercised by Congress under the ADEA. Although there was some argument that Congress acted pursuant to its enforcement powers under the Fourteenth Amendment in passing the ADEA, the Court in Gregory ultimately concluded that Congress had acted only pursuant to its Commerce Clause pow*759ers. See id. at —, 111 S.Ct. at 2404-06. The Court thus expressed its rationale for applying the plain statement rule when it said: “As against Congress’ powers ‘[t]o regulate Commerce ... among the several States,’ U.S. Const. Art. I, § 8, cl. 3, the authority of the people of the States to determine the qualifications of their government may be inviolate.” Id. at —, 111 S.Ct. at 2402-03.

The concerns underlying the Court’s application of the plain statement rule in Gregory do not, in our view, exist in this case. Section 2 of the Voting Rights Act does not impinge on the authority of the states to set qualifications for their elected officials. Rather, by its terms, Section 2 only impinges on the authority of states to employ certain election practices—namely, those which result in a denial or abridgment of the right to vote on account of race. Furthermore, it is clear that, when Congress amended Section 2, it was acting pursuant to its enforcement powers under the Fourteenth and Fifteenth Amendments. Congress specifically concluded that, “to enforce fully the Fourteenth and Fifteenth Amendments, it is necessary that Section 2 ban election procedures and practices that result in a denial or abridgment of the right to vote.” S.Rep. at 40, 1982 U.S.C.C.A.N. at 218 (emphasis added).

Moreover, even if we were to apply the plain statement rule in determining whether Section 2 provides for the balancing of state interests against proven vote dilution, we would be inclined to hold that Congress plainly intended that no such balancing should occur.12 After all, as previously discussed, Congress has indicated in the Senate Report accompanying the amended Section 2 that even if the challenged election procedure is consistently applied and premised on a racially neutral policy, such proof does “not negate a plaintiff’s showing through other factors that the challenged practice denies minorities fair access to the [political] process.” S.Rep. at 29 n. 117, 1982 U.S.C.C.A.N. at 207 n. 117. Along the same lines, the Senate Report further notes that the failure to prove that the state’s interests are tenuous “is not rebuttal evidence of non-dilution.” Id. at 29 n. 118, 1982 U.S.C.C.A.N. at 207 n. 118. Finally, after carefully considering Congress’ primary purpose in amending Section 2, we believe that examining a state’s motives beyond the limited inquiry into tenuousness would be inconsistent with that purpose. Congress has indicated that, when a plaintiff proceeds under the “results test” of Section 2, the court should “assess the impact of the challenged structure or practice on the basis of objective factors, rather than making a determination about the motivations which lay behind its adoption or maintenance.” S.Rep. at 27, 1982 U.S.C.C.A.N. at 205.

Nor are we prepared to hold, as Judge Entz would have us do, that Congress' failure to provide for a balancing of state interests against proven vote dilution ren*760ders Section 2 unconstitutional.13 In Mississippi Republican Executive Committee v. Brooks, 469 U.S. 1002, 105 S.Ct. 416, 83 L.Ed.2d 343 (1984), the Supreme Court summarily rejected the argument that Section 2, "if construed to prohibit anything other than intentional discrimination on the basis of race in registration and voting, exceeds the power vested in Congress by the Fifteenth Amendment." Id. at 1003, 105 S.Ct. at 416. We have similarly rejected a broad challenge to Congress' power to amend Section 2 and prohibit election practices that produce discriminatory results. In Jones v. City of Lubbock, 727 F.2d 364, 373 (5th Cir.1984), we stated that "Congressional power to adopt prophylactic measures to vindicate the purposes of the fourteenth and fifteenth amendments is unquestioned." We noted that Congress did not lightly interfere with state and municipal self-governance. Id. at 374 n. 5. Instead, we found that, in light of the legislative fact-finding accompanying the amendment and the limited nature of the intrusion into state and municipal self-determination, Congress had not offended any other provision of the Constitution. Id.; see also City of Rome v. United States, 446 U.S. 156, 173-75, 100 S.Ct. 1548, 1559-1560, 64 L.Ed.2d 119 (1980); South Carolina v. Katzenbach, 383 U.S. 301, 334, 86 S.Ct. 803, 822, 15 L.Ed.2d 769 (1966). But see Chisom v. Roemer, — U.S. —, —, 111 S.Ct. 2354, 2376, 115 L.Ed.2d 348 (1991) (Kennedy, J., dissenting) (clarifying that the Court was not addressing the constitutional validity of Section 2).

We recognize that in Gregory the Court indicated that "the Fourteenth Amendment does not override all principles of federalism." — U.S. at —, 111 S.Ct. at 2405. The Court noted, however, "that the principles of federalism that constrain Congress' exercise of its Commerce Clause powers are attenuated when Congress acts pursuant to its powers to enforce the Civil War Amendments." Id. The reason federalism principles are attenuated in this context, the Court explained, "is because those `Amendments were specifically designed as an expansion of federal power and an intrusion on state sovereignty.'" Id. (quoting City of Rome, 446 U.S. at 179, 100 S.Ct. at 1563). Thus, although Congress may not act to enforce the Civil Rights Amendments "in complete disregard of" the powers retained by the states, Gregory, — U.S. at —, 111 S.Ct. at 2405, concerns of federalism must not be allowed to emasculate Congress' power to adopt prophylactic measures to vindicate the purposes of those Amendments.

c. The Supreme Court’s decision in Houston Lawyers’ Association. Were it not for isolated passages in the Supreme Court’s decision in this case, we would likely conclude that the proposed balancing framework is not grounded in law. However, when the Supreme Court held that Section 2 applies to the election of state district court judges in Houston Lawyers’ Association, it also used language suggesting that state interests might have to be balanced against proven vote dilution. Relying on the Court’s suggestions, the State Defendants, Judge Wood, and Judge Entz argue that state interests must be balanced against proven vote dilution.

There are two passages in the Court’s opinion in Houston Lawyers’ Association that provide support for the proposed balancing framework. The first one reads as follows:

Even if we assume, arguendo, that the State’s interest in electing judges on a district-wide basis may preclude a remedy that involves redrawing boundaries or subdividing districts, or may even pre*761elude a finding that vote dilution has occurred under the “totality of the circumstances” in a particular case, that interest does not justify excluding elections for single-member offices from the coverage of the § 2 results test. Rather, such a state interest is a factor to be considered by the court in evaluating whether the evidence in a particular case supports a finding of a vote dilution violation in an election for a single-member office.

— U.S. at —, 111 S.Ct. at 2380-81 (emphasis added). In the second such passage, which follows the Court’s rejection of the argument that “the ‘single-member office’ theory automatically exempts certain elections from the coverage of § 2,” the Court states:

Rather, we believe that the State's interest in maintaining an electoral system — in this case, Texas' interest in maintaining the link between a district judge's jurisdiction and the area of residency of his or her voters — is a legitimate factor to be considered by courts among the totality of circumstances in determining whether a § 2 violation has occurred. A State's justification for its electoral system is a proper factor for the courts to assess in a racial vote dilution inquiry, and the Fifth Circuit has expressly approved the use of this particular factor in the balance of considerations. See Zimmer v. McKeithen, 485 F.2d 1297, 1305 (CA5 1973), aff'd sub nom. East Carroll Parish School Bd. v. Marshall, 424 U.S. 636, 96 S.Ct. 1083, 47 L.Ed.2d 296 (1976). Because the State's interest in maintaining an at-large, district-wide electoral scheme for single-member offices is merely one factor to be considered in evaluation the totality of the circumstances, that interest does not automatically, and in every case, outweigh proof of racial vote dilution.

Id. at —, 111 S.Ct. at 2381 (emphasis in original and emphasis added) (internal citations omitted). In both of these passages, the underscored language suggests that courts should balance state interests against proven vote dilution.

Other parts of the Court’s opinion in Houston Lawyers’ Association suggest that the Court was not substituting a new balancing framework for the one set forth in Section 2. Initially, the Court stated that it granted certiorari “for the limited purpose of considering the scope of the coverage of § 2.” — U.S. at —, 111 S.Ct. at 2380. The Court focused specifically on whether judicial elections are subject to Section 2, and its narrow holding was: “If a State decides to elect its trial judges, as Texas did in 1861, those elections must be conducted in compliance with the Voting Rights Act.” Id. In addition, before the Court made any statement suggesting that balancing of state interests might be required, it stated that it was “deliberately avoidpng] any evaluation” of the state interest arguments made by Judge Higginbotham in his concurrence below. Id. Such matters, the Court continued, “are relevant either to an analysis of the totality of the circumstances that must be considered in an application of the results test embodied in § 2, as amended, or to a consideration of possible remedies in the event a violation is proved, but not to the threshold question of the Act’s coverage.” Id. (emphasis added).

Further, in the second passage (quoted above) relied on to support a balancing framework, the Court begins by recognizing that the state interest in maintaining an electoral system is “a legitimate factor to be considered by courts among the ‘totality of circumstances’ in determining whether a § 2 violation has occurred.” Id. at —, 111 S.Ct. at 2381 (emphasis added). The Court then notes that this court “has expressly approved the use of this particular factor in the balance of consideration^]” and cites our opinion in Zimmer, 485 F.2d at 1305, which formed the basis for the treatment of state interests in the Senate Report. The inquiry into state interests in the Senate Report is limited to an inquiry into whether the state interest is tenuous, and if so, it is a factor supporting a finding of dilution. This reference to Zimmer, in our view, suggests that the Court did not *762intend to substitute a balancing framework.14

That the Court in Houston Lawyers’ Association did not mean to substitute a balancing framework is further supported by what is absent from the Court’s opinion. The Court’s decision does not mention, much less discuss, either the Senate Report or Thornburg v. Gingles. We question whether the Court would alter the Section 2 framework without reference to either of these authorities. Nor does the Court’s decision indicate why such a balancing framework would be required. There is no discussion in Houston Lawyers’ Association of congressional intent or of federalism principles. The absence of such a discussion, in our view, suggests that the Court did not mean to require courts to balance state interests against proven vote dilution.

Also, in Chisom v. Roemer, decided the same day as Houston Lawyers’ Association, the Court expressly disavowed any intention to alter or amend the elements of a vote dilution claim. The Court stated that it was not addressing “any question concerning the elements that must be proved to establish a violation of the [Voting Rights] Act or the remedy that might be appropriate to redress a violation if proved.” Id. at —, 111 S.Ct. at 2361. It is difficult to believe that the Supreme Court would issue one opinion expressly disavowing any intention to alter the elements of a Section 2 claim, and on the same day issue another opinion which, sub silentio, recasts the elements of such a claim.

Finally, prior to Houston Lawyers' Association, the Supreme Court has on several occasions rejected the argument that state interests may preclude relief from an election scheme violating federal law. In Mahan v. Howell, 410 U.S. 315, 93 S.Ct. 979, 35 L.Ed.2d 320, modified, 411 U.S. 922, 93 S.Ct. 1475, 36 L.Ed.2d 316 (1973), which involved a challenge to the apportionment plan for the Virginia legislature, the Court rejected the state's argument that its interest in drawing legislative districts to conform to county and city boundaries justified some population deviation between districts. Then-Justice Rehnquist wrote for *763the Court that "a State's policy urged in justification of disparity in district population, however rational, cannot constitutionally be permitted to emasculate the goal of substantial equality." Id. 410 U.S. at 326, 93 S.Ct. at 986. The Court rejected a similar argument in White v. Weiser, 412 U.S. 783, 93 S.Ct. 2348, 37 L.Ed.2d 335 (1973). In White, the State of Texas defended its congressional reapportionment plan on the ground that it had an interest in preserving the relationships between incumbent legislators and their constituents. Id. at 791, 93 S.Ct. at 2352. While the Court did not "disparage this interest," it held that district courts should defer to state policy "only where that policy is consistent with constitutional norms and is not itself vulnerable to legal challenge." Id. at 791, 797, 93 S.Ct. at 2353, 2355.

Accordingly, we conclude that Houston Lawyers’ Association is, at best, ambiguous about the proper role of evidence regarding state interests in Section 2 analysis. The passages relied on by the State Defendants, Judge Wood, and Judge Entz suggest that state interests may be relevant in determining whether there is a violation of Section 2. And, in at least one of those passages, the Court uses language suggesting that state interests might have to be balanced against proven vote dilution. In the same breath, however, the Court (1) emphasizes the limited scope of its holding, (2) refuses to address Judge Higginbotham’s concurring opinion regarding the importance of state interests, and (3) notes that it is appropriate to consider state interests when imposing a remedy for a Section 2 violation. Given this ambiguity, we cannot say that the proposed balancing framework is wholly unsupported by the law.

2. The Practical Problem

Section 2 demands a great deal from district courts. They must “consider the totality of the circumstances and ... determine, based upon a searching practical evaluation of the past and present reality, whether the political process is equally open to minority voters.” Gingles, 478 U.S. at 79, 106 S.Ct. at 2781 (internal quotation marks and citations omitted). “This determination is peculiarly dependent on the facts of each case, and requires an intensely local appraisal of the design and impact of the contested electoral mechanisms.” Id. (internal quotation marks and citation omitted). Moreover, because Section 2 cases require, close analysis of “extraordinarily fact-oriented issuesf,]” we have held that, under Rule 52(a) of the Federal Rules of Civil Procedure, a district court is “required to discuss all the substantial evidence contrary to its opinion.” Velasquez v. City of Abilene, 725 F.2d 1017, 1020 (5th Cir.1984).

The balancing framework possibly suggested in Houston Lawyers’ Association would further complicate this already complex undertaking. Because the state interests would not be a signal of dilution like the other Section 2 factors (including the tenuousness of the state interests), a court analyzing a Section 2 claim under the proposed balancing framework would no longer ask only whether, under the totality of the circumstances, there is an unequal opportunity to participate in the political process. The court would also ask whether, in light of the state interests, the unequal opportunity to participate is sufficiently unequal to support a finding of liability under Section 2. In other words, the state interests would have to be weighed against proven vote dilution. See Houston Lawyers’ Association, — U.S. —, 111 S.Ct. at 2381. The problem with such a framework, from a district court’s perspective, is the lack of standards for such an approach. Neither the Supreme Court’s decision in Houston Lawyers’ Association nor Section 2 provides any guidance for determining whether the state interests are substantial enough—or whether the proof of vote dilution is insubstantial enough—to excuse the state from Section 2 liability.

The proposed balancing framework becomes even more complicated when we consider the number of state interests and the possible remedies that might be implicated in any case. Although state election schemes generally embody several state policy choices, see, e.g., infra Part III.D.l., *764the only state interests threatened in a Section 2 case are those that would be trammeled by a proposed remedy. Thus, the only way to identify the threatened state interests would be to identify a remedy. This difficulty in identifying the threatened state interests is not present in the non-balancing framework described in Part II, where a court must only ask whether the state interest is tenuous. The proposed balancing framework, on the other hand, would apparently require a district court to assess the potential damage to each of the various state interests during the liability phase of a Section 2 case, before the need for a remedy has been absolutely determined. In short, a district court would have to weigh, against a showing of minority vote dilution, each of the state interests, as that interest is affected by each of the possible remedies. Such a complicated balancing act would be required, because it is the only way to fairly assess the strength of the asserted state interests.

3. Summation

In sum, although there are legal and practical problems with the proposed balancing framework, we need not decide whether such a framework is required in this case. After a careful review of the state interests asserted in this case, see infra Part III.D., we conclude that even if Houston Lawyers’ Association were read to require courts to balance state interests against proven vote dilution, the Plaintiffs in this case would still prevail in the eight counties in which they have demonstrated vote dilution under Section 2, see infra Part IV. Their evidence establishes substantial vote dilution in each of the eight counties, and as discussed below, the interests asserted by the state are little more than tenuous. Moreover, there are several potential remedies that could cure the proven vote dilution while simultaneously protecting the asserted state interests.

D. Applying the Proposed Balancing Framework in this Case: Evaluating Texas’ Asserted Interests

Before we can weigh the asserted state interests against proven vote dilution, see infra Part IV.C., we must evaluate Texas’ asserted interests. Our evaluation proceeds in three steps. First, we identify the threatened state interests. Second, we scrutinize those threatened state interests. And finally, we assign a weight to the threatened state interests.

1. Identifying the Threatened State Interests

In evaluating the state interests in a challenged election scheme, a court must first, as a matter of law, identify those interests that would be threatened by possible remedies. To identify state interests in an election scheme or practice, a court may -look to state statutes and constitutions. See White v. Weiser, 412 U.S. 783, 795, 93 S.Ct. 2348, 2355, 37 L.Ed.2d 335 (1973) (in the context of legislative reapportionment, a federal court “should follow the policies and preferences of the State, as expressed in statutory and constitutional provisions or in the reapportionment plans proposed by the state legislature,” to the extent possible under federal law) (emphasis added). A court may also look to evidence introduced at trial. To identify the subset of state interests that are threatened by a finding of vote dilution, however, a court must consider all possible remedies. As discussed above, only the interests that are threatened by a proposed remedy need be considered under the proposed balancing framework.

Texas has various interests in the current method of electing district court judges. Among those expressed in the state statutes and constitution are the following: an interest in an elected judiciary; 15 an interest in partisan judicial elections; 16 an interest in a qualified judicia*765ry;17 an interest in having district judges reside in the district in which they serve;18 and an interest in requiring judicial candidates to refrain from certain kinds of political activity.19 Based on testimony and evidence at trial, the district court identified other state interests in Texas’ current scheme of electing district judges. These include: an interest in an electorate informed about the qualifications of candidates for district court judgeships; an interest in preserving the administrative advantages of at-large elections; an interest in allowing judges to specialize in areas like family law; a “linkage” interest in electing district judges from the area over which they have “primary jurisdiction”; and finally, an interest in preserving the function of the district court judge as a sole decision-maker.20

Most of the interests listed above are not threatened by any of the possible remedies in this case. Neither subdistricting, cumulative voting, nor limited voting would threaten Texas’ asserted interests in having an elected and qualified judiciary or in having partisan judicial elections. Moreover, none of the possible remedies would threaten Texas’ interests in requiring that judges reside in the district in which they serve and refrain from certain forms of political activity. Finally, the state interest in an informed electorate would be enhanced by at least some of the proposed remedies and threatened by none of them.

Four of the interests listed above, however, appear to be threatened to some extent by some or all of the proposed remedies. Texas’ interest in preserving the administrative advantages of the current at-large election system is threatened to some extent by all of the possible remedies. Texas’ interest in allowing judges to specialize may also be threatened to a minimal degree, but only by a strict subdistricting remedial scheme. Texas’ linkage interest may also be threatened by a subdistricting remedy. Finally, we are told that Texas’ interest in preserving the fact and appearance of independence in district court judges—that is, in preserving the function of district court judges as sole decision-makers—is threatened by a subdistricting remedy.

2. Scrutinizing the Threatened State Interests

Having identified the interests asserted by Texas that would arguably be threatened by possible remedies, we now proceed to scrutinize those interests. It is not enough to evaluate the state interests as a matter of legal or political theory. See Gingles, 478 U.S. at 45, 106 S.Ct. at 2764 (in Section 2 case, court must ground its analysis in "a searching practical evaluation of the past and present reality" of the community and its politics, and must take a "functional view of the political process") (quoting S.REP. at 30 & n. 120, 1982 U.S.C.C.A.N. at 207 & n. 120). A court must also examine the practical realities of *766the challenged election scheme to determine the strength of the asserted interest. See White v. Weiser, 412 U.S. at 796, 93 S.Ct. at 2355 (in deciding whether district court erred in choosing one of two apportionment plan options, an appellate court considers the impact of each plan on state apportionment policy); see also Westwego III, 946 F.2d at 1123 (concluding, after scrutinizing district court's finding that at-large system was justified by concerns of administrative efficiency, that such a finding was no longer supported by the evidence because of a change in underlying state law). In addition, a court must determine whether the challenged election scheme "does in fact advance" the asserted state interests. See Mahan v. Howell, 410 U.S. 315, 329, 93 S.Ct. 979, 987, 35 L.Ed.2d 320 (emphasis added), modified, 411 U.S. 922, 93 S.Ct. 1475, 36 L.Ed.2d 316 (1973); see also id. 410 U.S. at 326, 93 S.Ct. at 986 ("The inquiry ... [is] whether it can reasonably be said that the state policy urged ... is, indeed, furthered by the plan adopted by the legislature...."). Finally, in scrutinizing the asserted state interests, a court must determine whether any of the proposed remedies will serve the asserted state interests as well as the challenged election scheme. See White v. Weiser, 412 U.S. at 791-92, 93 S.Ct. at 2353 (refusing to decide whether state interest was sufficient to justify legislative apportionment plan, where another proposed plan served the state interest just as well).

The first two state interests that are threatened in this case—namely, Texas’ interests in preserving the administrative advantages of the current system and allowing judges to specialize—have received little attention in the parties’ briefs. This is because the third and fourth threatened state interests—Texas’ interest in “linking the jurisdictional base of district judges directly to the electoral base” and Texas’ interest in preserving the function of district court judges—have threatened to assume a larger than life role in this lawsuit. We scrutinize each of these interests in turn.

a. Texas’interest in preserving the administrative advantages of the current at-large system. As a practical matter, we can agree that any of the proposed remedies will result in Texas losing some administrative advantages of its current method of electing judges. We can also agree that maintenance of the current method of electing Texas district court judges does to a considerable extent further the state’s administrative interest. We note, however, that elevating Texas’ administrative interest in the liability phase of a Section 2 case to a dispositive position would be contrary to our decision in Westwego I, where we held that the district court should not consider the potential disruption of reorganizing an administrative structure in determining whether an at-large electoral system violates Section 2. 872 F.2d at 1211. More importantly, however, it must be acknowledged that, in passing the 1985 constitutional amendment permitting the creation of judicial districts smaller than a county, see Tex. Const. art. V, § 7a(i), the Texas legislature and the voters considered the potential loss of administrative advantages and concluded that the advantages in such smaller districts outweighed any loss of administrative advantages under the current system. The voters have spoken to this issue.

b. Texas’ interest in allowing judges to specialize. Our scrutiny of Texas’ interest in allowing judges to specialize reveals one major weakness in this interest: Several of the proposed remedies would advance the interest just as well as the current election scheme. Under a cumulative voting scheme, for example, judges could still specialize. Voters would not be required to cast a vote for specialized judges, but would nonetheless have the opportunity to cast one or more of their votes for such judges. Moreover, a subdistricting remedy could also protect the interest in allowing judges to specialize. Under a flexible subdistricting remedy, specialized judges would not necessarily have to be elected from districts the same size as district court judges who do not specialize. And, if a subdistricting election scheme for a specialized judge could not be devised—e.g., if there were only one family law judge per *767county—that judge could still be elected on a county-wide basis.

c. Texas’ linkage interest. As we have previously noted, the so-called “linkage interest”—Texas’ interest in linking the elective base of district judges to the area in which they exercise “primary jurisdiction”—has threatened to take on a life of its own in this litigation. The genesis of this interest, it appears, is the testimony of Chief Justice Thomas Phillips of the Texas Supreme Court. At trial, in response to a question asking him whether “it is a good idea to require at least county-wide elections” for district court races, Chief Justice Phillips testified that, in his view, such a requirement was a good idea. In explaining his answer, Chief Justice Phillips stated:

In my opinion, the district judge should not be responsible to the voters over an area that is smaller than that area in which the district judge exercises primary jurisdiction. If we wanted to go to a system where the judge had primary venue responsibility over an area smaller than the county, I don’t know that I would have any objection to electing judges from [an area] smaller than a county. And our Justices of the Peace do ordinarily have jurisdiction primarily over a portion of the county, but as long as we are going to elect judges, as long as judges are going to be responsible for all the cases in the county and are going to exercise that kind of power and authority over persons[’] lives and as long as they are going to be responsible to the voters, I believe they should be responsible to all the voters of the county, or a multi-county district.

In this somewhat conflicting testimony, the so-called linkage interest was born.

Although Chief Justice Phillips first uses the phrase "primary jurisdiction,"21 it becomes clear that what he was attempting to describe was two-fold: first, a judge's elective base, and second, a judge's "primary venue responsibility." Thus, although the State Defendants repeatedly speak of the state's interest in terms of "linking the jurisdictional base of district judges directly to the electoral base," it is really only the district court's venue that has any link to its electoral base. The terms "jurisdiction" and "venue" are not synonymous under Texas law. See State v. Pounds, 525 S.W.2d 547, 550 (Tex.Civ. App. — Amarillo 1975, writ ref'd n.r.e.). Jurisdiction "is the power of the court to decide a controversy between parties and to render and enforce a judgment with respect thereto...." Id. In Texas, district courts have general, statewide jurisdiction — that is, they have the power to hear any case that may be filed in the state of Texas. See TEX. CONST. art. V, § 8 (district courts have jurisdiction of all actions and proceedings except where jurisdiction is conferred on another court or tribunal by the Texas Constitution or by statute). Moreover, a district court's order is effective throughout the state. Venue, on the other hand, "is the proper place where [a court's] power [to decide a specific controversy] is exercised." Pounds, 525 S.W.2d at 550. It concerns the propriety of "prosecuting a suit involving a given subject matter and specific parties in a particular county." Jeter-Millar Co. v. Kasch Bros. Inc., 466 S.W.2d 598, 600 (Tex.Civ.App. — Eastland 1971, no writ). Thus, because all but one of the district judges in this case are elected from county-wide judicial districts, their elective base can only be linked — if it is linked at all — with venue, not jurisdiction.22

Moreover, whatever link may have existed between a district judge’s elective base and venue disappeared in 1985. In that year, as we have seen, the state amended its constitutional provision tying a district judge to the county in which he or she was elected. Specifically, the Texas Constitu*768tion now permits district judges to be elected from areas smaller than a county when a majority of the voters in such a county approve of such an election scheme. See Tex. Const. art. V, §§ 7, 7a(i).23

The Texas legislature has also weakened the ties between district court judges and the counties which elect them by establishing a mechanism by which district judges can be compelled to periodically sit in other counties. The state of Texas is divided into nine administrative judicial regions, the smallest of which consists of eleven counties. See Tex.Gov’t Code Ann. § 74.042 (Vernon 1988). The presiding judge of each region may assign judges within the region “to hold special or regular terms of court in any county of the administrative region to try cases and dispose of accumulated business.” Id. § 74.056(a). The presiding judge of an administrative region may further “request the presiding judge of another administrative region to furnish judges” to aid in the disposition of litigation. Id. § 74.056(b). A judge assigned to sit in another county under these provisions “has all the powers of the judge of the court to which he is assigned.” Id. § 74.059(a). Texas also makes widespread use of retired judges, who are assigned across administrative region lines. See id. § 75.002. In view of these provisions, it is simply not possible to conclude that Texas insists on tying its district judges to the counties in which they are elected.24

Texas' interest in linking district judges' elective base to the respective areas in which they have venue is further weakened by the operation of the state's venue rules. The Texas venue rules have not been drafted to insure that parties appear before judges for whom they have had an opportunity to vote. Instead, the venue rules for lawsuits involving living persons — including the general venue rule,25 the mandatory venue exceptions,26 and the permissive venue exceptions27 — "were, in the main, manifestly adopted to prevent serious inconveniences and probable injury to defendants ...." Snyder v. Pitts, 150 Tex. 407, 241 S.W.2d 136, 142 (1951). Nor do the Texas venue rules operate evenly. When someone who fears being hailed into court as a defendant files an action under Texas' Declaratory Judgment Act, the action will *769be "subject to the general venue rule and its ... exceptions." Bracewell v. Fair, 638 S.W.2d 612, 613 (Tex.App. — Houston [1st Dist.] 1982, no writ). There is no mechanism under Texas' venue rules, it appears, for piercing the form of the lawsuit. See Charles T. Frazier, Jr., Note, Venue Procedure in Texas: An Analysis of the 1983 Amendments to the Rules of Civil Procedure Governing Venue Practice Under the New Venue Statute, 36 BAYLOR L.Rev. 241, 260 (1984) (recognizing that, under the current Texas venue rules, a party who anticipates being sued may "capture" venue by filing suit first).

The operation of Texas’ venue rules suggests that these rules have not been developed to promote the accountability of judges to the litigants appearing before them. Put differently, Texas’ venue rules, like the venue rules in the federal courts and in the courts of other states, are not designed to foster judicial accountability. The most that can be said for the so-called “linkage interest” is that there may be some rough correlation—a loose link indeed—between a district court’s elective base and venue. Perhaps it simply comes down, as Justice Phillips candidly put it, to a “good idea.”28

Finally, Texas’ asserted linkage interest is weakened by the availability of several remedies that could cure the dilution caused by the use of at-large elections while simultaneously protecting Texas’ linkage interest. For instance, a system of limited voting, in which each voter casts fewer votes than the total number of places to be filled, would allow Texas to continue to select its district judges through county-wide elections—and thereby preserve the “linkage interest”—while also giving minority voters a meaningful opportunity to elect candidates of their choice. See Sen. Gene Green, The Future of Judicial Selection in Texas: Election Methods, 55 Tex.B.J. 294, 295 (1992); Daniel R. Ortiz, Note, Alternative Voting Systems as Remedies for Unlawful At-Large Systems, 92 Yale L.J. 144, 148-50 (1982). A system of cumulative voting, in which each voter is entitled to cast as'many votes as seats to be filled and to give her preferred candidate more than one voté, would also protect Texas’ “linkage interest.” See Green, supra, at 295; Note, supra, at 153-60.

In sum, our scrutiny of Texas’ asserted linkage interest reveals several substantial weaknesses in the interest. The first is Texas’ refusal to insist that judges in fact be elected from districts no smaller than a county. The second is Texas’ willingness to direct judges elected from one county to preside periodically over lawsuits filed in other counties. The third weakness arises from the uneven operation .of Texas’ venue rules, which do not ensure, even as a general matter, that parties will appear before a judge for whom they have had the opportunity to vote. And fourth, the strength of Texas’ linkage interest is weakened by the availability of remedies that will protect it.

d. Texas’ interest in preserving the function of district court judges as sole decision-makers. Texas maintains that, even if it does not consistently link a district court’s elective base with venue, it nonetheless has a distinct, compelling state interest in the function of the district court *770judge as a sole decision-maker. The argument is essentially as follows: Although in any one county there may be more than one district judge with county-wide, overlapping jurisdiction, each district judge acts alone in wielding power. That is, “[bjefore any suits are filed, before any cases are assigned, there is a group of judges with concurrent jurisdiction----” League of United Latin Am. Citizens Council No. 4434 v. Clements, 914 F.2d 620, 649 (5th Cir.1990) (en banc) (Higginbotham, J., concurring) (LULAC II). Once a case is assigned, however, “It is decided by only one judge. The other judges have absolutely no say over the disposition of that case, and no influence over the deciding judge.” Id. Thus, the district court judge in Texas is not a member of a multi-member body like a city council or even an appellate court that engages in collegial decision-making. Rather, the district court judge in Texas is charged with making his or her decision alone and with exercising exclusively the full authority of the office.

The current method of electing judges, we are told, is necessary to preserve the function of the district court as a sole decision-maker. Electing district judges from an area no smaller than a county ensures that district judges are responsive to all litigants equally and that no particular group will have undue influence over the decisions that must be made alone by the judge. Moreover, under the current method of electing district court judges, minorities can participate in all judicial elections in each county. This opportunity to participate, the argument continues, gives minority voters the opportunity to influence all elections.

A subdistricting remedy, it is argued, poses two threats to Texas’ interest in preserving the function of the district court judge as sole decision-maker. First, such a remedy would destroy the appearance if not the fact of independence and responsiveness to all. Said another way: if elected from a subdistrict, a district judge would be prevented from acting alone, or making his or her decisions independently, because the judge would be beholden to the particular racial group who elected him or her. Second, a subdistricting remedy—perversely—would disenfranchise minority voters, because they would no longer be able to influence the outcome of every district court election, but only a few. Such a remedy, the argument goes, makes it probable that a minority litigant would appear “before a judge who has little direct political interest in being responsive to minority concerns.” LULAC II, 914 F.2d at 650. We address each of these contentions in turn.

Certain parties have suggested on remand that; if minorities are given greater voting strength in district court elections, the function of district court judges as sole decision-makers will be destroyed. Judge Wood argues that the subdistricting remedy sought by the Plaintiffs runs “directly counter to the historical preference of the citizens of Texas for an elected judiciary in which each judge is accountable to each voter and independent from special interest groups.” Judge Wood’s Brief on Remand at 31. Judge Wood further suggests that increasing minority voters’ ability to sponsor and elect district court candidates of their choice from single member districts would serve no “purpose other than to provide [minority] role models,” role models who would apparently serve “specialized interests different from those served by Texas’ majority judges.” Id. at 29. Judge Entz makes similar suggestions. He argues that, under a subdistricting remedy, a trial judge would still be required to exercise his or her authority alone, but the authority, which would necessarily be “derived from voters representing only a fraction of the judge’s [current] electoral base,” would somehow be tainted. Judge Entz’s Brief on Remand at 16.

The argument that any proposed subdistricting remedy would destroy the district judge’s ability to act alone or independently is premised on two flawed assumptions. The first flawed assumption is that district judges elected from smaller election districts will not abide by their oaths. The Texas Code of Judicial Conduct charges judges with applying the law and states that an honorable judiciary separated from *771the influence of others is “indispensable to justice in our society.” Tex.Code of Judicial Conduct, Canon 1. The Code also stresses that “[a] judge should be unswayed by partisan interests, public clamor, or fear of criticism.” Id. Canon 3, pt. A(l). These directives, by which we must assume judges will abide, effectively negate any concern about a judge’s ability, even one elected from a subdistrict, to continue making his or her decisions alone.29 The second flawed assumption is a more disturbing one. Underlying the argument that a subdistricting remedy would destroy a district judge’s ability to act independently is the following assumption: while district judges who are currently elected by white majorities in each of the counties at issue are responsive to the needs of all voters in the county—including minority voters—a judge who would be elected by minorities in a single-member district would be unresponsive to the current white majority. The flaw in this assumption being clear, we will not address it further.

More important, this argument has been implicitly rejected by the Texas legislature and Texas voters in 1985. By amending the Texas Constitution to allow for the election of district judges from an area smaller than a county, see Tex. Const. art. V, § 7a(i), the Texas legislature and Texas voters must have concluded that district judges who are elected from an area smaller than a county could continue to make their decisions alone and independently. In short, the voters have spoken to this concern.

The second contention raised by the parties in connection with the single-member office theory is that, because there is no share of district court judgeships, creating safe minority district court positions would disenfranchise minorities. Judge Entz argues that, under the current method of electing district court judges, voters of all races are permitted to influence every judicial election in the county. Judge Entz’s Brief on Remand at 17. An alternative system would, in Judge Entz’s view, “fix geographical and racial power and influence over judicial elections for all time— and actually limit the numbers of judicial benches possible for a growing minority population.” Id. According to Judge Wood, the subdistricting remedy sought by the Plaintiffs in this case would disenfranchise all the voters in each of the counties at issue, “thereby greatly decreasing minority voting power.” Judge Wood’s Brief on Remand at 31.

Again, the argument about the disastrous results of subdistricting is flawed. The argument that in subdistricting judicial positions, “[minority] voters will ... be sacrificing [an] extremely valuable political right—the right to vote for all of the judges who will be serving as judges in the [county] wherein they live,” see Southern Christian Leadership Conference of Alabama, 785 F.Supp. at 1478—smacks of paternalism. . It essentially tells minority voters in each of the target counties in this case that they must somehow be saved from themselves. In our view, it is not difficult to understand why minority voters would willingly “sacrifice” their ability to vote for all judges and elect none for the ability to vote for and elect a few judges. Indeed, Congress amended Section 2, at least in part, to allow minority voters to make such a sacrifice.

In sum, after scrutinizing Texas’ interest in preserving the function of district court judges as sole decision-makers, we con-*772elude that this interest is not actually-threatened by any of the proposed remedies—including subdistricting. There is no evidence in this record suggesting that judges elected from smaller districts, or judges elected from districts in which minority voters constitute a majority, would be any less independent in deciding cases. Nor is there any evidence suggesting that judges elected from such districts would be any less sensitive to the diverse interests of the litigants than the judges who are currently elected from larger districts.

3. Assigning a Weight to the Threatened State Interests

Assuming that courts are required to apply a new balancing framework to Section 2 claims, we think that the weight assigned to any state interests is probably a question of law.30 We note that, to a limited extent, the inquiry into the weight of the state interests is not novel. That is, from the beginning, vote dilution cases have focused on the tenuousness of the state interests. See, e.g., Cross v. Baxter, 604 F.2d 875, 878, 884 (5th Cir.1979); Zimmer, 485 F.2d at 1305, 1307. District courts deciding these cases have assigned tenuous or non-tenuous weights to the various interests asserted by the affected jurisdictions. See, e.g., Canton Branch, NAACP v. City of Canton, 472 F.Supp. 859, 876 (S.D.Miss.1978). Appellate courts appear to have looked at the weight of the state interests afresh without deference to the district courts’ decisions. See, e.g., Zimmer, 485 F.2d at 1307. In all cases, courts addressing the tenuous or non-tenuousness of the state interests have done so without explaining whether the decision is a question of law or fact. Given the legal sources that underlie the determination, we think that the strength of the state interests may be a legal question. But, in candor, we are not certain that it has made any difference what kind of a question it is. It is important, however, that the inquiry about the weight of a state interest, at least to the extent of determining whether it is tenuous, is a familiar question.

Even though the weight of Texas’ threatened interests may be a legal question which we could determine de novo, we find ourselves in agreement with much of what the district court had to say about Texas’ asserted interests. In particular, we agree with the district court that, while the threatened state interests underlying Texas’ current method of electing district judges are not tenuous, they are not compelling either. On the basis of our scrutiny of those interests, we conclude that Texas’ interests in (1) preserving the administrative advantages of the current at-large system, (2) allowing judges to specialize, (3) linking the elective base of district court judges with venue, and (4) preserving the function of district court judges are at best “legitimate state interests.” They are not, based on our review of Texas law and the practical realities of the Texas judicial system, compelling or substantial state interests. Accordingly, in any proposed balancing framework, they would be entitled to little weight.

IY. REVIEW OF THE DISTRICT COURT’S SECTION 2 LIABILITY FINDINGS

Having set forth the analytical framework (or frameworks) for analyzing Section *7732 claims, we now proceed to the question framed at outset of this opinion: Did the district court err in concluding that the method by which Texas elects district court judges—as that method operates in the nine target counties—violates Section 2 of the Voting Rights Act? Because of the uncertainty about the relevance of state interests in Section 2 analysis, we think it prudent, in answering this question, to evaluate the district court’s findings under both the accepted analytical framework and the proposed balancing framework. Before reviewing the district court’s findings under the two frameworks, however, we explain the standard of appellate review.

A. Standard of Appellate Review

The standard of appellate review for findings made in a Section 2 case is the clearly erroneous standard set forth in Rule 52(a) of the Federal Rules of Civil Procedure. The Supreme Court made this clear in Gingles by "reaffirm[ing]" that the ultimate Section 2 liability or vote dilution finding is reviewed for clear error. See 478 U.S. at 79, 106 S.Ct. at 2779. We have also recognized that the district court's subsidiary findings under the Gingles threshold inquiry and the totality of circumstances inquiry — with the possible exception of the tenuousness of the state interests, see supra Part III.D.3 — are also reviewed under the clearly erroneous standard. See Westwego I, 872 F.2d at 1203; Campos v. City of Baytown, 840 F.2d 1240, 1243 (5th Cir.1988), cert. denied, 492 U.S. 905, 109 S.Ct. 3213, 106 L.Ed.2d 564 (1989).

The reason we apply such a standard of review is due to the nature of the vote dilution inquiry. As the Supreme Court in Gingles recognized, the vote dilution determination under Section 2 is “peculiarly dependent upon the facts of each case” and requires “an intensely local appraisal of the design and impact of the contested electoral mechanisms.” Gingles, 478 U.S. at 79, 106 S.Ct. at 2779 (internal quotations and citations omitted). Moreover, “the application of the clearly-erroneous standard to ultimate findings of vote dilution preserves the benefit of the trial court’s particular familiarity with the indigenous political reality without endangering the rule of law.” Id.

Even under the clearly erroneous standard set forth in Rule 52(a), a district court’s factual findings may be set aside if they rest on an erroneous view of the law. Pullman-Standard v. Swint, 456 U.S. 273, 287, 102 S.Ct. 1781, 1789, 72 L.Ed.2d 66 (1982). As we stated in Armco, Inc. v. Armco Burglar Alarm Co., 693 F.2d 1155 (5th Cir.1982): “When a finding of fact is based on a misapprehension of governing legal standards, it loses the insulation of the ‘clearly erroneous’ rule.” Id. at 1162. Once a district court’s finding loses the insulation of the clearly erroneous rule, however, “a remand is the proper course unless the record permits only one resolution of the factual issue.” Pullman-Standard, 456 U.S. at 292, 102 S.Ct. at 1792 (emphasis added). A remand is generally proper, according to the Supreme Court, because it is the district court, and not the court of appeals, who is “charged with the task of factfinding in the first instance.” Id. at 293, 102 S.Ct. at 1792.

If the district court applies the appropriate legal standards in making a finding of fact, under Rule 52(a) we will not set aside that finding unless, based upon the entire record, we are "left with the definite and firm conviction that a mistake has been committed." Anderson v. City of Bessemer City, 470 U.S. 564, 573, 105 S.Ct. 1504, 1511, 84 L.Ed.2d 518 (1985) (quoting United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 542, 92 L.Ed. 746 (1948)). "If the district court's account of the evidence is plausible in light of the record viewed in its entirety," we will not reverse it — even if convinced that had we "been sitting as the trier of fact, [we] would have weighed the evidence differently." Anderson, 470 U.S. at 573-74, 105 S.Ct. at 1511. Moreover, we give "due regard ... to the opportunity of the trial court to judge of the credibility of the witnesses." FED.R.CIV.P. 52(a). This reluctance to reverse a district court's findings under the clearly erroneous standard *774persists "even when the district court's findings do not rest on credibility determinations, but are based instead on physical or documentary evidence or inferences from other facts." Anderson, 470 U.S. at 574, 105 S.Ct. at 1511-1512.

B. Review of the District Court’s Vote Dilution Findings Under the Accepted Section 2 Framework

Our review of the district court's vote dilution findings under the accepted Section 2 framework proceeds in three parts. In the first part, we discuss the methodology used by the Plaintiffs and accepted by the district court, and address the general attacks on that methodology by the State Defendants, Judge Wood, and Judge Entz. In the second part, we review the district court’s subsidiary fact findings under the Gingles threshold inquiry and the totality of circumstances inquiry, as well as the district court’s ultimate vote dilution finding in each of the nine counties. Finally, in the third part we consider the impact of the evidence of partisan voting patterns on the district court’s ultimate vote dilution findings.

1. Statistical Methodology

At trial, the Plaintiffs relied heavily on statistics to establish that Texas’ method of electing judges results in vote dilution in each of the nine target counties. To establish the first Gingles factor—the size and geographical compactness of the minority group—the Plaintiffs used voting age population data from the 1980 Census. To establish the second and third Gingles factors—the political cohesiveness of the minority group and legally significant white bloc voting—as well as the existence of racially polarized voting, the Plaintiffs performed certain statistical analyses on results from various elections. Finally, the Plaintiffs relied on statistics to establish certain factors that are relevant to the totality of circumstances inquiry—namely, the lingering socioeconomic effects of discrimination and the extent to which minorities have been elected to public office.

Because of the position of the second and third Gingles factors and racially polarized voting in Section 2 analysis, it is important to outline the Plaintiffs’ statistical methodology with regard to these factors. There are three aspects to this methodology: (1) the subset of elections that the Plaintiffs chose to analyze; (2) the manner in which the Plaintiffs isolated the race of the voters in the nine target counties; and (3) the types of analyses the Plaintiffs chose to perform on the particular election results.

With regard to the elections that were analyzed, the Plaintiffs’ experts, Dr. Engstrom and Dr. Brischetto, were guided by several principles. They only studied elections in which a minority candidate opposed a white candidate. Also, they preferred to analyze general elections as opposed to primary elections; however, they analyzed primary elections when no minority candidate made it past that stage of the electoral process. Finally, they preferred to analyze state district court elections. If, however, there were no district court elections or primaries in which a minority candidate opposed a white candidate, the Plaintiffs’ experts looked, in the following order of preference, to county court elections, justice of the peace elections, and statewide judicial races.

In isolating the independent variable for their statistical analyses, the racial composition of election precincts, the Plaintiffs’ experts took various approaches, depending on the availability of data. In counties where the Plaintiffs were proceeding only on behalf of Hispanics, they relied on Spanish surname counts done by the Texas Secretary of State to determine the percentage of Hispanic registered voters in any given precinct. In other instances—for example, in Harris and Dallas counties—the Plaintiffs’ experts used updated counts of Black and Hispanic total or voting age population in each precinct of the county. If such counts were not available, then the Plaintiffs’ experts relied on data from the 1980 Census. When relying on census data, the Plaintiffs’ experts calculated the number of non-minorities within precincts by subtracting the number of Hispanics and Blacks *775from the total number of persons within the precinct.

The Plaintiffs’ experts performed two types of analyses on the data sets that were obtained: ecological regression analysis and extreme case or homogeneous precinct analysis.31 “Ecological regression analysis” shows the relationship between the ethnic composition of voting precincts and voting behavior—or, to put it differently, which candidate receives how many votes from each race or ethnic group. See Overton v. City of Austin, 871 F.2d 529, 539 (5th Cir.1989). “Extreme case or homogeneous precinct analysis” focuses on precincts in which almost all of the voting age population belong to one ethnic group. Under this type of analysis, if the race or ethnicity of a voter predicts voting behavior, then election results in predominately minority precincts should differ from results in predominately white precincts.

The State Defendants, Judge Wood, and Judge Entz raise various, across-the-board challenges to the methodology used by the Plaintiffs and accepted by the district court. They first attack the Plaintiff’s reliance on 1980 Census figures to establish thé Cingles factors. They also challenge the Plaintiffs’ selection of elections. Finally, they complain about the Plaintiffs’ use of bivariate regression analysis in establishing racially polarized voting.

With regard to the Plaintiffs’ use of population statistics from the 1980 Census, the complaint is that the statistics are dated. Quoting this court’s decision in Houston v. Haley, 859 F.2d 341, 345 (5th Cir.1988), vacated on other grounds, 869 F.2d 807 (5th Cir.1989), Judge Entz argues that “[wjhatever the voting age population composition was [in 1980], given mobility, mortality, and coming of age, we cannot tell with any certainty what it is today.” Thus, Judge Entz concludes that the Plaintiffs failed to carry their burden of proof with regard to the Cingles factors, and that, therefore, the district court erred in relying on their statistical analyses. We disagree.

After examining the record and Fifth Circuit law on this issue, we conclude that the complaint about the Plaintiffs’ use of outdated statistical data is without merit. In Overton, we recognized that “absolute perfection of the base statistical data is not to be expected.” 871 F.2d at 539. It is difficult for us to fault the Plaintiffs—and the district court—for relying on the most recent census figures in this case. Moreover, we note that in Dallas and Harris counties, the Plaintiffs relied on updated census figures. Judge Entz would apparently require minority groups seeking relief under Section 2 to either bring their lawsuits within the first few years after census figures are released or engage in a costly re-estimation of population figures. Although we cautioned in Overton that “a trial court should not ignore the imperfections of the data used nor the limitations of the statistical analysis,” id., we do not think the trial court ignored the imperfections in the data set relied on by the Plaintiffs in Dallas County. Instead, the district court, after carefully considering those imperfections, accepted the data sets used by the Plaintiffs as being reliable. Accordingly, we conclude that the district court did not err in relying on the 1980 Census population data or on the analyses by the Plaintiffs’ experts who also relied on this data.

The State Defendants, Judge Wood, and Judge Entz also complain about the principles by which the Plaintiffs’ experts chose the relevant elections to be analyzed. They argue specifically that the Plaintiffs’ experts should have considered more elections—elections pairing white candidates against white candidates, elections pairing white candidates against other minority candidates, primary elections, and non-judicial elections. Because the Plaintiffs’ experts did not consider all of these types of elections, they conclude that the district court erred in relying on the Plaintiffs’ analyses. Again, we disagree.

*776The State Defendants, Judge Wood, and Judge Entz essentially complain that, in the battle of the experts, their expert—Dr. Taebel—lost. Dr. Taebel analyzed the various types of elections described above, and the district court considered his analyses in each of the counties at issue. We cannot fault the district court for giving more weight to the elections analyzed by the Plaintiffs’ experts, especially given our recognition .that “the evidence most probative of racially polarized voting must be drawn from elections including both [minority] and white candidates.” Westwego I, 872 F.2d at 1208 n. 7. Accordingly, we reject the complaint that, as a matter of law, the elections the Plaintiffs chose to analyze are insufficient to establish the second and third Gingles factors as well as the existence of racially polarized voting.

Finally, Judge Wood argues that, to establish racially polarized voting, the Plaintiffs were required to conduct a multivariate analysis on the election results—an analysis that would consider variables other than the race of the voter, such as partisan affiliation of the candidate, whether the candidate was endorsed by the local bar poll, and campaign expenditures. We reject this argument. As we have already explained, see supra Part II.B.l.b. and note 7, to establish racially polarized voting, a Section 2 plaintiff need only demonstrate that whites and minorities vote differently. Thus, bivariate analyses such as the ones used by the Plaintiffs are sufficient.

In sum, we conclude that the statistical methods used by the Plaintiffs, although not perfect, are sufficient for purposes of establishing the Gingles factors and racially polarized voting. The Plaintiffs and the district court could reasonably rely on 1980 Census statistics. It was also reasonable for the district court to give more weight to the elections analyzed by the Plaintiffs. And finally, the Plaintiffs’ were not required to use a multivariate analysis to demonstrate the existence of racially polarized voting.

2. Review of District Court’s Vote Dilution Findings

In Gingles, the Supreme Court stated that “[t]he inquiry into the existence of vote dilution caused by submergence in a multimember district is district specific.” 478 U.S. at 59 n. 28,106 S.Ct. at 2771 n. 28. In a case like this, where a number of vote dilution claims are raised in one case, the district court must rely on data “specific to each individual district” in making the appropriate findings under the Gingles threshold inquiry and the totality of circumstances inquiry. See id. The district court may not, according to the Court in Gingles, rely on data aggregated from all the challenged districts. Id.; see also id. at 101, 106 S.Ct. at 2793 (O’Connor, J., concurring) (voting statistics from one district are ordinarily irrelevant in assessing the totality of circumstances in another).

Because the Plaintiffs challenge Texas method of electing judges, as that method operates in nine target counties, we review the district court’s findings on a county-by-county basis. That is, with regard to each of the nine target counties—Bexar County, Dallas County, Ector County, Harris County, Jefferson County, Lubbock County, Midland County, Tarrant County, and Travis County—we review the district court’s specific findings under the Gingles inquiry, the totality of circumstances inquiry32 and the ultimate vote dilution inquiry33 for clear error.

*777a. Bexar County

In Bexar County, where nineteen (19) state district court judges are elected in county-wide elections, Plaintiffs proceed only on behalf of Hispanic voters. At trial, the district court found that Plaintiffs had introduced evidence sufficient to establish all three Gingles factors and five of the factors listed as relevant in the totality of circumstances inquiry. Based on this evidence, the district court found that Texas’ method of electing district judges dilutes Hispanic voting strength in Bexar. For the following reasons, we conclude that none of the district court’s findings with regard to Bexar County is clearly erroneous.34

(i) Gingles factors. With regard to the first Gingles factor, the district court found that, as a group, Hispanic voters in Bexar County are sufficiently large and geographically compact to constitute a majority in a single member district. In making this finding, the district court relied on evidence introduced by the Plaintiffs indicating that (a) Hispanics constitute 41.1% of the voting age population in Bexar County, (b) Hispanics primarily reside in the central and south central sections of Bexar County, and (c) it would be possible to create eight single-member districts in which voting-age Hispanics constitute majorities. Because the district court correctly based its determination on the voting age population in Bexar County, see supra Part H.A.I., and because its determination is plausible in light of the record, we conclude that its finding with regard to the first Gingles factor is not clearly erroneous.

The district court also found, under the legal standards required to establish the second Gingles factor, see supra Part II. A.2., that Hispanic voters in Bexar County are politically cohesive. In reaching this finding, the district court primarily relied on Dr. Brischetto’s statistical analysis of six district court elections in which Hispanic candidates opposed white candidates. This analysis indicated that, in Bexar County, Hispanic voters gave 73% to 93% of their votes to the same candidate. The district court also relied on a bivariate regression analysis performed by Dr. Brischetto, an analysis that demonstrated a “strong relationship between race and voting patterns in Bexar County.” In view of this evidence, we are not “left with a definite and firm conviction” that the district court was mistaken in finding that Hispanic voters in Bexar County are politically cohesive. Thus, this finding may not be set aside as clearly erroneous.

As for the third Gingles factor, the district court determined, after considering all the testimony offered by the Plaintiffs and the State Defendants, that there is legally significant white bloc voting in Bexar County. The district court first considered Dr. Brischetto’s statistical analysis of the six general elections in which Hispanic district court candidates opposed white candidates. This data indicated that the preferred candidate of Hispanic voters won only once and that the non-Hispanic support for the Hispanic-preferred candidate ranged from 18% to 35% in those six elections.35 The district court also considered *778statistical evidence offered by Dr. Taebel, the State Defendants’ expert, who analyzed twelve general elections, only one of which was a district court election, and eighteen primary elections. Dr. Taebel found that the Hispanic-preferred candidate won three of twelve general elections and six of eighteen primary elections. Finally, the district court considered testimony from Hispanic judicial candidates who attributed their losses to the lack of adequate support from white voters. Given the evidence presented to the district court, we cannot say that it clearly erred in finding persuasive Dr. Brischetto’s conclusion that the white bloc vote in Bexar county is sufficiently strong to usually defeat the Hispanic community’s preferred candidate.

(ii) Totality of circumstances factors. Inquiring into the totality of the circumstances in Bexar County, the district court found: that Hispanics had been subject to a history of discrimination in voting, that elections in Bexar County were characterized by racially polarized voting, that several aspects Texas’ method of electing district judges enhance the opportunity for discrimination in Bexar County, that Hispanics in Bexar County continue to bear the socioeconomic effects of past discrimination, and that Hispanics in Bexar County have not been successful in district court elections. The district court found, however, no evidence of a slating process and no use of racial appeals in Bexar County elections. None of these findings is clearly erroneous.

The district court effectively took judicial notice of the history of official discrimination against Hispanics and Blacks in Texas, finding that this history was “either well chronicled or undisputed.” This history of official discrimination, the district court continued, “touched many aspects of the lives of minorities in [the nine target counties,] including their access to and participation in the democratic system governing this State____” (emphasis added). Neither the State Defendants nor the intervenors challenge this finding on appeal. Indeed, the State Defendants acknowledge the “long, shameful history of racial discrimination in Texas.” Given this admission, we cannot conclude that the district court clearly erred in finding a history of discrimination touching the rights of Hispanics to participate in the Bexar County political process. See also Graves v. Barnes, 343 F.Supp. 704, 728 (W.D.Tex.) (three-judge court) (noting that “the Mexican-American population of Texas ... has historically suffered from, and continues to suffer from, the results and effects of invidious discrimination and treatment in the fields of education, employment, economics, health, politics, and others”), aff’d sub nom. Archer v. Smith, 409 U.S. 808, 93 S.Ct. 62, 34 L.Ed.2d 68 (1972), aff’d in part and rev’d in part sub nom. White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973) (Graves I).

The district court also found strong evidence of racially polarized voting in Bexar County. Dr. Brischetto’s statistical analysis of six district court elections from 1980 to 1988 revealed that, in all but one of the races, as the percentage of Hispanic voters in a precinct increased, support for the Hispanic preferred candidate also increased. According to Dr. Brischetto’s statistical analysis, race explains 64% to 77% of the variance in voting in these six elections. The district court further observed that, in twenty-eight of the twenty-nine judicial elections involving Hispanic candidates in Bexar County, whites and Hispanics voted differently. Because the record supports the district court’s finding of a strong relationship between race and voting patterns in Bexar County, we will not set that finding aside as being clearly erroneous.

Surveying the other factors relevant to the political landscape in Bexar County, the district court found that certain features of Texas’ scheme of electing district court judges interact with the at-large nature of that election scheme to further enhance the opportunity for discrimination in Bexar County. In particular, the district court found that the large size of Bexar County, the majority vote requirements in primary *779elections, and the use cf. what amounts to a numbered-post system in the county contribute to the dilution of minority voting strength. The potential of such features to dilute minority voting strength being well known, see supra Part II.B.l.c., we will not disturb the district court’s finding with respect to this factor.

The district court further found that the Hispanic population in Bexar County — as well as minorities throughout Texas — continue to bear the effects of past discrimination in such areas as education and employment. This finding is amply supported by evidence introduced at trial: Over 73% of Bexar County's Hispanic families have a yearly income below $20,000, while approximately 44% of Bexar County's white families have a yearly income below that level. Some 24.3% of Hispanic families in Bexar County — compared to 5.4% of white families in that county — live below the poverty line. Moreover, only 5.2% of Bexar County Hispanics over the age of 25 are college graduates, while 24.9% of whites in Bexar County hold college degrees. And, only 11.8% of those in Bexar County's Hispanic work force hold managerial or professional positions, while 31.7% of those in Bexar County's white work force hold such positions. See Appendix IV. There is also evidence in the record indicating that (a) Hispanics in Bexar County register to vote at a substantially lower rate than members of other ethnic or racial groups, and (b) Hispanics, who constitute only 11.7% of Bexar County lawyers, are underrepresented in the pool of qualified district court candidates. Finally, the district court's finding is supported by decisions from this court acknowledging the lingering effects of discrimination on Hispanics in Texas. See e.g., Campos v. City of Baytown, 840 F.2d at 1249-50; United States v. Texas Educ. Agency, 564 F.2d 162, 174 (5th Cir. 1977), cert. denied, 443 U.S. 915, 99 S.Ct. 3106, 61 L.Ed.2d 879 (1979). We therefore conclude that the district court did not clearly err in finding that Hispanics in Bexar County continue to bear the socioeconomic effects of past discrimination.

In considering the extent to which Hispanics have been elected to office in Bexar County, the district court determined that Hispanic district court candidates are not very successful. This finding is plausible in light of the record, and therefore, it is not clearly erroneous. The record reveals that from 1981 to 1989, while only 15.8% to 31.6% of the district judges sitting in Bexar County were Hispanics, Hispanics constituted some 46.61% of the total population in Bexar County. The record also reveals that Hispanics running for state district court judgeships won only one of six elections. Finally, the district court noted that, in the 1988 Republican primary for the 150th District Court, Arellano—an appointed Hispanic incumbent—was defeated by White, an Anglo candidate who had withdrawn from the primary and endorsed Arellano.

The State Defendants argue that “Hispanic judicial candidates in Bexar County have been very successful.” They point specifically to the fact that, at the time of trial, there were five Hispanic district court judges in Bexar County. The problem with relying solely on the number of district court judges who are Hispanic, at least according to the Plaintiff’s expert Dr. Brischetto, is that it does not give a “complete” picture of what happens, to a Hispanic candidate who runs against a white candidate. Several of the Hispanic judges that the State Defendants’ point to, according to Dr. Brischetto, got their position without having to run in a contested election. Faced with this conflicting evidence, the district court was warranted in finding that Hispanics running for state district court judgeships in Bexar County are not very successful.36

With regard to the remaining totality of circumstances factors listed in the Senate Report, the district court found: no evidence of a slating organization in Bexar County or of overt or subtle racial appeals *780in Bexar County political campaigns. We conclude that these findings, which are not challenged, are not clearly erroneous.

(iii) Ultimate vote dilution finding. Based on these subsidiary findings, the district court found that, under the totality of the circumstances, Texas’ current method of electing district judges operates in Bexar County to dilute the voting strength of Hispanic voters. This finding was based on a number of political and social realities in Bexar County, which together' provide strong evidence that Hispanic voters in Bexar County currently have an unequal opportunity to participate in district court elections on account of their ethnic origin. Assuming the district court’s ultimate vote dilution finding in Bexar County was based on the correct legal standards, see infra Parts IV.B.3. and IY.C., we conclude that the finding is not clearly erroneous.

b. Dallas County

Plaintiffs proceed on behalf of Black voters in Dallas County, where, at the time of trial, thirty-six (36) state district judges were chosen in county-wide elections.37 As with Bexar County, the district court found that the Plaintiffs satisfied all three factors of the Gingles threshold inquiry. The district court also determined that the Plaintiffs had sufficiently demonstrated six of the totality of circumstances factors listed in the Senate Report. The district court finally determined, based on the totality of circumstances, that the Plaintiffs had demonstrated vote dilution in Dallas County. We conclude that the district court did not clearly err in making these findings.

(i) Gingles factors. With regard to the first Gingles factor, the district court determined that Blacks in Dallas County were sufficiently large and geographically compact to constitute a majority in seven single-member districts. This determination was based partly on evidence that, of the 1,106,757 Dallas County citizens of voting age, 16% are Black. This determination was further based on evidence that Blacks primarily reside in the central and south central sections of Dallas County. Because it is supported by the record, this determination is not clearly erroneous.

Moving to the second Gingles threshold factor, the district court found that Blacks in Dallas County are politically cohesive when voting for state district court judges. In making this finding, the district court primarily relied on a regression analysis performed by Dr. Engstrom, an expert employed by the Plaintiffs. The regression analysis of seven Dallas County district court elections involving a Black candidate indicated that “Blacks consistently gave more than 97% of their vote to their preferred candidate." Dr. Engstrom’s statistical analysis was confirmed by Dr. Wiser, another expert employed by the Plaintiffs, who concluded that 98% of the vote in homogeneous Black precincts and 83% of the vote in precincts with a Black population of 50% to 90% went to the Black-preferred candidate. The district court correctly applied the legal standards for determining political cohesiveness, see supra Part H.A.2., and its finding that Black voters, in Dallas County are politically cohesive is plausible in light of the record. Accordingly, the finding does not constitute clear error.

With regard to the third Gingles factor, the district court found, on the basis of the exhibits and testimony of Dr. Engstrom and Dr. Wiser, that the “white bloc vote in Dallas County is sufficiently strong to generally defeat the choice of the Black community.” In the seven elections studied by Dr. Engstrom and Dr. Wiser—elections in which a Black district court candidate opposed a white district court candidate— some 61% to 71% of white voters in Dallas County voted against the candidate preferred by Black voters. Significantly, in five of the seven elections, the candidate preferred by Black voters was defeated. The highest percentage of white cross-over voting for the Black-preferred candidate was approximately 39%. Based on this evi*781dence we are not left with a “definite and firm conviction” that the district court was mistaken in finding legally significant white bloc voting in Dallas County district judge elections.

(ii) Totality of circumstances factors. In analyzing the totality of the circumstances in Dallas County, the district court found: a history of official discrimination against Blacks in Dallas County; racially polarized voting; the use of voting practices that enhance the opportunity for discrimination against Blacks; lingering effects of past discrimination in the Black population of Dallas County; the use of racial appeals in Dallas County judicial elections; and a lack of success by Black district court candidates in Dallas County. The district court refused to find, however, that the Republican Party in Dallas County operated as a white-dominant slating group.

We find no clear error in the district court’s first finding under the totality of circumstances inquiry—namely, that the history of official discrimination in Texas “touched many aspects of the life of minorities ... including their access to and participation in the democratic system governing this State.” The district court correctly recognized that the history of racial discrimination against Blacks in Texas is “either well chronicled or undisputed.”38 Moreover, cases decided in this circuit reveal that Dallas County shares Texas’ history of discrimination. As recently as 1970, an organization calling itself the “Dallas Committee for Responsible Government” employed racial campaign tactics in white precincts to defeat the candidates supported by Black voters. See Graves, 343 F.Supp. at 726-27. And during the late 1950’s, not one member of the Dallas County delegation voted against certain segregation measures introduced in the Texas House of Representatives. Id. Thus, we think the district court was warranted in taking judicial notice of the fact that Blacks in Dallas County historically have been the victims of official discrimination.39

The district court also determined, based on Dr. Engstrom’s homogeneous precinct and ecological regression analyses, that there is “a strong relationship between race and voting patterns in Dallas County.” According to Dr. Engstrom’s analyses, race explains at least 75% of the variance in voting in at least six of the seven elections studied. The State Defendants and Judge *782Entz do not dispute that voters in Dallas County vote along racial lines. Rather, they argue that the raw statistics do not explain why Dallas County voters vote along racial lines. Having refused the invitation to redefine the concept of racially polarized voting by including a causation requirement, see supra note 7, we conclude that the district court correctly found that racially polarized voting exists in Dallas County.40

As with Bexar County, the district court found that certain features of Dallas County district court elections—namely, the use of large election districts, majority vote requirements in primary elections, and the use of a numbered-post system—enhance the possibility of vote dilution. Judge Entz challenges the district court’s finding with regard to enhancement features, arguing (a) that Dallas County is not an unusually large election district, (b) that post-primary elections do not have a majority vote requirement, and (c) that, although a judicial candidate must run for a specifically numbered court, this requirement "is a reflection of the autonomy of the each court and its status in the true sense of the term as a single member judicial district.” These challenges are without merit. First, with regard to the size of Dallas County, we note that district judges elected in Dallas County, with a population of 1.5 million, have a larger constituency than sixteen governors and thirty-two United States senators. Second, although we can agree that the district court did not find a majority vote requirement in post-primary elections, we must take into account the district court’s finding of a majority vote requirement in primary elections. Contrary to Judge Entz’s suggestion, the absence of a majority vote requirement in general elections does not remove the dilutive potential of such a requirement in primary elections. See supra Part II.B.l.c. Third, the justification offered for the existence of a numbered post system in Dallas County undercuts neither the existence of such a system nor its potential to dilute. See id. Because the district court’s finding of enhancement features in Dallas County is amply supported by the record, it is not clearly erroneous.

The district court further found that Blacks in Dallas County continue to bear the effects of discrimination in education and employment. The district court specifically stated that “[t]he effect of past discrimination against Blacks and Hispanics in areas such as education, employment and health in most of the Counties in question is either well chronicled or undisputed.” With respect to Dallas County, this finding was plausible in light of the record. As the statistics in Appendix IV demonstrate, Blacks in Dallas County lag unreasonably behind whites in terms of income, education, and professional status. Conversely, Blacks are disproportionately represented among those who live below the poverty line: 21.2% of Black families in Dallas County, while only 3.9% of white families, live below the poverty line. Moreover, the record reveals that, while Blacks constitute some 16% of Dallas County’s voting age population, Black lawyers make up at most 2.2% of the lawyers residing in Dallas County. On the basis of this uncontroverted evidence, we conclude that the district court did not clearly err in finding that racial discrimination continues to affect the lives of Blacks in Dallas County.41

*783The district court also found—significantly—that in two recent Dallas County elections, white candidates made racial appeals. In a 1986 district attorney election between John Vance, a white candidate, and Royce West, a Black candidate, Mr. Vance inserted a picture of his Black opponent in campaign advertisements. In the 1988 Republican district court primary between Larry Baraka, a Black candidate, and Brook Busby, a white candidate, Ms. Busby was less subtle. She distributed campaign literature that referred to Mr. Baraka as a “Black Muslim.”

Judge Entz challenges the district court’s finding of racial appeals in Dallas County as being clearly erroneous. Judge Entz argues specifically that judicial elections in Dallas County “were not characterized by *784overt or subtle racial appeals” and that “the only incident the trial court found to be a racial appeal in a judicial race was actually a reference to the candidate’s religion.” We disagree. The district court did not clearly' err in finding that Ms. Busby’s reference to her opponent as a “Black Muslim” constituted more than a reference to Mr. Baraka’s religion.42 The district court’s finding with respect to Mr. Vance’s campaign literature is also supported by the record. The district court determined that Vance’s campaign literature sought to identify his opponent by race only after hearing testimony and reviewing the evidence in question. In view of this evidence, we will not set aside as clearly erroneous the district court’s finding that Dallas County judicial elections have been tainted by racial appeals.

As for the success of Black candidates in Dallas County, the district court was confronted with conflicting evidence. At the time of trial Black district court candidates had opposed white district court candidates only seven times in Dallas County general elections, and the Black candidate won two of the seven elections. In both of those elections, however, the Black candidate was not the candidate preferred by Blacks. The record also reveals that, from 1980 to 1989, the percentage of Black judges in Dallas County ranged from 0.0% to 8.3%; by contrast, Blacks constituted 18.5% of Dallas County’s total population during this time period. Faced with this conflicting evidence about the success of Black district court candidates in Dallas County, the district court found Black candidates to be relatively unsuccessful.

Judge Entz challenges the district court’s finding with respect to the success of Black candidates in Dallas County on two grounds. First, he argues that the baseline for determining minority candidate success is not the total minority population, but the number of minority lawyers in the county. He also contends that the district court should have considered the success of Black candidates in primary elections. We reject both of these challenges and hold that the district court’s finding on this totality of circumstances factor is not clearly erroneous.

We have already explained our serious reservations about gauging minority electoral success under the qualified applicant pool approach. See supra Part II. B.l.g. Such an approach would, in our view, “allow voting rights cases to be defeated at the outset by the very barriers to political participation that Congress sought to remove.” Id. (quoting Westwego I, 872 F.2d at 1209). . Moreover, the district court correctly recognized that, “even if there is some relationship between the low number of minority judges and the number of eligible minority lawyers, that fact does not explain why well-qualified eligible minority lawyers' lose judicial elections.” Accordingly, we conclude that the district court correctly refused to compare the proportion of Blacks on the Dallas County district court bench with the proportion of Black lawyers in the county.

Judge Entz’s contention that the district court should have considered the success of Black candidates in primary elections is similarly without merit. The Senate Report instructs courts to consider “the extent to which members of the minority group have been elected to public office in the jurisdiction.” S.Rep. at 29, 1982 U.S.C.C.A.N. at 207 (emphasis added). *785The report does not instruct courts to consider the extent to which members of the minority group have won primary elections. Thus, while it may be true that, in the nine contested primary and general district court elections with Black candidates, the Black candidate won four of those elections, this is not the success rate with which Section 2 is concerned. Rather, under Section 2 it is significant that (a) no Black district court candidate who has received any measurable support from the Black community has ever been elected in Dallas County, and (b) Blacks are significantly underrepresented on the Dallas County district court bench in comparison to their proportion in the Dallas County population.

Therefore, the district court did not clearly err in finding that Black district court candidates have been relatively unsuccessful in Dallas County. As we have explained, see supra Part II.B.l.g., in considering the success of minority candidates, courts should make a relative, cautious inquiry. Moreover, the election of a few minority candidates, which is all that has been demonstrated in Dallas County, should not be allowed to foreclose a vote dilution claim. See id.

With regard to the other totality of circumstances factors, the district court rejected testimony offered by the Plaintiffs depicting the Republican Party in Dallas County as a white-dominant slating group. The Plaintiffs do not challenge this finding on appeal.

(iii) Ultimate vote dilution finding. Based on these findings, the district court concluded that, under the totality of the circumstances, Texas’ method of electing district court judges in Dallas county operates to dilute Black voting strength. The district court’s vote dilution finding with regard to Dallas County is supported by strong evidence—evidence raising an inference that the at-large method of electing district judges in Dallas County interacts with the current and past racial climate and voting behavior, such that Blacks have an unequal opportunity to participate in Dallas County district court elections. The district court’s ultimate vote dilution finding in Dallas County, if based on the correct legal standards, see infra Parts IV.B.3. and IV.C., is not clearly erroneous.

c. Ector County

In Ector County, where the Plaintiffs proceed on behalf of Black and Hispanic voters combined, there are four (4) state district judges elected in county-wide elections. In accordance with the law of this circuit, see Campos v. City of Bay-town, 840 F.2d at 1244 (allowing Blacks and Hispanics to proceed as one minority group under Section 2),43 the district court *786found that Black and Hispanic voters combined satisfied the three Gingles factors. The district court further found that the Plaintiffs demonstrated five relevant factors under the totality of circumstances inquiry. Finally, the district court determined that Texas’ method of electing district court judges operates to dilute Hispanic and Black voting strength in Ector County. As explained below, none of the district court’s subsidiary findings with regard to Ector County is clearly erroneous.

(i) Gingles factors. The district court found that Hispanic and Black voters in Ector County were, as a group, sufficiently large and geographically compact to constitute a majority in one single member district. The district court in part based its finding on evidence that Ector County, with a total voting age population of 79,516, has a Black voting age population of 3,255 (4.1%) and a Hispanic voting age population of 14,147 (17.8%). The combined voting age population of Blacks and Hispanics in Ector County, according to evidence introduced at trial, is 17,402, or 21.9% of the total voting age population. The district court further relied on evidence indicating that, because minority residents in Ector County are concentrated in the southwest section of Ector County, it would be possible to draw one single-member district in which voting age minority residents constitute a majority. Such a single-member district could be created, the district court determined, even if non-citizen voting-age Hispanics are eliminated from the estimates of the voting age populations. Finally, the district court relied on the anecdotal testimony of Lawrence Barber, an attorney who has practiced law in Ector and Midland Counties for many years. Barber testified that, “[i]n Ector County, there are geographic or residential areas that are identifiably minority, in other words where most of the Blacks and Hispanics live.” Because-the district court’s finding on this first Gingles factor is plausible in light of the record, we conclude that it is not clearly erroneous.

With regard to the second Gingles factor, the district court found that “Blacks and Hispanics are cohesive as a group in Ector County judicial elections.” Because no Black or Hispanic has ever run for a district court seat in Ector County, the district court focused on the results of four exogenous judicial elections pairing Hispanic candidates against white candidates. A multiple regression analysis of these elections demonstrated that Blacks and Hispanics voted together. Moreover, in three of the four contests large majorities of Blacks and Hispanics voted together for the Hispanic candidate. The district court’s finding of political cohesion, which is supported by the record, is not clearly erroneous.

The district court also determined that the white bloc vote in Ector County is legally significant. In all four of the appellate judicial elections pairing a Hispanic candidate against a white candidate, substantial majorities of whites voted against the Hispanic candidate. And in three of the four elections, the Hispanic candidate lost. The State Defendants’ expert, Dr. Taebel, analyzed five general judicial elections that did not pair a white against a minority candidate. Minority and white voters voted differently in each of these elections, with the minority preferred candidate losing three times. Faced with this conflicting evidence, the district court gave more weight to the elections pitting minority candidates against white candidates. Because we have recognized that such elections are generally more probative in vote dilution cases, we will not set aside the district court's finding of legally significant white bloc voting as clearly erroneous.

(ii) Totality of circumstances factors. The district court found that Plaintiffs had demonstrated several factors that are relevant to the totality of circumstances analy*787sis: a history of discrimination against Blacks and Hispanics in Ector County; racially polarized voting in Ector County judicial elections; voting practices that enhance the opportunity for discrimination; lingering socioeconomic effects of past'discrimination; and a lack of success by minority judicial candidates in Ector County. We address each of these findings in turn and conclude that none is clearly erroneous.

The district court found that Ector County shares Texas’ history of discrimination against Blacks and Hispanics. The State Defendants—wisely—do not challenge this finding. As recently as 1982, Ector County was not only failing to dismantle its dual school system, but also was intentionally increasing the segregation of Black and Hispanic students from white students in its county-wide school district. See United States v. CRUCIAL, 722 F.2d 1182, 1184-85 (5th Cir.1983).

Relying on ecological regression and homogeneous precinct analyses performed by Dr. Brischetto on the four appellate judicial contests pairing a white and Hispanic candidate, the district court found a “strong relationship between race/ethnieity and voting patterns.” According to Dr. Brischetto’s analysis, there was a 78% correlation between race and voting patterns, with race explaining at least 61% of the variance in such patterns. This finding is further supported by the elections analyzed by Dr. Taebel. In the five general judicial elections that did not involve a white candidate running against a minority candidate, minorities and whites voted differently. The district court’s finding of strong racially polarized voting is therefore plausible in light of the record.

The district court also found that Ector County’s majority vote requirement in primaries and its requirement that judges run for a designated post enhance the opportunity for discrimination against minorities. Because of the relatively small number of people in Ector County, however, the district court refused to find that Ector County used “unusually large election districts.” These findings are not challenged on appeal, and we will not set them aside as being clearly erroneous.

The district court further considered the extent to which Blacks and Hispanics in Ector County bear the effects of discrimination in such areas as education, employment, and health. The evidence introduced by the Plaintiffs on this totality of circumstances factor is significant: Some 73% of Ector County’s Black families and some 67.6% of Ector County’s Hispanic families have a yearly income below $20,000. In comparison, only 40% of white families in Ector County have a yearly income level below $20,000. With regard to poverty line statistics, Plaintiffs introduced evidence showing that 30.6% of Black families and 18.3% of Hispanic families—as compared with 5.0% of white families—live below the poverty line. Only 6.1% of Blacks and 2.8% of Hispanics over the age of 25 in Ector County are college graduates. The comparable figure for whites in Ector County is 14.2%. Finally, only 10.6% of working Blacks and 6.9% of working Hispanics. in Ector County hold managerial or professional positions, while 21.3% of working whites hold such positions. See Appendix IV. On the basis of this evidence, as well as evidence indicating that only six Black and Hispanic attorneys reside in Ector County, we conclude that the district court did not clearly err in finding that racial discrimination continues to affect the lives of minorities in Ector County, “including their access to and participation in the democratic system governing this State.”

In analyzing the extent to which minority candidates have been elected to office in Ector County, the district court was again confronted with a sparsity of data. As noted above, no Black or Hispanic has ever run for a district court seat in Ector County. Nonetheless, the district court focused on the fact that, in the four other Ector County judicial elections pairing Hispanic candidates against white candidates, the Hispanic candidate won only once. Given this evidence, the district court was warranted in finding that minority candidates in Ector County are relatively unsuccessful in running for judicial office.

*788Concerning the other totality of circumstances factors, the district court found no evidence of a white-dominated candidate slating process in Ector County and no evidence of overt or subtle appeals to race in Ector County judicial elections. Neither of these findings is clearly erroneous.

(iii) Ultimate vote dilution finding. Based on the evidence discussed above, the district court found that Texas’ method of electing district judges operates in Ector County to dilute the voting strength of Blacks and Hispanics. This ultimate vote dilution finding is amply supported by the structural, behavioral, and racial, features of Ector County’s political landscape. Assuming this finding was based on the appropriate legal standards, see infra Parts IV.B.3. and IV.C., we conclude that it is not clearly erroneous.

d. Harris County

Harris County, with a population of some 2,409,544, is the largest county involved in this lawsuit. With a voting age population of 1,685,024, Harris County elects fifty-nine (59) state district judges in county-wide elections. Proceeding on behalf of Black voters in the county, Plaintiffs offered evidence that, in the district court’s view, was sufficient to satisfy the three Gingles factors. The district court also found, on the basis of the Plaintiffs’ evidence, that five of the totality of circumstances factors are present in Harris County. After considering the totality of the circumstances surrounding Harris County district court elections, the district court found that the Plaintiffs had demonstrated vote dilution. For the following reasons, we conclude that the district court’s findings with regard to Harris County are not clearly erroneous.

(i) Gingles factors. The district court first determined, in accordance with the proper legal standards, that Blacks in Harris County are sufficiently large and geographically compact to constitute a majority in at least one single-member district. This finding was based on evidence that Blacks constitute 18.2% of Harris County’s voting age population, that Blacks primarily reside in the north central, central, and south central sections of Harris County, and that nine single-member districts could be drawn in which Blacks constituted a majority of the voting age population. Because this finding is supported by the record, we conclude that it is not clearly erroneous.44

To establish the second and third Gingles factors, the Plaintiffs offered the testimony of Dr. Engstrom, who analyzed seventeen district court races between 1980 and 1988 pairing white and Black candidates. To determine the racial composition of the various precincts in Harris County, Dr. Engstrom used 1980 Census figures of total Black population by precinct for 1980 races. For his analysis of 1982, 1984, 1986, and 1988 election results, however, Dr. Engstrom used precinct voter registration estimates provided by Dr. Richard Murray, a non-testifying expert.45

*789The district court, relying on Dr. Engstrom’s statistical analysis and on anecdotal evidence, found that Blacks in Harris County are politically cohesive. Based on a statistical analysis of seventeen district court general elections in Harris County in which a Black candidate ran against a white candidate, Dr. Engstrom testified that, in sixteen of these seventeen elections, over 96% of Black voters cast ballots for the same candidates. The Plaintiffs also offered the testimony of Sheila Jackson Lee, a Black resident of Harris County for approximately eleven years. Based on her experience—and particularly her unsuccessful bids in judicial elections in Harris County—Lee testified that Blacks in Harris County are politically cohesive in judicial elections. The district court’s finding that Blacks in Harris County are politically cohesive is amply supported by the record, and accordingly, we will not disturb it on appeal.

As for the third Gingles threshold factor, the district court determined that “the white bloc vote in Harris County is sufficiently strong to generally defeat the choice of the Black community.” In making this determination, the district court relied primarily on Dr. Engstrom’s analysis of the seventeen district court elections in which Black candidates opposed white candidates. This analysis revealed that (a) in sixteen of the seventeen elections, between 61% and 71% of white voters voted against the candidate preferred by Black voters, and (b) in fifteen of the seventeen elections, the candidate preferred by Black voters was defeated. The district court also relied on testimony from several unsuccessful Black district court candidates, who attributed their losses to their inability to attract enough white support. Because the district court applied the correct legal standard for determining the existence of legally significant white bloc voting, and because its finding is plausible in light of the evidence before it, we conclude that its finding is not clearly erroneous.

Pointing to elections in which Hispanic judicial candidates defeated white candidates, the State Defendants argue that the district court’s finding of legally significant white bloc voting is not plausible in light of the record. Citing our decision in Campos v. City of Baytown, they contend that, because Blacks and Hispanics are politically cohesive in Harris County judicial elections, the district court was required to specifically consider and give equal weight to elections pairing white and Hispanic candidates in determining whether white voters will usually defeat the preferred candidate of Black voters. The problem with their argument is that it was not properly raised before the district court and therefore is not properly before us on appeal. Their argument requires a fact finding that Blacks and Hispanics in Harris County are politically cohesive. Yet, neither the State Defendants nor Judge Wood proposed any such finding with respect to Harris County. Indeed, the record in this case reveals that Judge Wood proposed a fact finding, one which was adopted by the State Defendants, that “Blacks and Hispanics together in Harris County do not constitute a politically cohesive minority group.” (emphasis added). Thus, because the State Defendants failed to present their fact-based argument to the district court, we cannot fault the district court for refusing to specifically consider and give equal weight to elections pitting Hispanic judicial candidates against white judicial candidates in Harris County.

(ii) Totality of circumstances factors. Moving to the totality of the circumstances *790surrounding Harris County’s political process, the district court first determined that Blacks in Harris County—like Blacks throughout Texas—were historically the victims of official discrimination that touched their rights to participate in the political process. Neither the State Defendants nor Judge Wood challenges this finding. Accordingly, we will not set it aside as clearly erroneous.

The district court also determined that Blacks and whites vote differently—which under the controlling law amounts to a finding of racially polarized voting. See supra Part II.B.l.b. This finding is amply supported by the Plaintiffs’ evidence that, in sixteen of the seventeen races studied by Dr. Engstrom, over 96% of Black voters cast ballots for the same candidates, while between 61% and 71% of white voters voted against those same candidates. According to Dr. Engstrom’s analysis of those elections, there was a 79% correlation between race and voting patterns, with race explaining 62% of the variance in such voting patterns. This finding is further supported by testimony from Dr. Taebel, the State Defendants’ expert. Dr. Taebel studied twenty-three district court elections in which minorities opposed white candidates, and in all of these elections, Dr. Taebel testified, Blacks and whites voted differently. We conclude, based on this evidence, that the district court’s finding of racially polarized voting is not clearly erroneous.

With regard to voting practices that enhance the opportunity for discrimination against minorities, the district court found three such practices operated in Harris County: the use of a numbered-post system, the use of a majority vote requirement in primary elections, and the use of unusually large election districts. The size of Harris County is particularly relevant. The evidence shows that every judge in Harris County—with its population of nearly 2.5 million—has a larger constituency than twenty governors and forty United States Senators. Because the district court’s finding of enhancement features is plausible in light of the record, we conclude that it is not clearly erroneous.

Contrary to Judge Wood’s suggestion, the district court also found that the effects of Texas’ history of severe racial discrimination are readily apparent in and continue to hinder Harris County’s Black population. The district court expressly found that “[t]he effect of past discrimination against Blacks and Hispanics in areas such as education, employment and health in most of the Counties in question is either well chronicled or undisputed.” This finding is amply supported by the evidence, which shows that some 62% of Black families in Harris County have a yearly income below $20,000, in comparison to 28.2% of white families. The evidence introduced at trial also reveals that 19.5% of Black families—as compared to 8.5% of white families—live below the poverty line in Harris County. In addition, only 11.7% of Harris County Blacks over the age of twenty-five are college graduates, while 28.0% of whites over the age of twenty-five have college degrees. The evidence at trial further demonstrates that, while 13.1% of the members in Harris County’s Black work force hold managerial or professional positions, 30.4% of those in Harris County’s white work force hold such positions. See Appendix IV. Finally, evidence introduced by the State Defendants indicates that Black lawyers, who make up at most 3.8% of the lawyers in Harris County, are significantly underrepresented in the pool of qualified district court candidates. On the basis of this uncontroverted evidence, we conclude that the district court did not clearly err in finding that Harris County Blacks continue to bear the effects of racial discrimination in education and employment—effects which hinder their full participation in the political process.

In considering the extent to which Blacks have been elected to office in Harris County, the district court found that Black district court candidates had been relatively unsuccessful. The district court relied primarily on the low success rate of Black candidates in the seventeen elections in which Black district court candidates opposed white district court candidates. As noted above, the Black district court candidate only won two of these elections. The *791district court’s finding is also supported by evidence demonstrating that, since 1981, Blacks have occupied—at most—three of the fifty-nine district court seats in Harris County (5.1%). By comparison, Blacks constitute 19.7% of Harris County’s total population. Judge Wood challenges the district court’s finding on three grounds.

First, pointing to the “qualified applicant pool” of Black lawyers in Harris County, Judge Wood maintains that Black candidates in Harris County have been successful. In particular, she notes that, while Blacks constitute only 3.8% of the lawyers qualified to run for state district judge in Harris County, Blacks constitute 5.1% of the district judges in Harris County. Thus, she argues that “Blacks are overrepresented on the bench, not underrepresented.” We disagree. As discussed supra Part II.B.l.g., it would be perverse to measure the success of Black candidates by referring to the pool of Black lawyers, especially when the paucity of Black lawyers “is a likely result of a racially discriminatory system.” (quoting McMillan v. Escambia. 748 F.2d at 1045). Accordingly, we hold that the district court correctly refused to measure minority candidate success in Harris County by referring to the qualified applicant pool of eligible Black lawyers.

Judge Wood also argues that, when uncontested elections involving Black candidates are taken into account, the success of minority candidates drastically improves. Since 1978, however, Judge Wood can point to only three uncontested races in which Black district court candidates prevailed. Given the low number of instances in which a Black candidate has been elected in an uncontested district court race, we cannot fault the district court for considering the success of minority candidates in the more likely scenario of a contested election. In any event, we cannot say that the three uncontested elections in Harris County involving Black candidates renders erroneous the district court’s conclusion that, as a general rule, Blacks are unsuccessful in their attempts to be elected to district court judgeships.

Finally, Judge Wood challenges the district court’s refusal to consider two elections from 1978 in which Black candidates won seats on the Harris County district court bench. We note that it is not at all clear that the district court refused to consider these races. In its memorandum opinion, the district court merely cites the fact that, since 1980, only two of seventeen Blacks running for district judge prevailed. Moreover, the district court could reasonably choose to rely on elections since 1980 for determining the success of Black candidates in Harris County.

Thus, we conclude that the district court did not clearly err in finding that Black district court candidates have been relatively unsuccessful. The district court relied on the appropriate measures of Black candidate success—namely, (a) the success rate of Black candidates when they run against white candidates, and (b) the ratio of Black district judges to total Black population. These measures indicated that Black district court candidates generally lost in contested elections and that Blacks are underrepresented on the Harris County district court bench. In view of this evidence, we are not left with the definite and firm conviction that the district court was mistaken in finding a lack of success by Black district court candidates in Harris County.

The district court found no evidence of any slating process and no evidence of racial appeals in Harris County judicial elections. Neither of these findings is challenged on appeal as being clearly erroneous.

(iii) Ultimate vote dilution finding. Based on these subsidiary findings, the district court found that the total circumstances surrounding Harris County elections operate to dilute the voting strength of Blacks. Again, this finding is strongly supported by the evidence of racially polarized voting, the lack of success by Blacks running for district court seats, the history of discrimination against Blacks in Harris County, and the extent to which Blacks continue to suffer the effects of past discrimination. Again, the district court’s *792finding, if based on the appropriate legal standards, see infra Parts IV.B.3. and IV. C., is not clearly erroneous.

e. Jefferson County

In Jefferson County, where (8) state district court judges are elected in countywide elections, the Plaintiffs proceed only on behalf of Black voters. The district court found that Plaintiffs established each of the three Gingles threshold factors and five relevant factors under the totality of circumstances inquiry. The district court further found, on the basis of these factors, that Texas’ method of electing district court judges operates to dilute Black voting strength in Jefferson County. We conclude that the district court’s findings in Jefferson County, being amply supported by the record, are not clearly erroneous.

(i) Gingles factors. With regard to the first Gingles factor, the district court found that, as a group, Black voters in Jefferson County are sufficiently large and geographically compact to constitute a majority in at least one single member district. In making this finding, the district court relied on evidence indicating that (a) Blacks constitute 24.6% of the Jefferson County voting population, (b) Blacks primarily reside in the central and southeastern portions of Jefferson County, and (c) it would be possible to draw two single-member districts in .which voting age Blacks constitute a majority. This finding is plausible in light of the entire record and, accordingly, we will not disturb it on appeal.

In considering the second and third Gingles factors, the district court was confronted with a sparsity of data. Because no Black candidate has ever run for district court judge in Jefferson County, the district court relied on the results of seven other judicial elections, occurring from 1972 to 1988, which pitted Black candidates against white candidates. In particular, the district court relied on: four Democratic primaries involving Justice of the Peace contests; one Democratic primary involving a county court at law post; and two Democratic judicial primary runoffs. The district court also considered results from the 1988 Democratic Presidential primary, in which Jesse Jackson was a candidate.

The State Defendants and, as amicus curiae, certain Jefferson County District Judges attack the district court’s decision to rely on these elections. They argue that the electoral results reviewed by the district court were dated and that the district court should have given more weight to two legislative races in the 1980’s, races in which Black candidates won. We disagree. When a district court is faced with a sparsity of data, it must remain flexible. See Gretna, 834 F.2d at 502 (citing Gingles, 478 U.S. at 57 n. 25, 106 S.Ct. at 2770 n. 25). Here, the district court decided that the results from the eight elections offered by the Plaintiffs—seven of which involved judicial elections—were more probative than the two legislative races relied on by the State Defendants. We cannot say that the district court clearly erred in considering results from judicial races to be more probative than the results from legislative races. Our conclusion is buttressed by the fact that the legislative races relied on by the State Defendants occurred in a district that (1) is not contiguous with Jefferson County, and (2) has a considerably higher percentage of Black voters than Jefferson County.

Based on the eight elections analyzed by the Plaintiffs’ expert, the district court found that Black voters in Jefferson County are politically cohesive. In six of the eight elections, over 70% of Black voters chose the Black candidates; in one election, 51% of Black voters chose the same candidate; and finally, in the other election relied on by the district court, there was a split in the Black vote between two Black candidates. Faced with conflicting evidence on the issue, the district court could reasonably determine that Black voters are politically cohesive in Jefferson County judicial elections. As an appellate court, we will not reweigh the evidence. Accordingly, we conclude that the district court did not clearly err in finding that Plaintiffs satisfied the second Gingles factor with respect to Jefferson County.

*793The district court also determined that, based on the seven judicial elections that were analyzed by the Plaintiffs’ expert, there was legally significant white bloc voting in Jefferson County. In five of the seven judicial elections, the district court noted, Blacks and whites voted differently and the candidate preferred by Blacks lost—even though the Black-preferred candidate received 70% to 93% of the Black vote in those five elections. Because the district court’s determination is plausible in light of the record, we conclude that it is not clearly erroneous.46

(ii) Totality of circumstances factors. As with the other counties involved in this case, the district court found that Jefferson County shares Texas’ history of official discrimination against Blacks. The history of official discrimination against Blacks in Jefferson County was particularly “well chronicled” in Graves v. Barnes, 378 F.Supp. 640 (W.D,Tex.l974) (three-judge court), vacated sub nom. White v. Regester, 422 U.S. 935, 95 S.Ct. 2670, 45 L.Ed.2d 662 (1975) (Graves II). There, the three-member court found that twenty years after the Supreme Court’s decision in Brown v. Board of Education,47 the Beaumont Independent School District continued to bus Black children away from their neighborhood schools and across town to all-Black schools. Id. at 648. Given this history, the State Defendants do not challenge the district court’s finding that Blacks in Jefferson County suffered from official discrimination and that this discrimination touched their rights to participate in the political process.

In considering the extent to which voting in Jefferson County is racially polarized, the district court focused on the eight elections described above. It found, based upon Dr, Brischetto’s regression analysis, a “strong relationship between race and voting patterns in Jefferson County.” This finding is supported by evidence that, in five of the seven judicial elections analyzed, Blacks and whites voted differently. It is further supported by Dr. Brischetto’s calculation of a 66% to 97% correlation between race and voting patterns in the eight elections studied. Accordingly, the district court’s finding of strong racially polarized voting is not clearly erroneous.

The district court also found that two features of the method of electing judges in Jefferson County enhanced the opportunity for racial discrimination against Blacks. The potentially dilutive effects of these features—namely, requiring candidates to run for a specifically numbered post and obtain a majority of votes in the primary elections—have already been established. See swpra Part II.B.1.c. Thus, we will not set aside the district court’s finding with regard to this totality of circumstances factor.

The district court further found that Blacks in Jefferson County continue to bear the effects of past discrimination in areas such as education and employment. The statistics introduced by the Plaintiffs and set forth in Appendix IV support this finding. Over 68% of Jefferson County’s Black families have a yearly income below $20,000, while less than 27% of Jefferson County white families have an income below that level. In addition, 26.4% of Black families in Jefferson County—as compared to 4.0% of white families—live below the poverty line. With regard to education, only 6.2% of Jefferson County Blacks over the age of twenty-five have college degrees. In comparison, 16.2% of Jefferson County whites over the age of twenty-five have such degrees. Finally, only 13% of *794working Blacks in Jefferson County hold managerial or profession positions, while 30% of working whites hold such positions. On the basis of this uncontroverted evidence and evidence indicating that between 14 and 17 Black lawyers reside in Jefferson County, we conclude that the district court did not clearly err in finding that racial discrimination continues to affect the lives of Blacks in Jefferson County.

As to the success of minority candidates in Jefferson County, the district court found that they were relatively unsuccessful. This finding was based in part on evidence that, in the eight elections analyzed, the Black candidate lost in all but one. This finding is also supported by the fact that no Black candidate has ever been elected to a state district court in Jefferson County. In fact, no Black candidate has ever run for a district court judgeship in Jefferson County. According to testimony at trial, the reason Black lawyers do not run for seats on the district court bench is because of the high probability of defeat. Given this testimony and the evidence from other elections involving black judicial candidates, we cannot say that the district court clearly erred in finding that Blacks have been relatively unsuccessful in their attempts to get elected to judicial positions in Jefferson County.

With regard to the other factors that may be relevant to the totality of circumstances analysis, the district court found no evidence of those factors. In particular, the district court determined that there was no evidence of a candidate slating process in Jefferson County and no evidence of racial campaign tactics. These findings are not challenged on appeal.

(iii) Ultimate vote dilution finding. Considering the total circumstances surrounding Jefferson County district court elections, the district court found that Texas’ method of electing district judges operates in that county to dilute the voting strength of Blacks. This finding is strongly supported by evidence indicating that the at-large method of electing district judges in Jefferson County interacts with social and political factors, including the racial climate, to deny Blacks an equal opportunity to participate in the district court political process. Moreover, because the State Defendants introduced no evidence suggesting that partisan politics explain this unequal opportunity to participate, see infra Part IV.B.3., there is no argument that, under the accepted Section 2 framework, the district court applied an incorrect legal standard with respect to its ultimate vote dilution finding. We therefore hold that, unless the district court was required to balance state interests against proven vote dilution, its Section 2 liability finding in Jefferson County is not clearly erroneous.

f. Lubbock County

Plaintiffs proceed on behalf of Black and Hispanic voters combined in Lubbock County. The voters in Lubbock County currently elect five (5) state district judges in county-wide elections and one (1) additional state district judge in a bi-county election with the voters of Crosby County. The district court determined, after hearing evidence-, that the Plaintiffs had satisfied the Gingles threshold factors and that five of the totality of circumstances were present in Lubbock County. The district court further determined, based on these factors, that Texas’ method of electing district court judges operates to dilute Hispanic and Black voting strength in Lubbock County. After reviewing these findings, we conclude that none is clearly erroneous.

(i) Gingles factors. The district court first found, under the correct legal standards, that Black and Hispanic voters in Lubbock County are sufficiently large and geographically compact to constitute a majority in at least one single-member district. In making this finding, the district court relied on evidence that: (a) Blacks and Hispanics combined constitute 21.6% of the voting age population of Lubbock County; (b) Blacks and Hispanics primarily reside in the northeastern, eastern, and southeastern sections of the Lubbock County area; and (c) it would be possible to create one single member district in which voting-age minorities constitute a majority. Because the district court’s finding with regard to the first Gingles factor is plausible in light *795of the record, we will not set it aside as being clearly erroneous.

Because no local judicial contest involved a minority candidate opposing a white candidate, Plaintiffs’ expert Dr. Brischetto relied on six exogenous appellate judicial contests to evaluate the political cohesiveness of Lubbock County’s Blacks and Hispanics and the strength of the white bloc vote in Lubbock County. In particular, Dr. Brischetto analyzed two Texas Supreme Court general elections, two Democratic primary elections, and two Democratic primary runoffs. According to Dr. Brischetto, no local justice of the peace races were included in the analysis because the precincts for those races were not at least as large as a major city.

Based on the analysis conducted by Dr. Brischetto and on anecdotal evidence, the district court concluded that Hispanics and Blacks are politically cohesive in Lubbock County judicial races. Dr. Brischetto’s analysis revealed that Blacks and Hispanics favored the same candidates in each of the elections analyzed and that those candidates received a clear majority of the combined Black and Hispanic vote. Testimony from Maria Luisa Mercado, a Hispanic attorney from Lubbock County, also supports the district court’s finding. She testified that Blacks and Hispanics work together in the county on many significant issues. Because the district court’s finding of political cohesiveness is supported by the record, it is not clearly erroneous.

The State Defendants challenge the district court’s finding of political cohesiveness in Lubbock County by pointing to evidence that, on several occasions, Black and Hispanic candidates opposed one another in Lubbock County races. The Supreme Court has made clear that Section 2 plaintiffs need not prove absolute or perfect cohesion in every race or on every issue; rather, minority political cohesion can be shown by proof that a significant number of minority group members usually vote for the same candidates. Gingles, 478 U.S. at 56, 106 S.Ct. at 2769; see also supra Part II.A.2. The district court’s finding of political cohesiveness is supported by the results from six judicial elections in Lubbock County—results which indicate that over 87% of Black and Hispanic voters usually vote for the same candidate. Accordingly, we conclude that the State Defendants’ complaint about the district court’s finding of political cohesiveness is without merit.

With regard to the third Gingles factor, the district court, again relying on statistical and anecdotal evidence, found that the white bloc vote in Lubbock County judicial elections is legally significant. The district court first focused on the statistical evidence analyzed by Dr. Brischetto, which indicated that (a) minorities and whites voted differently in each of the six elections analyzed, and (b) the candidate preferred by Blacks and Hispanics won only two of the six elections. The district court also focused on the elections analyzed by. Dr. Taebel, the State Defendants’ expert. In the judicial elections analyzed by Dr. Taebel—two appellate court contests and two county court contests—minorities and whites voted differently, and the minority-preferred candidate, lost. Finally, the district court credited the testimony Mrs. Mercado, who testified that, when she unsuccessfully ran for City Council in 1978, she carried all minority precincts and zero white precincts. Faced with this evidence, the district court could reasonably conclude that the white bloc vote in Lubbock County “is sufficiently strong to defeat the minority community’s preferred candidate.”

(ii) Totality of circumstances factors. Surveying the totality of circumstances surrounding district court elections in Lubbock County, the district court found several factors indicative of minority vote dilution. The district court specifically found: a history of discrimination against Blacks and Hispanics in Lubbock County; racially polarized elections; the use of voting practices that enhance the opportunity for discrimination; evidence that minorities in Lubbock County continue to bear the effects of past discrimination; and a lack of success by minority candidates in Lubbock County judicial elections. We review each of these findings in turn, seeking to deter*796mine whether they are plausible in light of the record as a whole.

The district court found that Lubbock County shares Texas’ history of racial discrimination against Blacks and Hispanics. This finding is amply supported by evidence chronicled elsewhere. In Graves II, a three-judge court recognized that authorities in Lubbock County maintained racially and ethnically segregated schools and public facilities until the 1970’s. 378 F.Supp. at 654-55. Indeed, the State Defendants do not challenge the district court’s finding with respect to the history of discrimination in Lubbock County. Instead, they candidly acknowledge the “long, shameful history of racial discrimination in Texas.” Accordingly, we will not set the district court’s finding aside as being clearly erroneous.

The district court further found a “strong relationship between race/ethnicity and voting patterns in Lubbock County.” This finding is supported by evidence that, in each of the six judicial elections that were analyzed by Dr. Brischetto, minorities and whites voted differently. Dr. Brischetto determined, after analyzing these results, that there was an 87% correlation between race and voting patterns and that race explains 76% of the variance in these voting patterns. This finding is further supported by evidence from county court elections that were analyzed by Dr. Taebel, who testified that whites and minorities voted differently in those elections. We are therefore not left with the definite and firm conviction that the district court erred in finding a strong degree of racially polarized voting in Lubbock County judicial elections.

As with the other target counties, the district court determined that certain features of the method by which state district court judges are elected in Lubbock County increase the risk of vote dilution. The two features that are present in Lubbock County district court elections—namely, the use of majority vote requirements in primary elections and the use of a numbered-post system—are also present in each of the other counties. We see no clear error in the district court’s finding with respect to this totality of circumstances factor and therefore decline to set it aside.

We similarly find no clear error in the district court’s determination that Blacks and Hispanics in Lubbock County continue to suffer the effects of past discrimination. As the statistics in Appendix IV demonstrate, Blacks and Hispanics lag unreasonably behind whites in Lubbock County in the areas of education and employment. Further, the evidence reveals that 26.8% of Black families and 23.7% of Hispanic families in Lubbock County live below the poverty line. By contrast, only 5.2% of white families live below the poverty line. Finally, evidence introduced by the State Defendants indicates that only 22 or 23 Black or Hispanic lawyers reside in Lubbock County. Thus, the record amply supports the district court’s finding that racial discrimination continues to affect the lives of Blacks and Hispanics in Lubbock County, “including their access to and participation in the democratic system governing this State.”

In examining the success of minority candidates in Lubbock County, the district court was again confronted^ with evidence that no minority has ever run for such office in Lubbock County. As discussed above, however, in the exogenous elections analyzed by the Plaintiffs’ expert, the Hispanic candidate lost four out of six times. Given the sparsity of data and the ambiguous results from exogenous elections, we cannot say that the district court clearly erred in finding that Black and Hispanic judicial candidates have been relatively unsuccessful in Lubbock County.

As to the other totality of circumstances factors, the district court found no evidence of a white-dominated slating organization and no evidence of racial campaign appeals in judicial elections. There being no challenge to these findings on appeal, we will accept them as being not clearly erroneous.

(iii) Ultimate vote dilution inquiry. Based on its subsidiary findings in Lubbock County, the district court determined that the method by which Texas elects district judges operates to dilute the voting *797strength of Blacks and Hispanics in that county. This finding was based on a number of social and political factors, which together provide strong evidence that Blacks and Hispanics in Lubbock County have an unequal opportunity to participate in the district court political process on account of their race. Assuming that the district court applied the correct legal standard, see infra Parts IV.B.3. and IV.C., we conclude that its ultimate vote dilution finding in Lubbock County is not clearly erroneous.

g. Midland County

Proceeding on behalf of Black and Hispanic voters in Midland County, where three (3) district court judges are elected in county-wide elections, the Plaintiffs offered evidence of the Gingles factors and several of the relevant factors under the totality of circumstances inquiry. The district court found this evidence persuasive and determined that the Plaintiffs satisfied the Gingles threshold inquiry and five of the relevant totality factors. The district court also determined that the Plaintiffs had demonstrated vote dilution in Midland County. Because the district court’s determinations are supported by the record, we hold that they are not clearly erroneous.

(i) Gingles factors. With regard to the first Gingles factor, the district court determined that Blacks and Hispanics in Midland County are sufficiently large and geographically compact to constitute a majority in at least one single member district. This determination was based in part on evidence that Blacks and Hispanics together constitute 19.7% of Midland’s voting age population. This determination was further based on evidence that, because Blacks and Hispanics primarily reside in the northeastern, east central, and southeastern sections of Midland County, it would be possible to draw one district in which minorities constitute a majority of the voting age population—even excluding non-citizen Hispanics who are of voting age. Because this finding is plausible in light of the record, we will not set it aside as being clearly erroneous.

As in Lubbock and Ector counties, no Black or Hispanic has ever run for a district court seat in Midland County. Therefore, the Plaintiffs offered the results from other judicial elections to establish the second and third Gingles factors. In particular, the Plaintiffs offered evidence from three races in Midland County—two exogenous appellate court races and one justice of the peace race that encompassed the entire City of Midland.

Relying on the results of these elections, and the elections analyzed by Dr. Taebel, the district court found that Blacks and Hispanics are politically cohesive in Midland County judicial elections. This finding is amply supported by the record: In the three races analyzed by the Dr. Brischetto, more than 85% of Blacks and Hispanics combined voted for the same candidate. Moreover, as the percentage of minorities in a precinct increased, so did the percentage of support for the minority-preferred candidate. Finally, in the four judicial contests analyzed by Dr. Taebel, minorities voted cohesively. The district court’s finding of political cohesiveness is not, therefore, clearly erroneous.

Again, the State Defendants challenge the district court’s finding of political cohesiveness among Black and Hispanic voters. Pointing to a 1985 Midland city council race, a 1988 Midland school board race, and the 1986 Democratic primary race for the Texas Court of Criminal Appeals—all races in which a Black candidate opposed a Hispanic candidate—the State Defendants argue that “in most instances ..., Black-supported candidates do not receive the support of a majority of Hispanic voters, or vice versa____” We cannot agree. As with Lubbock County, there is evidence in the record sufficient to support a finding that in judicial races in Midland County, a significant number of Blacks and Hispanics usually vote for the same candidate. Under Gingles, such evidence is sufficient to establish minority political cohesiveness. See supra Part II.A.2.

With regard to the third Gingles factor, the district court determined that white bloc voting usually operates to cancel out *798the vote of minorities in Midland County judicial elections. The district court again relied on election results studied by Dr. Brischetto. These results indicated that, in each of the three elections studied, the minority-preferred candidate lost despite the large percentages of combined minority support. The district court also relied on the testimony of Aguilla Watson, a Black woman who ran unsuccessfully for justice of the peace in Midland County in 1986. According to Watson, she won very few white votes and carried only four of the thirty-six precincts, only one of which included any significant amount of white cross-over voting. Finally, the district court relied on the election results analyzed by Dr. Taebel. These results showed that the minority-preferred candidate lost three of four times—despite some cross-over voting by whites. Because the district court’s finding of legally significant white bloc voting is plausible in light of the record as a whole, we conclude that it is not clearly erroneous.

(ii) Totality of circumstances factors. In considering the totality of the circumstances surrounding Midland County’s judicial races, the district court first found that Midland County shares Texas’ history of official discrimination against Blacks and Hispanics. The district court was not clearly erroneous in making this finding. As has been chronicled elsewhere, the history of Midland County includes segregated schools, theaters, pools and restaurants. See League of United Latin Am. Citizens v. Midland Indep. School Dist., 648 F.Supp. 596, 598 (W.D.Tex.1986), aff’d, 812 F.2d 1494, vacated, 829 F.2d 546 (5th Cir. 1987).

Nor will we set aside the district court’s finding of racially polarized voting in Midland County. The regression and homogeneous precinct analyses performed by Dr. Brischetto, and relied on by the district court, show a “strong relationship between race/ethnicity and voting patterns in Midland County.” In the three elections studied by Dr. Brischetto, the correlation between race and voting patterns exceeded 89%, and 79% of the variance in voting patterns was explained by race.48 Moreover, even in the elections analyzed by Dr. Taebel, minorities and whites voted differently. Given this evidence, we cannot say that the district court’s finding of racially polarized voting is clearly erroneous.

The district court’s finding of voting practices that enhance the opportunity for discrimination are similarly supported by the evidence. Midland County, like the other counties at issue, employs what amounts to a numbered-post system and requires candidates to obtain a majority of votes in primary elections. The potentially dilutive effects of these devices having been established elsewhere, see supra Part II.B.1.c., we will not set aside as clearly erroneous the district court’s finding with regard to this aspect of Lubbock County’s district court elections.

The district court was also warranted in finding that Blacks and Hispanics in Midland County continue to suffer the effects of past discrimination. After all, as the statistics in Appendix IV indicate, Blacks and Hispanics in Midland County are disproportionately represented among those who make less than $20,000 and those who live below the poverty line. Moreover, they are disproportionately under represented among those in Midland County who hold college degrees and managerial and professional positions. The record further indicates that only 10 Black and Hispanic lawyers reside in Midland County. The district court's finding on this totality of circumstances factor is therefore not clearly erroneous.

Finally, the district court did not clearly err in finding that minority judicial candidates have been unsuccessful in Midland County. No Black or Hispanic has ever run for a district court seat in Midland County. Moreover, in all of the judicial elections involving minority candidates in *799Midland County, the minority candidate lost. Accordingly, the district court’s determination is amply supported by the record.

With respect to the other totality of circumstances factors, the district court found no supporting evidence. In particular, the district court found no evidence of a white dominated slating process and no evidence of racial campaign tactics. These findings are not challenged on appeal and we therefore will not set them aside as being clearly erroneous.

(iii) Ultimate vote dilution finding. Based on the evidence described above, the district court determined that Texas’ method of electing district judges operates in Midland County to dilute the voting strength of Blacks and Hispanics. This finding is supported by strong evidence of racially polarized voting and the complete lack of success by minority judicial candidates in Midland County. Accordingly, we conclude that the district court’s ultimate vote dilution finding, if based on the appropriate legal standard, see infra Parts IV. B.3. and IV.C., is not clearly erroneous.

h. Tarrant County

In Tarrant County, where twenty-three (23) state district judges are elected in county-wide elections, the Plaintiffs proceed only on behalf of Black voters. The district court found that Plaintiffs had introduced evidence sufficient to establish all three Gingles factors and five of the factors listed as relevant in the totality of circumstances inquiry. The district court also found, based on these factors, that Texas’ method of electing district court judges dilutes Black voting strength in Tar-rant County. As explained below, we conclude that none of the district court’s findings with regard to Tarrant County is clearly erroneous.

(i) Gingles factors. With regard to the first Gingles factor, the district court found that Blacks in Tarrant County are sufficiently large and geographically compact to constitute a majority in at least one single-member district. This finding was based on evidence that (a) Blacks constitute 10.4% of the voting age population in Tar-rant County, (b) Blacks primarily reside in the center of Tarrant County, and (c) it would be possible to draw two single-member districts with a majority of Blacks in the voting age population. The finding is supported by the record and is therefore not clearly erroneous.

The district court also found that the Plaintiffs had satisfied the second Gingles factor with respect to Tarrant County. This finding, that Blacks in Tarrant County are politically cohesive, is not clearly erroneous. In the three general district court elections pairing a Black candidate against a white candidate, a clear majority of Blacks preferred the same candidate. According to the homogeneous precinct analysis done by Dr, Brischetto, more than 89% of Black voters in Tarrant County cast their votes for the same candidate. Moreover, in the 1988 Democratic presidential primary, 98% of Black voters in Tarrant County cast their vote for Jesse Jackson.

As for the third Gingles factor, the district court found that the white bloc vote in Tarrant County is sufficiently strong to usually defeat the Black community’s preferred candidate. The district court based its finding partially on evidence that, in each of the three district court elections that were analyzed, the candidate preferred by Black voters lost, despite the presence of significant white cross-over voting. The district court also relied on testimony from Mary Ellen Hicks, a sitting Black district court judge in Tarrant County who gained her position by being appointed to fill a vacancy. She testified that, the only time she ran against a white candidate in a county-wide election, she could not convince whites to vote for her and, accordingly, she lost the election. Because the district court’s finding of legally significant white bloc voting is plausible in light of the record as a-whole, we conclude that it is not clearly erroneous.49

*800(ii) Totality of circumstances factors.. Under the totality of circumstances inquiry, the district court determined that: there is a history of official discrimination against Blacks in Tarrant County; judicial elections in Tarrant County are racially polarized; certain other features of the election scheme in Tarrant County enhance the opportunity for discrimination; Blacks in Tarrant County continue to suffer the effects of past discrimination; and Black judicial candidates have been relatively unsuccessful in Tarrant County. We review each of these findings in turn and conclude that all of :the findings are supported by the record.

Tarrant County, the district court found, shares Texas’ history of official discrimination against Blacks. This finding is supported by evidence that has been chronicled elsewhere. In particular, white citizens in Tarrant County, in addition to employing racial campaign tactics, opposed desegregation of all-white neighborhoods and community facilities. See Graves II, 378 F.Supp. at 644-45. We cannot say that the district court clearly erred in finding that racial discrimination is a part of Tarrant County's history.

The district court also found that judicial elections in Tarrant County are characterized by some degree of racially polarized voting. This finding is supported by Dr. Brischetto’s regression analysis of the three district court elections in which Black candidates opposed white candidates—an analysis which demonstrated a 80% to 90% correlation between race and voting pattern. This finding is further supported by Dr. Taebel’s analysis of five primary and general judicial elections involving Black candidates. In all of these races, Blacks and whites voted differently, and there was a correlation of 60%-63% between race and voting patterns. Accordingly, we will not set aside the district court’s finding as being clearly erroneous.

The district court further found that three features of the current method of electing judges in Tarrant County enhance the opportunity for discrimination. These features are: the use of unusually large election districts, the use of majority vote requirements in primary elections, and the use of what amounts to a numbered post system. Because all these mechanisms increase the possibility of vote dilution, see supra Part II.B.l.c., the district court’s finding on this totality of circumstances factor is not clearly erroneous.

Looking at other factors that are typically relevant to the totality of circumstances inquiry, the district court determined that Blacks in Tarrant County continue to bear the effects of past discrimination. This finding is amply supported by the data set forth in Appendix IV, which indicates that Blacks in Tarrant County, lag unreasonably behind whites in the areas of income, education, and employment. Moreover, Blacks, who make up only 2.4% of the Tarrant County bar, are substantially underrepresented in terms of the pool of qualified district court candidates. Finally, Blacks are disproportionately represented among those living below the poverty line. Accordingly, we conclude that this district court finding is not clearly erroneous.

We next address the district court’s finding that Blacks have been relatively unsuccessful in bids to obtain a district court seat in Tarrant County. In making this finding, the district court was faced with ambiguous evidence. On one hand, there was evidence that, in the three elections pitting Black district court candidates *801against white candidates, the Black community’s preferred choice won only once. On the other hand, there was evidence that, from 1985 to 1989, the proportion of Black district judges in Tarrant County roughly mirrored the Black population in Tarrant County. This latter evidence, which at first glance looks like consistent, proportional minority electoral success, might appear inconsistent with a vote dilution claim. See supra Part II.B.l.g. (citing Gingles). Upon further probing, however, we conclude that the evidence is not inconsistent with a vote dilution claim. At least two of the minority candidates that were on the district court bench in Tarrant County during this period were appointed and did not gain the seat through a contested election.50 Thus, the district court could reasonably give less weight to this evidence and conclude that Blacks have been relatively unsuccessful in their bid to gain district court seats in Tarrant County.

Finally, with regard to the other totality of circumstances factors, the district court found no evidence of a white-dominated slating process and no evidence of racial campaign tactics. These findings are not challenged on appeal.

(iii) Ultimate vote dilution finding. Based on these subsidiary findings, the district court found that Texas’ method of electing district judges operates in Tarrant County to dilute the voting strength of Blacks. This finding was based on a searching and practical inquiry into the past and present realities of the political landscape in Tarrant County. Moreover, it is supported by the district court’s findings under the totality of circumstances analysis. We therefore conclude that the district court’s ultimate vote dilution finding in Tarrant County, assuming it is based on the correct legal standard, see infra Parts IV.B.3. and IV.C., is not clearly erroneous.

i. Travis County

In Travis County, the last county at issue in this case, thirteen (13) state district judges are elected in county-wide elections. Proceeding only on behalf of Hispanic voters in Travis County, the Plaintiffs offered evidence of the first three Gingles factors and five of the relevant factors under the totality of circumstances inquiry. For the following reasons, we conclude that the district court clearly erred in finding the third Gingles factor—namely, legally significant white bloc voting. Accordingly, the district court’s vote dilution finding with regard to Travis County must be reversed.

The district court, in addressing the first Gingles factor, reluctantly found that Hispanics in Travis County were sufficiently large and geographically compact to constitute a majority of the voting age population in at least one single-member district. This finding was based on evidence that Hispanics constitute 14.4% of Travis County’s voting age population, that a fairly large number of Hispanics reside in the eastern portion of Travis County, and that it was possible to draw one “minimally contiguous” single-member district in which Hispanic voters constituted a majority. Based on this evidence, it is at least debatable whether the district court clearly erred in finding that Plaintiffs satisfied the first Gingles factor; however, we need not decide this issue in view of our holding with respect to the third Gingles factor.

In analyzing the second and third Gingles factors, both the Plaintiffs’ and the State Defendants’ experts relied primarily on Democratic primary elections. Primaries were more important, according to the experts, because Travis County is the second most Democratic county in Texas. Indeed, at the time of trial, all of the district judges in Travis County were Democrats. Moreover, according to Dr. Taebel, most *802would agree that Travis County is “the most liberal county in the state."

Relying on three primary elections analyzed by the Plaintiffs’ expert, the district court found that Hispanics are politically cohesive in Travis County judicial elections. Dr. Brischetto’s analysis of three 1988 Democratic primary races—one district court race and two county-court-at-law races—demonstrated that Hispanics gave from 77% to 95% of their votes to the same candidates. Because this finding is plausible in light of the record, we conclude that it is not clearly erroneous.

Based on the same three primary elections, the district court also found legally significant white bloc voting in Travis County. We conclude that this finding is clearly erroneous. While in many cases it will be sufficient to analyze results from three (or conceivably, even fewer) races,51 the three elections analyzed by Dr. Brischetto do not provide a sufficient basis on which to judge whether the white bloc vote usually defeats the candidate preferred by Hispanics. In one of the three primaries, there was a three-way contest between a white candidate, a Black candidate, and a Hispanic candidate, and the Black candidate won with 23% of the Hispanic vote and 52% of the white vote.52 Moreover, in the district court primary pitting Gallardo, the Hispanic incumbent, against McCown, there was testimony indicating that the reason the Gallardo lost is because he was a “bad judge.” Such testimony, we have noted, see supra note 6, is probative of whether whites will consistently vote against the candidate preferred by the minority group.

Elections not considered by Dr. Brischetto further call into doubt the district court’s finding of legally significant white bloc voting. The State Defendants’ expert, Dr. Taebel, testified that “[w]hites don’t vote consistently as a bloc” in Travis County. He based this conclusion on an analysis of nine primary elections in Travis County with a Hispanic candidate. In those nine elections, the Hispanic candidate won three times, made the runoff twice, and lost four times. Thus, Dr. Taebel was correct when he observed that, Hispanic-preferred candidates in Travis County win some and lose some in Travis County primaries.

In addition, this court has recently acknowledged the willingness of white voters in Austin—who constitute a large majority of the voters in Travis County—to vote for Hispanic-preferred candidates. We stated:

Austin has repeatedly elected black and Mexican-American [city] council members during the past 17 years.... [T]he winning minority candidates frequently received well over fifty percent (50%) of the Anglo vote and were also the preferred candidates of the minorities. Minority candidates have routinely been elected to other posts in Austin and the surrounding Travis County.

Overton, 871 F.2d at 540.

In sum, after reviewing the evidence in the record, we are left with a definite and firm conviction that the district court was mistaken in finding legally significant white bloc voting in Travis County judicial races. While the Hispanic-preferred candidate lost in each of the three 1988 primaries that were analyzed by the Plaintiffs’ expert, we conclude that, in Travis County, the pattern does not establish legally significant white bloc voting—especially given the success of Hispanic-preferred candi*803dates in other judicial and legislative elections. Because Plaintiffs have failed to establish one of the Gingles threshold factors, the district court’s ultimate vote dilution finding cannot stand with respect to Travis County. We therefore do not need to address the other totality factors.

3. Effect of District Court’s Refusal to Consider Partisan Voting Evidence

Our next task in reviewing the district court’s vote dilution findings under the accepted Section 2 framework is to consider the district court’s treatment of partisan voting evidence. In particular, we must determine whether the district court’s apparent refusal to consider this evidence renders erroneous its ultimate vote dilution findings in Bexar, Dallas, Ector, Harris, Lubbock, Midland, and Tarrant counties. After carefully considering the partisan voting evidence offered in this case, we conclude that the district court’s refusal to consider such evidence does not render its vote dilution findings in those counties erroneous.

a. The Partisanship Evidence

At trial, the State Defendants, Judge Wood, and Judge Entz introduced expert testimony purporting to show that factors other than racially motivated voting determine the outcome of Texas district court elections. According to these experts—Dr. Taebel, Judge Davidson, and Dr. Champagne—the determinative “other factor” in such elections is not the race of the voters or candidates, but the partisan affiliation of the voters and candidates. The dynamics of this partisanship factor, in the view of these experts, play out as follows: In the counties involved in this case, most of which are large, there are many district court races, which appear low on the ballot. These races generate only minimal voter interest. When voters prepare to vote in district court races, there is a substantial anonymity factor—meaning that voters are generally unaware of the name or race of the candidate. Consequently, most of the voters are uninformed about district court candidates and either fail to vote in such races or, reflexively, vote along partisan lines.

(i) History of partisan politics in Texas. Before describing how partisan politics allegedly explains voting behavior in Texas district court elections, we shall briefly outline the history of partisan politics in Texas. According to evidence introduced at trial, Texas was, until recently, a one-party state in which the majority of voters called themselves Democrats. Now, however, Texas appears to be a two-party state or a state that in some areas leans strongly toward the Republican Party.

Dr. Taebel testified about the evolution of Texas partisan politics in general. He stated that, “[bjack in the ’50’s and ’60’s, the Democratic Party dominated Texas politics.” During this time period, Dr. Taebel continued, Texas “had a one-party state, and therefore, all questions evolving around polities focused on the Primary elections.” Dr. Taebel further stated that, “since the election of [Governor] Clements in about 1978, we have seen Texas become a very competitive two-party state.” Because of this transformation, Dr. Taebel concluded, the focus is more extensively on general elections.

Dr. Champagne described one area of Texas—namely, Dallas County—which has undergone a dramatic shift in partisan politics. He testified that, during the 1980’s, Dallas County went from “a one party Democratic system at the District Court level to essentially a one party Republican system.” In making this conclusion, Dr. Champagne relied primarily on graphs indicating that: (1) from 1977 to 1988, the percentage of Republican district judges in Dallas County increased from 0% to almost 100%; and (2) from 1976 to 1988, Republican district court candidates in Dallas County went from winning 0% of contested elections to winning almost 100% of such elections. The last time a Republican district court candidate lost a contested seat in a Dallas County district court election, according to Dr. Champagne, was in 1978.

(ii) How partisan politics operate in Texas district court elections. The expert testimony on partisanship revealed that *804partisan voting patterns are not identical in the seven counties still at issue in this case. In four of the counties, Dallas, Ector, Lubbock, and Midland counties, voters demonstrate a consistent preference for Republican candidates. In other counties, the partisan voting pattern appears to be “in transition.” And finally, in Harris County, there is evidence of a competitive, two-party system.

Dr. Taebel first testified about Dallas County, concluding that Republican candidates received strong support from voters. Based on his analysis of twenty-three elections, Dr. Taebel said, “what we see in Dallas County is very strong and solid support for Republican candidates.” He further testified that, in his view, these elections “clearly illustrate partisanship is the best way to describe what is going on in Dallas County.” When asked whether white voters would vote consistently as a bloc so as to usually defeat the Black-preferred candidate, Dr. Taebel responded:

There are two blocs of White voters. As I said, one is the Democrat White bloc voters and one is the Republican bloc White voters. Both of those two blocs vote consistently for the nominees of their parties. It just so happens in Dallas County there is more Republican voters than Democratic voters and the Republicans win all the time, or almost all the time.

Dr. Champagne also testified about Dallas County . partisan voting patterns. Based on his analysis of district court elections from 1980 to 1988, he concluded that the race of the candidate has no impact on a candidate’s ability to win elections. During that time period, there were nine primary and general elections combined in which a Black candidate opposed a white candidate, and the Black candidate won four of those contests. According to Dr. Champagne, when a Black candidate runs as a Democrat, he or she loses, but when a Black candidate runs as a Republican, he or she wins. Thus, Dr. Champagne concluded that the one-party nature of Dallas County’s political landscape explains elections results: “Republicans win, Democrats, lose.” -

In concluding that the race of the candidate has nothing to do with his or her ability to win, Dr. Champagne also relied on a random survey of 1,000 Dallas County registered voters. This survey, which was designed to determine voter awareness of Dallas County public officials and judges, asked respondents the following questions: whether they voted in certain elections; whether they recognized the names of certain public officials and judges; whether they knew the position of certain public officials and judges; whether they knew the race/ethnic origin of certain public officials and judges; and finally, what factors influenced their decisions in judicial races. The. results of this survey indicate that (a) registered voters could not generally recognize the names of district court judges, (b) registered voters did not know the race/ethnic origin of district judges, and (c) the reasons volunteered for voting decisions in district court races did not include the race of the candidate, but factors such as judicial integrity and honesty. This evidence, according to Dr. Champagne, supports his conclusion that Dallas County district court elections are characterized more by partisan voting than by racially motivated voting.

The other Republican counties, according to the testimony of Dr. Taebel, are Ector County, Lubbock County, and Midland County. With regard to Ector County, Dr. Taebel stated that “what we are seeing is that White Republicans constitute a, constitute a significant bloc vote,” such that “White Republicans win.” With regard to Lubbock County, Dr. Taebel observed that voters “tend to vote Republican” and that, in judicial races, there is “very extensive partisan voting patterns.” Finally, Dr. Taebel characterized Midland County as “Republican territory” and, when asked what best describes election results in that county, Dr. Taebel responded that partisanship prevails: “Republicans beat Democrats.”

Dr. Taebel also testified about two counties that appear to be undergoing a transition from a strictly. Democratic county to a *805Republican County: Bexar County and Tar-rant County. Bexar County, Dr. Taebel noted, has the smallest number of white Democratic voters, approximately 12%. Dr. Taebel then concluded, relying on results from twenty-eight elections, that “voting essentially follows party lines” in Bexar County. When asked specifically about district court general elections in Bexar County, Dr. Taebel responded that, “what you see in the General elections is that voters basically revert back to their partisan mode and vote for the candidate of their Party, irrespective of race or any other factors.” With regard to Tarrant County, Dr. Taebel stated that, in “the early part of 1980’s, it was primarily Democratic, or slightly Democratic, but in 1986, there was a major shift in Tarrant County toward the Republican Party.” Based on this recent shift, Dr. Taebel concluded that the “party preference of the voters” best describes why certain candidates win and others do not.

Finally two experts, Dr. Taebel and Judge Mark Davidson, testified about how partisan politics operate in Harris County. According to Dr. Taebel, Harris County “is probably the most competitive county we looked at.” Because of this competitiveness, Dr. Taebel explained, it is difficult to predict who will win elections. He noted, however, that “since 1980, only 75 percent of the voters actually vote in judicial races.”

The extent of straight-ticket voting in Harris County deserves special mention. Dr. Taebel testified that, of the voters who vote in Harris County judicial elections, almost 90% cast straight ticket votes. Judge Davidson testified to similar statistics. Such testimony is not surprising, since in any given election, voters in Harris County may be presented with as many as thirty (30) district court races. Further, according to both Dr. Taebel and Judge Davidson, approximately 45% of the straight ticket voters vote Democratic, while the other 45% of such voters vote Republican. Accordingly, 10% of the voters in Harris County,- most of whom are white, end up deciding who wins elections. And, although these “discretionary” voters appear to vote in a variety ways, Judge Davidson conceded that it is reasonable to conclude that white judicial candidates in Harris County receive more of the so-called discretionary vote than Black candidates.

(iii) The limitations of the partisanship evidence. The experts also testified about the limitations of relying on partisan voting patterns. Dr. Champagne, on cross-examination, stated that, based on his studies of Dallas County elections, he did not know if voters equate “voting Republican” with voting safely white. Moreover, he conceded the possibility that partisanship and race are the same thing in Dallas County. Judge Davidson made a similar concession with regard to Harris County. He stated that there might be reasons other than bar polls, incumbency, and the type of campaign run by a candidate that influence the decisions of discretionary voters in Harris County.

Dr. Taebel, however, was the most forthright about the limitations of statistical analyses of partisan voting patterns. First, he stated that there is a strong identification between being Black or Hispanic and participating in the Democratic Party. Dr. Taebel also recognized at several points in his testimony that “you are always going to find” racially polarized voting. Finally, Dr. Taebel emphasized that his statistical analysis of partisan voting patterns did not, and could not, explain why voters vote along party lines. He stated:

I was basically analyzing how one can realistically describe the political process that exists in the nine different counties. My intent was not to find out the precise reasons. why a candidate won or lost. My intent was to describe the overriding feature of partisan politics in nine counties. I think if you try, if I tried to get involved in campaign expenditures and incumbency, ratings by the Bar Association, it would be an impossible task to do.

(iv) Summation of partisanship evidence. The partisan voting evidence offered by the State Defendants, Judge Wood, and Judge Entz describes several features of Texas district court elections. *806First, it demonstrates that Texas district court elections are at least nominally partisan, in that the candidates are identified as being affiliated with either the Republican of Democratic Party. Second, the evidence demonstrates that, as a general rule, district court races appear at the bottom of a long ballot, and they are generally low profile elections. Third, because the races are low profile, the voters are generally unaware of either the names or the racial background of the various candidates. And finally, because the voters are relatively uninformed, those who actually vote tend to reflexively pull the straight ticket lever.

Thus, we are presented with evidence of predictable partisan voting patterns, evidence which purports to demonstrate that “partisanship” and not racially polarized voting best describes Texas district court elections. We are also told, however, that these predictable partisan voting patterns are not the result of an electorate informed about the views of the judicial candidates, but the result of voters being generally uninformed about' the names and racial background of the various candidates. As Dr. Taebel recognized, where ballots are long, “it becomes very difficult for voters to sort out each of the individual candidates. So straight party voting tends to take over in those types of elections.”

b. The District Court’s Treatment of the Evidence

It is not clear exactly how the district court treated the partisanship evidence offered by the State Defendants, Judge Wood, and Judge Entz. In at least one part of its memorandum opinion, the district court appeared to credit the partisanship testimony. In other parts, however the district court discredited the testimony.

And finally, in at least two instances, the district court stated that the partisanship evidence was legally irrelevant.

First, we note that the district court admitted the partisanship evidence and apparently credited some of the testimony. At the beginning of its memorandum opinion, the district court acknowledged the partisan features of Texas district judge elections. It stated:

It was brought to the Court’s attention that perhaps a majority of voters in a General Election, and for than matter, in Primary Elections, have no idea of the qualification of a judge for whom they vote. Their vote is cast because a straight ticket is being cast, and a straight ticket includes judicial nominees from a particular party.

At other places in its memorandum opinion, the district court appeared to discredit the partisanship testimony. With regard to the evidence of straight ticket voting in Dallas County, the district court referred to a statement by one of the Plaintiffs’ experts that “something other than just straight party voting is going on.” Moreover, in response to the State Defendants’ argument that partisan preference best accounts for electoral success in the various counties, the district court found that the totality of circumstances factors nonetheless “point to the continual effects of historical discrimination hindering the ability of minorities to participate in the political process.”

Ultimately, however, the district court labelled the partisanship evidence legally irrelevant.53 With regard to Judge Davidson’s testimony in Harris County, the district court stated that, although it found Judge Davidson to be a.credible witness, his testimony was legally irrelevant. And, *807in its conclusions of law, the district court stated that, “under the controlling law, party affiliation, straight party ticket voting and campaign factors do not constitute legally competent evidence.”

c. The District Court’s Error

In concluding that the partisan features of Texas district court elections are legally irrelevant, the district court apparently relied on a part of Justice Brennan’s opinion in Cingles, a part in which only a plurality of Justices joined. In particular, the district court relied on Justice Brennan’s statement that, “under the ‘results test’ of § 2, only the correlation between the race of voter and selection of certain candidates, not the causes of the correlation, matters.” 478 U.S. at 63, 106 S.Ct. at 2773. Thus, the district court rejected the partisanship evidence because it thought that such evidence was being offered to establish why minorities and whites vote differently. For the following reasons, we conclude that the district court erred in relying on the language from Justice Brennan’s opinion and refusing to consider the partisanship evidence.

First, we think that, even under Justice Brennan’s Gingles opinion, the partisanship evidence was relevant. After all, in that opinion, Justice Brennan recognized that “the question whether the political processes are equally open depends upon a searching practical evaluation of the past and present reality, ... and on a functional view of the political process.” 478 U.S. at 45, 106 S.Ct. at 2764 (internal quotes and citations omitted). Moreover, even the language relied on by the district court only purported to exclude evidence explaining why voters vote the way they do. In this case, Dr. Taebel expressly disclaimed any intent to explain why voters vote the way they do.

Second, the language relied on by the district court was not embraced by a majority of the Justices. Justice White, in a separate concurrence, disagreed with Brennan’s assumption that the race of the candidate does not matter in analyzing racially polarized voting patterns under Section 2. See id. at 83, 106 S.Ct. at 2782. Justice O’Connor was more specific in her disapproval of Brennan’s statement that the causes of racially polarized voting are irrelevant. In her view, “evidence that the divergent racial voting patterns may be explained in part by causes other than race, such as an underlying divergence in the interests of minority and white voters,” might in some cases “affect the overall vote dilution inquiry.” Id. at 100, 106 S.Ct. at 2792 (O’Connor, J., joined by three Justices, concurring in the judgment). Thus, in refusing to consider the partisan voting evidence, the district court appears to have erected, in Justice O’Connor’s words, 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’Connor, J., concurring in the judgment).

Accordingly, we hold that the district court should have considered the partisan features of Texas district court elections, including partisan voting patterns, in the totality of circumstances inquiry under the accepted analytical framework. To the extent the partisanship evidence purported to demonstrate a lack of racial animus towards specific district court candidates, it was unquestionably relevant to the totality of circumstances inquiry. See supra Part II.B.2. More important, the partisan features of Texas district court elections are part of the “context of the total circumstances of the local electoral process.... ” Id. (quoting S.Rep. at 16,1982 U.S.C.C.A.N. at 193). Such features are also relevant to the overall vote dilution inquiry of whether members of a minority group have an unequal opportunity to participate on account of race. Thus, the partisanship evidence should have been considered in the totality of circumstances inquiry—as part of the “searching practical evaluation of the ‘past and present reality’ ” in the counties at issue. See supra Part II.B.2. (quoting S.Rep. at 30, 1982 U.S.C.C.A.N. at 208).

d. The Effect of the Error; Whitcomb Considered

Having held that the district court erred in refusing to consider the partisan fea*808tures of Texas district court elections, we must now determine how that error affects the district court's ultimate vote dilution findings in the seven counties for which such evidence was offered. We have already éxplained that where, as here, “a finding of fact is based on the misapprehension of governing legal standards, it loses the insulation of the ‘clearly erroneous’ rule.” See supra Part IV.A. (quoting Armco Burglar Alarm, 693 F.2d at 1162). We have also explained, however, that where, as here, a district court’s finding loses the protection of the clearly erroneous rule, “a remand is the proper course unless the record permits only one resolution of the factual issue.” See supra Part IV.A. (quoting Pullman-Standard v. Swint, 456 U.S. at 292, 102 S.Ct. at 1792). Thus, as an appellate court, we must decide (a) whether to reverse and remand the case to the district court with instructions to consider the partisanship evidence; (b) whether to reverse and render judgment for the State Defendants; or (c) whether to affirm the district court’s vote dilution findings. Of course, we cannot pursue one of the latter two options unless, in light of the partisanship evidence, "the record permits only one resolution” of the ultimate vote dilution inquiry in the seven counties at issue.

The State Defendants, Judge Wood, and Judge Entz argue that we need not remand this case for further factfinding by the district court. They contend that, if we consider the partisan features of Texas district court elections, the record permits only one resolution of the ultimate vote dilution inquiry—a resolution in which we reverse the district court’s ultimate vote dilution findings and render judgment in favor of the State Defendants. Citing Whitcomb v. Chavis, they reason that the evidence of partisan voting patterns preeludes any finding of vote dilution under Section 2. We disagree.

In Whitcomb v. Chavis, 403 U.S. 124, 91 S.Ct. 1858, 29 L.Ed.2d 363 (1971), a case that continues to have relevance under the amended Section 2, see S.Rep. at 20-23, 32-33, 1982 U.S.C.C.A.N. at 197-201, 209-211, the Court rejected a challenge by Black voters in Marion County, Indiana to an at-large election scheme. In rejecting this challenge, the Court reasoned that interest group politics untainted by a racial minority’s unequal opportunity to participate in the political process will not support a vote dilution claim. The Court, after recognizing that there was no evidence of an intent to discriminate against Blacks in Marion County, stated:

Nor does the fact that the number of [Black] residents who were legislators was not in proportion to [the Black] population satisfactorily prove invidious discrimination absent evidence and findings that [Black] residents had less opportunity than did other Marion County residents to participate in the political processes and elect legislators of their choice.

Id. at 149, 91 S.Ct. at 1872. The Court in Whitcomb then observed that there was no evidence that Blacks in Marion County had an unequal opportunity to participate in the political process.54 Indeed, the plaintiffs in Whitcomb conceded that Black voters had an equal opportunity to participate in the political process. Because Blacks would have had “no justifiable complaints about representation” if the Democratic candidates had won all of the elections, the Court concluded that “the failure of [Blacks] to have legislative seats in proportion to [their] population emerges more as a function of losing elections than of built-in bias against poor [Blacks].” Id. at 152, 153, 91 S.Ct. at 1873. Ultimately, the Court in Whitcomb stated:

*809On the record before us plaintiffs’ position comes to this: that although they have equal opportunity to participate in and influence the selection of candidates and legislators, and although [Blacks] vote[ ] predominantly Democratic and that party slates candidates satisfactory to [them], invidious discrimination nevertheless results when [Blacks], along with all other Democrats, suffer[ ] the disaster of losing too many elections .... The mere fact that one interest group or another concerned with the outcome of Marion County elections has found itself outvoted and without legislative seats of its own provides no basis for invoking constitutional remedies where, as here, there is no indication that this segment of the population is being denied access to the political system.

403 U.S. at 153-, 155, 91 S.Ct.- at 1874, 1875 (emphasis added)

The State Defendants, Judge Wood, and Judge Entz read Whitcomb too broadly. The Court in Whitcomb did not hold that, by itself, statistical evidence of strong partisan voting patterns forecloses a vote dilution claim. Nor does the Court’s decision in Whitcomb preclude a finding of vote dilution absent evidence that partisan voters are motivated by racial animus toward particular candidates. Rather, Whit-comb stands for a much narrower proposition: Where there is evidence of partisan voting or interest group politics and no evidence that members of the minority group have an unequal opportunity to participate in the political process on account of race or color, the minority group’s vote dilution claim will fail.55

*810Thus, Whitcomb must be read in the context of the community in which the vote dilution claim was being brought—a community in which the Black ghetto residents conceded that they had an equal opportunity to participate in the political process. The Black ghetto residents of Marion County, Indiana presented evidence that they were politically cohesive, evidence that they failed to elect representatives in proportion to their population, and evidence that they currently suffered from socioeconomic disparities. The plaintiffs in Whitcomb did not, however, at least according to the Supreme Court opinion and the district court’s opinion, see Chavis v. Whitcomb, 305 F.Supp. 1364 (S.D.Ind.1969), present evidence of a history of official discrimination touching the rights of Blacks to vote and participate in the political process. Nor did the Blacks in Whitcomb present evidence that the socioeconomic disparities from which they suffered were the result of, or on account of, their race or color.

To suggest that Whitcomb would have been decided the same had the lawsuit challenged at-large election practices in, for example, Dallas County, Texas, is to ignore the official discrimination in voting that gave rise to the passage of the Voting Rights Act. As the Senate Report accompanying the amended Section 2 stated, “the purpose of the Voting Rights Act was ‘not only to correct an active history of discrimination, the denying to [Blacks] of the right to register and vote, but also to deal with the accumulation of discrimination.’ ” S.Rep. at 5, 1982 U.S.C.C.A.N. at 182. The Voting Rights Act was further designed, according to the Senate Report, to deal with the new and “broad array of dilution schemes [that were being] employed to cancel the impact of the new black vote.” Id. at 6, 1982 U.S.C.C.A.N. at 183. The state of Texas and the various counties at issue in this case, unlike—as far as we can tell— the state of Indiana and Marion County, have impeded the opportunity of Blacks and Hispanics to participate fully in the political process. See, e.g., Texas v. United States, 384 U.S. 155, 86 S.Ct. 1383, 16 L.Ed.2d 434 (1966) (invalidating Texas’ poll tax); Terry v. Adams, 345 U.S. 461, 73 S.Ct. 809, 97 L.Ed. 1152 (1953) (invalidating the use of racially discriminatory “pre-primary” elections by private organization associated with the Texas Democratic party); Beare v. Smith, 321 F.Supp. 1100 (invalidating Texas law requiring voters to register annually because it deprived minorities of the right to participate in the electoral process on an equal basis), aff’d sub nom., Beare v. Briscoe, 498 F.2d 244 (5th Cir. 1974). And, as the State Defendants concede in their brief on appeal, the “long, shameful history" of racial discrimination in Texas “finds its present day reflection in the disproportionately lower socioeconomic status of Blacks and Hispanics as groups, as compared to Anglos as a group.”

In this case, we are presented only with statistical evidence that correlates the party of candidates with voting behavior. We are not presented with evidence that partisan politics explains minority voters’ unequal opportunity to participate in the political process any better than, or more importantly, to the exclusion of, racial factors. Dr. Taebel specifically stated that his "intent was not to find out the precise reasons why a candidate won or lost.” He further suggested that “it would be an impossible task” to attempt, through a multivariate analysis, to determine the relative effect of the various factors that could influence voters’ decisions.56

The other witnesses also conceded that they ultimately could not exclude the possibility that race was at work. Judge David-, son, who testified about discretionary voters in Harris County, conceded that reasons other than bar polls, incumbency, and *811the type of campaign run by a candidate might influence the decisions of those discretionary voters. Dr. Champagne, who testified about partisan voting in Dallas County, admitted that race and partisanship could be the same thing.

Moreover, in several of the counties involved, there was evidence that racial factors continue to play a role in judicial elections. In Dallas County, for example, there were two instances of racial campaign appeals in recent elections. And Dr. Engstrom, the Plaintiffs’ expert, opined that general district court elections in Harris County are “more of a filter for Black candidates than Democratic candidates.” This testimony was based on evidence that, from 1980 to 1988, 52% of all white Democratic district court candidates won contested general elections, while only 12.5% of all Black Democratic district court candidates won contested elections during the same time period. Finally, Dr. Brischetto, the other expert employed by the Plaintiffs, stated that, in his studies of voting behavior, “ethnicity is still a major, the major determinant of how people vote.”

Even if we assume that partisan politics best describes voting behavior in the seven counties at issue, we are still faced with substantial evidence that minority voters in those counties have an unequal opportunity to participate in the political process on account of their race or color. In each of the counties at issue—in Bexar County, Dallas County, Ector County, Harris County, Lubbock County, Midland County, and Tarrant County—there is evidence of a history of discrimination against minorities that violated their rights to participate in the political process, evidence of strong to severe racially polarized voting,57 evidence that minorities continue to bear the effects of past discrimination, and evidence that minority candidates, especially those who are sponsored by the minority community, are unsuccessful in their bids to hold judicial offices like the district court bench. See supra Part IV.B.2. All of these factors point to an unequal opportunity to participate in the political process58 and, taken together, raise an inference that race or color is at least partly responsible.

Accordingly, we hold that the Supreme Court’s decision in Whitcomb does not preclude a finding of vote dilution in the seven counties at issue. The partisanship evidence that was presented to the district court simply does not sufficiently cancel out or negate the other factors that raise an inference that minority voters have an unequal opportunity to participate in the *812district court political process. Nor does the evidence negate the inference that race or color is at work in the various counties at issue.

The crucial question thus becomes, not whether to reverse and render judgment for the State Defendants, but rather, whether to remand the case to the district court for consideration of the partisan voting evidence. This question, in turn, depends on whether, despite the district court’s failure to consider the partisanship evidence, the record nonetheless requires us to affirm the district court’s ultimate vote dilution findings under the accepted Section 2 framework. In answering this question, we must determine specifically whether the partisanship evidence somehow undercuts or minimizes the district court’s vote dilution findings in each of the seven counties for which such evidence was offered.

In our view, the partisanship evidence does not undercut or minimize the district court’s findings that, under the totality of the circumstances, minority voters in the seven counties at issue have an unequal opportunity to participate in the political process on account of race. The partisanship evidence does not purport to establish that minority voters are able to participate in the political process and elect representative of their choice. Rather, the evidence purports to show that, because voters do not know the races or names of the candidates, voters do not vote with specific racial animus towards particular candidates. Such statistical evidence of partisan voting does not, and could not, detract from evidence suggesting that minority voters’ unequal opportunity to participate in the political. process is linked to race.

Moreover, not only do the partisan features of Texas district pourt elections fail to undercut or minimize the district court’s ultimate vote dilution findings, they also serve to reinforce or enhance the possibility of vote dilution. After all, the evidence of partisan voting patterns in this case is readily distinguishable from the kind of “partisan politics” governing races at the top of the ballot. As already discussed, the partisan voting patterns in this case result from several unique features of Texas district court races: There are many district court races in most of the counties at issue, and as a general rule, these races appear low on the ballot. There is also, as a general rule, low voter interest in such races and a resulting low level of voter knowledge about the qualifications, the race, and even the name of candidates. This anonymity factor combines with the mechanism in Texas allowing straight ticket voting to produce a substantial correlation between voting and partisan preference.59

When viewed in this light, the so-called partisanship evidence does not rebut or minimize the other strong evidence that minorities have an unequal opportunity to participate in the district court political process on account of race or color. Instead, it is just another feature of large, multimember election districts—a feature that reinforces the minority group’s unequal opportunity to participate. Because of the low profile nature of district court races, the high degree of anonymity involved in such races, and the resulting high incidence of straight-ticket voting, district court candidates who are sponsored by the minority community in the various counties at issue face an unusually high hurdle in winning elections: They generally cannot, due to their lack of resources, mount a campaign in which they could gain county-wide name recognition and thereby overcome the anonymity factor which dominates district court elections.

Accordingly, we conclude that, even though the district court erred in refusing *813to consider the partisan features of Texas district court elections, the record nonetheless permits only one resolution of the ultimate vote dilution inquiry in Bexar, Dallas, Ector, Harris, Lubbock, Midland, and Tar-rant counties. The partisanship evidence demonstrates only that voters in district court elections, because they are unaware of the race of the candidates, could not be motivated by specific racial animus towards those candidates. The evidence does not demonstrate that voters, knowing the race of the candidates involved, nonetheless vote without regard to race. Thus, this “partisanship” evidence does not demonstrate the absence of general racial animus in the electorate, and it does not rebut the district court’s other findings under the Gingles and totality of circumstances inquiry—findings which demonstrate an unequal opportunity to participate in the political process on account of race or color.

As we have explained, see supra Part II.C., the ultimate vote dilution inquiry does not focus narrowly on whether there is current racial animus in the electorate. Instead, the ultimate vote dilution inquiry asks whether, in the context of the particular electoral process, minorities have an unequal opportunity to participate on account of race or color. To demonstrate a lack of opportunity “on account of race or color,” minority voters challenging an at-large election scheme need not show racial animus in the electorate, but may rely on factors in the totality of circumstances inquiry to raise an inference that race is at work. While the presence of general racial animus is clearly significant to the question of whether minority voters’ unequal opportunity to participate in the political process is on account of race or color, the absence of general racial animus, as we have already explained, see supra Part II.B.2, will not weigh heavily against the minority group. More importantly, where, as here, evidence of partisan voting patterns suggests that the electorate, being generally uninformed, is not motivated by specific racial animus towards candidates, such evidence will not defeat or undercut a vote dilution claim—so long as the minority group demonstrates that its members have an unequal opportunity to participate in the political process and that this unequal opportunity is linked to race or color.

4. Summation

In sum, we hold that the district court did not err in finding, under the accepted Section 2 framework, that Plaintiffs proved their vote dilution claims in eight of the nine target counties. Because no partisanship argument was raised with regard to the district court’s ultimate vote dilution finding in Jefferson County, we affirm that finding under the clearly erroneous standard. With regard to the district court’s ultimate vote dilution findings in Bexar, Dallas, Ector, Harris, Midland, Lubbock, and Tarrant counties, we affirm the district court’s findings even though they are not protected by the clearly erroneous rule. Finally, because the Plaintiffs did not satisfy the Gingles threshold inquiry in Travis County, we reverse the district court’s ultimate vote dilution finding with regard to that county.

C. Review of the District Court’s Vote Dilution Findings Under the Proposed Balancing Framework

We now proceed to review, under the proposed balancing framework set forth in Part II of this opinion, the district court’s vote dilution findings in eight of the nine counties against the asserted state interests. Of course, the district court did not apply any kind of balancing framework, because back in 1989, no such framework had ever been suggested or contemplated. Thus, if a balancing framework is now required, the district court’s ultimate vote dilution findings in each of the counties lose the insulation of the clearly erroneous standard of review.

Under such a balancing framework, were it not for the weight of the state interests in this case, we would be required to remand this case to the district court with instructions to balance state interests against proven vote dilution. Given the uncompelling or insubstantial nature of the threatened state interests in this case, *814see supra Part III.D.3., however, this record only permits one result.60 The significant amount of vote dilution in Bexar, Dallas, Ector, Harris, Jefferson, Midland, Lubbock, and Tarrant counties outweighs Texas' legitimate state interests in preserving the administrative advantages of the current method of electing district judges, in allowing judges to specialize, and in “linking” a district court’s venue with its electoral base.

The vote dilution findings in each of the eight counties, as we have already explained, find strong support in the record. In each of those counties, there is substantial evidence that the voting strength of Blacks, Hispanics, or Blacks and Hispanics combined, is cancelled out by the use of an at-large election scheme. For example, in every county, elections are polarized along racial lines, and minority candidates who are sponsored by the minority community have been suspiciously unsuccessful in getting elected to district court seats. In every county, minorities who have suffered from a history of official discrimination continue to lag unreasonably behind whites in the areas of income, education, and employment.

Accordingly, we conclude that even if the district court was required to apply a balancing framework in making its Section 2 liability findings, we would still affirm those findings in Bexar, Dallas, Ector, Harris, Jefferson, Lubbock, Midland, and Tar-rant counties. The Plaintiffs proved substantial vote dilution in each of the eight counties, while the state interests asserted by the state of Texas are little more than tenuous. Thus, the record permits only one resolution even under the proposed balancing framework: the state interests do not outweigh this proven vote dilution.

V. REMEDY

Because we have affirmed the district court’s Section 2 liability findings in eight of the nine counties, we must now address questions about the appropriate remedy. We address first the question of which unit of government has priority in devising a remedial plan. We then address the question of the criteria that must govern any remedy imposed by the district court. And finally, we review the interim plan already devised by the district court.

With regard to the first question, it is clear that the affected political subdivision should be given the first opportunity to develop a remedial plan for Section 2 violations. See Westwego III, 946 F.2d at 1124; Cook v. Luckett, 735 F.2d 912, 918-19 (5th Cir.1984); see also East Jefferson Coalition for Leadership & Dev. v. Parish of Jefferson, 926 F.2d 487, 490 (5th Cir. 1991); Gretna, 834 F.2d at 504. The affected political subdivision in this case is the state of Texas. Accordingly, the state of Texas must be left to develop, in the first instance, a plan that will remedy the Section 2 violations found in this case. The federal district court will have the power to devise a remedy only if the state of Texas fails to develop a remedial plan in a timely manner or fails to develop a plan that fully remedies the current vote dilution. See Cook, 735 F.2d at 918-19 (recognizing that a district court may be forced to impose a remedial scheme submitted by a private litigant or formulate one of its own when plan submitted by legislature is “uncorrectably invalid”).

If the district court is required to develop its own remedial plan, it must proceed with extreme caution. As an initial matter any plan developed by the district court must: (a) cure the dilution of minority voting strength; and (b) intrude on state policy choices only where necessary. See Upham v. Seamon, 456 U.S. 37, 41-42, 102 S.Ct. 1518, 1521, 71 L.Ed.2d 725 (1982); White v. Weiser, 412 U.S. 783, 796-97, 93 S.Ct. 2348, 2355, 37 L.Ed.2d 335 (1973); Whitcomb v. Chavis, 403 U.S. 124, 160-61, 91 S.Ct. 1858, 1878, 29 L.Ed.2d 363 (1971). In this case, state policy choices may re*815quire the district court to carefully consider remedies such as cumulative voting, limited voting, or subdistricting. And, if it becomes apparent that a subdistricting remedy will be required, the district court should also consider whether multi-member subdistricts, at least in the larger counties, would be feasible.

We turn finally to the district court’s interim remedy, which requires district judges in Texas to be elected in non-partisan elections from single-member districts. This order cannot stand under the requirements set forth by the Supreme Court in Upham, White, and Weiser, because it unnecessarily brushes aside Texas’ expressed interest in electing judges through partisan elections. The district court gave no explanation for rejecting Texas’ system of partisan elections, nor did it hold a hearing on the issue. Because the equitable powers of the district court neither encompass nor justify the interim plan ordering nonpartisan elections from single-member districts, we vacate the district court’s interim remedial plan.

Thus, we remand this case to the district court with the following instructions concerning the remedial phase of this lawsuit: The district court is directed to give the state of Texas 180 days to develop and submit to the United States Attorney General, in accordance with Section 5 of the Voting Rights Act, a plan to remedy the current vote dilution in Bexar, Dallas, Ector, Harris, Jefferson, Lubbock, Midland, and Tarrant counties. If and when the United States Attorney General preclears the plan submitted by the state, the district court shall review the plan to determine whether it rectifies the current vote dilution in those counties and whether it otherwise complies with the law. And finally, if the state of Texas does not develop and preclear a remedial plan in a timely fashion, or if the precleared plan does not, in the district court’s view, comply with Section 2 of the Voting Rights Act or other provisions of law, the district court shall determine, in accordance with the guidelines set forth above, what remedial plan to impose.

VI. CONCLUSION

For the foregoing reasons, we conclude that the district court did not err in determining that Texas’ method of electing district court judges—as that method operates in Bexar, Dallas, Ector, Harris, Jefferson, Lubbock, Midland, and Tarrant counties— violates Section 2 of the Voting Rights Act. We conclude, however, that the district court did err in finding a Section 2 violation in Travis County, Accordingly we AFFIRM the district court’s judgment as to eight of the nine target counties, REVERSE the district court’s judgment as to Travis County, and REMAND for the imposition of a remedy that is consistent with this opinion.

APPENDIX I

Judicial Districts with Fewer than 40,000 People

[[Image here]]

*816[[Image here]]

*817[[Image here]]

APPENDIX II

Population of Potential Subdistricts

[[Image here]]

*818APPENDIX III

The First Gingles Factor:

[[Image here]]

APPENDIX IV

[[Image here]]

*819[[Image here]]

. An explanation for the length of this opinion is in order. This case was tried and decided in 1989. Since then, it has taken an unexpected detour to the Supreme Court and back. We recognize that this panel may well not have the last appellate word. The parties and amici have filed briefs totalling nearly 1200 pages and have raised an extraordinarily large number of issues *737in what is, in reality, nine separate Voting Rights Act cases. This opinion attempts to address all the issues raised, with the hope of facilitating the ultimate resolution of the case and minimizing the need for further consideration at our level of issues already briefed.

. Stated another way, the Texas Constitution was amended in 1985 to provide for the first time that a county may subdivide itself into smaller judicial districts than the full extent of the county; such action requires a majority vote of the county.

. The Senate Report is the authoritative expression of Congress’ intent in amending Section 2 of the Voting Rights Act. The House of Representatives did not draft its own report, nor did it negotiate a conference committee report. The House instead adopted the language of the Sen*742ate Report. Therefore, the Senate Report serves in place of a conference committee report and should be considered, next to the statute itself, the most persuasive evidence of congressional intent. See generally Thornburg v. Gingles, 478 U.S. 30, 43-46, 106 S.Ct. 2752, 2762-2764, 92 L.Ed.2d 25 (1986).

. Sushma Soni, Note, Defining the Minority-Preferred Candidate Under Section 2, 99 Yale LJ. 1651, 1653 (1990).

. Evidence of racially polarized voting does not, however, automatically establish minority political cohesiveness. If, in a certain community, white citizens vote only for candidates of type A, while minority citizens are split in voting for candidates of types X, Y, and Z, then there would be evidence of racially polarized voting— minority and white voters voting differently— but no evidence of minority political cohesiveness.

. Judge Higginbotham contends that five Justices in Gingles expressly held that the extent to which voting patterns are attributable to causes other than racial political considerations is an integral part of the racial bloc voting inquiry. Dissenting Op. at 831-32. Indeed, in Judge Higginbotham’s view, the failure to demonstrate racial political considerations, or racial animus, in the electorate automatically defeats a vote dilution claim, because the minority group will not be able to demonstrate either legally significant white bloc voting or racially polarized voting. After a careful reading of Gingles, we respectfully disagree.

We recognize that Justice O’Connor, in her Gingles concurrence, disagreed with Justice Brennan’s position that explanations of divergent racial voting patterns are "irrelevant’’ to the Section 2 inquiry. Gingles, 478 U.S. at 100, 106 S.Ct. at 2792 (O’Connor, J., joined by three Justices, concurring in the judgment). In expressing her views on the use of statistical evidence to demonstrate racially polarized voting, however, she never "expressly held” that racial political considerations are an "integral part,” much less a prerequisite, to a finding of racial bloc voting. Rather, she stated:

Insofar as statistical evidence of divergent racial patterns is admitted solely to establish that the minority group is politically cohesive and to assess its prospects for electoral success, I agree that defendants cannot rebut this showing by offering evidence that the divergent racial voting patterns may be explained in part by causes other than race, such as an underlying divergence in the interests of minority and white voters. I do not agree, however, that such evidence can never affect the overall vote dilution inquiry. Evidence that a candidate preferred by the minority group in a particular election was rejected by white voters for reasons other than those which made that candidate the preferred choice of the minority group would seem clearly relevant in answering the question whether bloc voting by white voters will consistently defeat minority candidates. Such evidence would suggest that another candidate, equally preferred by the minority group, might be able to attract greater white support in future elections.
I believe Congress also intended that explanations of the reasons why voters rejected minority candidates would be probative of the *746likelihood that candidates elected without decisive minority support would be willing to take the minority’s interests into account. In a community that is polarized along racial lines, racial hostility may bar these and other indirect avenues of political influence to a much greater extent than in a community where racial animosity is absent although the interests of the racial groups diverge. Indeed, the Senate Report clearly stated that one factor that could have probative value in § 2 cases was "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. S.Rep., at 29. The 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. Such a rule would give no effect whatever to the Senate Report’s repeated emphasis on "intensive racial politics,” on "racial political considerations," and on whether "racial politics ... dominate the electoral process” as one aspect of "the racial bloc voting” that Congress deemed relevant to showing a § 2 violation. Id. at 33-34. Similarly, I agree with Justice White that Justice Brennan’s conclusion that the race of the candidate is always irrelevant in identifying racially polarized voting conflicts with Whitcomb [v. Chavis, 403 U.S. 124, 91 S.Ct. 1858, 29 L.Ed.2d 363 (1971) ] and is not necessary to the disposition of this case.

478 U.S. at 100-01, 106 S.Ct. at 2792 (emphasis added).

How this passage amounts to, as Judge Higginbotham states, an "express holding” that racial political considerations are an integral part of, or a prerequisite to, a finding of racially polarized voting is not clear. A straightforward reading of this passage, in our view, suggests that Justice O’Connor would consider relevant anecdotal testimony indicating that, "in a particular election,” white voters rejected a minority-preferred candidate for "reasons other than those which made that candidate the preferred choice of the minority group.” Such anecdotal testimony would be relevant, in Justice O’Con-nor’s words, to "suggest that another candidate, equally preferred by the minority group,” might be able to attract greater white support in the future. By making this statement, Justice O’Connor recognized that the phenomenon of white bloc voting, as well as racially polarized voting, is only significant if it persists over time. Moreover, she also indicated that "explanations of the reasons why voters rejected minority candidates" would be probative of the totality of circumstances factor relating to the "responsiveness of elected officials.” See infra Part II. B.l.h. Finally, Justice O'Connor recognizes that in a community "polarized along racial lines,” the added circumstance of "racial hostility” may bar even "indirect avenues of political influence to a much greater extent than in a community where racial animosity is absent although the interests of racial groups diverge.” Thus, when read in context, Justice O’Connor’s Gingles opinion is much more consistent with our treatment of racial political considerations, or racial animus in the electorate, as an "other factor” that is relevant to the "overall vote dilution inquiry, ” rather than as a prerequisite to a finding of legally significant white bloc voting. 478 U.S. at 101-02, 106 S.Ct. at 2793 (emphasis added).

Finally, we note that Judge Higginbotham in another portion of his dissent appears to reject the idea that a minority group must demonstrate that racial animus is responsible for a white bloc voting pattern. Judge Higginbotham concedes that it would be unworkable to require plaintiffs to prove the absence of all non-racial causes of voting behavior in order to demonstrate legally significant white bloc voting. However, he clings to the idea that evidence of partisan voting is somehow different from other evidence suggesting the presence or absence of racial political considerations or of racial animus in the electorate. This differentiation between partisanship evidence and other evidence is patently illogical. First, this approach would appear to prevent an inquiry into the many reasons why people vote for the candidates of one party or another. Second, if evidence of partisan voting is sufficient to defeat a finding of white block voting, why not other evidence suggesting the absence of racial political considerations? If, as Justice O’Connor suggests, the reasons why people vote the way they do are relevant to the overall vote dilution inquiry, how can one stop that inquiry at statistical correlations between the party affiliations of the candidates and electoral success? Indeed, why would such a rule not amount to an "arbitrary rule” against consideration of all evidence concerning voting preferences similar to the rule condemned by Justice O’Connor? Ultimately, Judge Higginbotham’s approach would needlessly confuse the narrow Gingles factors with the broader inquiry into the totality of the circumstances. The logical solution is to follow the approach mandated by Congress and adopted by Justice O'Connor, i.e., to consider evidence suggesting that divergent voting patterns are not the result of racial political considerations in the totality of circumstances inquiry.

. Although the ultimate inquiry in a Section 2 case is whether minority voters have an unequal opportunity to participate in the political process on account of race or color, see infra Part Ü.C., we decline to redefine the concept of racially polarized voting to include a causation element. To establish racially polarized voting, a minority group need not prove that divergent racial voting patterns are motivated or caused by racial animus in the electorate. Moreover, we decline to hold that a multivariate analysis— i.e., an analysis that investigates more variables than the race of the voters—is necessary to show racially polarized voting. In our view, such a definition would be inconsistent with Congress' intent in amending Section 2 and directing courts to “assess the impact of the challenged structure or practice on the basis of objective factors, rather than making a determination about the motivations which lay behind its adoption or maintenance.” S.Rep. at 27, 1982 U.S.C.C.A.N. at 205 (emphasis added). See generally, Richard L. Engstrom, The Reincarnation of the Intent Standard: Federal Judges and At-Large Election Cases, 28 How.L.J. 495, 498 (1985) (disapproving the recent trend in federal courts to give a "second life" to the intent requirement); cf. Kirksey v. City of Jackson, 663 F.2d 659, 662 (5th Cir. Unit A Dec.1981) (holding that, because of First Amendment concerns, the motivations of voters are not subject to searching scrutiny by plaintiffs in a voting rights case), clarified, 669 F.2d 316 (1982).

In refusing to redefine the concept of racially polarized voting, however, we do not mean to say that evidence of the presence or absence of racial animus in the electorate is irrelevant to the totality of circumstances inquiry. See infra Part II.B.2.

. In his dissent, Judge Higginbotham, citing our opinion in Gretna, suggests that the reason elections with a minority candidate are most probative of racially polarized voting is because we are trying to determine whether such voting is caused by racial animus in the electorate. See Dissenting Op. at 832-33. We disagree. Our decision in Gretna indicates that the reason for focusing on these elections is—generally—to ensure that we are considering elections "that offer[ ] voters the choice of supporting a viable minority candidate,” 834 F.2d at 503, or more specifically, a minority candidate that is preferred by the minority group.

. Judge Higginbotham suggests that, because Congress explicitly mentioned the extent to *751which minorities have been elected as a totality of circumstances factor in the statute itself, it should somehow take precedence over the extent to which minorities have been able to elect candidates "of their choice.” Judge Higginbotham suggests that "it is quite possible” that minority plaintiffs will not be able to demonstrate a lack of electoral success,” an element that is crucial to their claim, "where black Republicans win and white Democrats lose.” Dissenting Op. at 827. Judge Higginbotham further states that the Senate Report precludes courts from considering—through a filter— whether particular minorities who have won elections have been the minority-preferred candidate. Dissenting Op. at 828. We disagree. The Senate Report directs courts to consider the "totality of the circumstances,” and where a minority candidate is elected with virtually no support from the minority community, a court may reasonably take that fact into account in resolving the ultimate question under Section 2 of whether minorities have an equal opportunity to elect "representatives of their choice." 42 U.S.C. § 1973(b).

. In his dissent, Judge Higginbotham recognizes that "[ejvidence of the electoral success of minority candidates should not, of course, be accepted uncritically.” Dissenting Op. at 828. He then asks us, however, to hold that where "several” Republican minority candidates have been elected—as opposed to a "few”—such successes must be accepted uncritically as evidence that white voters are not motivated by racial animus. Dissenting Op. at 828. We think such a distinction would be unwise and inconsistent with Supreme Court precedent.

First, it is not clear, and Judge Higginbotham does not tell us, what the difference is between a "few” and "several” Republican minority candidates. Moreover, he does not even set forth the number of minority Republican district court candidates that were elected in the eight years studied by the experts in this case, and for good reason. Our review of the record leaves us uncertain about that number. As best we can tell from the record, from 1980 to 1988 it appears that: (1) in Harris County, one Black Republican district court candidate won a contested general election; (2) in Dallas County, two Black Republican district court candidates won contested general elections; (3) in Bexar County, one Hispanic Republican won a contested general election; and (4) in Tarrant County, one Black Republican won a contested general election.

Based on our review of the record, we can only surmise that Judge Higginbotham reaches the conclusion that "several” minority Republican candidates have been elected by aggregating elections from the various counties. This case should not turn on whether the election of five minority Republican district court judges qualifies as “several” or "few." Besides, such aggregation is contrary to Gingles, which in this case requires that vote dilution inquiries be county-specific. And, focusing on the county-by-county analysis mandated by Gingles, we think it would be splitting hairs to hold that the election of two Black Republican district judges in Dallas County precludes a successful vote dilution claim— whereas the election of one such judge would not foreclose a claim.

Finally, while all members of the Court in Gingles agreed that consistent and virtually proportional minority candidate success would foreclose a Section 2 claim, see 478 U.S. at 77, 104-05, 106 S.Ct. at 2780, 2794, "proportional” success by minority candidates is a far cry from the success of “several” minority Republican candidates. At bottom, Judge Higginbotham urges that the election of "several” (perhaps five in the nine counties at issue here) Republican minority candidates, candidates who were not even supported by the minority community, indicates as a matter of law that minorities have an equal opportunity to elect representatives of *752their choice. This result is simply contrary to the plain language of Section 2.

. Judge Higginbotham would hold that evidence merely suggesting the absence of specific racial animus in the electorate—namely, evidence that "several" minority Republicans have won elections—by itself defeats a vote dilution claim. For the reasons we have already stated, see supra notes 6 and 10, we reject this argument as being inconsistent with Gingles and congressional intent.

. As Judge Higginbotham persuasively explained in his concurring opinion in League of United Latin American Citizens v. Clements:

We have insisted in other contexts that Congress clearly state its intent to supplant traditional state prerogatives. Judicial insistence upon clear statement is an important interpretative tool vindicating concern for separation of powers and federalism. See Atascadero State Hospital v. Scanlon, 473 U.S. 234, 105 S.Ct. 3142, 87 L.Ed.2d 171 (1985); Pennhurst State School and Hospital v. Halderman, 465 U.S. 89, 104 S.Ct. 900, 79 L.Ed.2d 67 (1984) (Pennhurst II). This insistence upon an unequivocal expression of congressional intent, Pennhurst II, 465 U.S. at 99, 104 S.Ct. at 907, is based upon the fundamental nature of the interests at stake, Atascadero, [473 U.S. at 242] 105 S.Ct. at 3147 (The `constitutionally mandated balance of power' between the states and the Federal Government was adopted by the Framers to ensure the protection of `our fundamental liberties.') (quoting Garcia v. San Antonio Metropolitan Transit Authority, 469 U.S. 528, 572, 105 S.Ct. 1005, 1028, 83 L.Ed.2d 1016 (1985) (Powell, J., dissenting)). These mighty principles do not carry us very far here. Congress has clearly expressed the [Voting Rights] Act's application to the states, and has clearly expressed its intent that violations of the Act be determined by a results test rather than an intent standard. By these actions, the Act, with all of its intrusive effect, has been made to apply to the states.

914 F.2d 620, 641-42 (5th Cir.1990) (Higginbotham, J., concurring in the judgment) (emphasis in original) (parallel citations omitted), rev'd and remanded, — U.S. —, 111 S.Ct. 2376, 115 L.Ed.2d 379 (1991).

. Judge Entz makes four arguments in this regard. First, he argues that applying the accepted Section 2 framework to state trial judges would violate Tenth Amendment federalism principles. He further argues, citing The Federalist No. 48 (J. Madison) (J. Cooke ed. 1961) and the Texas Constitution, that an application of the accepted Section 2 framework would violate separation of powers principles. Third, he arces that the 1982 amendments to Section 2 were not a valid exercise of Congress’ authority under the Fourteenth and Fifteenth Amendments. And finally, he argues that Section 2, if applied without balancing state interests, is unconstitutionally vague. For the reasons discussed in the text, we reject all of these arguments.

. Judge Higginbotham interprets the Court’s reference to Zimmer in another way and suggests, by quoting an isolated passage from our decision in Zimmer, that we held that “a strong state policy divorced from the maintenance of racial discrimination," could by itself preclude a holding of vote dilution. Dissenting Op. at 838. A careful reading of our Zimmer decision reveals that Judge Higginbotham’s partial quotation is flawed. In Zimmer, we stated:

The Supreme Court has identified a panoply of factors, any number of which may contribute to the existence of dilution. Clearly, it is not enough to prove a mere disparity between the number of minority residents and the number of minority representatives. Where it is apparent that a minority is afforded the opportunity to participate in the slating of candidates to represent its area, that the representatives slated and elected provide representation responsive to minority’s needs, and that the use of a multimember districting scheme is rooted in a strong state policy divorced from the maintenance cf. racial discrimination, Whitcomb v. Chavis, supra, would require a holding of no dilution. Whit-comb, would not be controlling, however, where the state policy favoring multi-member or at-large districting schemes is rooted in racial discrimination. Conversely, where a minority can demonstrate a lack of access to the process of slating candidates, the unresponsiveness of legislators to their particularized interests, a tenuous state policy underlying the preference for multi-member or at-large districting, or that the existence of past discrimination in general precludes the effective participation in the election system, a strong case is made. Such proof is enhanced by a showing of the existence of large districts, majority vote requirements, anti-single shot voting provisions and the lack of provision for at-large candidates running from particular geographical subdistricts. The fact of dilution is established upon proof of the existence of an aggregate of these factors. The Supreme Court’s recent pronouncement in White v. Regester, supra, demonstrates, however, that all these factors need not be proved in order to obtain relief.

485 F.2d at 305 (emphasis added). When this passage is read in its context, then, it becomes clear that our treatment of state interests in Zimmer, rather than supporting the balancing framework advocated by Judge Higginbotham, is consistent with Congress’ treatment of state interests in the Senate Report—namely, that the inquiry is whether the asserted state interests are tenuous. We have never held, in Zimmer or otherwise, that a "strong state policy divorced from the maintenance of racial discrimination,” could by itself “outweigh” other factors pointing towards vote dilution.

. The Texas Constitution requires that each district judge be elected. See Tex. Const. art. V, § 6.

. See generally Tex.Elec.Code Ann. §§ 161.001-182.007 (Vernon 1986).

. Under the Texas Constitution, all district judges must be licensed to practice law in the state and must have practiced law for at least four years prior to their election. See Tex. Const. art. V, § 6.

. The Texas Constitution requires that all district judges reside in the district during the term of office. See Tex. Const. art. V, § 6. Further, a district judge candidate must have resided in the district for two years prior to the election. See id.

. The Texas Code of Judicial Conduct prohibits judicial candidates from making statements that indicate an opinion on an issue that might be subject to judicial interpretation. See Texas Supreme Court, Code of Judicial Conduct, Canon 7 (Vernon Supp.1992) [hereinafter Tex.Code of Judicial Conduct]. The Code further prohibits district judge candidates from making pledges or promises of conduct in office. See id.

.On appeal, Judge Entz raises other state interests that he asserts are advanced by Texas’ current method of electing district court judges. According to Judge Entz, Texas’ current method of electing district judges also (1) reflects Texas' citizens’ expression of self-government, (2) protects litigants from disenfranchisement, (3) protects the rights of criminal defendants, (4) allows judges to specialize, and (5) protects minority representation. We do not separately address these asserted interests, because they are either components of the "linkage” or "sole decision-maker” interests being asserted by the State Defendants, were not presented to the district court, or are not threatened by any of the proposed remedies.

. Our search of Texas constitutional, statutory, and decisional law reveals no concept or definition of the phrase "primary jurisdiction.”

. Moreover, it bears noting that 27% of Texas’ district judges are elected from judicial districts that are larger than a county. See The American Bench 2138-54 (6th ed. 1991). With regard to these district courts, the link between elective base and venue becomes more tenuous.

. In his dissent, Judge Higginbotham places great weight on the fact that no county has voted to elect district judges from areas smaller than a county. Dissenting Op. at 839. This does not change our conclusion that the so-called “required link" between elective base and venue is non-existent. The fact is that the Texas Constitution does not presently require that district judges be elected from areas no smaller than a county. Thus, the state has taken a neutral position regarding the elective base of district judges: district judges may be elected from county-wide election districts or they may be elected from districts smaller than a county.

. For example, under these provisions, any of the fifty-nine district court judges elected in Harris County may be assigned, at any time, to hear and decide cases in any of the thirty-four other counties in the second administrative judicial region. Tex.Gov’t Code Ann. § 74.042(c) (Vernon 1988). Those same judges may as well be assigned to counties in other administrative judicial regions throughout the entire state. Any judge can be assigned to any county in any region of the state.

. The general venue provision in Texas provides: "Except as otherwise provided by [the mandatory venue rules] or [the permissive venue rules], all lawsuits shall be brought in the county in which all or part of the cause of action accrued-or in the county of defendant’s residence if defendant is a natural person." Tex.Civ.Prac. & Rem.Code Ann. § 15.001 (Vernon 1986).

. There are mandatory venue provisions for the following types of lawsuits: (1) actions involving real property, see Tex.Civ.Prac. & Rem. Code Ann. § 15.011 (Vernon 1986); (2) actions to stay proceedings in an already-pending lawsuit, see id. § 15.012; (3) actions to restrain execution of a judgment, see id. § 15.013; (4) actions for mandamus against the head of a department of state government, see id. § 15.014; (5) actions against a county, see id. § 15.015; and (6) actions involving libel, slander, or invasion of privacy, see id. § 15.017.

. There are permissive venue rules for the following types of actions: (1) suits against an executor, administrator, or guardian, see Tex. Civ.Prac. & Rem.Code Ann. § 15.031 (Vernon 1986); (2) suits against insurance companies, see id. § 15.032; (3) suits for breach of warranty by a manufacturer of consumer goods, see id. § 15.-033; (4) actions involving certain written contracts, see id. § 15.035; (5) suits against corporations and associations, see id. § 15.036; (6) suits against foreign corporations doing business in Texas, see id. § 15.037.

. Judge Higginbotham suggests that we understate the local nature of district courts by overlooking how venue operates in criminal law. Dissenting Op. at 840. A review of Texas' criminal venue rules, however, belies Judge Higginbotham's claim that they are somehow designed to foster judicial accountability and independence. First, it is important to note that Texas district courts have "jurisdiction" over any felony committed in whole or part anywhere in Texas. See Ex Parte Watson, 601 S.W.2d 350, 351 (Tex.Crim.App.1980). Second, there are over ten separate criminal venue rules in Texas, each depending on the particular offense committed. See Tex.Code.Crim.P. art. 13.01 et seq. And, although criminal venue may be "quasi-jurisdictional" in nature, venue can be acquired by "consent" of the parties. See id. art. 13.20. Moreover, criminal venue can be waived by a defendant's failure to object at trial. See Ex Parte Watson, 601 S.W.2d at 351. Finally, the Texas Code of Criminal Procedure provides various methods for obtaining a change of venue, some of which are meant to protect the rights of a criminal defendant. See id. art. 31.01, 31.02, 31.03. Thus, we fail to see how Texas' criminal venue rules operate to ensure that district judges remain accountable to the people who form their elective base.

. At trial, several parties contended that judges elected from smaller election districts would be more susceptible to undue influence from organized crime. In view of the vast differences in the size of Texas’ current judicial districts, such a contention is without merit. Texas has created numerous judicial districts with relatively small populations. In two districts, less than 12,000 people constitute the electorate. In six more districts, less than 20,000 constitute the electorate. Finally, in thirty-one judicial districts, less than 40,000 constitute the electorate. See Appendix I (table of judicial districts with fewer than 40,000 people in the electorate).

Even if a single-judge subdistricting remedy were imposed in this case, the resulting subdistricts would have populations ranging from 27,-545 to 52,042. See Appendix II (table of potential subdistricts). Given that every single-judge subdistrict would thus be larger than several existing districts, it is not possible to conclude that a single-judge subdistricting remedy would increase the risk of undue influence from organized crime.

. In analogous areas of the law — including substantive due process, equal protection, and first amendment law — the weight of an asserted state interest has always been resolved by the court as a legal question. For example, in Posadas de Puerto Rico Association v. Tourism Co. of Puerto Rico, 478 U.S. 328, 341, 106 S.Ct. 2968, 2977, 92 L.Ed.2d 266 (1986), the Court had "no difficulty in concluding that the Puerto Rico Legislature's interests in the health, safety, and welfare of its citizens constitutes a `substantial' governmental interest." In reaching this conclusion, the Court did not refer to the clearly erroneous rule, but instead made its own, de novo determination of the weight of the state interest. See also Midnight Sessions, Ltd. v. City of Philadelphia, 945 F.2d 667, 682 (3rd Cir.1991) (holding that determination of whether a legislative scheme is rationally related to a legitimate government interest is a question of law for the court to decide), cert. denied, — U.S. —, 112 S.Ct. 1668, 118 L.Ed.2d 389 (1992); Magill v. Lynch, 560 F.2d 22, 27 (1st Cir.1977) (reviewing de novo the district court's determination that municipality's interests were not sufficiently compelling to justify prohibiting city employees from engaging in a broad range of political activities), cert. denied, 434 U.S. 1063, 98 S.Ct. 1236, 55 L.Ed.2d 763 (1978).

. Both of these types of analyses are essentially forms of "bivariate regression analysis”—meaning that they correlate the race of voters and the level of support given to a particular candidate without considering other variables that might explain voters' choices.

. We do not separately address two of the district court’s findings under the totality of circumstances inquiry. First, with respect to the responsiveness of district judges in the target counties, the district court found nothing in the record "to suggest a lack of responsiveness ... to the particularized needs of members of the minority community.” This finding is not challenged on appeal and we do not address it further. Second, we do not address the tenuousness of the state interests, having already done so in Part II.D. supra and having agreed with the district court’s assessment.

. With respect, to the ultimate vote dilution inquiry, we assume, for purposes of this part of the opinion, that the district court’s ultimate vote dilution finding is protected by the clearly erroneous rule. We consider separately, in Part IV.B.3 infra, whether the district court’s refusal to consider evidence of partisan voting patterns renders erroneous its ultimate vote dilution findings in the seven counties for which such *777evidence was offered. We also consider separately, in Part IV.C. infra, whether the proposed balancing framework that may have been indicated by the Supreme Court in Houston Lawyers’ Association, see supra Part II, would affect the district court's ultimate vote dilution findings in the target counties.

. Before addressing the district court’s vote dilution findings in Bexar County, we consider the district court’s refusal to allow certain Bexar County district judges to intervene after trial. On appeal, these Bexar County district judges argue that the district court erred in denying their post-trial motion to intervene under Rule 24 of the Federal Rules of Civil Procedure. We disagree. A motion to intervene under Rule 24(a) (intervention as of right) or under Rule 24(b) (permissive intervention) must be timely. See Jones v. Caddo Parish School Bd., 735 F.2d 923, 926 (5th Cir. 1984) (en banc). Although the district court did not expressly state, in denying the Bexar County district judges’ motion to intervene, that their motion was untimely, it was well within the district court’s discretion to deny the motion on this ground.

. The State Defendants contend that the district court did not specifically assess the impact of white cross-over voting. The district court’s memorandum opinion refutes this contention. The district court expressly considered the extent of white cross-over voting in Bexar County, as well as in all the remaining counties. Thus, *778the State Defendants' contention is without merit.

. As discussed above, see supra Part II.B.l.g., the relevance of the number of Hispanic lawyers in Bexar County is doubtful—especially given the extent to which Hispanics in Bexar County continue to bear the effects of discrimination in education.

. On September 1, 1989, the Texas Legislature created a thirty-seventh state district court in Dallas County.

. The following examples of official discrimination against Blacks in Texas were cited by the court in Graves I, 343 F.Supp. at 725 n. 15, n. 16: a state statute prohibiting Blacks from participating in the Democratic primary; a city ordinance segregating parks; a city ordinance making it unlawful for whites and Blacks to have sexual intercourse within the city limits; a statute prohibiting the adoption of white children by Blacks and Black children by whites; a criminal statute prohibiting fights or wrestling matches between whites and Blacks; and a Texas appellate court decision holding that the plaintiff stated a valid cause of action by alleging that his wife was wrongfully excluded from a passenger elevator set aside for whites and made to ride an elevator set aside for the use of Blacks.

. On appeal, Judge Entz concedes that Dallas County and Texas "once had official discrimination against the voting rights of [B]lack voters.” He seeks to minimize this past reality of the Dallas County political landscape, however, by noting that the days of official discrimination are "thankfully long in the past." Judge Entz also suggests that the district court did not find that the unfortunate history of official discrimination has any present day lingering effect on the ability of Blacks to participate in the political process in Dallas County. Finally, Judge Entz points out that "there are certainly no present official impediments to [B]lack voting."

We reject Judge Entz’s invitation to discount the history of official discrimination against Blacks in Dallas County. First, we think that Judge Entz overstates the remoteness of this official discrimination. The year 1970 is not, in our view, "long in the past.” Moreover, we disagree with Judge Entz’s reading of the district court’s opinion with respect to the effects of official discrimination. The district court expressly found that “the effect of past discrimination against Blacks ... is either well-chronicled or undisputed," thereby recognizing that past discrimination continues to hinder Blacks' ability to participate in the political process. Lastly, we note that Judge Entz’s observation that there are no longer any official impediments to Black voting is beside the point. In amending Section 2, Congress indicated that it was concerned not only with overt prohibitions against casting a ballot, but also with the use of schemes that cancel out or minimize minority voting strength. See S.Rep. at 6, 1982 U.S.C.C.A.N. at 182.

. Contrary to Judge Entz’s suggestion, the district court did not reject this court’s test for racially polarized voting. As we have already explained, see supra Part II.B.l.b., the concept of racially polarized voting does not denote the tendency of citizens to vote for candidates of their own race. Rather, racially polarized voting denotes the tendency of whites and minorities to vote differently. Although the district court, at one point in its memorandum opinion, referred to a Fifth Circuit decision defining racially polarized voting as the tendency of a group to vote for candidates of a particular race, the district court also noted, at the same point in the opinion, that under Gingles, racially polarized voting exists when whites and minorities vote differently. Moreover, in its county-by-county analysis, the district court clearly applied the correct legal standard in determining the existence and extent of racially polarized voting.

. Judge Higginbotham suggests in his dissent that the history and lingering effects of discrimination against Blacks and Hispanics are entitled to no weight in the totality of circumstances analysis in any of the counties, because there was no evidence — other than evidence of low *783voter registration rates among Hispanics — of depressed political participation by those minority groups in the political process. Dissenting Op. at 826. Thus, he implies that depressed political participation may only be shown by low voter registration or low voter turnout rates.

Initially, we note that the district court, in addition to finding that Blacks and Hispanics continue to bear the effects of discrimination in education and employment, specifically found that "the continual effects of historical discrimination hinder[] the ability of minorities to participate in the political process."

We also note that this is not an issue that has been seriously pursued by the parties on appeal. The State Defendants do not raise this issue at all. Judge Entz devotes two sentences in each of his briefs to the argument that, insofar as Dallas County is concerned, the Plaintiffs offered "no testimony showing that, at this date, [Blacks'] lower [socioeconomic] status is directly a result of prior discrimination or that it hinders the ability of blacks to participate in the political process." Judge Entz candidly concedes in a footnote that "Dr. Brischetto offered such testimony with respect to other counties." Judge Wood devotes a fragment of one sentence to this issue on appeal, arguing only that the Plaintiffs produced "no evidence that lingering effects of discrimination hinder the ability of blacks to participate effectively in the political process in Harris County."

In addition, we note that, to the extent that the issue has even been raised with respect to Dallas and Harris counties, it is without merit. While evidence of depressed voter registration or turnout rates is one indicator of depressed political participation, it is by no means the only indicator. To say that there is no evidence in the record that "the level of black participation in politics is depressed," is to ignore the evidence repeatedly emphasized by Judge Entz, Judge Woods, and Judge Higginbotham with respect to the paucity of minority lawyers in each of the counties at issue. Before minorities can fully participate in district court elections, they must be qualified to run for such offices. And, to be qualified to run for a district court position, minorities must attend and graduate from law school, a privilege which was not equally accorded to Blacks and Hispanics in Texas until at least 1950. See Sweatt v. Painter, 339 U.S. 629, 70 S.Ct. 848, 94 L.Ed. 1114 (1950). Thus, we conclude that the evidence indicating that Blacks constitute, at most, 2.2% of the lawyers in Dallas County and 3.8% of the lawyers in Harris County, when viewed in conjunction with the district court's finding that the paucity of minority lawyers is "due in part to historical discrimination," supports the district court's conclusion about "depressed political participation" by Blacks.

The district court's conclusion that the lingering effects of discrimination continue to hinder minority participation in the political process is further supported by Dr. Brischetto's expert testimony. Dr. Brischetto specifically testified with respect to all counties in which minorities continue to suffer the lingering socioeconomic effects of discrimination:

Well, certainly having less of these socioeconomic resources or characteristics to draw on, we find that minority voters will participate less in the election system. Education is an important resource. For example, it enables people to feel like they are more a part of and take part in the election system to a greater extent. Lacking that they participate less. So it is important, it has an effect certainly on their participation when they are subordinate status in the stratification system.

Finally, there is also evidence suggesting that, because of the lingering socioeconomic effects of discrimination, Blacks and Hispanics are unable to successfully mount a county-wide campaign. Several witnesses expressed their concern that minority candidates, because they generally do not have the financial resources to mount a county-wide campaign, are hindered in their ability to fully participate in the current system of electing judges. In Harris County, for example, Bonnie Fitch stated that Black district court candidates usually find it harder to raise money. This concern was echoed by Adam Serrata in Bexar County. He stated that "Anglos generally can raise more money than minority candidates but it is not as expens[ive] to run in a single member district election as it is in a countywide election." Jim Coronado also testified that the problem minorities have in mounting a successful, county-wide district court campaign arises from the facts that "[i]t costs a bunch of money" to mount such a campaign and that minority candidates generally "don't have the financial resources to get out there and do that."

. The dissent, like Judge Entz, attempts to minimize this racial appeal by pointing out that it did not work for Ms. Busby, who lost the race to Mr. Baraka. Dissenting Op. at 821. While we recognize that the racial appeal was perhaps unsuccessful in this case, it nonetheless constitutes “some evidence" that racial politics are at work in Dallas County judicial elections. Moreover, the Senate Report does not instruct courts to consider whether candidates have won by using racial campaign appeals. Instead, the Senate Report instructs courts to consider "whether political campaigns have been characterized by overt or subtle racial appeals." S.Rep. at 29, 1982 U.S.C.C.A.N. at 206. Finally, although we hold that the district court correctly considered this racial appeal in the totality of the circumstances in Dallas County judicial elections, neither the district court nor this court would have been willing—contrary to the dissent’s disingenuous suggestion otherwise—to change the structure of all of Dallas County courts based solely on one racial campaign appeal.

. The impetus for two minority groups seeking to proceed as a coalition under Section 2 is apparently their inability, as separate groups, to overcome the first Gingles threshold factor. Thus, in locations where neither Blacks, by themselves, or Hispanics, by themselves, are sufficiently large and geographically compact to constitute a majority of the voting age population in a single-member district, Blacks and Hispanics combined may be able to satisfy this threshold requirement.

We recognize that the procedure of allowing Blacks and Hispanics to proceed as a "coalition" minority group in a Section 2 claim is fraught with risks. See generally Katharine I. Butler and Richard Murray, Minority Vote Dilution Suites and the Problem of Two Minority Groups: Can a “Rainbow Coalition” Claim the Protection of the Voting Rights Act?, 21 Pac.L.J. 619 (1990); Rick G. Strange, Application of Voting Rights Act to Communities Containing Two or More Minority Groups—When Is the Whole Greater Than the Sum of the Parts?, 20 Tex.Tech.L.Rcv. 95 (1989). First, there is a risk that members of one of the minority groups will increase their opportunity to participate in the political process at the expense of members of the other minority group. There is also the risk of running afoul of Congress’ intent in amending Section 2. After all, ”[t]he Voting Rights Act purpose was to eliminate racial discrimination— not to foster particular political coalitions.” Butler & Murray, supra, at 648. In view of these risks, we note that minority groups should be allowed to proceed as a coalition under Section 2 only after the district court is satisfied that the risks we have discussed are not present in the community at issue.

On appeal, no party takes issue with the legitimacy of allowing Blacks and Hispanics to proceed as a coalition under Section 2. Nor does any party argue that the interests of one minority group have been advanced at the expense of another. Instead, various parties level county-specific challenges to the district court’s finding that Hispanics and Blacks are politically cohe*786sive. We address such challenges, in the counties where they are made, in our discussion of the district court's finding of political cohesiveness.

Finally, we note that no party asked the district court to make a finding that Blacks and Hispanics were politically cohesive in any county other than the three counties where the Plaintiffs proceeded on behalf of Blacks and Hispanics.

. Judge Wood challenges the district court's finding with regard to the first Gingles factor. In particular, she argues that it is mathematically impossible to draw nine single member districts (or 22% of the total number of districts in Harris County) in which voting age Blacks, who constitute only 18.2% of the voting age population, are a majority. The problem with Judge Wood’s argument is that, to satisfy the first Gingles factor, the minority group must only demonstrate that "it is sufficiently large and geographically compact to constitute a majority in a single member district.” 478 U.S. at 50, 106 S.Ct. at 2766 (emphasis added). Thus, as long as Blacks in Harris County are sufficiently large and geographically compact to constitute a majority in at least one single member district, the first Gingles factor is satisfied. Because Judge Wood does not dispute that Blacks could constitute a majority in at least one district in Harris County, the district court’s finding with regard to this factor is not clearly erroneous.

. Judge Wood attacks Dr. Engstrom’s reliance on Dr. Murray’s voter registration estimates. She points out that Dr. Murray did not authenticate the estimates and that Dr. Engstrom only performed a limited check on the estimates. Thus, she reasons, it was error for Dr. Engstrom and the district court to rely on Dr. Murray’s voter registration estimates. We disagree. Judge Wood’s argument is that, under Rule 703 of the Federal Rules of Evidence, the data relied on by Dr. Engstrom is not "of a type reasonably relied upon by experts” analyzing voting patterns. At trial, however, Dr. Engstrom testified that estimates of the ethnic *789makeup of precincts, such as the ones provided by Dr. Murray, are of a type customarily relied on by voting rights experts. Dr. Engstrom further testified that, when he did perform a limited check on the estimates provided by Dr. Murray, the data was reliable. The district court apparently credited Dr. Engstrom’s testimony and found the data set used in his analysis of Harris County to be reliable. Moreover, the district court correctly noted that any imperfections in the estimates provided by Dr. Murray went to the weight of Dr. Engstrom’s testimony and not to its admissibility. In any event, we hold that the district court did not abuse its discretion in allowing Dr. Engstrom to testify about his analysis of election results. We also hold that the district court did not clearly err in relying on Dr. Engstrom’s testimony concerning the second and third Gingles factors.

. In their amicus brief, certain Jefferson County district judges request this court to take judicial notice of the fact that, in 1990, a Black county court at law candidate won the Democratic primary, defeating two white opponents, as well as the general election, defeating another white opponent. We decline to do so. The record in a Voting Rights Act case must close at some point, and in our view, that point is the end of trial. Moreover, we discern no principled way in which to limit such a procedure in Voting Rights Act cases. If we were to take judicial notice of elections occurring after trial that tend to undermine a vote dilution claim, we would also be compelled to judicially notice elections that tend to support a vote dilution claim.

. 347 U.S. 483, 74 S.Ct. 686, 98 L.Ed. 873 (1954).

. Dr. Brischetto’s testimony regarding racially polarized voting in Midland County is further supported by his analysis and testimony in LU-LAC v. Midland I.S.D., 648 F.Supp. 596, 600 (W.D.Tex.1986), aff'd, 812 F.2d 1494 (5th Cir. 1987), vacated, 829 F.2d 546 (5th Cir.1987), which was introduced at trial as evidence of racially polarized voting.

. The State Defendants again argue, relying on our decision in Campos, that because Blacks and Hispanics in Tarrant County are politically cohesive, the district court clearly erred in re*800fusing to specifically consider and give equal weight to elections pitting whites against Hispanics for purposes of determining whether the white bloc vote in Tarrant County is legally significant. Again, the State Defendants’ argument in this regard would require a fact finding that Blacks and Hispanics in Tarrant County are politically cohesive. Judge Higginbotham correctly observes in his proposed opinion that "the district court made no findings as to whether black and Hispanic voters were also a cohesive group" in Tarrant County. Dissenting Op. at 856. What he fails to mention, however, is that neither the State Defendants nor the defendant intervenors ever asked for such a finding. Thus, the State Defendants are raising an argument that was not raised before the district court and are requesting that we make a finding of fact on appeal. We decline to address their argument and hold that the error, if any, was not preserved for appellate review.

. According to testimony at trial, two of the three Black district judges were appointed and did not face opponents during the period from 1985 to 1988: Judge Clifford Davis and Judge Mary Ellen Hicks. In 1988, however, Judge Davis lost his bid to remain of the bench despite winning "every identifiable Black box by an overwhelming majority of the votes cast.” The third Black judge that remained on the district bench in Tarrant County during this period, Judge Louis Sturns, ran and won as a Republican in 1986, despite garnering only some 10% of the Black vote.

. In Gingles, the district court’s findings with regard to one of the locales were based on analysis of only three races. There the Court expressly noted that ”[t]he number of elections that must be studied in order to determine whether voting is polarized will vary according to pertinent circumstances.” 478 U.S. at 57 n. 25, 106 S.Ct. at 2770 n. 25.

. Contrary to Judge Higginbotham’s suggestion, we are not discounting this election solely because ’’it tends to demonstrate that the voting preferences of whites are not attributable to racial considerations." Dissenting Op. at 836. Instead, we question the reliability of this election because of evidence indicating that a Black candidate won with 23% of the Hispanic vote and only a bare majority of the white vote. This evidence, in Justice O’Connor’s words, "would suggest that another candidate, equally preferred by [Hispanics], might be able to attract greater white support in future elections.” Gingles, 478 U.S. at 100, 106 S.Ct. at 2792.

. We confess to being somewhat baffled by Judge Higginbotham’s reference to “dependence upon a winning Republican Party for proof of bloc voting," which he terms an “embarrassment to the findings of vote dilution in this case.” Dissenting Op. at 821. Because the district court considered the partisan voting to be legally irrelevant, the findings of vote dilution in this case are devoid of any connection to partisan voting patterns or a "winning Republican Party.” Moreover, no one has even argued here that "racial prejudice cements white majorities" in Republican strongholds but not in Democratic strongholds. Nor has anyone argued that "the Republican and Democratic Parties are mere proxies for racial and ethnic groups in Texas.” Dissenting Op. at 833. Finally, we expressly reject Judge Higginbotham's insistence that such views underlie this majority opinion.

. In particular, the Court noted that there was nothing in the record suggesting that Blacks were "not allowed to register to vote, to choose the political party they desired to support, to participate in its affairs or to be equally represented on those occasions when legislative candidates were chosen.” Whitcomb, 403 U.S. at 149, 91 S.Ct. at 1872. Moreover, there was no evidence that Blacks in Marion County “were regularly excluded from the slates of both major parties, thus denying them the chance of occupying legislative seats." Id. at 150, 91 S.Ct. at 1872. Instead, the record revealed that under the multimember district in Marion County, two Black senators and seven Black representatives had been elected over an eight year period.

. To buttress his claim that only "interest group politics” are at work in the Texas counties at issue in this case, Judge Higginbotham points to the Seventh Circuit’s decision in Baird v. Consolidated City of Indianapolis, 976 F.2d 357 (7th Cir.1992), and quotes Judge Easterbrook’s statement that ”[t]he 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.” Dissenting Op. at 832. What Judge Higginbotham fails to disclose, however, is the context in which Judge Easterbrook made this statement.

In Baird, Blacks challenged the method by which Marion County elected the twenty-nine members of its city-county council. Twenty-five of the city-county council seats were filled by elections from single-member districts, while four of the seats were filled by at-large elections. At the time of trial, Blacks constituted 21.28% of Marion County's population and approximately 60% of the population in seven of the twenty-five (or 24.14% of all twenty-nine seats). Thus, while Blacks had an "undiminished right to participate in the political process” and elect representative of their choice, see 976 F.2d at 359, they were not guaranteed the opportunity to elect representatives in the four at-large seats.

Faced with this evidence, the district court granted summary judgment in favor of the defendants, concluding that, ”[i]f [Black] voters are politically cohesive, they will elect candidates of their choice and obtain representation in the Council exceeding their numbers in the electorate; and if they are not cohesive, they cannot satisfy the requirements for relief under Gingles. Id. The Seventh Circuit agreed, noting that under the current election plan in Marion County, "losses by the candidates black voters prefer may have more to do with politics than with race." 976 F.2d at 361. The Seventh Circuit further observed, that ”[t]he 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.” Id. (emphasis added).

By citing Baird, Judge Higginbotham suggests that circumstances surrounding the election of city-council members in Marion County, Indiana are like the circumstances surrounding the election of district judges in the nine Texas counties at issue in this case. Even a cursory reading of the case, however, reveals the marked differences between the two political landscapes. Texas does not elect its district court judges from single member districts, and the minority groups in the various counties, even if they were 100% politically cohesive, would not be guaranteed the opportunity to elect representatives even approaching their numbers in the community.

Thus, .our conclusion about the significance of partisan voting patterns in this case says little about the significance of partisan voting—or interest group politics—under other circumstances. We are not suggesting that, in a situation like Baird, where a minority group was guaranteed through subdistricting the opportunity to elect representatives of its choice roughly in proportion to its population in the community, the minority group’s failure to achieve proportional representation—due to interest group politics—would support á vote dilution claim. In such a case, members of the minority group would simply be unable to demonstrate, in the words of Section 2, that they, had "less opportunity ... to participate in the political process and to elect representatives of their choice.” In the context of a case like Baird, we would agree with Judge Easterbrook’s statement that ”[t]he Voting Rights Act does not guarantee that nomi*810nees of the Democratic Party will be elected, even if black voters are likely to favor that party's candidates.” 976 F.2d at 361.

. In her brief on remand, Judge Wood argues that Dr. Taebel’s regression analysis, "which factored in the political party of the candidate," is the “only statistical proof that could accurately explain the outcomes in the partisan races being analyzed.” Judge Wood’s Brief on Remand at 26 (emphasis in original). Given Dr. Taebel’s express disclaimer that he was not explaining why voters vote the way they do, we reject this argument.

. In his dissent, Judge Higginbotham suggests that there was also racially polarized voting in Whitcomb. Dissenting Op. at 824. To get to this conclusion, of course, Judge Higginbotham has to combine the evidence that the "ghetto area"—an area which included whites and blacks—voted overwhelmingly Democratic with evidence that Republican candidates usually won. And, although he states, citing Whitcomb, 403 U.S. at 149-53, 91 S.Ct. at 1872-1874, that "white voters consistently supported Republicans,” the Supreme Court never mentioned this fact. The only voting behavior mentioned by the Court in Whitcomb was that the "ghetto voted overwhelmingly Democratic.” Thus, while the Court did focus on the fact that ghetto Blacks in Whitcomb were politically cohesive, the Court did not, under a straightforward reading of the case, focus on the existence or extent of racially polarized voting in Marion County elections.

. In her brief on remand, Judge Wood suggests that minority voters can show an unequal opportunity to participate in the political process only by producing evidence of formal barriers— i.e., that they are not allowed to register or vote, to choose the political party they desire to support, to participate in its affairs, or to be represented when candidates are chosen or slated. Judge Higginbotham would agree. See Dissenting Op. at 823. In response to this suggestion, we note only that, under the amended Section 2, the concept of unequal opportunity to participate in the political process is a broad concept. Indeed, the whole vote dilution principle is premised on the notion that “[t]here is more to the right to vote than the right to mark a piece of paper and drop it in a box or the right to pull a lever in a voting booth. The right to vote includes the right to have the ballot counted.... It also includes the right to have the vote counted at full value without dilution or discount." Reynolds v. Sims, 377 U.S. 533, 555 n. 29, 84 S.Ct. 1362, 1378 n. 29, 12 L.Ed.2d 506 (1964) (quoting Justice Douglas). Thus, the fact that minorities are not inhibited from becoming candidates or overtly hindered in casting their votes is not dispositive of whether they have an equal opportunity to participate in the political process. See S.Rep. at 30 n. 120, 1982 U.S.C.C.A.N. at 208 n. 120.

. As Dr. Taebel’s historical testimony alludes to, before 1980, minority-preferred candidates lost in Democratic primary elections—generally, to white Democrats. See White v. Regester, 412 U.S. at 767, 93 S.Ct. at 2340. In 1989, we are told, minority preferred candidates may make it to the general election, but once there, they lose to white Republicans. From the vantage point of minority voters—which is the vantage point of Section 2—it is difficult to see how the arrival of a two party system in Texas has altered their ability to participate in the political process and elect candidates of their choice.

. The State Defendants and Judge Entz suggest that, because of the state interests asserted in this case, the Plaintiffs were required to prove their vote dilution claims by clear and convincing evidence. We reject this suggestion. They offer no authority for increasing the Plaintiffs’ burden of proof.