Cottier v. City of Martin

Full title: PEARL COTTIER and Rebecca Three Stars, Plaintiffs, v. CITY OF MARTIN; Todd…

Court: United States District Court, D. South Dakota, Western Division

Date published: Mar 22, 2005

2005 SD 8, Civ. 2002-5021 (D.S.D. Mar. 22, 2005)

Opinion

2005 SD 8, Civ. 2002-5021.

March 22, 2005

Bryan L. Sells, Laughlin McDonald, American Civil Liberties Union Foundation, Atlanta, GA, Patrick K. Duffy, Duffy Duffy, Rapid City, SD, Attorneys for Plaintiffs.

Donald P. Knudsen, James S. Nelson, Sara Frankenstein, Gunderson, Palmer, Goodsell Nelson, LLP, Rapid City, SD, Attorneys for Defendants.

MEMORANDUM OPINION AND ORDER

KAREN E. SCHREIER, District JudgePage 2

INTRODUCTION

[¶ 1] Plaintiffs allege that the City of Martin Ordinance 122 dilutes the voting strength of Indians by fragmenting the Indian voters into three wards, which has the result and effect of denying the right of Indians to vote on account of race in violation of § 2 of the Voting Rights Act (VRA) of 1965. Plaintiffs also allege that Ordinance 122 was enacted and is being maintained with the discriminatory purpose of denying or abridging the right of Indians to vote on account of race or color or membership in a language minority in violation of plaintiffs' rights guaranteed by § 2 of the VRA, as amended, 42 USC § 1973, and the Fourteenth and Fifteenth Amendments to the Constitution of the United States. After considering the evidence admitted during an eleven-day court trial, the court determines by a preponderance of the evidence the following facts and conclusions of law.

I. Parties

[¶ 2] Plaintiffs Pearl Cottier and Rebecca Three Stars are Indians, qualified electors, members of the Oglala Sioux Tribe, and residents of Martin, South Dakota. T. I p. 238. T. III pp. 576, 578. Ex. 180 p. 7.

[¶ 3] Defendant City of Martin is a municipality located in southwestern South Dakota. Id. Defendants Todd Alexander, Rod Anderson, Scott Larson, Don Moore, Brad Otte, and Molly Risse are City of Martin council members. Don Moore has since been replaced by Ellis Ray Hicks. T. VII, p. 1458. Defendant Janet Speidel is the former City of Martin Finance Officer. She has since been replaced by Beth Strain. T. X, p. 2112.

II. History of the City of Martin Redistricting

[¶ 4] The city of Martin is in Bennett County, which is located in southwestern South Dakota near the Nebraska border. Bennett County is surrounded to the north and west by the exterior boundaries of the Pine Ridge Indian Reservation and to the east by the Rosebud Reservation. Although Bennett County was part of the Pine Ridge Reservation¹ under the Act of March 2, 1889, it was later opened up for settlement. As a result, only the unextinguished allotted lands in Bennett County are considered "Indian country" within the definition of 18 USC § 1151. Today, the Oglala Sioux Tribe, based on the Pine Ridge Reservation, extends its services to tribal members living within Bennett County. Two Oglala Sioux tribal council representatives are elected to represent the LaCreek District, which covers a significant portion of Bennett County.

1 The Act of March 2, 1889, set apart the Pine Ridge Reservation, encompassing what were later organized as three full counties (Bennett, Washabaugh, and Shannon). United States ex rel. Cook v. Parkinson, 396 FSupp 473, 477 (D.S.D.), aff'd, 525 F2d 120, 124 (8th Cir. 1975). By a subsequent act of Congress, the Pine Ridge Reservation was diminished and parts of Bennett County were opened up for settlement. Id. at 489. Thus, the Pine Ridge Reservation now consists of Shannon County and the area formerly known as Washabaugh County, now known as the portion of Jackson County that lies south of the White River.

[¶ 5] Martin is a small city, which according to the 2000 census had a total population of 1078 persons and a voting-age population of 737 persons. The city covers an area slightly greater than one-half square mile. The Indian population in Martin is 485, which is 44.71 percent² of the total population and 36 percent of the voting-age population according to the 2000 census.

2 The 2000 Census was the first federal census to allow respondents to identify themselves with more than one racial group. This court will consider all individuals who identify themselves as Native American, including those who identify with more than one group in light of the Supreme Court decision in Georgia v. Ashcroft, 539 U.S. 461, 123 S. Ct. 2498, 2507 n. 1, 156 L. Ed. 2d 428 (2003).

[¶ 6] Historically, the residents of the city of Martin have elected a mayor who ran at-large for a two-year term on a non-partisan ballot. In addition, Martin was divided into three wards, which each elected two city council members to staggered two-year terms on a non-partisan ballot. The record is unclear as to when the ward lines were initially drawn, but both parties agree the ward lines had not changed for at least 47 years. By 2001, the wards within the city were not within the requisite variation of population.

[¶ 7] The Martin City Council has the power and duty under South Dakota law to redistrict ward boundaries following the decennial federal census. The city contracted with the Black Hills Council of Local Governments (BHCLG) to refigure the wards so as to be in compliance with the one-person-one-vote requirement. BHCLG initially used incorrect population data when drawing the new wards. The city council, unaware of the mistake made by the BHCLG, adopted the redistricting recommendations submitted by BHCLG in Ordinance 121 on January 16, 2002.

[¶ 8] Upon publication of the new boundaries in the local newspaper, city residents suspected that the boundaries were flawed and contacted their attorneys for assistance. The attorneys analyzed Ordinance 121 and concluded that the new ward boundaries were severely malapportioned in violation of the one-person-one-vote principle of the Fourteenth Amendment and that the wards unlawfully fragmented the Indian population in Martin in violation of Section 2 of the Voting Rights Act. These concerns were communicated to BHCLG by letter dated March 7, 2002, with a copy to Martin's Mayor Kuxhaus. The city council requested BHCLG to redraw the wards to correct the one-person-one-vote problem. A new map was submitted to the city council. On March 12, 2002, plaintiffs' attorneys received a copy of the revised redistricting plan drafted by BHCLG. Plaintiffs believed that this plan did not correct the fragmentation problem, and they communicated that concern to Mayor Kuxhaus in a letter dated March 12, 2002.

[¶ 9] The City Council, although aware of plaintiffs' fragmentation concerns, moved ahead with the adoption of the March 8 plan as Ordinance 122. Like its predecessor plan, Ordinance 122 divides the City into three wards, none of which contains an Indian majority. The total population and voting-age population (VAP) figures under Ordinance 122 are summarized as follows:

Ordinance 122 Statistics

Ward Total Indian Percent VAP Indian VAP % Indian Population Population Indian VAP

1 352 165 46.88% 236 90 38.14% 2 361 177 49.03% 237 86 36.29% 3 365 143 39.18% 264 90 34.09%

Ordinance 122 took effect on May 8, 2002, and is the plan currently in effect in Martin. A map of the adopted Ordinance 122 follows as Figure 1.

Indian voters submitted a petition to have Ordinance 122 referred to the voters as a ballot issue. City Finance Officer Speidel reviewed the petition, determined that the petition did not have enough valid signatures, but waited to notify those submitting the petition of the defect until the deadline for petitioning for ballot initiatives had passed.

[¶ 10] Plaintiffs brought suit on April 3, 2002, alleging that Ordinance 121 violated the one-person-one-vote requirement under the Equal Protection Clause of the Fourteenth Amendment. After trial, the court dismissed the complaint as moot. The court found that Ordinance 121 had been repealed by Ordinance 122, which equally redistributed the population into three wards, and that plaintiffs no longer had an interest in an actual ongoing case or controversy. Plaintiffs then moved to supplement or amend their complaint to include the allegations currently pending before the court regarding Ordinance 122. The court granted plaintiffs' motion to supplement their complaint.

III. Section 2 of the Voting Rights Act

[¶ 11] Section 2 of the Voting Rights Act of 1965, as amended, prohibits the use of any voting practice which "results in a denial or abridgement of the right of any citizen of the United States to vote on account of race or color" or membership in a language minority. 42 USC §§ 1973(a), 1973b(f)(2); Thornburg v. Gingles, 478 US 30, 44, 106 SCt 2752, 2763, 92 LEd2d 25 (1986). A violation of § 2 is established "if, based on the totality of the circumstances, it is shown that . . . [members of a protected minority group] have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice." 42 USC § 1973(b). The voting strength of a politically cohesive minority group can be diluted either "by fragmenting the minority voters among several districts where a bloc-voting majority can routinely outvote them, or by packing them into one or a small number of districts to minimize their influence in the districts next door." Johnson v. De Grandy, 512 US 997, 1007, 114 SCt 2647, 2655, 129 LEd2d 775 (1994). Both the dispersal of Indians into districts in which they constitute an ineffective minority of voters or the concentration of Indians into districts where they constitute an excessive majority may dilute racial minority voting strength. Voinovich v. Quilter, 507 US 146, 154, 113 SCt 1149, 1155, 122 LEd2d 500 (1993).

[¶ 12] The Supreme Court has established a test to prove vote dilution through the use of multimember districts under § 2 of the Voting Rights Act:

First, the minority group must be able to demonstrate that it is sufficiently large and geographically compact to constitute a majority in a single-member district. . . . Second, the minority group must be able to show that it is politically cohesive. . . . Third, the minority 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 US at 51. Upon satisfying these three factors, the court must then consider the totality of the circumstances "to determine, based upon a searching practical evaluation of the past and present reality whether the political process is equally open to minority voters. This determination 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." Id. at 2781. A violation of § 2 is established "if, based on the totality of the circumstances, it is shown that . . . [members of a protected minority group] have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice." 42 USC § 1973(b).

A. Sufficiently Large and Geographically Compact

[¶ 13] Under the first Gingles factor, plaintiffs must demonstrate that the minority is sufficiently large and geographically compact to constitute a majority in a single-member district. Gingles, 478 US at 50. Small and dispersed minority groups undermine the ability to create a district that would remedy the grievance. Sanchez v. Colorado, 97 F3d 1303, 1311 (10th Cir. 1996). It considers whether the court can "fashion a permissible remedy in the particular context of the challenged system." Sanchez, 97 F3d at 1311. When requiring proof of this factor, the Court noted that:

[u]nless minority voters possess the potential to elect representatives in the absence of the challenged structure or practice, they cannot claim to have been injured by that structure or practice. . . . Thus, if the minority group is spread evenly throughout a multimember district, or if, although geographically compact, the minority group is so small in relation to the surrounding white population that it could not constitute a majority in a single-member district, these minority voters cannot maintain that they would have been able to elect representatives of their choice in the absence of the multimember electoral structure.

Gingles, 478 US at 50 n. 17.

[¶ 14] Plaintiffs introduced into evidence three illustrative redistricting plans which were drawn by their demographic expert, William Cooper. Cooper's Plan A and Plan B create at least one additional majority-Indian ward. Ex. 180. Illustrative Plan A is shown here:

Illustrative Plan A divides the city into three dual-member wards, one of which is a majority Indian ward. Thus, two majority-Indian city council seats based on VAP have been created. The following table summarizes the total population and VAP for Cooper's Illustrative Plan A.

Population Percent Figure Dual-Race Indian

Ward 1(VAP) 209 54.55% (total population) 351 66.95% Ward 2(VAP) 248 29.03% (total population) 356 37.64% Ward 3(VAP) 280 28.21% (total population) 371 30.46%

Illustrative Plan B is shown here:

Illustrative Plan B divides the city into six single-member wards, two of which are majority-Indian. Thus, a total of two majority-Indian seats based on VAP have been created under Illustrative Plan B. The table summarizes the total population and VAP for Illustrative Plan B.

Population Percent Figure Dual-Race Indian

Ward 1(VAP) 114 53.51% (total population) 180 67.78% Ward 2(VAP) 110 52.73% (total population) 187 63.64% Ward 3(VAP) 124 25.81% (total population) 173 32.95% Ward 4(VAP) 119 30.25% (total population) 180 40.56% Ward 5(VAP) 128 27.34% (total population) 181 30.94% Ward 6(VAP) 142 30.28% (total population) 177 31.07%

[¶ 15] Illustrative Plan C eliminates all ward boundaries and implements an alternative voting system called "limited voting." Under this system, three members of the city council would be elected at-large in each election, and voters would be allowed to case a single vote for their candidate of choice. Under such a system, the threshold of election when three seats are available is 25 percent. See, e.g., Douglas J. Amy, Behind the Ballot Box: A Citizen's Guide to Voting Systems, 125-31 (2000). Thus, the potential for Indians to elect two members of the city council exists under Illustrative Plan C.

[¶ 16] Defendants argue that plaintiffs have failed to satisfy the first Gingles factor because any proposed remedy requires an "effective" majority of at least 60 percent Indian VAP if a majority-minority ward were to be created. The court disagrees. First, the law does not definitively require establishing more than a "majority" district. Gingles, 478 US at 50 (requiring proof of a majority); Valdespino v. Alamo Heights Ind. Sch. Dist., 168 F3d 848, 852-53 (5th Cir. 1999) (requiring proof that the minority group exceeds 50 percent of the relevant population); Solomon v. Liberty County, Fla., 899 F2d 1012, 1013, 1018 (11th Cir. 1990) (holding that plaintiffs satisfied the first factor of Gingles where minority voters made up 49 percent of the total population, 51 percent of the VAP and 46 percent of the registered voters).

[¶ 17] Second, any necessary VAP above a majority is required only at the remedial stage of litigation. Indeed, a 60 percent guideline is a general remedial goal and "is irrelevant to the first part of the Thornburg tripartite threshold test for liability." Magnolia Bar Ass'n v. Lee, 793 FSupp 1386, 1397 (S.D. Miss. 1992). See also Neal v. Coleburn, 689 FSupp 1426, 1438 (E.D. Va. 1988) ("Contrary to defendants' contention, the general 65% guideline for remedial districts is not a required minimum which the plaintiffs must meet before they can be awarded any relief under § 2 of the Voting Rights Act. Rather, the 65% standard is a flexible and practical guideline to consider in fashioning relief for a § 2 violation."). In Dickinson v. Indiana State Election Board, the Seventh Circuit noted that although several cases have recognized a need for a supermajority of minority voters in the proposed district:

the Supreme Court requires only a simple majority of eligible voters in the single-member district. The court may consider, at the remedial stage, what type of remedy is possible based on the factors traditionally examined in single-member districts, such as minority voter registration and turn-out rates. . . . But this difficulty should not impede the judge at the liability stage of the proceedings.

933 F2d 497, 503 (7th Cir. 1991) (citations omitted). Because the current case is at the liability stage of the proceedings, the court concludes that proof of a simple majority is all that is required.

[¶ 18] Defendants rely on African Americans Voting Rights Legal Defense Fund, Inc. v. Villa, 54 F3d 1345, 1348 n. 4 (8th Cir. 1995) to support their contention the majority-minority district must have at least 60 percent Indian VAP before minority voters are afforded an effective majority. Villa describes a "safe ward" as a ward in which a minority has a practical opportunity to elect the candidate of its choice and notes that something in the vicinity of 65 percent of the total population or 60 percent VAP are the target percentage. Id. The question before the court in Villa, however, was not whether a 60 percent VAP supermajority was necessary to meet the first Gingles precondition. In fact, in Villa the parties did not dispute that the three Gingles preconditions were satisfied. Rather the dispute centered around whether proportionality had been established and if so, whether proportionality defeated the plaintiffs' § 2 claim despite satisfaction of the three Gingles factors. Id. at 1352. The issue here is not one of proportionality which may be relevant as part of the totality-of-circumstances review, but rather whether the first Gingles precondition has been met, namely whether the minority voters constitute "a majority in a single-member district." Gingles, 478 US at 50 n. 17.

[¶ 19] Even if Villa did require proof of an effective majority at the first Gingles factor, plaintiffs meet this requirement. The court in Villa concluded that " either 60% of the voting age population or 65% of the total population is reasonably sufficient to provide black voters with an effective majority." Villa, 54 F3d at 1348 n. 4 (emphasis added). Here, Illustrative Plan A creates a dual-member ward that exceeds 65 percent Indian total population. And Illustrative Plan B creates one single-member ward that exceeds 65 percent Indian total population and a second single-member ward that is close to 65 percent Indian total population. Because Villa does not require proof of both a 60 percent VAP and 65 percent total population, plaintiffs have met their burden of providing at least one illustrative plan with a minority population that constitutes an effective majority in at least one additional single member district.

[¶ 20] Defendants also argue that plaintiffs' proposed wards are too fragile to constitute a workable remedy. Defendants rely on the Eighth Circuit opinion in Stabler v. County of Thurston, Neb., 129 F3d 1015, 1025 (8th Cir. 1997), which found that plaintiffs' proposed plans were too fragile because if four or five Indians moved from the proposed majority-minority districts, and they were replaced by non-Native Americans, the majority-minority composition would be destroyed. Id. Defendants allege that if only ten Indian VAP moved out of Illustrative Map A Ward 1 and were replaced by white VAP, Ward 1's majority-minority status would be destroyed.

[¶ 21] In Stabler, plaintiffs were challenging an at-large method of election for the school and village boards and sought the implementation of either two three-member districts or six single-member districts for the village board and two single-member districts and one three-member district for the school board. The district court found that "the challenged jurisdictions are of such small size that no feasible districting plan can be drawn. The influx of as few as one or two households into or out of a district would disturb the incredibly delicate balance demanded by such a districting scheme." Stabler v. County of Thurston, Nebraska, CIV 93-00394, p. 20 (unpublished opinion).

[¶ 22] Unlike Stabler, this court is not faced with a challenged jurisdiction that is so small that no feasible districting plan can be drawn. In fact, Martin has been operating under a three-ward system for over 40 years. And unlike Stabler, it would take more than the influx of one or two households to disturb the balance under the proposed districting scheme. In fact, it would take 60 Indians of the general population (or 10 VAP) to move out of Ward I and be replaced by non-Indians. This is unlikely because based on census data, the net effect of mobility in Martin is that the Indian population is increasing while the non-Indian population is declining. Between 1990 and 2000, the overall population of the city of Martin declined by 24 persons, while the Indian population in Martin increased by 111 persons. Ex. 180, at 11. Thus, the court finds that Plans A and B are neither fragile nor unworkable.

[¶ 23] Defendants also argue that plaintiffs did not satisfy the first Gingles factor because they failed to propose majority-minority districts that are compact. Defendants contend that plaintiffs' proposed districts are irregularly and bizarrely shaped. and that race was a primary concern in drawing the maps. The first Gingles factor does not require "some aesthetic ideal of compactness," but rather looks at whether the minority population is sufficiently compact to constitute a majority in a single-member district. Clark v. Calhoun County, Miss., 21 F3d 92, 95 (5th Cir. 1994) ( Clark I). The constitution does not require regularity of district shape. Bush v. Vera, 517 US 952, 963, 116 SCt 1941, 1953, 135 LEd2d 248 (1996); Sanchez, 97 F3d at 1312. "Nor . . . is the decision to create a majority-minority district objectionable in and of itself." Vera, 116 SCt at 1953. The court must determine "whether the affected minority is diffused and thus politically ineffective, not whether the area by which it is bound is geographically dense." Sanchez, 97 F3d at 1312.

[¶ 24] Cooper's Plans A and B follow census blocks, follow marked streets, exhibit more than point contiguity, are of equal population, and recognize the community of interest known as "Snob Hill."³ Cooper testified that he considered these traditional race-neutral districting principles when designing Plans A and B. These traditional redistricting principles have been recognized by the United States Supreme Court. See, e.g., Miller v. Johnson, 515 US 900, 916, 115 SCt 2475, 132 LEd2d 762 (1995) (identifying "respect for political subdivisions or communities defined by actual shared interests" as traditional districting principle); Shaw v. Reno, 509 US 630, 651-52, 113 SCt 2816, 2828-29, 125 LEd2d 511 (1995) (identifying population equality as a traditional districting principle). Furthermore, no point within a ward in either Plan A or B is more than a mile from the farthest point within that same ward. And similar to the map under Ordinance 122, Plans A and B do not take incumbency protection into consideration. Although the proposed wards in Plans A and B are not as regular in shape as the wards under Ordinance 122, Plans A and B are not so irregular that one could differentiate them from many existing electoral districts throughout the state and nation.

3 "Snob Hill" is the popular name for the area in the southeast corner of Martin.

[¶ 25] The proposed Plans A and B are not so irregular on their face that they appear to be solely an effort to segregate races for the purposes of voting. The United States Supreme Court has indicated that examples of bizarre shape that would raise concerns include district shapes that look "like a jigsaw puzzle . . . in which it might be impossible to get the pieces apart;" and a district that looks like "a sacred Mayan bird, with its body running eastward . . . [s]pindly legs reach south . . . while the plumed head rises northward . . . an open beak appears to be searching for worms . . . [and] [h]ere and there, ruffled feathers jut out at odd angles." Vera, 116 SCt at 1958-59. Plans A and B do not match these descriptions.

[¶ 26] Defendants contend that Plan A appears virtually identical to Cooper's proposed plan found to be a racial gerrymander in Stabler, 129 F3d at 1025. In Stabler, the Eighth Circuit found that the "bizarre shape of the proposed districts, considered in combination with the racial and population densities of the proposed districts, support the district court's finding that race was the predominant factor in drawing the proposed districts to create a majority-minority single-member district and that the proposed redistricting resulted in gerrymandered districts." Id. at 1025. Thus, the Eighth Circuit did not rely solely on the shape of the proposed districts to conclude that race was the predominant factor in drawing the proposed districts, which is what defendants are asking this court to do.

[¶ 27] The court accepts Cooper's explanation that he applied traditional districting principles. Cooper has been a geographic information system (GIS) consultant for over 15 years. He works primarily on mapping for voting rights legislation, including providing testimony in 24 federal trials and declarations or depositions in an additional 19 cases. He has extensive experience in drawing redistricting maps, having drafted almost 500 using GIS technology and about as many prior to the availability of such technology. Cooper has prepared "thousands and thousands" of maps for almost 500 jurisdictions. He has drawn maps for about 50-75 small jurisdictions (population under a thousand), including one town with a population of 220. He has qualified as an expert witness on the topics of demographics, redistricting, and census data analysis. T. II p. 346-50; Ex. 180 p. 1-3. A review of Plans A and B reveals that the plans follow census blocks, follow marked streets, exhibit more than point contiguity, create wards of equal population, recognize the community of interest known as "Snob Hill," and create compact and contiguous wards. The court finds Cooper's testimony credible and probative.

[¶ 28] Furthermore, plaintiffs need not propose a complete remedy at this stage. See Gingles, 478 US at 50 n. 17 (plaintiffs must show that minority voters "possess the potential to elect representatives in the absence of the challenged structure or practice") (emphasis added). Plaintiffs' proposed districts must simply demonstrate the feasibility of drawing a majority-minority district and are not cast in stone. Houston v. Lafayette County, Miss., 56 F3d 606, 611 (5th Cir. 1995). See also Dickinson, 933 F2d at 503 (completeness of remedy considered at the remedial stage of litigation). "If a § 2 violation is found, the [city] will be given the first opportunity to develop a remedial plan." Id. See also Clark v. Calhoun County, Miss., 88 F3d 1393, 1407 (5th Cir. 1996) ( Clark II) (county's challenge to the remedy was not ripe for review because the county was "free, within limits, to develop a different remedial plan from those proposed by the plaintiffs"); Sanchez, 97 F3d at 1315 ("drawing the necessary district is not [plaintiffs'] onus because the [city] must be given the first opportunity to fashion a remedy").

[¶ 29] Finally, defendants contend that plaintiffs' plans are race based, and that as a result, strict scrutiny applies. The Supreme Court has recognized that "drawing racial distinctions is permissible where a governmental body is pursuing a `compelling state interest.'" Shaw v. Hunt, 517 US 899, 908, 116 SCt 1894, 1902, 135 LEd2d 207 (1996) (Shaw II). The Supreme Court has assumed without deciding that compliance with the results test of § 2 of the VRA is a compelling state interest. See Shaw II, 116 SCt at 1905; Miller, 515 US at 920-21, 115 S. Ct. at 2490-91; Vera, 517 US at 977. Justice O'Connor in a separate concurrence found that compliance with VRA in fact is a compelling state interest. Vera, 116 SCt at 1968. A city may then pursue that compelling state interest and create a district that is narrowly tailored to remedy the VRA liability. Id. at 1970. Thus, the consideration of race when proposing a plan does not necessarily invalidate the plan. The court finds as a matter of law that plaintiffs have met their burden of showing that a permissible remedy in the context of the challenged system can be fashioned, which is the first Gingles factor.⁴

4 Plaintiffs proffer Illustrative Plan C as an alternative which would allow Indians to elect preferred candidates in numbers more proportional to the Indian population. Although Illustrative Plan C's limited voting plan cannot be a racial gerrymander, the court does not believe it has authority to order such a remedy. Cane v. Worcester County, Maryland I and II, 35 F3d 921 (4th Cir. 1994), 59 F3d 165 (4th Cir. 1995). See also SDCL 9-11-5 (requiring voters to choose form of government).

B. Minority Political Cohesiveness

[¶ 30] The second Gingles factor requires plaintiffs to show that the minority group is politically cohesive. Gingles, 478 US at 51. This ensures that the minority group at issue has distinctive minority group interests. Id. Without such distinct interests, unequal opportunity in the political arena cannot harm the minority group. Id. "A showing that a significant number of minority group members usually vote for the same candidates is one way of proving the political cohesiveness necessary to a vote dilution claim, . . . and, consequently, establishes minority bloc voting within the context of § 2." Id. at 2769. Voting patterns are the central focus. Campos v. City of Baytown, Tex., 840 F2d 1240, 1244 (5th Cir. 1988). See also Ruiz v. City of Santa Maria, 160 F3d 543, 552 (9th Cir. 1998) (candidate who received sufficient votes to be elected if the election were held among the minority group was considered the minority-preferred candidate even if he received less than 50 percent of the minority vote).

[¶ 31] Proving this factor typically requires statistical evaluation of elections. Campos, 840 F2d at 1244-45. "The number of elections that must be studied in order to determine whether voting is polarized will vary according to pertinent circumstances." Gingles, 478 US at 57 n. 25. Courts have relied on various statistical methods. See, e.g., Houston, 56 F3d at 611 (use of bivariate ecological regression and extreme case analysis); Clark II, 88 F3d at 1397 (expert employed regression and homogenous precinct analysis); Sanchez, 97 F3d at 1317-18 (court considered both ecological and multivariate regression analysis). This court will examine each method used by the experts in this case.

[¶ 32] The first method is homogeneous precinct analysis (HPA) or extreme case analysis. This technique examines voting behavior in precincts that are closest to being racially or ethnically homogeneous in population, typically 90 percent or more. The vote in the most heavily minority precincts is used as an estimate of minority voting behavior and the voting behavior in the most heavily majority precincts is used as an estimate of majority voting behavior. HPA is based directly on voter behavior and requires no statistical inference. Ex. 186 p. 7-8.

[¶ 33] A second technique is bivariate ecological regression analysis (BERA). Under this technique, election results are correlated with census data or some other measure of the racial or ethnic composition of the electorate to generate estimates of the voting behavior of majority and minority voters. Ex. 186 p. 5.

[¶ 34] A third technique is ecological inference (EI), or the King method. It assumes that the actual votes of two groups for two particular candidates are based on fixed underlying propensities, but vary from precinct to precinct in random ways. It estimates the underlying propensity of each group to turn out for an election and to vote for a particular candidate using the estimation technique of maximum likelihood. Ex. 186, p. 12-13. EI can be used to generate estimates of voting behavior within each ward as well as across all wards in a voting district. See, e.g., Gary King, A Solution to the Ecological Inference Problem: Reconstructing Individual Behavior from Aggregate Data (1997), pg. 91.

[¶ 35] Certain elections are more probative of unequal electoral opportunity than others. Interracial elections are generally more probative than racially homogeneous elections because voters have a racial choice. See Gingles, 478 US 30 at 80-82 (relying exclusively on interracial legislative contests to determine whether a legislative redistricting plan diluted the black vote); United States v. Blaine County, Mont., 363 F3d 897, 911 (9th Cir. 2004) (contests between white and Indian candidates are most probative of bloc voting). Endogenous elections, contests within the jurisdiction and for the particular office that is at issue, are more probative than exogenous elections. See Sanchez, 97 F3d at 1317 (greater weight to endogenous elections). Recent elections are more probative than elections in the distant past. See Uno v. City of Holyoke, 72 F3d 973, 990 (1st Cir. 1995) (recent elections more probative); Meek v. Metropolitan Dade County, Fla., 985 F2d 1471, 1482-83 (11th Cir. 1993). When endogenous election data is sparse or unavailable altogether, reliance upon exogenous election data is appropriate. Gingles, 478 US at 57 n. 25 (in the absence of ideal data, "courts must rely on other factors that tend to prove unequal access to the electoral process" and "the fact that statistics from only one or a few elections are available for examination does not foreclose a vote dilution claim"). Exogenous elections for offices with comparable levels of importance within the community are entitled to more weight than dissimilar elections. See Black Political Task Force v. Galvin, 300 FSupp 2d 291, 308 (D. Mass. 2004). Here, the court will give more weight to local county-wide elections than state legislative, statewide or federal elections.

[¶ 36] Although interracial elections are highly probative of minority voting patterns, the court recognizes that the minority preferred candidate is not always a minority. See Lewis v. Alamance County, N.C., 99 F3d 600, 605-06 (4th Cir. 1996) (minority preferred candidate may be white). Thus, the court will not limit its consideration to interracial elections.

1. Dr. Cole's Analysis

[¶ 37] Plaintiffs' expert, Dr. Steven Cole, used HPA, BERA, and EI. Ex. 186 p. 4; 188, p. 8-9; T. III p. 617-18. Dr. Cole also conducted and evaluated exit polls of endogenous elections in the city of Martin. T. III p. 619-20, 667-68. He applied the dual-race method of identifying Indians. Dr. Cole examined federal, state, and county elections from 1996 through 2002. Plaintiffs' counsel provided identification of Indian candidates and voters within the city of Martin. Dr. Cole used countywide election data to estimate the behavior of city of Martin voters using HPA, BERA, and EI. Ex. 186 p. 3

[¶ 38] Dr. Cole generated tables to demonstrate his results. In Table 1, Dr. Cole used a regression analysis, which examines the relationship between the ward's racial composition and the candidate's vote share. The R-squared value demonstrates what percentage of the variables in a candidate's vote share can be predicted by race alone. It measures how close the precincts fall to the regression line and estimates Indian cohesion and white crossover voting. T. III p. 623-25; Ex. 186 p. 5. The P-value represents an analysis of variants by testing how well the regression model fits the data. If the P-value is less than .05, it is considered statistically significant. T. III, p. 641-42; Ex. 186 p. 9-12.

[¶ 39] The third column estimates the percentage of Indian voters voting for a candidate and the fourth column estimates the percentage of non-Indian voters voting for that candidate. Dr. Cole used two regression equations adjusted for turnout differences in groups to determine these amounts. The final column is the number of votes received by that candidate, which is obtained from election returns. Ex. 186 p. 28-31.

[¶ 40] In Table 2, Dr. Cole employed HPA. For this analysis, he relied on virtually all-white or all-Indian wards, generally over 90 percent. Wards with 90 percent or more non-Indian VAP provide an estimate of white crossover voting while wards with 90 percent of more Indian VAP measure Indian cohesion. T. III, p. 618-19; Ex. 186, p. 7-8. When drawing conclusions about cohesion, Dr. Cole averaged estimates of minority support across elections and relied on the overall pattern of results. Ex. 186, p. 24-27. When defining "cohesion," Dr. Cole does not employ a strict numerical threshold. Rather, according to Dr. Cole, he measures political cohesiveness on a "continuum that begins at 51 percent and goes all the way up to 100." T. III p. 647.

[¶ 41] Defendants argue that Dr. Cole erroneously finds cohesion if 50.1 percent of the Indian voters vote for the same candidate, noting that cohesion is absent only when two candidates tie. The court disagrees with this characterization of Dr. Cole's testimony. Dr. Cole explained that 50.1 percent represents the very beginning stages of cohesion but would be evidence of very weak cohesion. He repeatedly indicated that cohesion does not have a specific "cutoff point." T. III p. 696. In Dr. Cole's opinion, polarization exists in contests involving two candidates "when a majority of the voters of one race would elect a different candidate than would the majority of voters of the other race. . . . In head to head contests with more than two candidates, significant racial polarization is exhibited when a majority/plurality of the voters of one race would elect a different candidate than would a majority/plurality of voters of the opposite race." Ex. 186 p. 6.

[¶ 42] Dr. Cole did not analyze any endogenous contests for mayor or city council using HPA, BERA or EI. In Dr. Cole's opinion, because only three wards vote in the mayor contest and only one ward votes in each of the three city council contests, regression analysis is precluded. In addition, because none of the three city wards is largely all Indian or all non-Indian, HPA is also precluded. Ex. 186, pg. 11. The court agrees that city of Martin endogenous election data cannot be analyzed in a scientifically valid manner using HPA, BERA or EI. Thus, the court will examine exogenous election data.

[¶ 43] Dr. Cole analyzed four interracial, multi-candidate, county-wide, exogenous contests: the 2002 Bennett County school board election (5 candidates vying for 2 seats), 2002 county commissioner democratic primary (6 candidates vying for 3 slots on the general election ballot), 2002 county commissioner general election (9 candidates vying for 3 seats), and the 2000 Bennett County school board election (3 candidates vying for 2 seats). Ex. 186, p. 24, 28, 30-31. Dr. Cole found cohesion amongst Indian voters in all four contests using a BERA analysis. In the 2002 County school board election, Indian political cohesion was 100 percent for Flye and 64 percent for Burritt, both of whom were Indian candidates. In the 2002 county commissioner democratic primary election, Indian political cohesion was 89 percent for Sharp, 70 percent for Bettelyoun, and 69 percent for Ruff, all of whom were Indian candidates. In the 2002 general election contest for county commissioner, Indian political cohesion was 85 percent for Bettelyoun, 86 percent for Sharp (both of whom were Indian candidates), and 53 percent for Hammond. And in the 2000 County School Board contest, Indian political cohesion was 100 percent for Three Stars, an Indian candidate. Thus, the average estimate of Indian political cohesion was 79 percent for interracial, multi-candidate, county races.

[¶ 44] Dr. Cole analyzed four interracial, head-to-head, county-wide, exogenous contests: multiple races from the 2002 general election — county sheriff, county register of deeds, and county coroner, and the 2001 election for Bennett County school board. Using EI, Dr. Cole estimated the level of Indian political cohesion for Cummings (an Indian candidate) at 95 percent in the 2002 general election for sheriff; for Sterkel (the non-Indian candidate) at 56 percent in the 2002 general election for county Register of Deeds; for Mesteth (an Indian candidate) at 76 percent in the 2002 general election for county coroner, and for Three Stars (an Indian candidate) at 58 percent in the 2001 county school board election. Dr. Cole found Indian political cohesion in all four elections under an EI analysis. Thus, the average estimate of Indian political cohesion is 71 percent for interracial, single-candidate, county races.

[¶ 45] Three interracial, head-to-head, state exogenous contests were analyzed by Dr. Cole: the 1998 general election for governor,⁵ the 2002 democratic primary for governor, and the 2002 general election for attorney general. Using EI, Dr. Cole estimated the level of political cohesion for Hunhoff/Meeks (Meeks is an Indian candidate) at 55 percent in the 1998 general election for governor, for Volesky (an Indian candidate) at 51 percent in the 2002 Democratic primary for governor; and for Volesky (an Indian candidate) at 73 percent in the 2002 general election for Attorney General. Dr. Cole found Indian political cohesion in all three races under an EI analysis. The average estimate of Indian political cohesion is 59 percent for interracial, single-candidate, state races.

5 Although the race for governor itself was not interracial, this particular election involved an Indian lieutenant governor candidate running on the ticket withthe governor. The court finds that the analysis of this contest as interracial was appropriate under these special circumstances.

[¶ 46] Dr. Cole analyzed three general election contests for county-wide office involving only white candidates. Ex. 186, p. 25-27, 38-61. In the 2002 general election for county auditor, Dr. Cole estimated the level of political cohesion for Hudson (an Indian-preferred candidate) at 62 percent. In the 2000 general election for county sheriff, Dr. Cole estimated the level of political cohesion for Duke (the Indian-preferred candidate) at 61 percent. And in the 1998 general election for county auditor, Dr. Cole estimated the level of political cohesion for Williams (the Indian-preferred candidate) at 80 percent. The average estimate of Indian political cohesion for white-white, single-candidate, county offices is 67 percent.

[¶ 47] Dr. Cole analyzed 30 general election contests for 32 state or federal offices for the time period of 1996 to 2002 involving only white candidates. Ex. 186 p. 25-27, 38-61. Using Dr. Cole's EI analysis, the average level of Indian political cohesion was 82.2 percent for these contests.

[¶ 48] Across the 11 exogenous interracial elections (to select 16 candidates) that Dr. Cole analyzed, the average estimate of Indian political cohesion for the top preferred candidates was 73.7 percent. Across the 35 exogenous white-candidate only elections that Dr. Cole analyzed, the average estimate of Indian cohesion for the top preferred candidates was 81.5 percent. All of Dr. Cole's statistical results in every category demonstrate that Indians in the city of Martin are politically cohesive.

[¶ 49] In addition to his statistical analysis of exogenous races, Dr. Cole also conducted an exit poll at the combined municipal and school board election held on June 3, 2003. T. III 651:5-10. Two seats on the city council (Ward I and Ward III) and two at-large seats on the school board were on the ballot. To conduct the poll, Dr. Cole hired Indian and non-Indian poll takers from the local college and personally trained them regarding the proper procedures for conducting a poll. Dr. Cole's poll used a one-page instrument mimicking the ballot with questions added for demographic data. If a voter agreed to participate in the poll, the voter was given a form to complete in confidence. The poll takers attempted to achieve participation by every voter. Out of 293 total voters, of whom 87 were Indian voters as determined by plaintiffs, 114 voters participated in the exit poll for an overall response rate of 40.3 percent. Five responses were not valid, however, reducing the overall participation rate to 38.5 percent. Of the 109 exit poll participants, 68 self-identified themselves as Indian. T. III 664:4-16. Of the 87 Indian voters, 68 (78.2 percent) participated in the exit poll, which Dr. Cole found to be a representative sample. Non-Indian participation in the poll, however, was only (20.9 percent). Because of the low non-Indian participation, Dr. Cole recognized that the poll did not produce a representative sample of non-Indian voters. Dr. Cole had confidence in the estimates of Indian voter preference, but not as much confidence in the estimates of non-Indian voter preference. To compensate for the low non-Indian voter participation, he opted to use an arithmetic technique to estimate maximum non-Indian crossover voting. This technique assumes that the difference in votes between the actual number of votes for a candidate and the number of votes reported by Indian voters in the exit poll were all from non-Indian voters. In Dr. Cole's opinion, because it is unlikely that all remaining votes were from non-Indian voters, this estimate of non-Indian support for an Indian-preferred candidate likely results in an overestimate of non-Indian crossover voting.

[¶ 50] Both city council races were head-to-head, nonpartisan, and not interracial.⁶ The Ward I race involved two Indian candidates, where one candidate was the Indian-preferred candidate (Gotheridge), who according to the poll received 100 percent of the Indian vote, and the other Indian candidate (Risse) received 0 percent of the Indian vote. The Indian-preferred candidate received 35 votes, with 19 of those votes coming from Indian voters according to actual polling responses. Dr. Cole found the Ward I contest demonstrated cohesion. Ex. 188, p. 10-11, 15.

6 Both candidates for Ward I were Indian and both candidates for Ward III were non-Indian.

[¶ 51] In Ward III, the Indian-preferred candidate (Justus) received 85.7 percent of the Indian vote (12 votes out of 14 votes) while the other non-Indian⁷ candidate (Alexander) received 14.3 percent of the Indian vote (2 votes out of 14 votes). Of the 107 Ward III voters, 84 were non-Indian. Dr. Cole found the Ward III contest also demonstrated Indian voter cohesion. Ex. 188, p. 11-12, 16.

7 Defendants argue that Todd Alexander is an Indian candidate. Although Alexander testified during the trial that he considers himself to be an Indian person, under the test articulated by the court in United States v. Driver, 755 FSupp 885, 888 (D.S.D. 1991), he would not legally be considered an Indian person. Under thetest, tribal enrollment is the most important factor, although the court should also consider whether the individual receives assistance reserved for Indians, enjoys the benefits of tribal affiliation, or lives on the reservation and participates in Indian social life. Id. at 888-89. While Alexander has some Indian blood from a Cherokee great-great-grandmother, he is not enrolled in any tribe. T. VII 1450:20-1451:8. Alexander admitted that he does not receive any benefits usual to tribal affiliation.T. VII p. 1395-96, 1453. And the city of Martin, where Alexander resides, is not located within the exterior boundaries of the Pine Ridge Reservation. While Alexander does attend powwows and memorial services, these public events are not sufficient to support a finding that he is an Indian person.

[¶ 52] Dr. Cole also analyzed the exogenous county school board election in 2003 using exit polling. This contest was interracial and nonpartisan involving six candidates, two of whom were Indian, vying for two seats. The two Indian candidates were the Indian-preferred candidates, receiving 82.7 and 58.8 percent of the Indian vote. Dr. Cole opined the race reflected cohesive Indian voting. Ex. 188, p. 10-11, 14.

[¶ 53] The court finds that Dr. Cole qualifies as an expert in this case. He holds a Ph.D. in human experimental psychology. He has taught courses in research methods, computer analysis, scientific data, and psychology. He currently teaches at Emory University and has taught courses at the law school. He served as the director of research for Research Designs Associates, Inc., since 1982. He has worked as a consultant for numerous schools and organizations. He has published numerous papers and has testified about voting behaviors in many cases. T. III p. 614; Ex. 186 Attachment. The court finds that Dr. Cole's education, experience, knowledge, and skill qualify him to testify as an expert in this case.

2. Dr. Weber's Analysis

[¶ 54] Defendants' expert, Dr. Weber, stated in his supplemental declaration that neither he nor Dr. Cole was able to conduct standard BERA or HPA analysis for endogenous elections held within the three wards of the city of Martin, because of the lack of sufficient variation on the Indian percentage of the VAP across the three wards. Ex. 448, p. 25. Dr. Weber opines that exit poll data "would be the most appropriate data to employ to determine the preferences of Indian and non-Indian voters in the City of Martin." Ex. 449, p. 15. Dr. Weber concluded that he was unable to determine registration, turn-out, participation, cohesion, or polarization for the city of Martin. Ex. 448 p. 24-26.

[¶ 55] Dr. Weber did analyze Martin city council election history as provided by defendants. Dr Weber opined that since 1981, 27.2 percent of elections have been contested and Indian candidates or white candidates married to Indians have been elected "sometimes." Dr. Weber reported that Ward One has elected two Indian candidates, one in 1984 and one in 2000. Dr. Weber opined that one Indian candidate has won in Wards II and III, but both the candidates ran unopposed. Ex. 448, p. 26-37. Dr. Weber also noted that no Indian candidate has competed for the mayoral elections since 1980, although one non-Indian married to an Indian was appointed and then re-elected without opposition. Ex. 448, p. 38-41. Dr. Weber also opined that no expert could identify the Indian-preferred candidate and, consequently, no expert could demonstrate vote dilution. Ex. 448, p. 42.

[¶ 56] Dr. Weber replicated Dr. Cole's statistical analyses presented in his initial and rebuttal reports. In doing so, Dr. Weber identified that Dr. Cole used the wrong VAP data in his estimates, but that this did not significantly affect the analyses.⁸ Dr. Weber identified seven interracial, exogenous contests: the 1998 general election for governor, the 2001 school board election, the 2002 democratic primary for governor, the 2002 general election for attorney general, the 2002 general election for sheriff, the 2002 general election for register of deeds, and the 2002 general election for coroner. Ex. 450, Table 3. Under Dr. Weber's EI analysis, the average estimate of Indian political cohesion in the seven interracial elections is 70.6 percent.

8 Dr. Weber stated that it is clear that wards outside Martin are driving all of Dr. Cole's assumptions regarding the city of Martin because when Dr. Weber removed the city of Martin wards from his analysis of Table 1, the slope of the regression line is unchanged. Ex. 449 p. 3-15.

[¶ 57] Dr. Weber then analyzed 28 elections for county, state, and federal office involving only white candidates. Under Dr. Weber's EI analysis, the average estimate of Indian political cohesion in these 28 white candidate only elections is 88.2 percent. Dr. Weber's EI analysis of both interracial and white-only contests supports a conclusion that Indian's are politically cohesive in the city of Martin.

[¶ 58] The court finds that Dr. Weber qualifies as an expert. Dr. Weber has a doctorate in political science and has taught in that field for 36 years. He currently teaches at the University of Wisconsin where he was asked to fill an endowed faculty position. He has published numerous articles and was the editor of several political science journals. Dr. Weber has testified in several trials on the issue of voting rights. Ex. 451; T.V p. 956-974. The court finds that Dr. Weber's education, experience, knowledge, and skill qualify him as an expert in this case.

3. Reliability of Each Method

[¶ 59] In Sanchez, the court evaluated both experts' analyses and results to determine which reached the most reliable results. 97 F3d at 1316-1319. The court is not obliged to accept any parties' statistical evidence. Clark I, 21 F3d at 96.

[¶ 60] Dr. Cole employs four techniques: HPA, BERA, EI, and exit polling. Numerous courts, including the United States Supreme Court, have accepted HPA and BERA methods as reliable in § 2 cases. See Gingles, 478 at 52-53 (relying on single regression analysis, which the Court considered "standard in the literature for the analysis of racially polarized voting"). See, e.g., Old Person v. Cooney, 230 F3d 1113, 1123 (9th Cir. 2000) (relying on BERA); Rural West Tennessee African-American Affairs Council v. Sundquist, 209 F3d 835, 839 (6th Cir. 2000) (considering Dr. Cole's BERA and HPA); Harvell v. Blytheville Sch. Dist. No. 5, 71 F3d 1382, 1386 (8th Cir. 1995) (relying on regression analysis). The prevalence of both district and circuit courts relying on these methods demonstrates the wide acceptance of these analyses. See Teague v. Attala County, Miss., 92 F3d 283, 290 (5th Cir. 1996) (district court erred by disregarding the "established acceptance of regression analysis as a standard method for analyzing racially polarized voting"). Other courts have found that discounting this statistical analysis amounts to reversible error. In Sanchez, the Tenth Circuit reversed the district court for rejecting plaintiffs' BERA and HPA. 97 F3d at 1321. The Fifth Circuit also required the district court to consider this method and noted that the Supreme Court used this analysis. Teague, 92 F3d at 291. See also Houston, 56 F3d at 606. Precedent, therefore, supports the acceptance of Dr. Cole's and Dr. Weber's analyses.

[¶ 61] With regard to EI, the method employed by Dr. Cole and Dr. Weber, the court finds that EI is also a reliable method of analysis. Courts recently have recognized EI as a reliable improvement on ecological regression analysis. Rodriguez v. Pataki, 308 FSupp 2d 346, 387-88) (S.D.N.Y. 2004) (citing Georgia v. Ashcroft, 195 F. Supp. 2d 25, 69 (D.D.C. 2002), vacated on other grounds, 539 US 461, 123 SCt 2498, 156 LEd2d 428 (2003)) (recognizing King's ecological inference as having the "prospect of improving on ecological regression" despite its recency in voting rights litigation); United States v. Alamosa County, Colo., 306 FSupp 2d 1016, 1023 (D. Colo. 2004) (noting use of King's EI by experts Weber and Engstrom). Dr. Weber admits that EI is being used by experts today. T. V, p. 1029.

[¶ 62] Defendants contend that Dr. Cole improperly used EI to generate precinct level data from countywide data. Defendants contend that in Stabler, 129 F3d at 1025, the Eighth Circuit rejected such evidence. The Eighth Circuit, however, actually found that the district court did not err in rejecting plaintiffs' proffer of evidence of county-wide voting behavior to prove voter cohesiveness and bloc voting in the School Board and the Village Board elections. Id. The district court found that "the parties were unable to perform regression or homogenous precinct analysis on either School Board or Village Board elections. And while the analysis of county-wide exogenous elections was probative of the claims against Thurston County, the Court has serious reservations about drawing conclusions about racial voting behavior in the school district and Walthill from this analysis." Stabler, CV93-00394, p. 21. At the time Stabler was decided, the EI statistical method had not been developed by Dr. King. The EI method produces valid estimates of voting behavior at the precinct level, and Dr. Cole's figures are estimates of voting in the city of Martin wards only, not the county as a whole. Dr. Richard Engstrom explained that Dr. Cole's use of EI to extrapolate city voting behavior based partially on county-wide data was "something that [EI] allows you to do." T. VII, p. 1279, 1280-82. The court finds the estimates of city voting behavior at the ward level extrapolated partially from county-wide data analyzed under EI is a generally accepted method within the scientific community.

[¶ 63] Defendants also question Dr. Cole's expertise in the use of EI because he has previously testified that he was not an expert in EI. While Dr. Cole admits that he did not derive the equations for EI, he testified that he has used EI for some time, he understands it conceptually, and would be comfortable teaching EI to graduate students. The court finds that Dr. Cole is sufficiently familiar with the application of EI to properly apply it as an expert in this case.

[¶ 64] Defendants also allege that the experts used differing definitions of legal terms such as cohesion and polarization, and thus Dr. Cole's analysis should be found invalid. A court can accept numerical calculations of an expert, however, without adopting his legal conclusions regarding polarization. See Askew v. City of Rome, 127 F3d 1355, 1367 n. 2 (11th Cir. 1997) (although court accepted Dr. Cole's numerical estimates, it did not necessarily accept his legal conclusions regarding polarization because the court must base that determination on the relevant law). Proving a pattern of voting behavior does not require complete accuracy in Dr. Cole's numerical estimates. A "pattern will not be fatally altered if a few of his percentages are somewhat inaccurate." Id.

[¶ 65] Defendants contend that a finding of cohesion requires that a "significant number" of voters, namely 60 percent or more, vote for the same candidate. In support of this conclusion, defendants rely on Clay v. Board of Education of City of St. Louis, 90 F3d 1357, 1362 (8th Cir. 1996), aff'g Clay v. Board of Education of City of St. Louis, 896 FSupp 929, 935-36 (E.D. Mo. 1995). The court does not need to reach the issue as to whether a finding of cohesion requires an Indian vote of 60 percent or more because the statistical analysis of both Dr. Cole and Dr. Weber exceed this threshold.

[¶ 66] With regard to HPA, the court acknowledges that HPA has some limitations, especially here where no wards in the city of Martin are homogenous. While there are limitations on the conclusions that can be drawn from HPA results regarding voters that are not part of the majority, HPA remains a statistical analysis that is commonly used by other experts in the field as a reliable indicator of how the majority will vote. The court gives the HPA analysis less weight in its considerations, but does note its general consistency with the other data.

[¶ 67] The court finds that all three methods employed by the parties' experts in this case generated sufficiently similar results on the issue of Indian cohesion. Even though Dr. Weber, by applying his 60 percent threshold, found cohesion in fewer races than Dr. Cole, the court finds that both experts' analyses demonstrate significant cohesion among Indian voters. In light of the similar estimates and the general acceptance within the scientific community of all three statistical analysis methods, the court accepts the portions of both experts' testimony set forth herein as reliable and probative of cohesion existing among Indians. The court finds that the statistical evidence demonstrates political cohesiveness among Indians.

[¶ 68] Plaintiffs also admitted exit poll survey data to demonstrate political cohesiveness. Generally, a properly conducted exit poll survey is the most reliable method for determining racial polarization. Romero v. City of Pomona, 665 FSupp 853, 860 (C.D. Cal. 1987). An exit poll permits the court to examine how voters voted and which candidates were preferred by minority voters. Id. Exit polls, however, are "prone to high nonresponse rates which can seriously bias estimates and distort inferences, because people who do not respond may vote differently than those who do." Aldasoro v. Kennerson, 922 FSupp 339, 352 (S.D. Cal. 1995). In addition, exit poll respondents may lie. Id. at 353. A truly representative poll of the votes actually cast should logically demonstrate some consistency between the responses to the poll and the actual returns. Hall v. Holder, 757 FSupp 1560, 1577 (M.D. Ga. 1991).

[¶ 69] Here, by Dr. Cole's own admission, the exit poll sample under-represented non-Indians, over-represented Indians, and slightly over-represented females. Ex. 188, pp. 7-8. Thus, it was not a representative sample of voters as a whole. Furthermore, the exit poll fails to demonstrate consistency between the responses to the poll and the actual returns. For example, the results of the exit poll for Ward I show Gotheridge receiving 21 votes and Risse receiving 4 votes, when in fact Risse won the election. The results of the exit poll for Ward III show Alexander receiving 10 votes and Justus receiving 17, when in fact Alexander won the election. And the results of the exit poll for the school board show Dillon receiving 71 votes, Red Bear receiving 50 votes, Knecht receiving 20 votes, and Kocourek receiving 19 votes, when in fact Knecht and Kocourek won the election and Dillon and Red Bear came in fourth and fifth respectively. The high nonresponse rates of non-Indians seriously distorted inferences that could be drawn from the exit poll. When one compares the preferences expressed in the poll with those actually expressed in the election, it is evident that the poll overemphasized the electoral support received by some candidates and underestimated the support harnessed by other candidates. While Dr. Cole attempted to draw inferences from the poll results and to make adjustments based on the high response rates of Indians, the court finds the exit poll evidence to be inconclusive and unconvincing as to whether voting in the 2003 city and school board elections was racially cohesive or racially polarized. Therefore, the court gives no weight to the exit poll evidence.

[¶ 70] Additionally, defendants allege that Dr. Cole's exit polling methods were suspect because some of the poll takers were related to one of the plaintiffs. The court too is troubled by the fact that some of the poll takers were related to one of the plaintiffs; however, defendants offer no actual proof that the poll takers knew for whom the poll was being conducted except that it was for Dr. Cole's firm. Because the court gives no weight to the exit poll evidence, the court does not reach the issue of whether this fact alone is sufficient to disregard the exit poll data.

4. Non-Statistical Evidence of Cohesiveness

[¶ 71] The inquiry, however, "does not stop with bare statistics." Whitfield v. Democratic Party of Ark., 890 F2d 1423, 1428 (8th Cir. 1989). "The experiences and observations of individuals involved in the political process are clearly relevant to the question of whether the minority group is politically cohesive. This testimony would seem to be required if the court is to identify the presence or absence of distinctive minority group interests." Sanchez v. Bond, 875 F2d 1488, 1494 (10th Cir. 1989). Evidence that "a specified group of voters share common beliefs, ideals, principles, agendas, concerns, and the like such that they generally unite behind or coalesce around particular candidates and issues," demonstrates cohesion. League of United Latin American Citizens, Council No. 4434 v. Clements, 986 F2d 728, 744 (5th Cir. 1993). See Askew, 127 F3d at 1377 (both empirical and anecdotal evidence demonstrated cohesion).

[¶ 72] Political parties and candidates have courted Indian voters with targeted advertisements and other appeals. Ex. 65, 71. Several Indian-oriented newspapers are circulated in South Dakota, including Indian Country Today, the Lakota Journal, and the Black Hills Peoples' News. These papers cover a wide variety of social, cultural, and political topics, including voting rights, redistricting, and political campaigns. Ex. 65, 71, 80, 84, 88, 91-94, 99-103, 106, 110. Indians have organized a network of tribal colleges with courses that teach traditional tribal arts, language, and culture. T. I, p. 242-43.

[¶ 73] The Lakota Nation Invitational basketball tournament brings together Indians from across the state every year to watch and participate in athletic and cultural activities. Ex. 26. Tribes and tribal members from across the state frequently come together at social, political, and economic conferences and seminars. Ex. 24, 27, 31, 34, 38.

[¶ 74] Examples of Indian people coming together regarding common interests include first, the holding of several protest marches to highlight issues of Indian concern, including abuse of Indian rights. T. IV, p. 912-13. Second, the LaCreek Civil Rights Commission, which is a group of Indian people who reside in Bennett County, endorsed candidates and have used print and broadcast media to encourage tribal members to support the slate. Third, Indian people have worked together through their tribal governments on a host of political issues including education, tribal housing, and tribal sovereignty. T. III, p. 552-54; T. VIII, p. 1499-1501.

[¶ 75] Dr. McCool, one of plaintiffs' experts, opined that "there is a strong sense of cohesion and loyalty" within the Indian community. He testified that "[t]here are two distinct political communities in Martin, South Dakota, divided by race, history, and culture." Ex. 185, p. 50. The court accepts this testimony of Dr. McCool as reliable. The court finds that he qualifies as an expert to testify in this case. Dr. McCool has a doctorate degree in political science. He has taught courses at the college level since the late 1970s and currently teaches political science at the University of Utah. He is the director of the American West Center, a unit of the University of Utah that researches issues affecting the West, and he teaches classes about western issues, including classes on American Indians. Dr. McCool has published numerous articles and books about the relationship between American Indians and non-Indians. Some of this research specifically discussed Indian issues and political relationships in South Dakota. He has testified as an expert in several other cases on the political relationship between Indians and non-Indians in the western United States.

[¶ 76] The court finds that Dr. McCool has the education, training, skill, and knowledge necessary to make him a reliable expert. The court also finds that his research and methods were reliable and of the type typically practiced in his field. His extensive publications, peer-reviewed work, and involvement in western issues makes his research credible and reliable. For these reasons and the analysis contained in the court's May 27, 2004 order, the court finds that Dr. McCool satisfies the requirements of Rule 702 and Daubert. Ex. 185; Court Order 5/27/04 (Docket 305).

[¶ 77] The court gives little weight, however, to the portions of Dr. McCool's report that relied on interviews of various people living on the Pine Ridge and Rosebud Reservations, including the conclusions he drew from those interviews. When comparing evidence, the court finds that the testimony of witnesses at trial is entitled to more weight than the personal history information acquired by Dr. McCool in unstructured interviews.

[¶ 78] Lay evidence also demonstrates Indian cohesiveness with regard to recruitment of candidates, registration of voters, and get-out-the-vote efforts. Indians in Bennett County, in and around Martin, established a grass roots political organization in 2001 called the Lacreek District Civil Rights Committee. Ex. 101, 106, 211; T. VIII, p. 1336. Jesse Clausen, who lives in Bennett County, testified that for several years, certain incidents occurred that made him feel like local public officials violated his civil rights. Indians felt as though law enforcement, particularly the Bennett County sheriff, unfairly singled them out. A group of people who live within the Lacreek tribal council district of the Oglala Sioux Tribe began meeting to discuss concerns and ideas. Clausen became the spokesperson or chairman of the group. Eventually the Lacreek District Civil Rights Committee was formed and they met with the city council, county commissioners, and the mayor. T. VIII, p. 1502-03. In the Committee's view, neither the sheriff, the city council, nor the mayor took action to remedy the situation.

[¶ 79] When elected officials did not adequately respond to Indian concerns about law enforcement, the Committee turned its focus to the political process. The Committee determined that by registering voters and getting people to the polls, they could elect a different sheriff and different representation in the county and city commissions. Ex. 94; TIII, p. 528; TIV, p. 911-12; T. VIII, p. 1496-97, 1502-03. In the spring of 2002, the Committee organized a peaceful march in Martin to protest against the actions of the sheriff's office. Between 500 and 1000 people participated. A second march took place that fall. Ex. 46; T. IV, p. 912-13.

[¶ 80] The Committee registered Indian voters and recruited candidates for local offices. It sponsored a slate of candidates in the municipal, school board, and county elections in 2002 and supported those candidates with campaign materials and advertisements in print media and on the local tribal radio station. The Committee met on a weekly basis, organized a phone tree, and registered voters, often by going door to door. On election day, the Committee transported voters to the polls on election day and set up poll watchers. The Committee compared a list of registered voters with those who had voted at the polls, contacted people who had not yet voted, and brought them to the polls. T. III, p. 529-30, 533-34; T. IV, p. 912; T. VII, p. 1344-45; T. VIII, p. 1502-04; Ex. 65, 71, 92, 148, 211.

[¶ 81] The Committee met with some success. In the primary, they unseated three county commissioner incumbents. During the general election, the Committee succeeded in electing the sheriff candidate they supported and in electing one county commissioner. The actions and results of the Committee's efforts were publicized nationally. Ex. 146; T. III, p. 532-33.

[¶ 82] The court finds the testimony of Clausen and Craig Dillon, who both described the efforts of the Lacreek District Civil Rights Committee, to be reliable and credible and gives their testimony great weight. The court finds that the actions of the Lacreek District Civil Rights Commission is strong evidence of cohesion among Indians. It demonstrates that a significant portion of Indian voters support the same candidates and are concerned with the same issues.

[¶ 83] Defendants point to testimony of internal division among tribal members and between tribes as evidence that Indians are not politically cohesive. Defendants quote Molly Risse, Gwen Ward, Gayle Kocer, and Joyce Wilson to support their contention that not all Indians are cohesive. T. VIII, p. 1580-81, 1618; T. IX, p. 1801-03; T. X p. 2064-67. The court agrees that there is some division among tribal members and between various tribes on certain issues, particularly in relation to internal tribal matters. Cohesion on inter-tribal matters, however, is not relevant to the current case. Indeed, a § 2 violation does not require proof that all members of the minority think alike. The evidence establishes, moreover, that any division among Indians is far less prominent when applied to external factors affecting the tribes, such as relations with the city, county, state or federal government. The court finds that cohesiveness among Indian people exists, specifically regarding outside influences toward Indians. The law, moreover, has historically recognized Indians as members of "distinct political communities." See Morton v. Mancari, 417 US 535, 554, 94 SCt 2474, 41 LEd2d 290 (1974); Santa Clara Pueblo v. Martinez, 436 US 49, 55, 98 SCt 1670, 56 LEd2d 106 (1978).

5. Partisanship

[¶ 84] Defendants maintain that partisanship, rather than race, accounts for Indian voting behaviors. While causation may be relevant to the totality-of-circumstances review, it is not relevant in the inquiry into the three Gingles factors. See Goosby v. Town Bd. of Town of Hempstead, N.Y., 180 F3d 476, 493 (2d Cir. 1999) (holding causation irrelevant to the three Gingles factors); Milwaukee Branch of the N.A.A.C.P. v. Thompson, 116 F3d 1194, 1199 (7th Cir. 1997) (explanation for the defeat of black-preferred candidates should only be considered in the totality-of-circumstances inquiry); Lewis v. Alamance County, 99 F3d 600, 616 n. 12 (4th Cir. 1996) (causation relevant to the totality-of-circumstances inquiry, but not in the Gingles analysis); Uno, 72 F3d at 983 (totality-of-circumstances inquiry may examine non-racial reasons for voting patterns). Cf. Nipper v. Smith, 39 F3d 1494, 1524 (11th Cir. 1994) (en banc). Accordingly, partisanship has no bearing on the Gingles factors.

[¶ 85] If partisanship is proper for the court to consider, however, the court finds that the evidence does not support that partisanship explains Indian voting patterns more than race. For example, the slate of candidates supported by the Lacreek District Civil Rights Committee were neither all Indian nor all Democrat. In the 2002 Democratic primary election for Bennett County Commission, Indians favored the three Indian candidates over all three Democratic incumbents by greater than a 4-to-1 margin. Ex. 186 p. 28. In the 1998 general election for US House, Indians favored a non-Indian Republican over his non-Indian Democratic opponent by a margin of almost 2-to-1. Ex. 186, p. 42. And in the 2000 general election for State House District 26, Indians voted for the Independent candidate by a 2-to-1 margin over his closest opponent. Ex. 186, p. 41.

[¶ 86] Defendants also argue that in a majority of the races analyzed by Dr. Cole, a greater percentage of Indians voted for Democrats over Republicans than voted for Indians over non-Indians. Defendants contend that this demonstrates that party affiliation best explains Indian voting patterns. The court disagrees. The record in the current case does not "indisputably [prove] that partisan affiliation, not race, best explains the divergent voting patterns among minority and white citizens in the contested counties." League of United Latin American Citizens, Council 4434 v. Clements, 999 F2d 831, 850 (5th Cir. 1993). Furthermore, minority voters have a right to an equal opportunity to elect representatives of their choice, whatever the basis for those choices. 42 USC § 1973(b). In Goosby, the Second Circuit stated, "The Town's argument implies that if blacks registered and voted as Republicans, they would be able to elect the candidates they prefer. But they are not able to elect preferred candidates under the Republican Party regime that rules in the Town. Moreover, blacks should not be constrained to vote for Republicans who are not their preferred candidates." 180 F3d at 495-96. Likewise, the court will not accept defendants' argument that partisan politics explains the Indian preference. The court, therefore, does not adopt defendants' partisanship argument.

[¶ 87] After considering the statistical evidence, historical evidence, and current-day lay testimony, the court finds that plaintiffs have shown by a preponderance of the evidence that Indians in Martin are politically cohesive. The court finds that plaintiffs have satisfied the second Gingles factor.

C. Usual Defeat of Indian-Preferred Candidates

[¶ 88] The third Gingles factor requires plaintiffs 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 US at 51. The presence of racially polarized voting "will ordinarily be the keystone of a vote dilution case." Buckanaga v. Sisseton Indep. Sch. Dist. No. 54-5, S.D., 804 F2d 469, 473 (8th Cir. 1986). Voting along racial lines "deprive[s] minority voters of their preferred representative . . . [and] allows those elected to ignore minority interests without fear of political consequences, leaving the minority effectively unrepresented." Gingles, 478 US at 47 n. 14. Unless the minority has substantial difficulty electing representatives of their choice, it is impossible to prove that a challenged electoral mechanism impairs their ability to elect. Id. at 48 n. 15.

[¶ 89] The Supreme Court adopted the definition of racial polarization as existing "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 black voters and white voters vote differently." Id. at 51 n. 21. There is no simple test to determine the existence of legally significant racial bloc voting. Id. at 58. In inquiring into the existence of racially polarized voting, the court should ascertain whether minority group members constitute a politically cohesive unit and whether whites vote sufficiently as a bloc usually to defeat the minority's preferred candidates. Id. at 56. Political cohesiveness can be shown by evidence that a significant number of minority group members usually vote for the same candidates, and by a white bloc vote that "normally will defeat the combined strength of minority support plus white `crossover' votes." Id. In order for white bloc voting to be legally significant, however, it had to be high enough to "normally defeat the combined strength of minority support plus white crossover votes." Id. at 56. Gingles, however, "does not require an absolute monolith" in the majority's bloc vote. Sanchez, 97 F3d at 1319.

[¶ 90] The level of white bloc voting sufficient to defeat a minority preferred candidate varies according to a variety of factual circumstances. Thus, no mathematical formula or simple doctrinal test is available to determine whether plaintiffs satisfied the third factor. Id. at 57-58; Ruiz, 160 F3d at 554. The inquiry therefore focuses on statistical evidence to discern the way voters voted. Gingles, 478 US at 57; Sanchez, 97 F3d at 1315. "The surest indication of race conscious politics is a pattern of racially polarized voting extending over time." Buckanaga, 804 F2d at 473.

[¶ 91] The third Gingles factor considers not the size of the bloc but considers the bloc's effect on minority voters' ability to fully participate in the political process and to elect their representatives of choice. Jenkins v. Red Clay Consol. Sch. Dist. Bd. of Educ., 4 F3d 1103, 1122 (3d Cir. 1993). When determining whether a pattern of usual defeat exists, the court must conduct "a searching practical evaluation of the past and present reality [with] a functional view of the political process." Gingles, 478 US at 45. In examining the third Gingles factor, the inquiry should center on districts with a "white majority." Gingles, 478 US at 50. All three of the wards in the city of Martin under Ordinance 122 are white majority.

[¶ 92] To ascertain the existence of white bloc voting in a particular contest, the court should determine: (1) the candidate who the Indian voters preferred; and (2) whether whites voted as a bloc to defeat the Indian-preferred candidate. Old Person, 230 F3d at 1122. In analyzing election contests, certain contests are more probative of bloc voting than others. Endogenous elections between white and minority candidates are the most probative in determining the existence of legally significant white bloc voting. See Gingles, 478 US at 80-82; Blaine County, 363 F3d at 911. The court gives greater weight to the most recent elections because they are more probative. Uno, 72 F3d at 990; Meek, 985 F2d at 1482-83.

[¶ 93] The only endogenous races examined for the city of Martin involved two city council races in Ward I and Ward III. The Ward I race involved two Indian candidates and the Ward III race involved two non-Indian candidates. The analysis was conducted with exit polling by Dr. Cole. Dr. Weber did not analyze this race, but criticized the results of the exit poll. Dr. Cole found:

ENDOGENOUS CONTESTS Contest 2003 non-Indian vote for IPC ⁹ Indian vote for Result Source IPC

9 IPC stands for Indian-preferred candidate.

Ward I (Indian 33.3% according to exit poll 100% exit poll IPC (Gotheridge) Ex. 188, p. only) 25% as adjusted lost 10-11, Table 2 Ward III 65.2% according to exit poll 85.7% exit poll IPC (Justus) lost Ex. 188, p. (white only) 36.7% as adjusted 11-12, Table 3 In Ward I, the Indian (and Indian-preferred) candidate lost to an Indian, but not Indian-preferred, candidate. In Ward III, the Indian-preferred candidate lost. As previously set forth, the court gives no weight to the exit poll. While the election results show that the Indian-preferred candidates lost, this is not sufficient proof of polarization and the court draws no conclusions regarding endogenous contests.

[¶ 94] The court next examines election contests between interracial candidates in exogenous races. Exogenous races are less probative than the endogenous races and thus the court gives them less probative value. See Sanchez, 97 F3d at 1317. Interracial contests are entitled to more probative weight than homogeneous contests. Gingles, 478 US at 80-82. And the court gives more weight to county-wide elections than state and federal because they are of comparable levels of importance to city elections. See Black Political Task Force, 300 FSupp 2d at 308.

[¶ 95] The following charts, which are set forth in order of the weight given by the court, summarize the exogenous election data. Dr. Cole examined eight county-wide interracial elections using HPA and BERA and four of these same contests using EI. Dr. Weber examined four of those contests using BERA and EI attempting to replicate Dr. Cole's EI findings. Their results are summarized as follows: EXOGENOUS INTERRACIAL Multi-Candidate County-Wide Offices Contest Method - non-Indian Indian vote for Result Source Expert vote for IPC IPC 2002 Nonpartisan school board (2 seats) 2002 Dem. Prim. county comm'r (3 seats) 2002 Gen. election county comm'r (3 seats) 2000 nonpartisan school board (2 EXOGENOUS INTERRACIAL Head-to-Head County-Wide Offices Contest Method - non-Indian Indian vote Result Source Expert vote for IPC for IPC 2002 Gen. election sheriff (fn11) 2002 Gen. election reg. of deeds 2002 Gen. election coroner 2001 school board - nonpartisan

BERA — Cole Flye 11% Flye 100% One IPC (Flye) Ex. 186, p. 28 won and one IPC Burritt 14% Burritt 64% (Burritt) lost HPA — Cole Flye 15% Flye N/A Ex. 186, p. 32 Burritt 7% Burritt N/A BERA — Cole Sharp 3% Sharp 89% 3 IPCs won Ex. 186, p. 26 Bettleyoun Bettleyoun 31% 70% Ruff 18% Ruff 69% HPA — Cole Sharp 12% Sharp N/A Ex. 186, p. 132 Bettleyoun Bettleyoun N/A 20% Ruff N/A Ruff 12% BERA — Cole Bettleyoun 3% Bettleyoun One IPC Ex. 186, p. 30 86% (Bettleyoun) won Sharp 0% Sharp 86% and two IPCs lost Hammond 7% Hammond 53% HPA — Cole Bettleyoun 2% Bettleyoun N/A Ex. 186, p. 34 Sharp 4% Sharp N/A Ruff 3% Ruff N/A BERA — Cole 26% Three Stars IPC (Three Stars) Ex. 186, p. 31 100% lost seats) (fn10) HPA — Cole 11% N/A Ex. 186, p. 35 BERA — Cole 14% 99% IPC (Cummings) Ex. 189, p. 2 won HPA — Cole 8.9% N/A Ex. 189, p. 3 EI — Cole 20% 94% Ex. 189, p. 3 BERA — 14.0% 99.1% Ex. 449, A-5 Weber EI — Weber 24.9% 91.7% Ex. 450, Table 3 BERA — Cole 5% Johnson 51% Johnson Court unable to Ex. 186, p. 29 determine IPC HPA — Cole 10% Johnson N/A candidate. Sterkel Ex. 186, p. 33 won. EI — Cole 96% Sterkel 56% Sterkel Ex. 186, p. 36 BERA — 5.2% Johnson 51.6% Johnson Ex. 449, p. A-6 Weber EI — Weber 2.5% Johnson 52% Johnson Ex. 450, Table 3 BERA — Cole 0% 92% IPC (Mesteth) Ex. 186, p. 30 lost HPA — Cole 5% N/A Ex. 186, p. 34 EI — Cole 1% 76% Ex. 186, p. 36, Ex. 450, Table 3 BERA — 0% 93% Ex. 449, p. A-7 Weber EI — Weber 1.8% 98% Ex. 450, Table 3 BERA — Cole 18% 81% IPC (Three Stars) Ex. 186, p. 30 lost HPA — Cole 17% N/A Ex. 186, p. 34 EI — Cole 11% 58% Ex. 186, p. 36, BERA — 18.2% 100% (fn12) Ex. 449, p. A-9 Weber EI — Weber 27.1% 30.0% Ex. 450, Table 3 The court notes that in most cases, Dr. Cole and Dr. Weber are consistent in their opinion of the Indian and non-Indian vote. The court finds that with regard to multi-candidate county-wide interracial elections, the Indian-preferred candidates won six contests and lost three contests. The court finds that with regard to head-to-head county-wide interracial elections, the Indian-preferred candidates won one contest and lost two contests. In the fourth contest in this category, which is the 2002 general election for register of deeds, the court finds that the election data is so conflicting that it is difficult to even determine the Indian-preferred candidate. Because the court cannot determine the Indian-preferred candidate, by definition the election is not polarized. Thus, among all interracial county-wide races, in five contests the white voters voted sufficiently as a bloc to defeat the Indian-preferred candidate and the court finds those contests to be racially polarized. In eight contests, the Indian-preferred candidate won and as a result the court finds those eight contests were not racially polarized. The court finds that within the category of interracial county-wide elections, white voters did not vote sufficiently as a bloc to "usually" defeat the Indian-preferred candidate.

[¶ 96] The court gives the next greatest weight to interracial, state office elections. Here, Dr. Cole analyzed three elections with BERA, HPA and EI. Dr. Weber analyzed all three elections also applying BERA and EI. They found: EXOGENOUS INTERRACIAL Head-to-Head State Offices Contest Method - non-Indian vote Indian vote for Result Source Expert for IPC IPC 1998 Gen. election governor 2002 Dem. primary governor 2002 Gen. election AG

BERA — Cole 25% 74% IPC (Hunhoff/ Ex. 186, p. 31 Meeks) lost HPA — Cole 18% N/A Ex. 186, p. 35 EI — Cole 29% 55% Ex. 186, p. 37 BERA — Weber 25.3% 73.7% Ex. 449, p. A- 11 EI — Weber 34.7% 64.4% Ex. 450, Table 3 BERA — Cole 10% 60% IPC (Volesky) Ex. 186, p. 29 lost HPA — Cole 32% N/A Ex. 186, p. 33 EI — Cole 10% 51% Ex. 186, p. 36 BERA — Weber 10.5% 60.6% Ex. 449, p. A-3 EI — Weber 6.4% 55.1% Ex. 450, Table 3 BERA — Cole 9% 74% IPC (Volesky) Ex. 186, p. 29 lost HPA — Cole 12% N/A Ex. 186, p. 33 EI — Cole 4% 73% Ex. 186, p. 36, BERA — Weber 9.0% 74.9% Ex. 449, p. A-4 EI — Weber 8.6% 63.1% Ex. 450, Table 3 In these races, the Indian-preferred candidate lost 100 percent (3 out of 3 candidacies). The court finds that the white voters voted sufficiently as a bloc to defeat the Indian-preferred candidate in those contests. The court finds that although all of the data does not support Indian cohesion at the 60 percent level, most races are well over or very close to that level. In most cases, Dr. Cole's and Dr. Weber's numbers are largely consistent. The court finds that within the category of interracial state office elections, white voters did vote sufficiently as a bloc to "usually" defeat the Indian-preferred candidate.

[¶ 97] Third, the court next examines election contests between white candidates in county-wide exogenous races. Dr. Cole examined three contests using HPA, BERA and EI. Dr. Weber examined those three contests using BERA and EI. Their results are summarized as follows:

EXOGENOUS WHITE ONLY County-Wide Offices

Contest Method - non-Indian vote Indian vote for Result Source Expert for IPC IPC 2002 Gen. Election County Auditor 2000 Gen. Election County Sheriff 1998 Gen. Election County Auditor

Cole — BERA 86% (Williams) 53% Court unable to Ex. 186, p. 40 (Williams) determine IPC candidate Cole — HPA 86% (Williams) N/A Ex. 186, p. 48 (Williams won) Cole — EI 6% (Hudson) 62% (Hudson) Ex. 186, p. 55 Weber — BERA 85.5% (Williams) 52.6% Ex. 449, p. D-12 (Williams) Weber — EI 6.0% (Hudson) 57.8% Ex. 450, Table (Hudson) 4 Cole — BERA 15% 91% IPC (Duke) lost Ex. 186, p. 41 Cole — HPA 16% N/A Ex. 186, p. 49 Cole — EI 15% 61% Ex. 186, p. 56, Weber — BERA 15.5% 90.5% Ex. 449, p. D-17 Weber — EI 8% 97.1% Ex. 450, Table 4 Cole — BERA 78% 71% IPC (Williams) Ex. 186, p. 43 won Cole — EI 79% 80% Ex. 186, p. 57 Weber — BERA 78.4% 70.9% Ex. 449, p. D-26 Weber — EI 72.1% 91.2% Ex. 450, Table 4 The court notes that in most cases, Dr. Cole and Dr. Weber are consistent in their opinion of the Indian and non-Indian vote. The court finds that with regard to county-wide white-only elections, the Indian-preferred candidate won one contest and lost one contest. In the third contest in this category, which is the 2002 general election for auditor, the court finds that the election data is so conflicting that it is difficult to even determine the Indian-preferred candidate. Because the court cannot determine the Indian-preferred candidate, by definition the election is not polarized. Thus, among all white-only county-wide races, in one contest the white voters voted sufficiently as a bloc to defeat the Indian-preferred candidate and the court finds this contest to be racially polarized. In two contests, the Indian-preferred candidate won or an Indian-preferred candidate could not be determined and as a result the court finds these two contests were not racially polarized. The court finds that within this category, white voters did not vote sufficiently as a bloc to "usually" defeat the Indian-preferred candidate.

[¶ 98] Fourth, the court next examines election contests between white candidates for state and federal office. Dr. Cole examined 25 contests using HPA, BERA, and EI. Using BERA, he examined four contests for the state house of representatives in which two candidates were elected in each election. Dr. Weber examined the initial 25 contests using BERA and EI. Their results are summarized as follows:

EXOGENOUS WHITE ONLY State and Federal Offices

Contest Method - non-Indian Indian vote for Result Source Expert vote for IPC IPC 2002 Gen. Election US Senate 2002 Gen. Election US House 2002 Gen. Election Governor 2002 Gen. Election Secretary of State 2002 Gen. Election State Auditor 2002 Gen. Election State Treasurer 2002 Gen. Election Comm'r of S PL 2002 Gen. Election PUC 6 yr 2002 Gen. Election PUC 4 yr 2002 Gen. Election State Sen. District 26 2002 Gen. Election State House District 26 26 (2 seats) 2000 Gen. Election President 2000 Gen. Election US House 2000 Gen. Election PUC 2000 Gen. Election State House Dist. 26 (two seats) 1998 Gen. Election US Senate 1998 Gen. Election US House 1998 Gen. Election Secretary of State 1998 Gen. Election State Treasurer 1998 Gen. Election Comm'r of S PL 1998 Gen. Election PUC 1998 Gen. Election State Senate Dist. 26 1998 Gen. Election State House Dist. 26 (two seats) 1996 Gen. Election President 1996 Gen. Election US Senate 1996 Gen. Election US House 1996 Gen. Election PUC 1996 Gen. Election State Senate Dist. 26 1996 Gen. Election State House Dist. 26 (two seats)

Cole — BERA 24% 94% IPC (Johnson) Ex. 186, p. 38 won Cole — HPA 18% N/A Ex. 186, p. 46 Cole — EI 29% 95% Ex. 186, p. 54 Weber — BERA 24.6% 94.3% Ex. 449, p. D-1 Weber — EI 35.7% 93.1% Ex. 450, Table 4 Cole — BERA 26% 91% IPC (Herseth) Ex. 186, p. 38 Cole — HPA 19% N/A won Ex. 186, p. 46 Cole — EI 30% 91% Ex. 186, p. 54 Weber — BERA 26.0% 91.8% Ex. 449, p. D-2 Weber — EI 28.7% 92.5% Ex. 450, Table 4 Cole — BERA 25% 86% IPC (Abbott) Ex. 186, p. 38 Cole — HPA 18% N/A won Ex. 186, p. 46 Cole — EI 32% 86% Ex. 186, p. 54 Weber — BERA 25.6% 86.7% Ex. 449, p. D-3 Weber — EI 36.6% 88.8% Ex. 450, Table 4 Cole — BERA 36% 87% IPC (Looby) Ex. 186, p. 38 Cole — HPA 32% N/A won Ex. 186, p. 46 Cole — EI 43% 89% Ex. 186, p. 54 Weber — BERA 36.2% 87.7% Ex. 449, p. D-4 Weber — EI 31.4% 92.9% Ex. 450, Table 4 Cole — BERA 36% 76% IPC (Butler) Ex. 186, p. 39 Cole — HPA 35% N/A won Ex. 186, p. 47 Cole — EI 39% 92% Ex. 186, p. 54 Weber — BERA 36.1% 91.3% Ex. 449, p. D-5 Weber — EI 35.5% 92.8% Ex. 450, Table 4 Cole — BERA 26% 95% IPC (McGregor) Ex. 186, p. 39 Cole — HPA 30% N/A won Ex. 186, p. 47 Cole — EI 29% 98% Ex. 186, p. 54 Weber — BERA 25.9% 95.8% Ex. 449, p. D-6 Weber — EI 23.1% 93.4% Ex. 450, Table 4 Cole — BERA 24% 90% IPC (Healy) Ex. 186, p. 39 Cole — HPA 29% N/A won Ex. 186, p. 47 Cole — EI 28% 92% Ex. 186, p. 55, Weber — BERA 24.1% 90.8% Ex. 449, p. D-7 Weber — EI 21.5% 95.8% Ex. 450, Table 4 Cole — BERA 29% 92% IPC (Nelson) Ex. 186, p. 39 Cole — HPA 24% N/A won Ex. 186, p. 47 Cole — EI 36% 94% Ex. 186, p. 55 Weber — BERA 30.0% 92.8% Ex. 449, p. D-8 Weber — EI 30.1% 94.6% Ex. 450, Table 4 Cole — BERA 35% 88% IPC (Johnson) Ex. 186, p. 39 Cole — HPA 34% N/A won Ex. 186, p. 47 Cole — EI 39% 91% Ex. 186, p. 55 Weber — BERA 35.4% 88.9% Ex. 449, p. D-9 Weber — EI 33.0% 90.7% Ex. 450, Table 4 Cole — BERA 36% 90% IPC (Reis) won Ex. 186, p. 40 Cole — HPA 34% N/A Ex. 186, p. 48 Cole — EI 39% 82% Ex. 186, p. 55 Weber — BERA 35.6% 90.8% Ex. 449, p. D- 10 Weber — EI 50.6% 60.6% Ex. 450, Table 4 Cole — BERA 14% 83% IPC (Heller) Ex. 186, p. 40 won Cole — BERA 21% 75% IPC (Gore) lost Ex. 186, p. 40 Cole — HPA 14% N/A Ex. 186, p. 48 Cole — EI 29% 56% Ex. 186, p. 55 Weber — BERA 21.0% 74.9% Ex. 449, p. D- 13 Weber — EI 12.1% 88.2% Ex. 450, Table 4 Cole — BERA 15% 62% IPC (Hohn) lost Ex. 186, p. 41 Cole — HPA 13% N/A Ex. 186, p. 49 Cole — EI 8% 58% Ex. 186, p. 56 Weber — BERA 15.1% 61.7% Ex. 449, p. D- 14 Weber — EI 10.5% 65.8% Ex. 450, Table 4 Cole — BERA 36% 88% IPC (Laubach) Ex. 186, p. 41 won Cole — HPA 28% N/A Ex. 186, p. 49 Cole — EI 41% 75% Ex. 186, p. 56 Weber — BERA 36.3% 88.0% Ex. 449, p. D- 15 Weber — EI 30.6% 95.0% Ex. 450, Table 4 Cole — BERA 35% 74% IPC (Jorgeson) Ex. 186, p. 41 lost (only 1 IPC candidate received 50% or greater) Cole — BERA 57% 100% IPC (Daschle) Ex. 186, p. 42 won Cole — HPA 46% N/A Ex. 186, p. 50 Cole — EI 65% 82% Ex. 186, p. 56 Weber — BERA 56.7% 100% Ex. 449, p. D- 18 Weber — EI 51.6% 95.6% Ex. 450, Table 4 Cole — BERA 80% 61% IPC (Thune) Ex. 186, p. 42 won Cole — HPA 82% N/A Ex. 186, p. 50 Cole — EI 77% 71% Ex. 186, p. 56 Weber — BERA 79.6% 61.7% Ex. 449, p. D- 19 Weber — EI 86.3% 27.5% Ex. 450, Table 4 Cole — BERA 25% 77% IPC (Green) lost Ex. 186, p. 42 Cole — HPA 17% N/A Ex. 186, p. 50 Cole — EI 25% 53% Ex. 186, p. 56 Weber — BERA 24.7% 77.6% Ex. 449, p. D- 20 Weber — EI 13.0% 93.9% Ex. 450, Table 4 Cole — BERA 43% 88% IPC (Butler) lost Ex. 186, p. 42 Cole — HPA 36% N/A Ex. 186, p. 50 Cole — EI 36% 65% Ex. 186, p. 57 Weber — BERA 42.6% 77.9% Ex. 449, p. D- 21 Weber — EI 35.6% 90.9% Ex. 450, Table 4 Cole — BERA 41% 63% IPC (Johnson) Ex. 186, p. 42 lost Cole — HPA 34% N/A Ex. 186, p. 50 Cole — EI 30% 78% Ex. 186, p. 57 Weber — BERA 41.3% 62.8% Ex. 449, p. D- 22 Weber — EI 30.6% 85.5% Ex. 450, Table 4 Cole — BERA 47% 87% IPC (Burg) won Ex. 186, p. 42 Cole — HP 42% N/A Ex. 186, p. 51 Cole — EI 40% 85% Ex. 186, p. 57 Weber — BERA 46.8% 87% Ex. 449, p. D-23 Weber — EI 36.1% 98.8% Ex. 450, Table 4 Cole — BERA 27% 100% IPC (Kindle) Ex. 186, p. 43 lost Cole — HPA 37% N/A Ex. 186, p. 51 Cole — EI 30% 57% Ex. 186, p. 57 Weber — BERA 27.0% 100% Ex. 449, p. D-24 Weber — EI 34.7% 94.8% Ex. 450, Table 4 Cole — BERA 31% Bartlett 95% Bartlett IPC (Bartlett) Ex. 186, p. 43 45% Jorgenesen 57% Jorgensen and Jorgensen lost Cole — BERA 23% 100% IPC (Clinton) Ex. 186, p. 44 lost Cole — HPA 26% N/A Ex. 186, p. 52 Cole — EI 24% 82% Ex. 186, p. 57 Weber — BERA 23.0% 100%* Ex. 449, p. D-27 Weber — EI 39.5% 97.5% Ex. 450, Table 4 Cole — BERA 31% 100% IPC (Johnson) Ex. 186, p. 44 lost Cole — HPA 30% N/A Ex. 186, p. 52 Cole — EI 30% 79% Ex. 186, p. 58, Weber — BERA 31.4% 100%* Ex. 449, p. D-28 Weber — EI 32.3% 97.4% Ex. 450, Table 4 Cole — BERA 20% 98% IPC (Weiland) Ex. 186, p. 44 lost Cole — HPA 21% N/A Ex. 186, p. 52 Cole — EI 14% 83% Ex. 186, p. 58 Weber — BERA 19.9% 100%* Ex. 449, p. D-29 Weber — EI 28.7% 89.3% Ex. 450, Table 4 Cole — BERA 23% 82% IPC (Nelson) Ex. 186, p. 44 lost Cole — HPA 23% N/A Ex. 186, p. 52 Cole — EI 22% 62% Ex. 186, p. 58 Weber — BERA 23.1% 82.4% Ex. 449, p. D-30 Weber — EI 30.3% 69.5% Ex. 450, Table 4 Cole — BERA 53% 100% IPC (Nelson) Ex. 186, p. 45 won Cole — HPA 66% N/A Ex. 186, p. 53 Cole — EI 54% 86% Ex. 186, p. 58, Weber — BERA 52.4% 100%* Ex. 449, p. D-31 Weber — EI 60.8% 93.8% Ex. 450, Table 4 Cole — BERA Good 56% Good 82% IPC (Good) won Ex. 186, p. 45 Risseeuw 10% Risseeuw 81% and IPC (Risseeuw) lost The court notes that in most cases, Dr. Cole and Dr. Weber are consistent in their opinion of the Indian and non-Indian vote. The court finds that with regard to county-wide white-only elections for state or federal office, the Indian-preferred candidates won 17 contests and lost 14 contests. Thus, in this category, in 14 contests the white voters voted sufficiently as a bloc to defeat the Indian-preferred candidate and the court finds these contests to be racially polarized. In 17 contests, the Indian-preferred candidate won and as a result the court finds these 17 contests were not racially polarized. The court finds that within this category, white voters did not vote sufficiently as a bloc to "usually" defeat the Indian-preferred candidates.

[¶ 99] Only in the category of interracial county-wide elections did whites vote sufficiently as a bloc to usually defeat the Indian-preferred candidate. This category is not given the greatest weight, however, and the other categories do not support the proposition that white bloc voting "usually" results in the defeat of the Indian-preferred candidate.

[¶ 100] Plaintiffs also offer non-statistical evidence to demonstrate the usual defeat of the Indian-preferred candidate. A number of lay witnesses testified that they could identify the Indian-preferred candidates and that those candidates consistently lost city council elections. T. I, p. 254-57; T. III, p. 530-33; T. IV, p. 845, 869, 873-74, 916-17. Plaintiffs contend that because virtually all the population in Martin is either white or Indian, white bloc voting is the only possible explanation for the defeat of Indian-preferred candidates. None of this testimony eliminates other considerations for candidate losses, including campaign efforts, platform popularity, or candidate characteristics. In light of the overwhelming statistical evidence, this lay testimony is not sufficient to meet plaintiffs' burden of demonstrating the usual defeat of the Indian-preferred candidate.

[¶ 101] After considering all the categories of elections and giving the greatest weight to exogenous county-wide interracial elections, the court finds that plaintiffs have not met their burden to prove that the white majority in Martin "votes sufficiently as a bloc to enable it . . . usually to defeat the [Indian] preferred candidate" as is required to prove the third Gingles factor. Gingles, 478 US at 51.

D. Totality of the Circumstances

[¶ 102] If a plaintiff satisfies the three Gingles factors, then the court must next consider the totality of the circumstances. Gingles, 478 US at 79. Plaintiffs successfully meet Gingles factors one and two; however, plaintiffs do not meet the third Gingles factor. Because plaintiffs have not satisfied the third Gingles factor, the court need not consider the totality of the circumstances. Clay, 90 F3d at 1362.

IV. Intent Claim

[¶ 103] Plaintiffs allege that Ordinance 122 was adopted and is being maintained for a discriminatory purpose in violation of § 2 of the Voting Rights Act, 42 USC § 1973, and the Fourteenth and Fifteenth Amendments. Supplemented Complaint (Doc. 61). The Eighth Circuit Court of Appeals has held that an "intent" claim requires proof of a discriminatory effect. Villa, 54 F3d 1345, 1357 n. 18. In Villa, after finding against plaintiffs on their § 2 "effects" claim, the court of appeals upheld the district court's grant of summary judgment on plaintiffs' intent claims stating "where there is no discriminatory effect, there is no intent violation." Id. Similarly here, the foregoing analysis rejected plaintiffs' § 2 effects claim. As a result, plaintiffs' intent claims also fail.

[¶ 104] Even if the court were to consider plaintiffs' intent claim, the court finds no evidence of discriminatory intent in the passage of Ordinance 122. The United States Supreme Court identified the relevant factors to consider in determining whether intent to discriminate existed in Village of Arlington Heights v. Metropolitan Housing Development, 429 US 252, 266-68, 97 SCt 555, 50 LEd2d 450 (1977).

The impact of the official action — whether "it bears more heavily on one race than another," may provide an important starting point. Sometimes a clear pattern, unexplainable on grounds other than race, emerges from the effect of the state action even when the governing legislation appears neutral on its face. . . . Absent a [stark pattern] # 8230; impact alone is not determinative, and the Court must look to other evidence.

The historical background of the decision is one evidentiary source, particularly if it reveals a series of official actions taken for invidious purposes. . . . The specific sequence of events leading up the challenged decision also may shed some light on the decision maker's purposes. . . . Departures from the normal procedural sequence also might afford evidence that improper purposes are playing a role. Substantive departures too may be relevant, particularly if the factors usually considered important by the decision maker strongly favor a decision contrary to the one reached.

The legislative or administrative history may be highly relevant, especially where there are contemporary statements by members of the decision making body, minutes of its meetings, or reports. In some extraordinary instances the members might be called to the stand at trial to testify concerning the purpose of the official action, although even then such testimony frequently will be barred by privilege. . . .

The foregoing summary identifies, without purporting to be exhaustive, subjects of proper inquiry in determining whether racially discriminatory intent existed.

Id. (citations omitted). Here, the city council adopted a redistricting plan based on a map provided by BHCLG. After a malapportionment problem was identified with the map, BHCLG provided a new map to the city council, which created three wards of nearly equal population. There is no evidence that BHCLG knew the ethnicity of the residents of the city of Martin when they were drafting either map. Upon visual inspection, the wards under Ordinance 122 appear to be compact and regular in shape. There are no contemporary statements by members of the city counsel, minutes of the meetings, or reports that reflect that the council acted with a racially discriminatory intent in the enactment of Ordinance 122.

[¶ 105] Although City Finance Officer Janet Speidel's testimony revealed a lack of truthfulness, she is not a member of the body that passed the city ordinance and there is no evidence attributing her actions to the city council. Although it would have been common courtesy, Speidel had no duty to instruct Fogg, a citizen who circulated petitions to refer Ordinance 122 to a public vote, that his petitions were invalid before the filing deadline. While Speidel's testimony evidenced a departure from the normal sequence of passing and enacting ordinances, that departure is readily explainable by the population error in the first map and the rapidly approaching election necessitating the quick adoption of the second map. [¶ 106] Plaintiffs also contend Speidel's untruthful testimony regarding the city's law enforcement contract with the county provides overall support for their claims of broader discriminatory intent on the part of city officials. While Speidel testified falsely in her affidavit regarding the timing of the city council's plans to withdraw from a contract with the Bennett County sheriff's department, it is undisputed that Sheriff Cummings, a Native American, misused county funds while in office. T. IV, p. 862. Had Cummings' departure from the sheriff's department not involved malfeasance, the court would be more likely to infer that the abrupt change in policy was racially motivated. Because it did involve malfeasance, the court finds that this is not sufficient evidence to support a finding of broader discriminatory intent on the part of city officials.

[¶ 107] Accordingly, it is hereby

[¶ 108] ORDERED that judgment be entered for defendants on the issue of vote dilution in violation of § 2 of the Voting Rights Act of 1965; and

[¶ 109] IT IS FURTHER ORDERED that judgment be entered for the defendants on plaintiffs' intent claim alleging violations of § 2 of the Voting Rights Act of 1965, the Fourteenth Amendment, and the Fifteenth Amendment of the United States Constitution.